Luminous Landscape Forum

Equipment & Techniques => Digital Cameras & Shooting Techniques => Topic started by: dreed on May 23, 2013, 11:31:14 PM

Title: ETTR vs ISO
Post by: dreed on May 23, 2013, 11:31:14 PM
ETTR is a mantra that I often see spouted on the Internet as the cure for better IQ and at ISO 100, it makes complete sense.

But if I'm at ISO 3200 (say) and shooting at 1/200 +1 EV, would I get better IQ if I shot at ISO 1600 and used 1/200 at +0 EV?

Or does this depend on the sensor?
Title: Re: ETTR vs ISO
Post by: Tony Jay on May 23, 2013, 11:40:31 PM
As you say, at ISO 100 (or, in fact, base ISO) it makes sense.
If not at base ISO it is (non)sense...

Tony Jay
Title: Re: ETTR vs ISO
Post by: ErikKaffehr on May 24, 2013, 12:42:47 AM
It depends on sensor, or more exactly signal readout electronics. But, it is really exposure that matters.


Sony sensors: Iso matters little
Canon: Better shadow noise when increasing ISO, but little difference in mid tones. My guess.
Nikon: Depends, D800, D600 use Sony sensor D4 is more like Canon

Best regards
Erik


ETTR is a mantra that I often see spouted on the Internet as the cure for better IQ and at ISO 100, it makes complete sense.

But if I'm at ISO 3200 (say) and shooting at 1/200 +1 EV, would I get better IQ if I shot at ISO 1600 and used 1/200 at +0 EV?

Or does this depend on the sensor?
Title: Re: ETTR vs ISO
Post by: Ray on May 24, 2013, 02:54:33 AM
What a coincidence. I'm in the process of gradually reorganizing my images, and just yesterday came across some images stored on an old hard drive, of a test I'd conducted in April 2006 to satisfy myself that increasing ISO on a Canon 5D to create an ETTR, as opposed to underexposing at ISO 100, really did have a substantial benefit regards general image quality.

My first reaction on coming across these images was to junk them, because the information they provide is now 'old hat' for me, and also I didn't have the RAW images stored in the same folder. I'll probably discover them later on some other old hard drive.

Anyway, I decided to keep the conversions in my newly created 'Technical issues and Lens Tests" folder and have attached below a couple of comparisons crops.
Title: Re: ETTR vs ISO
Post by: jrsforums on May 24, 2013, 03:10:58 AM
ETTR is a mantra that I often see spouted on the Internet as the cure for better IQ and at ISO 100, it makes complete sense.

But if I'm at ISO 3200 (say) and shooting at 1/200 +1 EV, would I get better IQ if I shot at ISO 1600 and used 1/200 at +0 EV?

Or does this depend on the sensor?

Dreed, this subject has been covered in quite a few threads over the last couple years.  The best summary I have seen came from ejmartin

http://www.luminous-landscape.com/forum/index.php?topic=56947.msg463071#msg463071

Quote
What is the appropriate mantra?  I would prefer "Maximize Exposure"; maximize subject to three constraints:

(1) maintaining needed DoF, which limits how much you can open up the aperture;
(2) freezing motion, which limits the exposure time;
(3) retaining highlight detail, by not clipping wanted highlight areas in any channel. 

Note that ISO is not part of exposure.  Exposure has only to do with aperture and shutter speed.  Maximizing exposure guarantees that one captures as many photons as possible subject to photographic constraints, and therefore optimizes S/N.

How does ISO enter?  It enters as a subsidiary aspect of optimizing S/N.  On many cameras (those with CCD sensors, and the newer Sony Exmor sensors), there is little or no advantage to raising the ISO, which aids point (3) -- leaving the ISO at a low value may leave the histogram "to the left" for your chosen exposure, it will give more highlight headroom but will not degrade S/N; such cameras can safely be operated at close to their lowest ISO (the precise optimal ISO depends on the details of a given camera design).  On the other hand, for many other CMOS sensor'd cameras, such as Canon's offerings, and Nikons with Nikon-designed CMOS sensors (D3/D700/D3s, for example), noise relative to exposure is improved by increasing the ISO; after you have maximized the exposure (ie by satisfying criteria (1) and (2)), you have a tradeoff to make for (3) -- raising the ISO lowers shadow noise (up to a camera-specific point of diminishing returns, usually about ISO 1600), therefore improving S/N, but reduces highlight headroom for your chosen exposure, so one has to decide how high the ISO can go and still keep wanted highlights unclipped. 

Anyway, the prescription is to set the exposure (shutter speed and aperture only) according to (1) and (2); back off the exposure if at base ISO and you are compromising (3).  If you are compromising (3) with your chosen exposure and you are not at base ISO, then you should have started with a lower ISO.  Afterward, depending on the specifics of the camera's noise profile, further optimization results from raising the ISO, up to the limit specified by (3), or the camera's ISO point of diminishing returns, whichever is arrived at first.

So, it's (almost) all about ME.     
Title: Re: ETTR vs ISO
Post by: RFPhotography on May 24, 2013, 07:23:43 AM
3 or 4 articles on ETTR on this site, http://chromasoft.blogspot.ca/2009/09/why-expose-to-right-is-just-plain-wrong.html

Schewe has a short article on it as well, http://schewephoto.com/ETTR/

Emil has an intensive article on his site, http://theory.uchicago.edu/~ejm/pix/20d/tests/noise/noise-p3.html, scroll down to the S/N and Exposure Decision section.
Title: Re: ETTR vs ISO
Post by: digitaldog on May 24, 2013, 09:32:29 AM
ETTR is a mantra that I often see spouted on the Internet as the cure for better IQ and at ISO 100, it makes complete sense.

It does depend on the camera system. In this example, using a Canon 5DMII, a HIGHER ISO produces less noise due to ETTR:

(http://digitaldog.net/files/100vs800iso.jpg)

ETTR is nothing more than Exposure 101, in this case for processing raw data in your converter (Exposure + Development). You still need to decide an appropriate shutter and what DOF you wish and how much light you can work with. ETTR is about optimal exposure, not over exposure. If you increase the signal data for less noise only to end up with camera shake or the incorrect DOF you desire, it's a waste! You have linear raw data (your neg), you need to test your camera system to uncover the ideal way to expose for that data which many call ETTR. It should just be called ideal exposure for the media (raw) which is quite different than ideal exposure for the camera generated JPEG. Then you have to test the development (normalize for ideal rendering from ETTR data). Again, exposure and photography 101 so to speak, just different equipment and data to work with.
Title: ETTR vs ISO: how high is it worth pushing ISO speed with various sensors?
Post by: BJL on May 24, 2013, 10:07:50 AM
Dreed, this subject has been covered in quite a few threads over the last couple years.  The best summary I have seen came from ejmartin
http://www.luminous-landscape.com/forum/index.php?topic=56947.msg463071#msg463071
Agreed! The only thing that I can see being worth adding is details on how high it is worth pushing up the ISO speed setting on a particular camera before you stop gaining anything with noise and instead are just saving the effort of adjusting levels up in post-processing, but at the added risk of blowing highlights due to "ISO amplifier clipping" even when the photosites themselves were not overfilled.

Some answers seem to be:
I am sure that others in this forum can add quantitative details (and corrections), based on all the testing that has been reported in other discussions here.
Title: Re: ETTR vs ISO
Post by: BJL on May 24, 2013, 10:19:02 AM
3 or 4 articles on ETTR on this site, http://chromasoft.blogspot.ca/2009/09/why-expose-to-right-is-just-plain-wrong.html
That article is mostly rather silly, with all its comparisons at higher ISO settings and between images taken at different ISO setting. In the end it almost entirely contradicts its link-bait "completely wrong" headline with this
Quote
there is one situation where ETTR can help - when you're already at the lowest ISO setting you camera offers
and this
Quote
Bottom line - ETTR offers improved image quality in only one specific situation - where you can use a lower ISO setting than your camera has. In all other situations, ETTR will only ever decrease image quality.
But working at base ISO speed is a key part of the "classic ETTR" as described by Michael Reichmann in the article that this article quotes, so the bottom line is to agree that ETTR as originally described and advocated by Michael Reichmann (and Thomas Knoll?) is a good approach! All the rest is a straw man argument.
Title: Re: ETTR vs ISO: how high is it worth pushing ISO speed with various sensors?
Post by: jrsforums on May 24, 2013, 10:29:18 AM
Agreed! The only thing that I can see being worth adding is details on how high it is worth pushing up the ISO speed setting on a particular camera before you stop gaining anything with noise and instead are just saving the effort of adjusting levels up in post-processing, but at the added risk of blowing highlights due to "ISO amplifier clipping" even when the photosites themselves were not overfilled.

Some answers seem to be:
  • With the combination of a CCD and a high quality ADC, as in DMF backs, there is no IQ benefit to increased "ISO amplification", because all the noise enters before the amplification.
  • With Canon's combination of CMOS sensors (with on-sensor ISO amplification?) and off-board ADC, there can be benefits all the way up to about ISO 1600; see Ray's post for example.
  • With recent Sony and Nikon system cameras, there can be some benefit to going one stop above base ISO speed, to about ISO 400, but no significant benefit beyond that.
  • Similarly with the current generation of micro Four Thirds sensors, there can be some benefit to going one or even two stops above base ISO setting, maybe to ISO 800, but not beyond.
I am sure that others in this forum can add quantitative details (and corrections), based on all the testing that has been reported in other discussions here.

As I remember, I believe it was Emil who "tied" the increased ISO capability to the DxoMark DR curve.  That is, as long as the DR curve was relatively flat, you could get some ETTR gain from increased ISO.

In the chart below you can see your examples of the 5D3 and MFT sensors and the Sony sensor in the D800.
Title: Re: ETTR vs ISO
Post by: jrsforums on May 24, 2013, 10:33:05 AM
That article is mostly rather silly, with all its comparisons at higher ISO settings and between images taken at different ISO setting. In the end it almost entirely contradicts its link-bait "completely wrong" headline with thisand thisBut working at base ISO speed is a key part of the "classic ETTR" as described by Michael Reichmann in the article that this article quotes, so the bottom line is to agree that ETTR as originally described and advocated by Michael Reichmann (and Thomas Knoll?) is a good approach! All the rest is a straw man argument.

BJL...The articles major focus seemed to be on ISO benefit (or lack of) for ETTR.  Unfortunately the camera used was a Canon G10, which is CCD....so in Emil's terms....ISO-less.
Title: Re: ETTR vs ISO
Post by: BJL on May 24, 2013, 10:42:26 AM
@jrsforums thanks for the graphs. My understanding of the math of it is that there is no further benefit once the curve hits the straight line "one stop per stop" slope section at right. So, unravelling the misguided and confusing horizontal axis shift on those graphs, the rule seems to be:
Title: Re: ETTR vs ISO
Post by: RFPhotography on May 24, 2013, 11:03:02 AM
That article is mostly rather silly, with all its comparisons at higher ISO settings and between images taken at different ISO setting. In the end it almost entirely contradicts its link-bait "completely wrong" headline

Don't agree.  What he's doing is showing why ETTR doesn't apply at higher ISO settings with many cameras.  Basically, cameras with sensor response like the D7000 and Pentax K5 - I think those were the first well-known about 'ISOless' sensors don't benefit from ETTR at other than base ISO.  Some other cameras, that have a more curved response will benefit, to a point, using ETTR at higher than base ISO.  But even in those cameras, after a point there is no longer a benefit because the response becomes like those with 'ISOless' sensors.  I'm not completely convinced by his hue twist argument but the ISO comparisons are valid.

Quote
But working at base ISO speed is a key part of the "classic ETTR" as described by Michael Reichmann in the article that this article quotes, so the bottom line is to agree that ETTR as originally described and advocated by Michael Reichmann (and Thomas Knoll?) is a good approach! All the rest is a straw man argument.

Neither the original article from 2003 nor the updated article from 2011 mention shooting at base ISO.  I just looked at both again and didn't see anything about base ISO. 

The problem with the 'ETTR Mantra' is that there are some who will claim that ETTR is the only way to shoot in all circumstances.  That's simply not the case.  That's what the Chromasoft articles point out and, if I recall, I think Schewe's article talks about using base ISO (would have to go back and reread to be sure).
Title: Re: ETTR vs ISO
Post by: digitaldog on May 24, 2013, 11:42:49 AM
  • 5D Mk 3: worth going as far as the ISO 1600 setting, and maybe even 3200
According to Eric Chan, if you shoot Raw, there’s no reason to shoot above ISO3200 on the 5DMII, not sure about the III. The noise levels will either be the same or might actually be worse if you don’t simply pin the max at ISO 3200 and adjust that Raw using the Exposure slider in ACR. ISO’s above 3200 are only useful for 5DMII users shooting JPEGs. That was back when we were using PV2003/2010 but I don't know that PV2012 would make that any different.

Quote
That article is mostly rather silly,
I agree!
Title: Re: ETTR vs ISO
Post by: digitaldog on May 24, 2013, 11:46:51 AM
The problem with the 'ETTR Mantra' is that there are some who will claim that ETTR is the only way to shoot in all circumstances.  That's simply not the case.  That's what the Chromasoft articles point out and, if I recall, I think Schewe's article talks about using base ISO (would have to go back and reread to be sure).

It's only a problem when the photographer doesn’t understand the basics of exposure and the role it plays on other aspects of photography which I mentioned (ETTR that causes camera shake, undesired DOF etc). Otherwise ETTR is simply an attempt at producing the best quality data. In the old days, we could pop ISO 100 film OR ISO 800 film based on what we knew or thought we knew about exposing a scene. We used ISO 100 when we knew we wanted lower noise (grain) than ISO 800 and we had abundant light to produce the desired results (san's camera shake etc). IF you have enough light, why would you not setup the camera system to provide the most noise free image? Then you can go out and buy a plug-in that adds noise to look like film <g>.
Title: Re: ETTR vs ISO
Post by: BJL on May 24, 2013, 11:57:57 AM
Fair enough: the article by MR does not explicitly talk about working at base ISO; it is just that in the context of what MR has said elsewhere on the same site, I tend to take it as "so obvious as to not need stating" that the first step to improving IQ is to use the lowest viable ISO setting, before moving on to refinements like ETTR.

Rather than debate his blatantly exaggerated "just plain wrong" criticism, based mostly on the inadequate example of a now almost irrelevant CCD sensor, let me just restate what might in the end be points of agreement:

Title: Re: ETTR vs ISO
Post by: digitaldog on May 24, 2013, 12:04:35 PM
Fair enough: the article by MR does not explicitly talk about working at base ISO; it is just that in the context of what MR has said elsewhere on the same site, I tend to take it as "so obvious as to not need stating" that the first step to improving IQ is to use the lowest viable ISO setting, before moving on to refinements like ETTR.

But it's not obvious as I tried to illustrate with the Canon. A higher ISO produces less noise than a lower one (again, due to ETTR).

Quote
- as MR notes in followups to his original article, the preview histograms that cameras give us, based on a JPEG preview, are not ideal. For example, on many cameras, that preview shows clipping one or more stops before there is any clipping of raw data. So optimal ETTR can involve being willing to push exposure a bit into "highlight red-lining".
We had no such LCD on film cameras yet many photographers learned to expose properly (ideally) for their film and development and didn't take any ISO setting at face value. ETTR is just ideal exposure for raw data. That the camera LCD is lying to us just forces us to learn to expose as we all did prior to digital capture. IF you capture raw data and use the LCD as your guide, you are basically under exposing that data which isn't ideal exposure. Just as we could under expose film and push it (to a degree), it wasn't considered an ideal exposure and development process.
Title: Re: ETTR vs ISO
Post by: jrsforums on May 24, 2013, 12:52:38 PM
We had no such LCD on film cameras yet many photographers learned to expose properly (ideally) for their film and development and didn't take any ISO setting at face value. ETTR is just ideal exposure for raw data. That the camera LCD is lying to us just forces us to learn to expose as we all did prior to digital capture. IF you capture raw data and use the LCD as your guide, you are basically under exposing that data which isn't ideal exposure. Just as we could under expose film and push it (to a degree), it wasn't considered an ideal exposure and development process.

I agree...sorta.

Just like with film, we can "calibrate" our results, so we "know' how to "get it right in the camera"....which means optimally exposed....which is different by media...whether it is different film stocks, digital jpeg, or RAWs from different digital sensors.

Doing my own testing with RawDigger, I found, for the 5D3, that I could spot meter the highest significant highlight and then place it at +3 to +3.5 stops over metered.  This gave me an image exposed as far to the right as possible, without any channels clipped.

Jim Kasson, on his blog, The Last Word, did a lot of testing on different cameras...many which did not have spot meters, to arrive at means to adjust the histogram to more closely mimic a RAW histogram.
Title: Re: ETTR vs ISO
Post by: bjanes on May 24, 2013, 12:54:19 PM
It does depend on the camera system. In this example, using a Canon 5DMII, a HIGHER ISO produces less noise due to ETTR:

(http://digitaldog.net/files/100vs800iso.jpg)

ETTR is nothing more than Exposure 101, in this case for processing raw data in your converter (Exposure + Development). You still need to decide an appropriate shutter and what DOF you wish and how much light you can work with. ETTR is about optimal exposure, not over exposure. If you increase the signal data for less noise only to end up with camera shake or the incorrect DOF you desire, it's a waste! You have linear raw data (your neg), you need to test your camera system to uncover the ideal way to expose for that data which many call ETTR. It should just be called ideal exposure for the media (raw) which is quite different than ideal exposure for the camera generated JPEG. Then you have to test the development (normalize for ideal rendering from ETTR data). Again, exposure and photography 101 so to speak, just different equipment and data to work with.

In this example, the exposure (if I read the exposure data properly--it is somewhat blurred) is the same: 1/60 sec @ f/5.6. As Emil states:

"Note that ISO is not part of exposure.  Exposure has only to do with aperture and shutter speed.  Maximizing exposure guarantees that one captures as many photons as possible subject to photographic constraints, and therefore optimizes S/N."


In the case of the Canon used here, increasing the ISO will give a better looking histogram, but the real reason for increasing the ISO in this case is to reduce the read noise. Photon noise will be unchanged, since the number of photons collected is the same. With an ISO-less sensor, one could leave the ISO setting at base and increase exposure with the raw converter and get the same results, while providing more highlight headroom.

Bill
Title: Re: ETTR vs ISO
Post by: digitaldog on May 24, 2013, 01:10:56 PM
In this example, the exposure (if I read the exposure data properly--it is somewhat blurred) is the same: 1/60 sec @ f/5.6.

Yes. Exposure settings for aperture and shutter are identical, the only difference is the ISO setting. And the differences in the amount of noise is significant as one would expect with ETTR (optimal exposure for raw).


Quote
As Emil states:

"Note that ISO is not part of exposure.  Exposure has only to do with aperture and shutter speed.  Maximizing exposure guarantees that one captures as many photons as possible subject to photographic constraints, and therefore optimizes S/N."


In the case of the Canon used here, increasing the ISO will give a better looking histogram, but the real reason for increasing the ISO in this case is to reduce the read noise. Photon noise will be unchanged, since the number of photons collected is the same. With an ISO-less sensor, one could leave the ISO setting at base and increase exposure with the raw converter and get the same results, while providing more highlight headroom.

I don't care about the Histogram! But what I see is a vast difference in noise whereby the ONLY setting change is ISO. So it's obviously affecting the degree of noise.

You saying increasing the ISO in this case is to reduce the read noise, this isn't also ETTR? Seems the number of photons collected should be the same since what he's calling exposure is fixed in both examples. The results however are clear in terms of the differences in noise and I understand that not all camera sensors respond as this Canon does.
Title: Re: ETTR vs ISO
Post by: digitaldog on May 24, 2013, 01:13:45 PM
I agree...sorta.
I don't see anything but full agreement <g>. What part did I mess?

Quote
Doing my own testing with RawDigger, I found, for the 5D3, that I could spot meter the highest significant highlight and then place it at +3 to +3.5 stops over metered.  This gave me an image exposed as far to the right as possible, without any channels clipped.
On my 5DMII using an incident meter, I was about 1.5+ stops but less than 2 stops before I clipped highlights (using the BableColor tile) with ACR/LR PV2010.
Title: using the lowest "viable" ISO speed setting
Post by: BJL on May 24, 2013, 01:27:01 PM
But it's not obvious as I tried to illustrate with the Canon. A higher ISO produces less noise than a lower one (again, due to ETTR).
I think we agree, but with some extra complications for the special case of Canon's sensors: read on to my third bullet point. By "lowest viable ISO setting", I did not mean the camera's base ISO setting, but more like the lowest one that gives an acceptable combination of shutter speed and aperture when exposing "on meter".
Title: Re: ETTR vs ISO
Post by: RFPhotography on May 24, 2013, 01:30:11 PM
It's only a problem when the photographer doesn’t understand the basics of exposure and the role it plays on other aspects of photography which I mentioned (ETTR that causes camera shake, undesired DOF etc). Otherwise ETTR is simply an attempt at producing the best quality data. In the old days, we could pop ISO 100 film OR ISO 800 film based on what we knew or thought we knew about exposing a scene. We used ISO 100 when we knew we wanted lower noise (grain) than ISO 800 and we had abundant light to produce the desired results (san's camera shake etc). IF you have enough light, why would you not setup the camera system to provide the most noise free image? Then you can go out and buy a plug-in that adds noise to look like film <g>.

I'm not denying that, Andrew nor am I arguing that point.  I'm simply stating that it's not, as some people believe, the only way to expose in all situations.  We're actually saying the same thing, just in a different way. 
Title: Re: ETTR vs ISO
Post by: RFPhotography on May 24, 2013, 01:33:42 PM
But it's not obvious as I tried to illustrate with the Canon. A higher ISO produces less noise than a lower one (again, due to ETTR).


And that is a given with some cameras depending on how the sensor responds to changes in ISO.  With the so-called ISOless sensors the higher ISO and ETTR won't produce lower noise.  With some cameras, including some Canons it will - up to a point.  Where that point is will vary from camera to camera.  With the 5D it was about, if I recall, ISO 800.  With the newer 5D models it's a stop or so higher.
Title: Re: ETTR vs ISO
Post by: digitaldog on May 24, 2013, 01:48:12 PM
And that is a given with some cameras depending on how the sensor responds to changes in ISO.  With the so-called ISOless sensors the higher ISO and ETTR won't produce lower noise.  With some cameras, including some Canons it will - up to a point.  Where that point is will vary from camera to camera.  With the 5D it was about, if I recall, ISO 800.  With the newer 5D models it's a stop or so higher.

Totally agree. The Canon's here are kind of unique. What I'm not fully understanding are the results and they differentiation between read noise and these settings versus exposure as defined by aperture/shutter to increase photon's which result in less noise. IOW, upping ISO on this Canon without adjust exposure lower results in less noise, this isn't ETTR? 
Title: Re: ETTR vs ISO
Post by: jrsforums on May 24, 2013, 01:56:14 PM
Totally agree. The Canon's here are kind of unique. What I'm not fully understanding are the results and they differentiation between read noise and these settings versus exposure as defined by aperture/shutter to increase photon's which result in less noise. IOW, upping ISO on this Canon without adjust exposure lower results in less noise, this isn't ETTR? 

I do not think it is....as the "classical" definition of ETTR is to increase the amount of photons captured, thus reduce SNR.  Increasing ISO does not increase the light (photons) captured.
Title: Re: ETTR vs ISO
Post by: digitaldog on May 24, 2013, 01:59:48 PM
I do not think it is....as the "classical" definition of ETTR is to increase the amount of photons captured, thus reduce SNR.  Increasing ISO does not increase the light (photons) captured.

What's it doing to reduce the noise?

Good to know that at least some are agreeing that this process isn't increasing the photon count, I've been scratching my head trying to understand what the ISO on the Canon is doing to reduce the noise. Not implementing ETTR and setting the Canon to ISO 100 certainly doesn't produce the lowest noise in a capture. Why?

I can understand the argument that exposure is Aperture + Shutter but then ISO, at least in this case, plays a role. Do we need a separate acronym for Canon cameras whereby we increase ISO, don't adjust 'exposure' based on the increase and end up with less noise?
Title: Re: ETTR vs ISO
Post by: jrsforums on May 24, 2013, 02:01:43 PM
In the case of the Canon used here...the real reason for increasing the ISO in this case is to reduce the read noise. Photon noise will be unchanged, since the number of photons collected is the same...

Bill


Bill...

With a camera such as the Canon, how do we know when it is beneficial to increase ISO to reduce noise?
Title: Re: ETTR vs ISO
Post by: RFPhotography on May 24, 2013, 02:04:45 PM
Totally agree. The Canon's here are kind of unique. What I'm not fully understanding are the results and they differentiation between read noise and these settings versus exposure as defined by aperture/shutter to increase photon's which result in less noise. IOW, upping ISO on this Canon without adjust exposure lower results in less noise, this isn't ETTR? 

But it sort of is.  At least in the context of the higher ISO.  If you shoot at ISO 100, f4 and 1/125 and that gives you a 'normal' exposure, then you switch to ISO 200 and shoot at the same aperture & shutter speed you're going to push the exposure 1 stop higher.  That's, essentially, what ETTR is doing.  Let's take it as a given for the moment that the 1 stop pushes the histogram to the right edge without clipping. 
Title: Re: ETTR vs ISO
Post by: BartvanderWolf on May 24, 2013, 02:10:03 PM
Bill...

With a camera such as the Canon, how do we know when it is beneficial to increase ISO to reduce noise?


Hi John,

Jim Kasson devised an illustrative way (http://www.luminous-landscape.com/forum/index.php?topic=76446.msg615106#msg615106) to determine the maximum useful ISO setting, beyond which it is better to just underexpose and boost with the Raw converter's exposure setting.

See the attachment where I used a similar method to determine it for my camera.

Cheers,
Bart
Title: Re: ETTR vs ISO
Post by: digitaldog on May 24, 2013, 02:12:24 PM
But it sort of is.  At least in the context of the higher ISO.  If you shoot at ISO 100, f4 and 1/125 and that gives you a 'normal' exposure, then you switch to ISO 200 and shoot at the same aperture & shutter speed you're going to push the exposure 1 stop higher.  That's, essentially, what ETTR is doing.  Let's take it as a given for the moment that the 1 stop pushes the histogram to the right edge without clipping. 
What I initially thought too but now I'm more confused <g>.

My main purpose for the test I did and shown was to point out to those that write "Upping ISO always increases noise". Clearly that's not the case, at least with the Canon's.

IS the reduction I see in noise the result of ETTR or how Canon deals with this ISO (Read Noise)? IF our goal is to test ISO and exposure such we get the lowest noise, and altering ISO up provides less, is this or isn't this ETTR? Going back to film days, when I'd buy film in bricks, fill a refrigerator, run exposure, ISO and processing tests (with CC gel packs) to nail everything. I feel we need to do the same today with digital. It appears that with the Canon sensor, bringing ISO into the mix is necessary. Not fully sure why.
Title: Re: ETTR vs ISO
Post by: jrsforums on May 24, 2013, 02:17:06 PM
Hi John,

Jim Kasson devised an illustrative way (http://www.luminous-landscape.com/forum/index.php?topic=76446.msg615106#msg615106) to determine the maximum useful ISO setting, beyond which it is better to just underexpose and boost with the Raw converter's exposure setting.

Cheers,
Bart

Bart....

Jim did a great piece of work.  However, none of his testing included the Canons which have more of the ability to increase ISO.  Nor, if I remember correctly, a correlation to the DxoMark DR charts, which does seem to be valid.
Title: Re: ETTR vs ISO
Post by: BartvanderWolf on May 24, 2013, 02:19:14 PM
Jim did a great piece of work.  However, none of his testing included the Canons which have more of the ability to increase ISO.  Nor, if I remember correctly, a correlation to the DxoMark DR charts, which does seem to be valid.

John,

I just added a chart for my 1Ds3 to the earlier post, maybe that helps?

Cheers,
Bart
Title: Re: ETTR vs ISO
Post by: RFPhotography on May 24, 2013, 02:20:54 PM
What I initially thought too but now I'm more confused <g>.

My main purpose for the test I did and shown was to point out to those that write "Upping ISO always increases noise". Clearly that's not the case, at least with the Canon's.

IS the reduction I see in noise the result of ETTR or how Canon deals with this ISO (Read Noise)? IF our goal is to test ISO and exposure such we get the lowest noise, and altering ISO up provides less, is this or isn't this ETTR? Going back to film days, when I'd buy film in bricks, fill a refrigerator, run exposure, ISO and processing tests (with CC gel packs) to nail everything. I feel we need to do the same today with digital. It appears that with the Canon sensor, bringing ISO into the mix is necessary. Not fully sure why.

It may be partly how Canon deals with noise.  There may be some sort of 'always on' noise reduction.  
Title: Re: ETTR vs ISO
Post by: jrsforums on May 24, 2013, 02:24:41 PM
I don't see anything but full agreement <g>. What part did I mess?
On my 5DMII using an incident meter, I was about 1.5+ stops but less than 2 stops before I clipped highlights (using the BableColor tile) with ACR/LR PV2010.

The "sorta" was that with film many photogs had handheld incident or spot meters to support their testing and field work....today we often need to rely on what the cameras give us....which is why I mentioned what I d and Kasson's work.

With digital, I do not believe that an incident meter is key to optimizing exposure.  Middle grey is interesting; the significant highlight is what is most important as we want to get that properly placed without clipping.

 This is why I use the camera spot meter and Kasson used other methods to arrive at a histogram which mimic the Raw.

On Lightroom, as you know, 2010 and 2012 PV are significantly different on highlight clipping.  On 2012, a highlight above 96-97% will have at least one channel clipped and recovery applied.  this is why RawDigger was needed to calibrate the 5D3 readings...
Title: Re: ETTR vs ISO
Post by: digitaldog on May 24, 2013, 02:39:38 PM
The "sorta" was that with film many photogs had handheld incident or spot meters to support their testing and field work....today we often need to rely on what the cameras give us....which is why I mentioned what I d and Kasson's work.
IF they used such equipment in the past, why wouldn’t they do so today? Exposure is exposure. The only differences I see are: 1. We have an LCD to look at, which in this context is wrong. 2. We have linear encoded raw data which isn't like a JPEG or film which has a curve (film being H&D). So we treat this like a different kind of 'film' if you will.

Quote
With digital, I do not believe that an incident meter is key to optimizing exposure.
 
The same is true for film. The differences are that an incident meter is less fooled than a reflective meter in an inexperienced users's hands. If they point that reflective meter at a white dog on snow and take what it tells them as a fact for proper exposure, they are in for a rude awaking. Not the case with the incident meter. But which ever you use, you have to use the tools properly. And no meter is useful until we decide on the ISO and that means we also have to put the processing (development) into the mix.

Quote
Middle grey is interesting; the significant highlight is what is most important as we want to get that properly placed without clipping.
You can do this with either meter once you know how they behave and how your exposure and development behave.

Quote
This is why I use the camera spot meter and Kasson used other methods to arrive at a histogram which mimic the Raw.
And if you point that at the white dog and understand it will inform you how to end up with a gray dog, then adjust, all is fine.

Quote
On Lightroom, as you know, 2010 and 2012 PV are significantly different on highlight clipping.  On 2012, a highlight above 96-97% will have at least one channel clipped and recovery applied.  this is why RawDigger was needed to calibrate the 5D3 readings...
Yes but nothing brings back clipped data so you still need to figure out the limits here. In that context, this part of the test is somewhat raw processing agnostic. If you blow out all data at plus 2 stops over what the meter recommends, nothing will bring that back. Yet I agree that you can't separate the development further on, so again, we have to test the process, just as we did with film. That means finding the proper ISO, along with proper metering and development.
Title: Re: ETTR vs ISO
Post by: jrsforums on May 24, 2013, 03:03:59 PM
IF they used such equipment in the past, why wouldn’t they do so today? Exposure is exposure. The only differences I see are: 1. We have an LCD to look at, which in this context is wrong. 2. We have linear encoded raw data which isn't like a JPEG or film which has a curve (film being H&D). So we treat this like a different kind of 'film' if you will.
 
The same is true for film. The differences are that an incident meter is less fooled than a reflective meter in an inexperienced users's hands. If they point that reflective meter at a white dog on snow and take what it tells them as a fact for proper exposure, they are in for a rude awaking. Not the case with the incident meter. But which ever you use, you have to use the tools properly. And no meter is useful until we decide on the ISO and that means we also have to put the processing (development) into the mix.
You can do this with either meter once you know how they behave and how your exposure and development behave.
And if you point that at the white dog and understand it will inform you how to end up with a gray dog, then adjust, all is fine.
Yes but nothing brings back clipped data so you still need to figure out the limits here. In that context, this part of the test is somewhat raw processing agnostic. If you blow out all data at plus 2 stops over what the meter recommends, nothing will bring that back. Yet I agree that you can't separate the development further on, so again, we have to test the process, just as we did with film. That means finding the proper ISO, along with proper metering and development.

First, there are a whole class of people, today, who never had all those tools...or want/need to carry them around if not necessary.  They need to understand how to use the tools they have available to them.

Second, I think you missed my point.  An incident reading will not help you find the significant highlight in a scene....and place it as far to the right as possible.  With testing, you can assume the range of the scene...or, if like you, you are really experienced, you can adjust the EC by knowing if bright tones are or are not present.

I agree that a wide reflective meter can be easily fooled.  However, I am talking about the spot meter, which is fairly narrow on a 5D3.

Reading the reflected light from the significant highlight with a spot meter places it (with knowledge and pretesting) in the right of the histogram....without any clipping (of significant tones).  All the other tones fall where they may.  No guessing on tonal values is needed.

This maximizes the dynamic range of the camera.  If the histogram shows shadows being clipped, the only choice is to bracket and blend.
Title: Re: ETTR vs ISO
Post by: jrsforums on May 24, 2013, 03:10:11 PM
Yes but nothing brings back clipped data so you still need to figure out the limits here. In that context, this part of the test is somewhat raw processing agnostic. If you blow out all data at plus 2 stops over what the meter recommends, nothing will bring that back. Yet I agree that you can't separate the development further on, so again, we have to test the process, just as we did with film. That means finding the proper ISO, along with proper metering and development.

We agree.

Just to be clear, this is why I mentioned "calibrating" with RawDigger.  The setting of for the 5D3 of about 3.5 stops over the spot reading of the significant highlight places that highlight at about 96%  in LR.  Above that is about 2 stops of potentially "recovered" highlights....which is nice, but will never have the texture and colors of the unrecovered highlights....a nice "buffer" to have, but not optimal to best image quality.
Title: Re: ETTR vs ISO
Post by: bjanes on May 24, 2013, 03:10:27 PM
Yes. Exposure settings for aperture and shutter are identical, the only difference is the ISO setting. And the differences in the amount of noise is significant as one would expect with ETTR (optimal exposure for raw).

The exposures are the same, so how do you call the shot at higher ISO as optimal exposure for raw? To obtain the same exposure while varying the ISO, one would likely set the camera to expose manually, set the aperture and shutter speed, and then vary the ISO. If you are on auto exposure, increasing the ISO will result in reduced exposure and reduced number of photons collected.

I don't care about the Histogram! But what I see is a vast difference in noise whereby the ONLY setting change is ISO. So it's obviously affecting the degree of noise.

That is where confusion sets in. The appellation "ETTR" implies that the histogram is a determining factor and the histogram gives some indication of the number of photons collected only at base ISO. Certainly, at any ISO it is not wise to leave space at the right of the histogram is one is interested in the best SNR.

You saying increasing the ISO in this case is to reduce the read noise, this isn't also ETTR? Seems the number of photons collected should be the same since what he's calling exposure is fixed in both examples. The results however are clear in terms of the differences in noise and I understand that not all camera sensors respond as this Canon does.

I wouldn't call it exposure to the right, since the exposure is the same in both instances. However, this is all semantics and the important point it to obtain maximal image quality.

Bill
Title: Re: ETTR vs ISO
Post by: digitaldog on May 24, 2013, 03:18:40 PM
First, there are a whole class of people, today, who never had all those tools...or want/need to carry them around if not necessary.  They need to understand how to use the tools they have available to them.
That's fine. But it doesn't dismiss the need for proper understanding of exposure and use of tools for producing proper exposure. Or they don't care, then neither do I <g>.

Quote
Second, I think you missed my point.  An incident reading will not help you find the significant highlight in a scene....and place it as far to the right as possible. 

It will give you a very good indicator of scene exposure if the proper ISO is setup and one adjusts properly to produce ETTR. Yes, a reflective spot meter is useful to gauge where highlights specifically fall. Or for determining scene contrast. But you can use an incident meter for ETTR easily and this tool was the ONLY way to get a meter reading for the article I did where I wanted to shoot flash and figure out where I could pin white just below a very specular white. I couldn't use nor didn't need a reflective meter here.

Quote
I agree that a wide reflective meter can be easily fooled.  However, I am talking about the spot meter, which is fairly narrow on a 5D3.
It's a useful tool, no question. Again, the user needs to understand what the tool is telling them, again this is exposure 101, it's been this way since meter's were invented I suspect.

Quote
Reading the reflected light from the significant highlight with a spot meter places it (with knowledge and pretesting) in the right of the histogram....without any clipping (of significant tones).  All the other tones fall where they may.  No guessing on tonal values is needed.
What Histogram? The camera's is pretty much useless here if precision is your goal. Why even mess with a Histogram? One need to understand how their meter works I think we agree. We need to understand how the 'film' in your DSLR reacts too. We didn't have nor need histograms till very, very recently in the time frame of photography. We really don't need it today although if the camera gave us a raw histogram, that be useful to a point. It still has no idea about the processing part of this.
Title: Re: ETTR vs ISO
Post by: digitaldog on May 24, 2013, 03:27:26 PM
The exposures are the same, so how do you call the shot at higher ISO as optimal exposure for raw?
The combo of setting provided a superior capture. ISO has no role here? I have to at the very least enter some value for the meter.
Quote
The appellation "ETTR" implies that the histogram is a determining factor and the histogram gives some indication of the number of photons collected only at base ISO. Certainly, at any ISO it is not wise to leave space at the right of the histogram is one is interested in the best SNR.
Where and why does ETTR have to do anything with a Histogram? Which Histogram? In the raw processor (which is now part of the development process)? On the camera (not useful)?
In my mind, ETTR is using the various tools and settings we have to produce the optimal capture data. I don't see how the Histogram is even necessary expect to understand you didn't expose to the right enough or you exposed too far and you clipped. You could find this out without a Histogram too.
Quote
I wouldn't call it exposure to the right, since the exposure is the same in both instances. However, this is all semantics and the important point it to obtain maximal image quality.
Yes it's semantics but that's all we have at this point. Maybe ETTR is the wrong term. Expose to the right of an incorrect histogram? How about just Expose Optimally? Expose to right implies that there's this 'recommended' exposure but that's 'wrong' so adjust everything over (to the right). The recommendation prior to moving to the right is incorrect in terms of an ideal exposure, would you agree? If so, ETTR probably is the wrong term and I think most agree, it is confusing to a lot of users. They think they are over exposing when they are simply exposing optimally and not using the recommendation of a meter that's been designed for film and later, a JPEG. IF one tests where the sensor clips data they wish to capture, and move a tad to the left, would this not be correct and further ideal exposure? Maybe the term should be ETTL (after you figure out what not to clip).
Title: Re: ETTR vs ISO
Post by: digitaldog on May 24, 2013, 03:39:55 PM
This entire ETTR, exposure, ISO, testing takes me back to the mid 1980's while at photo school. I think 2nd trimester, we had one of our first assignments shooting color (first trimester we only shot 4x5 B&W film, it could not be burned or dodged in the darkroom for any assignment). This new color assignment was to shoot 4x5 color neg, I think it was Kodak VPS II or something. Stated ISO 160 on the box. The subject had to be stationary and have an 18% gray card in the center filling up at least 20%. We were to shoot at the recommended ISO, then 1, 2 and 3 stops under and over. We processed all the film together. Then in the darkroom, we had to make prints whereby the gray card on the print matched the actual gray card. Once all 7 prints were mounted and viewed, the results were very surprising! The plus 1 and 2 stops were easily printable and both produced a better print quality than the "normal" ISO 160 image. The under exposed prints all suffered compared to the 'normal' exposure.

ISO 160 for VPS worked just fine! But rating it at ISO 80 (or 40) and treating the rest of the process the same produced a very visible and beneficial result! No Histograms. But we did expose to the right if you will.
Title: ETTR vs ISO vs "expose for the highlights"
Post by: BJL on May 24, 2013, 03:56:30 PM
Where and why does ETTR have to do anything with a Histogram?
The very phrase "Expose To The Right" seems to originate in Michael Reichmann's article Expose (to the) Right (http://www.luminous-landscape.com/tutorials/expose-right.shtml), and refers to a strategy that is explicitly involved the camera's preview histogram:
Quote
The simple lesson to be learned from this is to bias your exposures so that the histogram is snugged up to the right, but not to the point that the highlights are blown.

So at the risk of getting into a semantic debate, I agree that for what you are talking about with external meters and such,
[Maybe] ETTR is the wrong term.
Maybe your point is just the advantages of exposing for the highlights, and the possible advantage (mostly with Canon sensors) of pushing the analog gain (ISO) as high as possible without producing amplifier clipping (at least up to about ISO 1600 or 3200).
Title: Re: ETTR vs ISO
Post by: jrsforums on May 24, 2013, 04:00:57 PM
Hi John,

Jim Kasson devised an illustrative way (http://www.luminous-landscape.com/forum/index.php?topic=76446.msg615106#msg615106) to determine the maximum useful ISO setting, beyond which it is better to just underexpose and boost with the Raw converter's exposure setting.

See the attachment where I used a similar method to determine it for my camera.

Cheers,
Bart

Thanks, Bart

I must have missed that thread.  Only had seen Jim's work from his blog.  From the length of the thread, I got a lot of reading to do :-)  ....and some testing...
Title: Re: ETTR vs ISO
Post by: jrsforums on May 24, 2013, 04:35:33 PM
I have mentioned Jim Kasson's series of blog posts on ETTR and using in-camera histograms.  For those interested, here is a link to them:
http://blog.kasson.com/?page_id=2387

Title: Re: ETTR vs ISO
Post by: jrsforums on May 24, 2013, 04:46:42 PM
i had discussed with Jim my method of using the 5D3's built in 1.5% spot meter to place the significant highlight.

He did not have a 5D3, but did test with a hand held spot meter.  His results are here:
http://blog.kasson.com/?p=2215

Even using this method, I still use the histogram to make sure the shadows are includd and no bracketing is required.  If bracketing is required, I use the spot meter to capture the highlights in the first shot and subsequent shots are +EC to capture the shadows.

I do believe the histogram is a tool to use, unfortunately it does not give true readings of the RAW image.  However, using this or other methods Jim tested (and have been written up other places) one can, using Emil's term, Maximize Exposure...and it can be done mostly without external tools, handheld meters, etc.,  of the bygone film era.
Title: Re: ETTR vs ISO
Post by: Jim Kasson on May 24, 2013, 04:53:49 PM
Jim Kasson devised an illustrative way (http://www.luminous-landscape.com/forum/index.php?topic=76446.msg615106#msg615106) to determine the maximum useful ISO setting, beyond which it is better to just underexpose and boost with the Raw converter's exposure setting.

Thanks, Bart. After working with the photon-compensated SNR curves for a while, I now believe that they ought to be used in conjunction with the Unity-Gain ISO to determine the point where you should stop turning up the ISO and start using the Exposure control in your favorite raw developer instead. Conventional UG ISO wisdom is that you should stop turning up the ISO a stop or two above the UG ISO. The photon-compensated SNR curves usually, but not always, bear that out. I'll show some examples further on in this post.

There's one other ISO you should consider: it's the one where your camera just throws in the towel and stops increasing the analog gain between the sensor and the analog-to-digital converter (ADC) and starts shifting the data to the left, first by one bit, and then, as the ISO goes up, by more. You certainly don't want to crank the ISO up past that point, unless you can't see the image in the back of the camera when you're chimping.

How do you find out when your camera stops amplifying more and starts shifting bits? The easiest way is to take a series of pictures of a featureless subject exposed per meter across a range of ISOs, and bring the raw data into RawDigger, and look at the histograms. When every other bucket is empty, the camera has shifted the data one bit-position to the left.

If you're new to Unity-Gain ISO, here's a place to start, and a method to find out what the UG ISO of your camera is. (http://blog.kasson.com/?p=2896)

OK, now for the examples. Here's one for a Nikon D4:

(http://www.kasson.com/ll/D4.PNG)

As you can see, the D4 is essentially ISO-less until about 3200, which is about 2 stops above the UG ISO. Both US ISO and SNR agree that you shouldn't use ISOs above 3200 unless you have some special reason. Up to ISO 6400, the D4 does no bit shifting.

 
With the D800E, the situation is similar:

(http://www.kasson.com/ll/D800e.png)

Above ISO 1600, a little over two stops above the UG ISO, the SNR starts to degrade. At ISO 3200, the D800E becomes a 13-bit camera, and turns into a 12-bit camera at ISO 6400. Best to stop at 1600. A friend who is into astrophotgraphy did extensive subjective testing with his D800E and decided that you shouldn't go any higher than 1250.

With the Sony RX-1, here's the situation:

(http://www.kasson.com/ll/rx1.PNG)

The reason for the two UG ISO arrows is that, although Sony advertises the RX-1 as a 14-bit camera, my copy has every other histogram bucket empty, indicating that it's really a 13-bit camera. The SNR curves show that less than two stops above the 13-bit UG ISO, the party is pretty much over. Up to ISO 6400, the RX-1 does no bit shifting.

With the Sony NEX-7, the curves look like this:

(http://www.kasson.com/ll/nex7.PNG)

These curves give a different result from the ones above with respect to UG ISO. They indicate that you shouldn't use the NEX-7 above its UG ISO. Because the NEX7 is a 12-bit camera, the UG ISO is two stops higher than it would be if it were a 14-bit camera. If it were a 14-bit camera, the two-stops-above-UGISO rule would apply.

Now for something completely different. Here are the curves for a camera with a non-Sony sensor. In fact, the sensor in the Leica M9 isn't even a CMOS sensor:

(http://www.kasson.com/ll/M9.PNG)

We see dramatic reductions in SNR above 640, which is about a stop and a half abouve the UG ISO. The M9 does no bit shifting.

I don't have results for any Canons, and I have been told that they do better than the five cameras for which I've shown data here as the ISO increases.

Jim
Title: Re: ETTR vs ISO
Post by: bjanes on May 24, 2013, 05:58:33 PM
The combo of setting provided a superior capture. ISO has no role here? I have to at the very least enter some value for the meter. Where and why does ETTR have to do anything with a Histogram? Which Histogram? In the raw processor (which is now part of the development process)? On the camera (not useful)?

ISO does have a role, but as Emil points out here (http://www.luminous-landscape.com/forum/index.php?topic=56947.msg463071#msg463071), ISO is not exposure. He explains it better than I can and with more authority, so why don't you go to the link and read up on the matter?

The histogram that one uses when actually deciding on exposure in the field is the camera histogram. You could ask Michael what he meant by ETTR, but I think he was using the camera histogram to set ETTR exposure, hence the term right refers to the appearance of that histogram. The camera histogram is useful if you now how to use it and have done your homework with Rawdigger. Jim Kasson has some useful tips. One can also use UniWB to get an approximation of the raw histogram. The ACR/LR histogram has problems, since it uses a baseline exposure. Also, PV2012 has image adaptive exposure and tends not to clip highlights. RawTherapee does have a raw histogram available with the click of the mouse.

In my mind, ETTR is using the various tools and settings we have to produce the optimal capture data. I don't see how the Histogram is even necessary expect to understand you didn't expose to the right enough or you exposed too far and you clipped. You could find this out without a Histogram too. Yes it's semantics but that's all we have at this point. Maybe ETTR is the wrong term. Expose to the right of an incorrect histogram? How about just Expose Optimally? Expose to right implies that there's this 'recommended' exposure but that's 'wrong' so adjust everything over (to the right). The recommendation prior to moving to the right is incorrect in terms of an ideal exposure, would you agree? If so, ETTR probably is the wrong term and I think most agree, it is confusing to a lot of users. They think they are over exposing when they are simply exposing optimally and not using the recommendation of a meter that's been designed for film and later, a JPEG. IF one tests where the sensor clips data they wish to capture, and move a tad to the left, would this not be correct and further ideal exposure? Maybe the term should be ETTL (after you figure out what not to clip).

I agree. However, the term ETTR does imply the use of a histogram and this leads to confusion. What do you think right is referring to? Nevertheless, a properly interpreted histogram is helpful at base ISO. Measuring the highlight reading with a spot meter where the sensor saturates is a good method, but then you have to determine the location in the scene where the brightest highlight you don't want to clip is, take the reading with a 1 degree spot meter, and transfer it to the camera. When you do all this you may miss action in other than a static scene. The spot metering on the camera is often does not have a sufficiently narrow angle of view.

With today's high performance sensors such as the Sony Exmoor, ETTR is less important than it used to be and it is often more expeditious to use matrix metering and adjust the exposure if the camera histogram is too far to the left. However, bit of highlight headroom is sometimes advisable and one can make corrections with the raw converter. Remember that SNR varies with the square root of the exposure, so underexposing by one whole stop (not advisable in most situations) reduces the SNR by only a factor of sqrt(2).

Bill
Title: Re: ETTR vs ISO
Post by: RFPhotography on May 24, 2013, 06:50:12 PM
ISO is part of exposure.  Emil hints at in in his response linked above.  He says that the idea is to maximise exposure within the constraints of DOF and the shutter speed necessary for the situation.  Depending on the shooting circumstances, ISO has to be taken into consideration and, thus, is part of exposure.  At least insofar as the response of the sensor is concerned.  But also with respect to the needs of the photographer and the ability to post-process.

Shooting, for example, fast moving sports like hockey or motor racing requires faster shutter speeds to get the necessary action-stopping movement.  PJs shooting in these conditions don't have the luxury of time to fiddle with sliders in post.  They shoot and fire the images off to their publication or agency in order to get the pics in the public eye as quickly as possible.  A little extra noise isn't of significant concern so hiking up the ISO to get the required shutter speed for a 'normal' exposure is what will be done; as opposed to shooting at base ISO and screwing around in post while the competition is getting images onto the web for distribution or the competing newspaper is getting a jump on a story.  This doesn't apply just to sports, of course, but any situation where conditions merit cranking the ISO to get a 'normal' exposure.
Title: Re: ETTR vs ISO
Post by: jrsforums on May 24, 2013, 07:10:44 PM
ISO is part of exposure.  Emil hints at in in his response linked above.  He says that the idea is to maximise exposure within the constraints of DOF and the shutter speed necessary for the situation.  Depending on the shooting circumstances, ISO has to be taken into consideration and, thus, is part of exposure.  At least insofar as the response of the sensor is concerned.  But also with respect to the needs of the photographer and the ability to post-process.

Shooting, for example, fast moving sports like hockey or motor racing requires faster shutter speeds to get the necessary action-stopping movement.  PJs shooting in these conditions don't have the luxury of time to fiddle with sliders in post.  They shoot and fire the images off to their publication or agency in order to get the pics in the public eye as quickly as possible.  A little extra noise isn't of significant concern so hiking up the ISO to get the required shutter speed for a 'normal' exposure is what will be done; as opposed to shooting at base ISO and screwing around in post while the competition is getting images onto the web for distribution or the competing newspaper is getting a jump on a story.  This doesn't apply just to sports, of course, but any situation where conditions merit cranking the ISO to get a 'normal' exposure.

Please reread Emil's quote on the mantra "Maximize Exposure" in Reply#4 above.  I think it takes in all the conditions you mention.  If ISO must be raised to get the aperture and shutter speed you need, that's it....they are the first consideration.

...but exposure, per Emil, is the amount of light reaching the sensor.  Increasing ISO does not increase exposure, it increase the amplification of the signal.
Title: Re: ETTR vs ISO
Post by: digitaldog on May 24, 2013, 07:19:58 PM
ISO is part of exposure. 
Agreed, has to be (part).

bjanes:
Quote
ISO does have a role, but as Emil points out here, ISO is not exposure.
So if I set exposure then ISO, no change? Doesn't seem that way. It plays a role, always has. Let's put it this way, exposure and ISO have to both be under equal consideration. Or... exposure without any recognition of ISO is hard!

Quote
However, the term ETTR does imply the use of a histogram and this leads to confusion.
Seems that way if we are to interpret ETTR to mean Expose to the Right of the incorrect camera Histogram. I'd prefer we call it ER (Expose Right). Or ERFR (Expose Right For Raw). IF we had a true raw histogram we'd be foolish to ETTR. But my point is we don't need a histogram, we never did to expose right. It be great if we could use one that actually showed us the real data. I guess I'm old school. I've spent more years properly exposing film than shooting digital and old habits die hard. I question why I'd look at a Histogram that's flat out wrong, then expose to it's right instead of testing the sensor and exposure/development then get the correct exposure without looking at incorrect information.

If I could start over again, I'd prefer not to use ETTR but rather tell people that for raw capture, the current histogram is a hack. Use tools and techniques to expose right just as we did prior to the first camera histogram. Just as we did with film after testing. IOW, base ISO 100 for JPEG may very well be IOS 25 and it is easy to adjust for this when exposing. Just as I did with the color neg film (san's histogram <g>).

Quote
Nevertheless, a properly interpreted histogram is helpful at base ISO.
Agreed, now all we need is the properly displayed histogram because interpretation on the one provided isn't close to an exact science.

Quote
With today's high performance sensors such as the Sony Exmoor, ETTR is less important than it used to be
I'm sure that's true but I'd hate to suggest that due to this, we can and should be sloppy with exposure. Wasn't the case years ago when we shot transparency although we had some tricks: Snip tests, push/pull processing (more than ˝ a stop, you messed up), Polaroids (and back to interpretation), fill an A12 back with a roll and shoot 12 test shots, no snip tests etc. I suppose with digital, we can be far more off on exposure unlike chrome but that isn't ideal. But if you're shooting street scenes or animals at the zoo, getting a half decent exposure is fine. In a studio setup or controlled situations, I'd prefer to nail exposure because there's no reason not to.
Title: Re: ETTR vs ISO
Post by: digitaldog on May 24, 2013, 07:26:24 PM
Please reread Emil's quote on the mantra "Maximize Exposure" in Reply#4 above.  I think it takes in all the conditions you mention.  If ISO must be raised to get the aperture and shutter speed you need, that's it....they are the first consideration.

Absolutely agree! Lower ISO to reduce noise that add's camera shake isn't a good move! Just as in the old days, if we expected a lot of light, got the scene and found we had to replace the ISO 100 transparency for ISO 400 (despite more grain), we did so. This again is all photography 101. We're a bit spoiled that we can dial in any ISO, capture by capture as we wish. A noise free but blurry image isn't something most people desire. Sharp and a bit of grain, far better IMHO.

Quote
...but exposure, per Emil, is the amount of light reaching the sensor.  Increasing ISO does not increase exposure, it increase the amplification of the signal.

The net results are what's important. The ISO on a camera and the ISO rated on film (which can be wrong) should not be considered 100% identical or on parity should they? What if instead of calling the dial ISO it was named something else, would not altering the settings alter the results (and exposure)? More semantics. Can we really separate ISO from exposure when making a capture? If I use ISO 100 film and push it 1 stop, but shoot another roll at ISO 100 and process it normal, don't we expect to see a difference in the two rolls of film (all else being equal)? Same amount of light hit the film.
Title: Re: ETTR vs ISO
Post by: jrsforums on May 24, 2013, 08:12:17 PM
.

The net results are what's important. The ISO on a camera and the ISO rated on film (which can be wrong) should not be considered 100% identical or on parity should they? What if instead of calling the dial ISO it was named something else, would not altering the settings alter the results (and exposure)? More semantics. Can we really separate ISO from exposure when making a capture? If I use ISO 100 film and push it 1 stop, but shoot another roll at ISO 100 and process it normal, don't we expect to see a difference in the two rolls of film (all else being equal)? Same amount of light hit the film.

The net results are important....but you have to understand the perspective

For a photographer, the exposure net result includes the consideration of ISO.

For the sensor scientist, the exposure net result is ONLY the light hitting the sensor, which does not include ISO. 

This is important to understand for CCD and Exmor where the sensors are called ISOless, because you will not get any result from increasing ISO in camera than you can get in the RAW conversion.  In these cases, ISO for the photographer becomes the same as the sensor scientist...meaningless...except for the jpeg he sees in the LCD on the back of the camera.

Title: Re: ETTR vs ISO
Post by: digitaldog on May 24, 2013, 08:15:31 PM
For a photographer, the exposure net result includes the consideration of ISO.
Agreed!

Quote
For the sensor scientist, the exposure net result is ONLY the light hitting the sensor, which does not include ISO.  
I'm not a scientist (sensor or otherwise), don't even play one on TV. ;-)

Quote
This is important to understand for CCD and Exmor where the sensors are called ISOless, because you will not get any result from increasing ISO in camera than you can get in the RAW conversion.  In these cases, ISO for the photographer becomes the same as the sensor scientist...meaningless...except for the jpeg he sees in the LCD on the back of the camera.
Valid and a necessary consideration for such devices. All my cameras don't work that way. ISO plays a role as I illustrated. Now why do such systems even have an ISO setting?
Title: Re: ETTR vs ISO
Post by: jrsforums on May 24, 2013, 08:22:54 PM
Now why do such systems even have an ISO setting?

Because you and I would yell and scream if it didn't.  :-)

Not to mention the chimpers of the world...which often includes me.
Title: Re: ETTR vs ISO
Post by: jrsforums on May 24, 2013, 08:39:29 PM

With today's high performance sensors such as the Sony Exmoor, ETTR is less important than it used to be and it is often more expeditious to use matrix metering and adjust the exposure if the camera histogram is too far to the left. However, bit of highlight headroom is sometimes advisable and one can make corrections with the raw converter. Remember that SNR varies with the square root of the exposure, so underexposing by one whole stop (not advisable in most situations) reduces the SNR by only a factor of sqrt(2).

Bill

Bill,

Why is ETTR (Optimizing/Maximizing Exposure) less important? 

Is it the greater dynamic range from the new sensors giving us greater "slop" range?  (I don't really mean that as slanted as it sounds.)

I understand that there are different conditions we shoot under...and there are time where it is more important to get the shot then it is to maximize the dynamic range.  But there are still times where we want (need) to maximize it.  How has that changed?
Title: Re: ETTR vs ISO
Post by: Jim Kasson on May 24, 2013, 08:40:32 PM
Now why do such systems even have an ISO setting?

1) To make the jpegs come out right.
2) So you can chimp in dim light.
3) tradition
4) for the amusement of the photographers.

Jim
Title: Re: ETTR vs ISO
Post by: RFPhotography on May 24, 2013, 09:20:14 PM
For a photographer, the exposure net result includes the consideration of ISO.


This is a photography discussion board and despite the heavy dose of scientific discussion on this board, I think many of are photographers so those practical considerations are important.
Title: Re: ETTR vs ISO
Post by: jrsforums on May 24, 2013, 09:35:45 PM
This is a photography discussion board and despite the heavy dose of scientific discussion on this board, I think many of are photographers so those practical considerations are important.

Then I think you should include the rest of the quote...

This is important to understand for CCD and Exmor where the sensors are called ISOless, because you will not get any result from increasing ISO in camera than you can get in the RAW conversion.  In these cases, ISO for the photographer becomes the same as the sensor scientist...meaningless...except for the jpeg he sees in the LCD on the back of the camera.
Title: Re: ETTR vs ISO
Post by: digitaldog on May 24, 2013, 10:00:17 PM
1) To make the jpegs come out right.
2) So you can chimp in dim light.
3) tradition
4) for the amusement of the photographers.

IOW, for this audience who I assume captures raw data, nothing useful. OK, I do like a little amusement.
Title: Re: ETTR vs ISO
Post by: RFPhotography on May 25, 2013, 07:05:21 AM
Then I think you should include the rest of the quote...

This is important to understand for CCD and Exmor where the sensors are called ISOless, because you will not get any result from increasing ISO in camera than you can get in the RAW conversion.  In these cases, ISO for the photographer becomes the same as the sensor scientist...meaningless...except for the jpeg he sees in the LCD on the back of the camera.

Yes, I understand that.  But you miss my point.  I'll reiterate.  There are circumstances when screwing around processing images isn't an option.  In such circumstances raising the ISO to generate a 'normal' exposure is necessary.  In those circumstances ISO is part of exposure. 
Title: Re: ETTR vs ISO
Post by: jrsforums on May 25, 2013, 09:25:24 AM
Yes, I understand that.  But you miss my point.  I'll reiterate.  There are circumstances when screwing around processing images isn't an option.  In such circumstances raising the ISO to generate a 'normal' exposure is necessary.  In those circumstances ISO is part of exposure. 

OK...didn't read that in your comment, as I thought it was implicit in the original statement, "For a photographer, the exposure net result includes the consideration of ISO", as well as Bill Janes post.
Title: Re: ETTR vs ISO
Post by: bjanes on May 25, 2013, 11:38:54 AM
This entire ETTR, exposure, ISO, testing takes me back to the mid 1980's while at photo school. I think 2nd trimester, we had one of our first assignments shooting color (first trimester we only shot 4x5 B&W film, it could not be burned or dodged in the darkroom for any assignment). This new color assignment was to shoot 4x5 color neg, I think it was Kodak VPS II or something. Stated ISO 160 on the box. The subject had to be stationary and have an 18% gray card in the center filling up at least 20%. We were to shoot at the recommended ISO, then 1, 2 and 3 stops under and over. We processed all the film together. Then in the darkroom, we had to make prints whereby the gray card on the print matched the actual gray card. Once all 7 prints were mounted and viewed, the results were very surprising! The plus 1 and 2 stops were easily printable and both produced a better print quality than the "normal" ISO 160 image. The under exposed prints all suffered compared to the 'normal' exposure.

ISO 160 for VPS worked just fine! But rating it at ISO 80 (or 40) and treating the rest of the process the same produced a very visible and beneficial result! No Histograms. But we did expose to the right if you will.

Historical perspective is always helpful and many concepts learned earlier can be carried over into digital photography. However, the response of negative film and the digital sensor are quite different. Negative film has a toe on the H&D curve and exposures that do not place shadows above start of the toe will result in blank film base and all shadow detail will be lost. From my film days when using negative film, I had a sinking feeling when the shadows recorded as base density on an important shot that could not be recovered. Digital is linear and any shadow luminance will be recorded but any image detail may be lost in noise. The digital sensor response is linear without any shoulder and highlights will be clipped abruptly when the sensor saturates. Some highlight recovery may be possible with Bayer CFA sensors, since the red and blue channels may still have some detail, but the color information in the green channel that clips first (with daylight and most artificial illumination) will be lost.

The H&D curve has a shoulder that rolls off gradually and negative film is resistant to overexposure for this reason. With ETTR, one exposes for the highlights, placing then just short of clipping, and overall tonality can be brought back in post processing. What about negative film: the usual advice is to expose for the shadows and develop for the highlights, the exact opposite of digital.

Going back to your experiment with VPS the effect of exposure will depend on what you have surrounding your gray card. This is discussed in Ansel's The Negative, page 67 in my 1981 edition. For what he describes as a short-scale subject, one can place the important shadows on Zone III and the highlights will fall on Zone VI and less than the full range of tones will be exposed. This gives the option of placing the shadows higher and getting better shadow definition. Ansel states: "...Therefore, in the absence of compelling reason to give the additional one stop of exposure, use the Zone III placement of the shadow area; the resulting negative will have less grain and higher acutance since there is no 'wasted' density, and will still have full detail in the important areas." He then proceeds to talk about development and printing procedures to give the desired tonalities in the print and the handling of full scale subjects where there is less "exposure latitude".

This is the exact opposite of your findings with VPS, and we would call Ansel's exposure in this case as to the "Left". In his exposure scales, he placed Zone I to the left just as we do with histograms in the digital era. The concepts have not changed: one exposes to get optimal image quality, taking into account the characteristics of the medium.

Regards,

Bill
Title: Re: ETTR vs ISO
Post by: RFPhotography on May 25, 2013, 11:52:21 AM
OK...didn't read that in your comment, as I thought it was implicit in the original statement, "For a photographer, the exposure net result includes the consideration of ISO", as well as Bill Janes post.

I pay virtually no attention to anything Janes writes.  And yes, 'implicit'.  As in implied.  Which is what I said originally when I noted 'Emil hints at it'.  All I doing was trying to make it clearer. 
Title: Re: ETTR vs ISO
Post by: bjanes on May 25, 2013, 11:53:08 AM
What's it doing to reduce the noise?

Good to know that at least some are agreeing that this process isn't increasing the photon count, I've been scratching my head trying to understand what the ISO on the Canon is doing to reduce the noise. Not implementing ETTR and setting the Canon to ISO 100 certainly doesn't produce the lowest noise in a capture. Why?

I can understand the argument that exposure is Aperture + Shutter but then ISO, at least in this case, plays a role. Do we need a separate acronym for Canon cameras whereby we increase ISO, don't adjust 'exposure' based on the increase and end up with less noise?

The answer for your Canon camera is discussed by Emil here (http://theory.uchicago.edu/~ejm/pix/20d/tests/noise/noise-p3.html#ETTR) in Figure 12a and the accompaning text:

"There are several ways to interpret this graph that are useful to keep in mind for making exposure choices. If one has the option to lower the ISO and the shutter speed (or widen the aperture), the highest S/N for the image is obtained by increasing the exposure, pushing the right end of the histogram right up to the upper edge of the range of exposure on the horizontal axis. This is the usual ETTR philosophy. Lowering the ISO one stop pushes the upper end of the dynamic range one stop to the right in absolute exposure, and pushing the histogram to the right climbs the rising S/N curve to better overall image quality.

If on the other hand, one is limited by the subject matter (freezing motion, depth of field requirements, etc) to a given maximum EV, then it makes sense to raise the ISO to pull the top end of the camera's dynamic range down to the top end of the histogram; this has little benefit at that upper end, since all the curves are on top of one another in that regime. Nevertheless it improves image quality by raising the S/N ratio on the shadow end of the curves."


In the latter case when one is limited by the subject matter, one gives the required exposure in terms of f/stop and shutter speed, and raises the ISO. In this case exposure (measured in lux-seconds) is fixed at the required level.

Bill
Title: Re: ETTR vs ISO
Post by: bjanes on May 25, 2013, 12:01:49 PM
I pay virtually no attention to anything Janes writes.  And yes, 'implicit'.  As in implied.  Which is what I said originally when I noted 'Emil hints at it'.  All I doing was trying to make it clearer. 

I usually ignore your posts too, since any reasonable discussion with you is fruitless. You should brush up on basic definitions. Exposure (http://en.wikipedia.org/wiki/Exposure_%28photography%29) is measured in lux seconds and is independent of the ISO. ISO (film speed on see Wikipedia (https://en.wikipedia.org/wiki/Film_speed)) does affect our exposure decisions, since it describes how the sensor responds to exposure in terms of saturation.

Regards
Title: Re: ETTR vs ISO
Post by: Jim Kasson on May 25, 2013, 12:05:26 PM
Digital is linear and any shadow luminance will be recorded but any image detail may be lost in noise. The digital sensor response is linear without any shoulder and highlights will be clipped abruptly when the sensor saturates. Some highlight recovery may be possible with Bayer CFA sensors, since the red and blue channels may still have some detail, but the color information in the green channel that clips first (with daylight and most artificial illumination) will be lost.

[snip]

The concepts have not changed: one exposes to get optimal image quality, taking into account the characteristics of the medium.

Well said, Bill -- as usual. Thanks.

There's something else that hasn't changed. Photographers can use knowledge of how their tools work to make better images, as you are proposing. They can also become fixated on technical minutia to a degree that it interferes with, or even prevents, making better images (http://blog.kasson.com/?p=9).

Jim

Title: Re: ETTR vs ISO
Post by: digitaldog on May 25, 2013, 12:27:58 PM
Historical perspective is always helpful and many concepts learned earlier can be carried over into digital photography.
That's why I posted it <g>. The point was that ideally, photographers test their processes. Digital doesn't change that a lick.

Quote
However, the response of negative film and the digital sensor are quite different.
Of course it is (hence more reason to test). I'm well aware of H&D curves versus linear capture.

Quote
Going back to your experiment with VPS the effect of exposure will depend on what you have surrounding your gray card.
I'm going back nearly 30 years but I don't recall that being an issue when this one class of 30-ish students presented their work. I can't recall any presented assignment where the decrease in of ISO 160 didn't produce better results.

Quote
This is the exact opposite of your findings with VPS, and we would call Ansel's exposure in this case as to the "Left".
OK, but it doesn’t jive with the results from this class and I'm pretty sure, this assignment was given for years and years to other's. Maybe their results were different, I don't know. But again, the entire point was: test the process. Don't assume the ISO setting is ideal. Learn to expose for ideal (in this case) color neg. Digital doesn't change the need to understand and properly control exposure + development.
Title: Re: ETTR vs ISO
Post by: RFPhotography on May 25, 2013, 12:35:17 PM
I usually ignore your posts too, since any reasonable discussion with you is fruitless. You should brush up on basic definitions. Exposure (http://en.wikipedia.org/wiki/Exposure_%28photography%29) is measured in lux seconds and is independent of the ISO. ISO (film speed on see Wikipedia (https://en.wikipedia.org/wiki/Film_speed)) does affect our exposure decisions, since it describes how the sensor responds to exposure in terms of saturation.

Regards

I feel exactly the same way and the rest of your statement is a perfect example of why. You are simply too caught up in the science and pedantry of dictionary definitions to make any discussion of practical photography possible.  To borrow fro Jim's blog essay, you spend too much time sharpening pencils.
Title: Re: ETTR vs ISO
Post by: jrsforums on May 25, 2013, 12:43:20 PM
I feel exactly the same way and the rest of your statement is a perfect example of why. You are simply too caught up in the science and pedantry of dictionary definitions to make any discussion of practical photography possible.  To borrow fro Jim's blog essay, you spend too much time sharpening pencils.

Bob, in my opinion, understanding the definition of terms used....and by whom...is very important.  This is why I bothered to post the difference in how photographers use 'exposure' and how "sensor scientist" (the actual definition) use it.

If we do not all agree on the definition of the terms we are trying to discuss, we just talk past each other....which is what often happens on these forums.
Title: Re: ETTR vs ISO
Post by: digitaldog on May 25, 2013, 02:41:33 PM
Exposure (http://en.wikipedia.org/wiki/Exposure_%28photography%29) is measured in lux seconds and is independent of the ISO. ISO (film speed on see Wikipedia (https://en.wikipedia.org/wiki/Film_speed)) does affect our exposure decisions, since it describes how the sensor responds to exposure in terms of saturation.

That same page goes on to say:

Quote
An appropriate exposure for a photograph is determined by the sensitivity of the medium used. For photographic film, sensitivity is referred to as film speed and is measured on a scale published by the International Organization for Standardization (ISO). Faster film, that is, film with a higher ISO rating, requires less exposure to make a good image

Again, maybe semantics, but the sentence above seems to link ISO and exposure together. Or as I said earlier, how can one expose properly, even within the ballpark without taking ISO into account?
Title: Re: ETTR vs ISO
Post by: ErikKaffehr on May 25, 2013, 03:55:35 PM
Hi,

Getting back to the original question, ETTR is about utilizing the capability of the sensor. We want to have maximum exposure possible as this minimimises the noise in the image. Setting a higher ISO means underexposure, that is some of the capacity of the sensor is thrown away. What we really discuss is if it it better to underexpose at nomina ISO or set a higher ISO and let camera electronics play a few less or more dirty tricks.

Best regards
Erik

That same page goes on to say:

Again, maybe semantics, but the sentence above seems to link ISO and exposure together. Or as I said earlier, how can one expose properly, even within the ballpark without taking ISO into account?
Title: Re: ETTR vs ISO
Post by: RFPhotography on May 25, 2013, 04:26:56 PM
Bob, in my opinion, understanding the definition of terms used....and by whom...is very important.  This is why I bothered to post the difference in how photographers use 'exposure' and how "sensor scientist" (the actual definition) use it.

If we do not all agree on the definition of the terms we are trying to discuss, we just talk past each other....which is what often happens on these forums.

I don't disagree.  However, it's also possible to get too caught up in the precision of dictionary definitions and in doing so give too short shrift to, let's call them, working definitions.  It's sort of like book learning vs practical experience.  Book knowledge is important but what many learn early on in their working lives is that what's in the books bears little resemblance to how things actually work.  Ask many lawyers, for example, and they'll tell you that what they learned in law school has little do do with the practice of law.
Title: Re: ETTR vs ISO
Post by: bjanes on May 25, 2013, 04:36:49 PM
That same page goes on to say:

Again, maybe semantics, but the sentence above seems to link ISO and exposure together. Or as I said earlier, how can one expose properly, even within the ballpark without taking ISO into account?

I agree 100% that ISO must be taken into account since the ISO rating of the sensor determines how it responds to a given exposure (in lux seconds or a given shutter speed and aperture). That was what I was trying to convey to Bob Fisher in post 65 above, but he ignores me as well as generally accepted definitions. Without an ISO rating the light meter would be useless. The term "exposure" is used by some in a less restrive manner and to them merely taking the picture is an exposure.

I agree that testing is also absolutely necessary just as in Ansel's time. The ISO saturation standard allows 0.5 EV of headroom and if you are exposing to the right, you may not want to allow that. Those who use the camera histogram for exposure decisions must test to determine how much headroom it allows. For my Nikons it is about 0.5 EV. I don't shoot Phase One MF, but I understand that its sensors are rated to allow even more headroom.

Bill
Title: Re: ETTR vs ISO
Post by: bjanes on May 25, 2013, 04:46:12 PM
Well said, Bill -- as usual. Thanks.

There's something else that hasn't changed. Photographers can use knowledge of how their tools work to make better images, as you are proposing. They can also become fixated on technical minutia to a degree that it interferes with, or even prevents, making better images (http://blog.kasson.com/?p=9).

Jim

Jim, masterfully stated, a compliment countered with constructive criticism. Too bad we sometimes do not have more in this vein on the forum. :(

I hope I don't spend too much time sharpening my pencils. :)

Bill
Title: Re: ETTR vs ISO
Post by: Jim Kasson on May 25, 2013, 05:08:54 PM
Jim, masterfully stated, a compliment countered with constructive criticism.

Bill, I didn't mean it as any kind of criticism. I don't think that shoe fits you.

Jim
Title: Re: ETTR vs ISO
Post by: jrsforums on May 25, 2013, 05:31:39 PM
I don't disagree.  However, it's also possible to get too caught up in the precision of dictionary definitions and in doing so give too short shrift to, let's call them, working definitions.  It's sort of like book learning vs practical experience.  Book knowledge is important but what many learn early on in their working lives is that what's in the books bears little resemblance to how things actually work.  Ask many lawyers, for example, and they'll tell you that what they learned in law school has little do do with the practice of law.

I think you totally missed my point or just choose to ignore it.

There is a difference between book learning (definitions) and practical experience (how you best apply what you learn).

However, you can not use practical learning to redefine terms.....unless you carefully ensure that you understand the (book) term and how you have modified it.  It does not allow you to reject the book learned definitions.

To use your analogy, if a lawyer is using case law, he would not state that it meant something that was not documented.  However, after stating what the case law "defined", he might state how it could be "interpreted" to mean something different...and used that way for the current case he was working on.

Back to what we have been talking about....if one person talks about exposure...and defines it not to include ISO...and the other talks about exposure...but believes it includes ISO...and they do not make sure about this when incuding it in there points, there will be confusion and much angst which would not happen if they carefully used the terms....as shown in this thread.
Title: Re: ETTR vs ISO
Post by: RFPhotography on May 25, 2013, 06:46:14 PM
I didn't miss your point, John nor did I choose to ignore it.  Look, forget anything I've said.  Clearly I have no credibility here and frankly, I don't care.  But another, highly regarded photographer, author and speaker participating in this thread also feels that ISO is part of exposure in some manner and in some circumstances.  So I'm not alone in my position.  I understand the dictionary definition.  I understand the technical points.   But from a practical, on the ground standpoint, ISO plays a part in determining exposure.  Even Emil 'intimates' that in the article of his that has been referenced several times; as you and I have both agreed.

You're saying the same thing I am with your legal analogy.  What's in the book is open to interpretation.  No different here.

I'm not rejecting the book learning.  Not by any means.  All I'm saying is that the book knowledge isn't always completely relevant in real world conditions.  That's the only point I'm making.  Test bench jockeys won't understand that.  Photographers will.  The 'angst' in this thread has, seemingly, been a result of the complete repudiation by some of the idea that, in some cases, 'exposure' will include ISO.  I think the people who support the idea that exposure 'can' include a consideration of ISO have been abundantly clear in their position.
Title: Re: ETTR vs ISO
Post by: digitaldog on May 25, 2013, 07:27:53 PM
Quote
We want to have maximum exposure possible as this minimimises the noise in the image.

Agreed!

Quote
Setting a higher ISO means underexposure
Expect in the example I posted from the Canon <g>.

Quote
What we really discuss is if it it better to underexpose at nomina ISO or set a higher ISO and let camera electronics play a few less or more dirty tricks.
Seems to depend on the camera and sensor. And takes us back to the need to test the process. I'm sure everyone in this conversations understands you'd treat a camera JPEG differently than a raw in terms of exposure. I would suspect we came to this understanding by testing ETTR while capturing a JPEG.

It does concern me that a Canon for example deals with ISO settings differently than other camera systems. But at due to some testing, I at least expect this and can say the advise that raising ISO always increases noise isn't always true.
Title: Re: ETTR vs ISO
Post by: Ray on May 25, 2013, 09:05:50 PM
I have to say that my interest in photographic matters leans strongly towards the practical application of the theory, which is why I frequently carry out my own tests on my own equipment, not necessarilly to prove or disprove the theoretical claims that are being made, but to find out how significant such claims for particular technical qualities may be in practice, and in what circumstances they may be of some advantage to my efforts in producing an image or print to my own satisfaction.

However, in order to make such tests one really has to understand clearly what it is one is testing, and therefore a precise definition of the fundamental concepts related to the features one is testing, is necessary.

I remember well the occasion when this phenomenon of improved image quality resulting from the use of an increased ISO first came to my attention. It was on the old Rob Galbraith photographic forum around the year 2005/6. A poster by the name of John Sheehy made a comment to the effect, "It is better to use an exposure at ISO 200 than ISO 100 because you get better shadow detail."

Now, being the sort of guy who refuses to accept anything that doesn't make sense, I chimed in and disputed his claim. How can an exposure at ISO 200 possibly be better than an exposure at ISO 100? At ISO 100 the sensor receives double the amount of light. SNR is bound to be better.

John quickly dispelled my confusion by explaining that he was referring to circumstances where the same exposure was used at both ISO settings. My misunderstanding was due to my conflating ISO settings with actual exposure, which is quite understandable because we always have to either deliberately choose a specific ISO everytime we take a shot, or allow the camera to automatically choose it for us.

In order to avoid such confusion, I think it is better to consider exposure as something which is determined only by F/stop, shutter speed, and that quality we often ignore, T/stop, which relates to the transmissive qualities of the glass in the lens.

The ISO setting is merely an instruction to the camera's internal processing electronics to process an exposure in a particular way.
Title: Re: ETTR vs ISO
Post by: Jack Hogan on May 26, 2013, 05:59:58 AM
the same exposure was used at both ISO settings.

In order to avoid such confusion, I think it is better to consider exposure as something which is determined only by F/stop, shutter speed

The ISO setting is merely an instruction to the camera's internal processing electronics to process an exposure in a particular way.

Well said Ray.  'Processing' in the sense of amplifying the signal rolling off the sensor before writing it to the Raw data (making it 'brighter').  Amplification can be done analogically and/or digitally.  If done digitally, there is no difference whether the signal is increased in-camera (through a higher ISO) or later in the Raw converter (through a +'EV' correction) - so might as well keep the lower ISO and the extra Dynamic Range and correct brightenss later in the raw converter.  For instance the D800e applies only digital amplification from about 1600 ISO on up as can be seen in these Raw histograms:

(http://3.static.img-dpreview.com/files/g/TS1600x1600~2558720.jpg)

The ISO 800 Raw histogram of the green channel is fully populated, the ISO 1600 is showing a small amount of digital amplification, the 3200 and 6400 are simply the ISO 1600ish ADC integer output multiplied digitally by 2 and 4 - with every second and every fourth value present respectively.  There is no difference between doing the multiplication in-camera or later in the raw converter through +EC correction, except that in-camera one may clip desirable highlights which one may regret later.  The Raw converter needs to be well behaved for extreme (>3 stops?) recoveries to work as expected (RT apparently is while LR apparently is not).

That's digital amplification.  What about the analog component, that is when increasing ISO will amplify analogically the signal rolling off the sensor before feeding it to the ADC (think of it as turning up the volume on your stereo)?  Some cameras' shadow SNR, in particular those whose ADC's noise floor is higher than the sensor's (e.g. Canons), will benefit from boosting the signal from the sensor before it is fed to the ADC, so in their case it may be worth compromising DR in exchange for better shadow SNR.

This suggests the following strategy, as mentioned by emil and others: First max out Exposure at base ISO by choosing the longest shutter speed and widest aperture allowed by your technical/artistic constraints, thereby gathering the largest number of photons possible for a given blur, DOF and highlight retention objective.  Then you may choose to raise ISO as a compromise between minimizing shadow SNR and maximizing DR.  Bill Claff has worked out where increasing ISO no longer helps shadow SNR for various cameras (http://home.comcast.net/~NikonD70/Charts/PDR_Shadow.htm#D800e,EOS 5D Mark III).  No point increasing ISO beyond 1600 for the D800e (really ISO 400 if one is willing to live with a 1/6 of a stop penalty in shadow SNR for the extra two stops of DR), although if you own a 5DIII you may not want to stop before ISO 2500 or higher - if needed.  If the above strategy for the D800e results in a darker image OOC than usual, no sweat: simply correct for pleasing brightness/tones in the raw converter, confident that  you have the best IQ possible.

Once one realizes that Exposure is all about getting the most photons possible given one's technical/artistic constraints, and ISO is an independent variable about processing those photons into a raw file while maximizing SNR and DR without clipping desirable highlights one is ready to go.

Jack 
PS Ray, I sent you a PM on a separate matter.
Title: Re: ETTR vs ISO
Post by: dreed on May 26, 2013, 08:55:15 AM
Well said Ray.  'Processing' in the sense of amplifying the signal rolling off the sensor before writing it to the Raw data (making it 'brighter').  Amplification can be done analogically and/or digitally.  If done digitally, there is no difference whether the signal is increased in-camera (through a higher ISO) or later in the Raw converter (through a +'EV' correction) - so might as well keep the lower ISO and the extra Dynamic Range and correct brightenss later in the raw converter.  For instance the D800e applies only digital amplification from about 1600 ISO on up as can be seen in these Raw histograms:

Do you know if that graph has been formulated for cameras other than the D800?

Quote
The ISO 800 Raw histogram of the green channel is fully populated, the ISO 1600 is showing a small amount of digital amplification, the 3200 and 6400 are simply the ISO 1600ish ADC integer output multiplied digitally by 2 and 4 - with every second and every fourth value present respectively.

This I understand.

Quote
That's digital amplification.  What about the analog component, that is when increasing ISO will amplify analogically the signal rolling off the sensor before feeding it to the ADC (think of it as turning up the volume on your stereo)?  Some cameras' shadow SNR, in particular those whose ADC's noise floor is higher than the sensor's (e.g. Canons), will benefit from boosting the signal from the sensor before it is fed to the ADC, so in their case it may be worth compromising DR in exchange for better shadow SNR.

This I also understand.

Quote
This suggests the following strategy, as mentioned by emil and others: First max out Exposure at base ISO by choosing the longest shutter speed and widest aperture allowed by your technical/artistic constraints, thereby gathering the largest number of photons possible for a given blur, DOF and highlight retention objective.  Then you may choose to raise ISO as a compromise between minimizing shadow SNR and maximizing DR.  Bill Claff has worked out where increasing ISO no longer helps shadow SNR for various cameras (http://home.comcast.net/~NikonD70/Charts/PDR_Shadow.htm#D800e,EOS 5D Mark III).  No point increasing ISO beyond 1600 for the D800e (really ISO 400 if one is willing to live with a 1/6 of a stop penalty in shadow SNR for the extra two stops of DR), although if you own a 5DIII you may not want to stop before ISO 2500 or higher - if needed.  If the above strategy for the D800e results in a darker image OOC than usual, no sweat: simply correct for pleasing brightness/tones in the raw converter, confident that  you have the best IQ possible.

However I'm not sure I understand how to read the graph?
Title: Re: ETTR vs ISO
Post by: RFPhotography on May 26, 2013, 09:10:04 AM
Dreed, the article linked at the bottom of the page Jack links titled 'Sensor Analysis Primer - Photographic Dynamic Range Shadow Improvement' has the information on how to read the chart.  Basically, where the curve flattens out is the point where increasing ISO no longer has any benefit on shadow dynamic range.
Title: Re: ETTR vs ISO
Post by: Vladimirovich on May 26, 2013, 11:21:39 AM
Amplification can be done analogically and/or digitally. 
or there might be no amplification at all - just a tag (as a hint/instruction for a raw converter) in raw file.
Title: Re: ETTR vs ISO
Post by: Jack Hogan on May 26, 2013, 11:35:29 AM
Do you know if that graph has been formulated for cameras other than the D800?

Dreed, you can take a peak at your own Raw files thanks to the histogram function in free RawDigger (http://www.rawdigger.com), which is what I did.  Look at the first couple of hundred Raw values to spot the pattern, if any.

Vladimirovich, that would be a smart way to handle the digital component of ISO brightening.  Do you know any current DSLR that does that?

Jack
Title: Re: ETTR vs ISO
Post by: RFPhotography on May 26, 2013, 11:51:05 AM
or there might be no amplification at all - just a tag (as a hint/instruction for a raw converter) in raw file.

I can see that working if the OEM conversion utility were used. But it may not work with third party converters that don't pick all the tags.
Title: ETTR vs ISO: agreeing on the meaning of terms before we debate?
Post by: BJL on May 26, 2013, 11:57:47 AM
In order to avoid such confusion, I think it is better to consider exposure as something which is determined only by F/stop, shutter speed, and that quality we often ignore, T/stop, which relates to the transmissive qualities of the glass in the lens.

The ISO setting is merely an instruction to the camera's internal processing electronics to process an exposure in a particular way.
+1
To avoid arguments that are more about language than photographic technique, let us agree to use words like "exposure" in standard agreed way (like "the amount of light reaching the sensor") and move on the the practical issues.

Unfortunately "ETTR" has acquired two different and widely used meanings, so we should distinguish between
a) maximizing the exposure received by the sensor (stopping just short of blowing highlights by overfilling any photosites).
b) maximizing the output levels of the ADC, by some combination of exposure to the sensor and analog amplification of the sensor output before anaog-to-digital conversion.
Title: Re: ETTR vs ISO: agreeing on the meaning of terms before we debate?
Post by: digitaldog on May 26, 2013, 01:20:08 PM
+1
To avoid arguments that are more about language than photographic technique, let us agree to use words like "exposure" in standard agreed way (like "the amount of light reaching the sensor") and move on the the practical issues.
Unfortunately "ETTR" has acquired two different and widely used meanings, so we should distinguish between
a) maximizing the exposure received by the sensor (stopping just short of blowing highlights by overfilling any photosites).
b) maximizing the output levels of the ADC, by some combination of exposure to the sensor and analog amplification of the sensor output before anaog-to-digital conversion.

Call it optimal exposure. I know there's been one for transparency film which you probably want to nail within ⅓ to ˝ a stop (some would argue that is too much). Using the same (and good) E6 line.

Speaking of the full process to get the image, the labs or E6 lines could differ that amount although not ideal. Kind of like this digital ISO discussion. Amplification or lack of chemistry? I stuck with one very good lab for years (A&I) and tested every bulk purchase of E6 with them, it was part of the entire 'expose film' workflow.
Title: Re: ETTR vs ISO
Post by: jrsforums on May 26, 2013, 02:03:53 PM
The Raw converter needs to be well behaved for extreme (>3 stops?) recoveries to work as expected (RT apparently is while LR apparently is not).


Jack, thanks for a great, informative post.

What makes LR poorly "behaved"?

John
Title: Re: ETTR vs ISO
Post by: Jack Hogan on May 26, 2013, 02:18:32 PM
What makes LR poorly "behaved"?

Hey John,

It (and ACR) apparently do the twist (http://dcptool.sourceforge.net/Hue%20Twists.html). 

I suggest trying a few with your intended image type to see what works for you.  You may not notice or you might like the non-neutral rendition.  Most are fine with linear brightness increases (+EC) of a few stops so I wouldn't worry about it.  If you find undesirable hue twists  or other raw-converter introduced artifacts at high ISOs I understand that RawTherapee  with the 'neutral' profile is as 'clean' as it gets, and it lets you correct up to 10 stops.

Jack
Title: Re: ETTR vs ISO
Post by: RFPhotography on May 26, 2013, 06:28:50 PM
The solution there is to make your own DNG profiles.
Title: Re: ETTR vs ISO: agreeing on the meaning of terms before we debate?
Post by: Ray on May 26, 2013, 11:06:33 PM
+1
To avoid arguments that are more about language than photographic technique, let us agree to use words like "exposure" in standard agreed way (like "the amount of light reaching the sensor") and move on the the practical issues

Glad you agree with me on this point, BJL, because we've have had a few disagreements in the past.  ;)

Another issue which I think is also important is getting a grasp on the practical significance of any differences of exposure in conjunction with differences in ISO settings.

One may understand in terms of a theoretical concept that A is better than B, but in order to determine the practical significance of such differences, one might need to do one's own experiments and comparisons, rather than just accept the theoretical model and behave accordingly.

An example would be the way one uses a Canon 50D. This camera is specified as having an ISO range from 100 to 12,800, with ISOs above 3200 being expanded. What this means is that there is nothing to be gained regards fundamental image quality, by using ISO 12,800.
If one uses the same exposure (as defined by f/stop and shutter speed) at ISO 3200, as what one would consider to be the correct, or optimal, or ETTR exposure at ISO 12,800, then the camera's LCD screen and histogram will show an apparent underexposure of 2 stops.
The image will look dark, but after appropriate processing in the RAW converter the ISO 3200 shot should have equally good shadow detail as the 12,800 shot. Or to be more precise, equally bad shadow detail.  ;D

So what are the practical consequences here, one might ask. As I see it, the ISO 12,800 shot produces a review on the camera's LCD screen which can be seen more clearly. Have you captured what you want, or do you have to take another shot?
The ISO 3200 image will likely be a bit hopeless for such assessments. The main advantage will be, there is far less risk of blowing highlights unintentionally if one uses the same exposure at ISO 3200.

Another example of the importance of conducting one's own tests, which is illustrated by the peculiarities of the 50D, is the practical significance of using the ISO 100 setting as opposed to ISO 200 on the 50D. There's no doubt that the same exposure (as defined by F/stop and shutter speed) at ISO 200 will provide lower shadow noise than you would get at ISO 100. But what happens if we compare an ETTR exposure at ISO 100 with an 'apparent' ETTR exposure at ISO 200? In these circumstances the sensor receives twice the amount of light at ISO 100. Shadow detail must surely be better.

For a whole year or more, after buying a Canon 50D, taking photos in places like museums, churches and art galleries where flash and tripods were not allowed, I sometimes struggled to hold the camera steady, often preferring to use ISO 100 with the widest aperture on the lens instead of the sharpest aperture on the lens, thus compromising resolution at least a bit, because I imagined I'd be getting better shadow detail and better SNR at ISO 100 with the 50D.

Some time later, after DXO had made its test results freely available to the public, and I was in the market for an upgraded camera, I happened to compare the performance of the Canon 50D with other more recent models, using the DXOMark graphs for ISO Sensitivity, SNR and DR etc.

I got quite a surprise when I noticed that DXO seemed to be claiming that the ISO sensitivity of the 50D at ISO 100 is the same as at ISO 200, and that as a consequence their graphs for SNR and DR etc did not even include the results for ISO 100.
They seemed to be implying that ISO 200 was the real base ISO for the 50D, and that ISO 100 is an 'expanded' ISO, as it is in the Nikon D3 and D700. Yet there is no mention of this in the Canon handbook and no menu-setting for expanding ISO 200 to ISO 100.

Of course, my immediate reaction was not to fall off my chair, but to go out and take a few shots of high-dynamic-range scenes, at ISO 100 and 200 for the purpose of comparison.

I was very surprised to discover that I could not see any difference in any significant respect at all between shots taken at an exposure of twice the shutter speed at ISO 200, compared with half the shutter speed at ISO 100.
At half the shuitter speed, at ISO 100, the sensor receives twice the number of photons. What has happened to those photons, I asked myself? Why are the highlights in the ISO 100 shots not blown when both shots at both ISO's are ETTR? If the highlights are not blown in both shots compared, why does the ISO 100 shot not have cleaner shadows? What electronic processes have Canon employed? It's as though they have created the equivalent of an electronic neutral density filter.

I'm completely mystified as to what's going on here, and a bit annoyed with myself for not discovering this situation sooner. I've got lots of photos taken in places like the Hermitage in St Petersburg, which I now realise could have been a bit sharper if I'd used either a faster shutter speed at ISO 200, or an aperture of F4 at ISO 200 instead of F2.8 at ISO 100.  >:(


Title: Re: ETTR vs ISO
Post by: Tony Jay on May 27, 2013, 12:19:59 AM
An interesting and intriguing post Ray - thanks for sharing the info.
I need to investigate this a bit - not so much for the principle but to get the low-down for the 5D mark III sensor to be able to really leverage its characteristics when shooting.

Tony Jay
Title: Re: ETTR vs ISO
Post by: Jack Hogan on May 27, 2013, 06:21:49 AM
to get the low-down for the 5D mark III sensor to be able to really leverage its characteristics when shooting.

Hi TJ,

I don't shoot Canon, but from what I can tell the 5DIII uses post-ADC digital gain only after ISO 12800, so no point raising ISO beyond that unless you need 'properly' bright images SOOC. Looking at Bill Claff's graph, it seems that you do get a benefit in shadow SNR by raising ISO in-camera from 100 all the way up to 12800, although the law of diminishing returns needs to be considered after about ISO 2500: raising it from 2500 to 12800 only nets you a 0.2 stop improvement in shadow SNR at the expense of 2.3 stops of reduced DR: a decision best evaluated each time based on the situation (blips in the curve above 12800 are suspicious, as the data up there is often 'cooked').

The other take-away from the curve is that if you are after best SNR/DR it appears that you are better off increasing ISO one stop at a time (as opposed to in 1/3 or 1/2 stop increments) in correspondence of the steps in the graph: IOW if you are sitting at ISO 300 and you feel the urge to increase ISO a little, there is no point going to 400 or 500.  Just leave it there with the benefit of slightly better DR until you feel the need to go to 600.

Jack

PS the 5DIII appears to have an effective QE of about 14% and PRNU of about 0.4%, plus a fairly noisy analog amplifier contributing about 8 ADU random noise throughout the range, so the last three bits never contain much information.

(http://i.imgur.com/ZO7DC2t.jpg)
Title: Re: ETTR vs ISO
Post by: Tony Jay on May 27, 2013, 07:10:10 AM
Thanks for the confirmation Jack.
Perusal of Bill Claff's graphs does seem to point that way.
Of course it isn't a one-way street because DR is important, to me anyway, so knowing where the optimal trade-offs are, one for the other, is where I am at.
It is interesting how many of these curves see-saw around with variations from stop to stop and which ISO's become the preferred stop-off points (pardon the pun) and the fact they are approximately one full stop apart.

It is also interesting seeing how differently Canon versus Nikon (Sony) sensors behave.
Obviously how one shoots in practice is just so different.
For me, I have never really used any Nikon cameras, late or early models, but have assisted others with their Nikon cameras.
Clearly once the fundamentals are mastered for me to be of further help would require knowing the fundamentals of those sensors since they behave so differently to my Canon cameras.

This thread is a handy reminder that there is no such thing as a 'generic' sensor anymore and we haven't even really started talking about Sigma Foveon sensors yet!
Worse yet is the fact that the next generation of cameras (sensors) that we buy may make a lot (all?) of the current recommendations (sensor for sensor of course) redundant.
Still, batting this information around helps to seat it soundly in one's brain.

Tony Jay
Title: Re: ETTR vs ISO
Post by: jrsforums on May 27, 2013, 09:14:00 AM
Jack,

What is happening "north" of 12800 on the 5D3?  (or did you answer that as "suspicious")

Also, when you speak of 2500ISO, it is really 3200 as long as you understand Canon's "game" of +/-EC, right?

Title: Re: ETTR vs ISO
Post by: Jack Hogan on May 27, 2013, 10:55:08 AM
Jack,

What is happening "north" of 12800 on the 5D3?  (or did you answer that as "suspicious")

Hi John,

To see at what in-camera ISO the 5DIII starts to use digital amplification I downloaded DPR's Studio Scene Comparison Raw files (http://www.dpreview.com/reviews/canon-eos-5d-mark-iii/28) and looked at their histograms with RawDigger (http://www.rawdigger.com).  It is clear that at ISO 12800 every value is present, at 25600 every other Raw value is missing and at 51200 three out of four are missing - so it would seem that at those ISO levels the output of the sensor is amplified analogically to ISO 12800, then fed to the ADC and the ADC's output is then multiplied by two (25600) or four (51200) before being written to the Raw data.

(http://i.imgur.com/Xw0Hnb3.jpg)

Everything out of the ADC gets multiplied by two or four, signal and noise alike, but the signal fed to the ADC is the same as at ISO 12800.  So the SNR should be pretty well the same as well, all else (including Exposure) being equal.   But per Bill's graph the SNRs don't stay the same.  They get better.  A little noise reduction before writing to the file?  That would be a fair guess :)

Quote
Also, when you speak of 2500ISO, it is really 3200 as long as you understand Canon's "game" of +/-EC, right?

Not being a Canon guy I am not sure what the game is.  But it means the ISO setting that is 1/3 of a stop below what is labelled as ISO 3200 in-camera.

Jack
Title: Re: ETTR vs ISO
Post by: Jim Kasson on May 27, 2013, 11:34:59 AM
PS the 5DIII appears to have an effective QE of about 14% and PRNU of about 0.4%, plus a fairly noisy analog amplifier contributing about 8 ADU random noise throughout the range, so the last three bits never contain much information.

Jack, I'm thinking about trying to model this camera and compare it to the D800. Can you tell me the full well capacity and the unity gain ISO? The standard deviation of the ADU random noise would be nice -- I'll assume shot noise if you have no information. I'll assume the standard deviation of the pre-amp noise is that of shot noise, and I can calculate the pre-amp mean noise from your graph once I have the unity gain ISO.

Jim
Title: Re: ETTR vs ISO
Post by: RFPhotography on May 27, 2013, 11:46:43 AM
Q:  How does one reconcile the Unity Gain ISO figures with the Photographic Dynamic Range Shadow Improvement?
Title: Re: ETTR vs ISO
Post by: jrsforums on May 27, 2013, 12:03:10 PM
Q:  How does one reconcile the Unity Gain ISO figures with the Photographic Dynamic Range Shadow Improvement?

Bob, do you have UG ISO numbers/chart for the 5D3? 

I have not seen any.  Was thnking of doing it, but have not yet found the time (famous procrastinator).
Title: Re: ETTR vs ISO
Post by: jrsforums on May 27, 2013, 12:11:50 PM
Hi John,


Not being a Canon guy I am not sure what the game is.  But it means the ISO setting that is 1/3 of a stop below what is labelled as ISO 3200 in-camera.

Jack

Thanks for the RD chart.

The game, as I understand it, is that for Canon 5D series (and other Canons...not some 1D's) they only really have true ISOs at full stops.

The 1/3 stops are "fake".  That is, if you set 160 ISO, it is really 200 ISO, but metering done as 160...therefore 1/3 stop overexposed...as in slight ETTR.  Therefore the shadows are slightly better.

For jpeg OOC, the RAW is pulled back 1/3 in the OOC post processing.  For 1/3 stop ISOs over the full stops, the image is pushed up in post.

I, personally, just set my camera to full stop ISO.  Since I shoot RAW only, I want to control whatever "ETTR" I am doing.

John
Title: Re: ETTR vs ISO
Post by: jrsforums on May 27, 2013, 12:21:03 PM
Jack, another interesting datapoint, which goes with your comment about diminishing returns at higher ISOs for the 5D3 is the DR chart from DxoMark

Title: Re: ETTR vs ISO
Post by: RFPhotography on May 27, 2013, 12:29:49 PM
Bob, do you have UG ISO numbers/chart for the 5D3? 

I have not seen any.  Was thnking of doing it, but have not yet found the time (famous procrastinator).

No, John, I don't.  I don't shoot Canon and don't have the tools to do it.  

My question more more general in nature, though.  If the two figures are different, how does one choose which to follow or which is better?  I haven't compared figures for different cameras to see if or how the two may be different.  But if the two present differing or contravening information on choosing the best max. ISO setting it would be helpful to know how to reconcile the two.
Title: Re: ETTR vs ISO
Post by: Jack Hogan on May 27, 2013, 12:42:45 PM
Jack, I'm thinking about trying to model this camera and compare it to the D800. Can you tell me the full well capacity and the unity gain ISO? The standard deviation of the ADU random noise would be nice -- I'll assume shot noise if you have no information. I'll assume the standard deviation of the pre-amp noise is that of shot noise, and I can calculate the pre-amp mean noise from your graph once I have the unity gain ISO.

Jim

Hi Jim,

Reading off the chart in the earlier post, FWC appears to be around 67500 e- and UG around 400.  I get the curve in the graph assuming 8.4 ADU of random noise contributed by the amplifier at the output of the ADC.  It's not bad but it's not a perfect fit to the calculated read noise.  If I remember correctly the 5DIII has a couple of amplifying stages, which I treated as one in order to estimate this last figure.

Jack
Title: Re: ETTR vs ISO
Post by: jrsforums on May 27, 2013, 12:47:59 PM
No, John, I don't.  I don't shoot Canon and don't have the tools to do it.  

My question more more general in nature, though.  If the two figures are different, how does one choose which to follow or which is better?  I haven't compared figures for different cameras to see if or how the two may be different.  But if the two present differing or contravening information on choosing the best max. ISO setting it would be helpful to know how to reconcile the two.

Sorry...misinterpreted you question.

Probably Jim Kasson would best to answer this.  I could try, but would only be trying to interpret info he has provided me....and would probably do it poorly :-)
Title: Re: ETTR vs ISO
Post by: Jack Hogan on May 27, 2013, 12:50:05 PM
Q:  How does one reconcile the Unity Gain ISO figures with the Photographic Dynamic Range Shadow Improvement?

Bob, I am no expert on unity gain, but I believe that Jim proved a few threads ago that Unity Gain is in practice only meaningful in very specific applications such as when one does heavy duty stacking for astrophotography.  As far as I understand today PDR Shadow Improvement is as good as it gets in helping to make the proper SNR/DR compromise.

Jack
Title: Re: ETTR vs ISO
Post by: Jack Hogan on May 27, 2013, 12:57:29 PM
I, personally, just set my camera to full stop ISO.  Since I shoot RAW only, I want to control whatever "ETTR" I am doing.

John

Ok, assuming that Bill's data for the 5DIII is correct and that you are shooting in full manual mode using the exposure/ISO strategy discussed in this thread, for maximum IQ and control you would want to keep doing what you are doing but always choosing ISO values 1/3 of a stop less than the standard ones (i.e. 159, 318, 636 etc. instead of 200, 400, 800...)

Jack
Title: Re: ETTR vs ISO
Post by: Jim Kasson on May 27, 2013, 01:06:50 PM
Reading off the chart in the earlier post, FWC appears to be around 67500 e- and UG around 400.  I get the curve in the graph assuming 8.4 ADU of random noise contributed by the amplifier at the output of the ADC.  It's not bad but it's not a perfect fit to the calculated read noise.  If I remember correctly the 5DIII has a couple of amplifying stages, which I treated as one in order to estimate this last figure.

Jack

Thanks, Jack. I'm going to use Clark's numbers (http://www.clarkvision.com/articles/evaluation-canon-5diii/): FWC = 68900, UG ISO = 500.

From his read noise numbers, I fitted a curve, and came up with a pre-gain read noise mean of 2.5 electrons, and a post-gain read noise mean of 7.5 ADC counts. This gave me a pretty good fit to his read noise data. I'm going to assume the variance of the two components of the read noise is the same as the mean (a la shot noise).

Then I'll do some modeling and see if the camera differences that people are talking about can be explained by the model parameters, or if there's something going on in the Canon that I'm not modeling.

However, the results may take me a day or two to generate. Some actual photography is intervening.

Jim
Title: Re: ETTR vs ISO
Post by: jrsforums on May 27, 2013, 01:10:47 PM
Ok, assuming that Bill's data for the 5DIII is correct and that you are shooting in full manual mode using the exposure/ISO strategy discussed in this thread, for maximum IQ and control you would want to keep doing what you are doing but always choosing ISO values 1/3 of a stop less than the standard ones (i.e. 159, 318, 636 etc. instead of 200, 400, 800...)

Jack

...or be in a mettered mode, with +1/3 EC...if i wanted to mimic the exposure of -1/3 ISO  :-)
Title: Re: ETTR vs ISO
Post by: Jim Kasson on May 27, 2013, 01:13:00 PM
If I remember correctly the 5DIII has a couple of amplifying stages, which I treated as one in order to estimate this last figure.

Jack, that looks like a good assumption. Here's the result of my curve fitting:

(http://www.kasson.com/ll/5DIIRN%20model.PNG)

Jim
Title: Re: ETTR vs ISO
Post by: Jim Kasson on May 27, 2013, 01:15:34 PM
Sorry...misinterpreted you question.

Probably Jim Kasson would best to answer this.  I could try, but would only be trying to interpret info he has provided me....and would probably do it poorly :-)

Not having an actual Canon, I am somewhat at a disadvantage here. Maybe my simulated 5DIII will shed some light on the subject when I get it implemented.

Jim
Title: Re: ETTR vs ISO
Post by: Vladimirovich on May 27, 2013, 01:20:21 PM


Vladimirovich, that would be a smart way to handle the digital component of ISO brightening.  Do you know any current DSLR that does that?



Panasonic GH3 (Sony sensor, as it turned out) does for ISOs after ISO6400 (so they combine both amplification, be it analog and/or digital, _and_ tag)...

1) you can see that if you convert .rw2 to .dng using Adobe DNG converter and compare baseline exposure tags there, for example ISO6400 = 0, ISO12800 = 1, ISO25600 = 2

2) I did PDR test using Claff's utilities - here is GH3 PDR graph

(http://imageshack.us/a/img29/187/gh3megapdravg3.jpg)

Title: Re: ETTR vs ISO
Post by: Jack Hogan on May 27, 2013, 01:23:26 PM
Thanks, Jack. I'm going to use Clark's numbers (http://www.clarkvision.com/articles/evaluation-canon-5diii/): FWC = 68900, UG ISO = 500.

FWC: my figure comes out of the DxO full SNR curves - and they average the three channels (I am not sure whether they weigh them).  Are Clark's derived just from the green channel?  And I agree about the UG ISO: I had forgotten that Canon's Raw data is offset by a couple of thousand ADUs.  510 it is.

Quote
From his read noise numbers, I fitted a curve, and came up with a pre-gain read noise mean of 2.5 electrons

I used 2.4 e- so that looks good.  Look forward to the results.

Jack
Title: Re: ETTR vs ISO
Post by: Vladimirovich on May 27, 2013, 01:27:46 PM
I can see that working if the OEM conversion utility were used. But it may not work with third party converters that don't pick all the tags.

majority of commercial 3rd party raw converter either are communicated by manufacturer or will know that from simple test or reverse engineering... some like RPP will not honor the tag intentionally, by design... so it is not an issue... and I guess at some point raw files from MFDB (PhaseOne) were like that ? just ISO by tag ?
Title: Re: ETTR vs ISO
Post by: RFPhotography on May 27, 2013, 01:56:07 PM
Sorry, refresh my memory, what's RPP?

So Adobe does pick up that tag then it appears?
Title: Re: ETTR vs ISO
Post by: RFPhotography on May 27, 2013, 01:57:10 PM
Bob, I am no expert on unity gain, but I believe that Jim proved a few threads ago that Unity Gain is in practice only meaningful in very specific applications such as when one does heavy duty stacking for astrophotography.  As far as I understand today PDR Shadow Improvement is as good as it gets in helping to make the proper SNR/DR compromise.

Jack

Jim, do you concur about the use of UG and PDR Shadow Improvement?
Title: Re: ETTR vs ISO
Post by: Vladimirovich on May 27, 2013, 02:00:05 PM
Sorry, refresh my memory, what's RPP?

a raw converter valued by an alternative lifestyle of raw conversion crowd


So Adobe does pick up that tag then it appears?

certainly = otherwise why do you think Adobe's own DNG converter writes the proper values ?

as it turned out the hidden expocorrection that Adobe's raw converters (ACR/LR) do is the sum of 2 values : one hardcoded in their code (that is what Adobe DNG converter shall reveal when you convert a native raw into Adobe's DNG) and another in their .dcp profiles
Title: Re: ETTR vs ISO
Post by: Jim Kasson on May 27, 2013, 02:39:38 PM
Jim, do you concur about the use of UG and PDR Shadow Improvement?

The metric that I've been using is photon-noise-corrected signal-to-noise ratio at a particular average value that corresponds to the shadow tone that you're most concerned about. I would think that this would be strongly correlated with PDR Shadow Improvement. However, I don't have enough information of the specifics of Claff's measurement of PDR Shadow Improvement to do a rigorous comparison. Any pointer would be appreciated.

Thanks,

Jim
Title: Re: ETTR vs ISO
Post by: RFPhotography on May 27, 2013, 02:52:24 PM
a raw converter valued by an alternative lifestyle of raw conversion crowd

Your ass isn't as smart as you think it is.   ::)  What's the name of the program?


Quote
certainly = otherwise why do you think Adobe's own DNG converter writes the proper values ?


And that's fine.  Just confirming.
Title: Re: ETTR vs ISO
Post by: bjanes on May 27, 2013, 02:55:28 PM
Your ass isn't as smart as you think it is.   ::)  What's the name of the program?
And that's fine.  Just confirming.

A google search turns up this link (http://www.raw-photo-processor.com/RPP/Overview.html).

Bill
Title: what benefits to substantial "underexposing" through lower ISO in low light?
Post by: BJL on May 27, 2013, 03:47:08 PM
In low light photography where you cannot give the sensor enough exposure to get close to blowing any highlights through full wells, I can only see a case for erring a bit on the low side with analog gain (ISO speed setting) to avoid clipping of highlights by amplifying them beyond the maximum ADC output level. I cannot see any benefit beyond a bit of a safety margin, and the idea of under-amplifying so much that three or more stops of digital amplification are needed in post-processing sounds unwise --- more so if many widely-used raw convertors will then introduce a color shift when you push the levels up by three or more stops.

One thing to beware of is the blind pursuit of higher engineering DR values, like the talk of increasing the engineering DR by one stop for each stop of lowered ISO amplification: beyond some point, that extra DR is just unused numerical levels above any level present in the signal. If one ISO level avoids clipping any highlights in any color channel, then each subsequent halving of the ISO gain adds an extra zero leading bit before the first significant bit, which adds a stop of engineering DR, measured from the maximum ADC output level down, but adds nothing to signal quality, and moves you closer to increased shadow noise due to the increased ADC quantization noise.

With the metering of most modern system cameras producing raw files in which the metered level is placed far enough below the maximum level that there is almost always a significant margin of unused levels near the top, under-amplifying is in most cases going to offer no IQ benefit, while adding avoidable complications to the raw conversion process --- especially if you then need to construct custom profiles to avoid color shifts.
Title: Re: what benefits to substantial "underexposing" through lower ISO in low light?
Post by: jrsforums on May 27, 2013, 05:05:38 PM
In low light photography where you cannot give the sensor enough exposure to get close to blowing any highlights through full wells, I can only see a case for erring a bit on the low side with analog gain (ISO speed setting) to avoid clipping of highlights by amplifying them beyond the maximum ADC output level. I cannot see any benefit beyond a bit of a safety margin, and the idea of under-amplifying so much that three or more stops of digital amplification are needed in post-processing sounds unwise --- more so if many widely-used raw convertors will then introduce a color shift when you push the levels up by three or more stops.

One thing to beware of is the blind pursuit of higher engineering DR values, like the talk of increasing the engineering DR by one stop for each stop of lowered ISO amplification: beyond some point, that extra DR is just unused numerical levels above any level present in the signal. If one ISO level avoids clipping any highlights in any color channel, then each subsequent halving of the ISO gain adds an extra zero leading bit before the first significant bit, which adds a stop of engineering DR, measured from the maximum ADC output level down, but adds nothing to signal quality, and moves you closer to increased shadow noise due to the increased ADC quantization noise.

With the metering of most modern system cameras producing raw files in which the metered level is placed far enough below the maximum level that there is almost always a significant margin of unused levels near the top, under-amplifying is in most cases going to offer no IQ benefit, while adding avoidable complications to the raw conversion process --- especially if you then need to construct custom profiles to avoid color shifts.

Thanks. BJL.

Took me a couple readings, but I think to old brain finally got it.

In Emil's statement (which I posted very early in this thread), he was making the point that, with ISOless sensors, it bought nothing to raise ISO to set the histogram further to the right, ETTR.  What you are saying, I think, is that doing so does not cost you anything either....the "increased" DR has no value as it is highlights of no value.

...and, with sensors such as the Canons, increasing the ISO could get you less noise in the shadow areas.

John
Title: Re: what benefits to substantial "underexposing" through lower ISO in low light?
Post by: Jack Hogan on May 27, 2013, 05:15:03 PM
One thing to beware of is the blind pursuit of higher engineering DR values, like the talk of increasing the engineering DR by one stop for each stop of lowered ISO amplification: beyond some point, that extra DR is just unused numerical levels above any level present in the signal.

Yes Bill, although I fail to see how that would be undesirable.  Once you have maxed out Exposure (that is the longest shutter speed and the lowest f/number that your technical/artistic constrainsts allow), raise ISO from base to give you the best SNR possible without blowing desirable highlights and stop there.  This strategy allows you to capture all of the relevant information from the scene.  It makes zero difference to IQ whether the last few bits are unused or the first few bits are filled with random noise.  What matters is that you got all of the information you wanted.  In fact with this strategy not only do you get all of it, you need to worry about one less variable (ISO and blowing highlights) and you get smaller files to boot.  What's there not to like?

As for twists, no point blowing things out of proportion.  Here is a capture by a D800e at base ISO that the guy says requires a 6 stop brightness push for pleasing tones (http://www.dpreview.com/forums/post/51399489) (don't try this with Canon, though): I don't know that I'd go that far, but even then it's nice to know you can.  Click through to see whether you can spot the twist in LR: I can barely see it.

Cheers,
Jack
Title: Re: what benefits to substantial "underexposing" through lower ISO in low light?
Post by: Jim Kasson on May 27, 2013, 06:19:39 PM
Once you have maxed out Exposure (that is the longest shutter speed and the lowest f/number that your technical/artistic constraints allow), raise ISO from base to give you the best SNR possible without blowing desirable highlights and stop there.  This strategy allows you to capture all of the relevant information from the scene.  It makes zero difference to IQ whether the last few bits are unused or the first few bits are filled with random noise.  What matters is that you got all of the information you wanted.  In fact with this strategy not only do you get all of it, you need to worry about one less variable (ISO and blowing highlights) and you get smaller files to boot. 

Once you have given up on ETTR because you don't have enough light to make an image with the f-stop and shutter speed you want at base ISO, and you start letting the histogram slide to the left, something magical happens: you no longer have to worry about exposure. The light changes; you ignore it, knowing it will make no difference to you. You don't have to continue to consult the histogram; you don't care what it looks like, as long as it hasn't crept alarmingly leftward, and you'd know if that were happening. It's remarkably freeing. You work faster, and you are more able to concentrate on your subject.

Jim
Title: Re: ETTR vs ISO
Post by: RFPhotography on May 27, 2013, 07:10:50 PM
Bob, I am no expert on unity gain, but I believe that Jim proved a few threads ago that Unity Gain is in practice only meaningful in very specific applications such as when one does heavy duty stacking for astrophotography.  As far as I understand today PDR Shadow Improvement is as good as it gets in helping to make the proper SNR/DR compromise.

Jack

Thanks, Jack and Jim.  I'd think that, like PDR/SI, it would be useful in any situation where you were having to increase ISO to get a 'normal' exposure (or even an ETTR'd exposure in cases where ETTR above base ISO is useful).  Knowing what point to stop bumping the ISO to prevent digital multiplication and potential blowing of highlights would be good, no?  That assumes that one doesn't need a 'normal' looking exposure and has the time to make adjustments after the fact.  This article seems to make a reasonable point in that regard, http://www.ishootshows.com/2009/01/28/push-processing-and-unity-gain/.

Based on the small sampling in that article and comparing to the information on Claff's website, it does appear that the two are quite close, as Jim surmises.
Title: Re: what benefits to substantial "underexposing" through lower ISO in low light?
Post by: BJL on May 27, 2013, 08:39:56 PM
Once you have given up on ETTR because you don't have enough light to make an image with the f-stop and shutter speed you want at base ISO ... you no longer have to worry about exposure. The light changes; you ignore it, knowing it will make no difference to you ...
Jim, yes that approach makes sense in low and variable light where the traditional approach of "on meter exposure" would involve jumping around the ISO speed setting: just use a value that is sure not to be above the low end of the range, so that some shots might be "under-amplified" (underexposed with default JPEG conversion) but none are over-amplified to the point of highlight clipping.
Title: Re: what benefits to substantial "underexposing" through lower ISO in low light?
Post by: BJL on May 27, 2013, 08:53:24 PM
... I fail to see how that would be undesirable. 
To answer, I will quote my post that you are are quoting:
"adding avoidable complications to the raw conversion process" like having to adjust the levels individually for each conversion whereas a high proportion of on-meter exposures do not need this, and "especially if you then need to construct custom profiles to avoid color shifts." So only inconvenience, and Jim Kasson has just pointed out one case where the extra PP effort can be well worth it, but when you can easily enough use the ISO that "fits" with your choice of shutter speed, aperture and the light levels, why make work for yourself with post-processing manipulations and run at least a risk of color twist at more extreme levels under-amplification? My man warning is that some people are claiming that this extra effort gives an IQ advantage that does not in fact exist in this particular situation.
Title: Re: ETTR vs ISO
Post by: Jim Kasson on May 27, 2013, 09:20:50 PM
I've finished my Canon 5DIII simulation, and I think that the advantages of increasing the ISO that people are seeing can be explained by the post amplification portion of the read noise, which I put at 7.5 analog-to-digital converter (ADC) counts (aka ADUs). The effect of this relatively large noise can be countered by increasing the gain of the amplifier ahead of the ADC, which is what you do when you turn up the ISO.

Here are the photon corrected signal-to-noise ratio (SNR) results for the simulated 5DIII when presented with light levels that produce constant mean ADC counts a factor of two (one stop) apart from near-saturation to near the noise floor. I did not simulate any digital gain; I assumed it is all analog. The results are for one of the (assumed identical) green channels only.

(http://www.kasson.com/ll/Canon3DIIIsimulatedCorSNR.PNG)

You can see that increasing the ISO helps all the way to ISO 12800, especially in the darkest tones. The reason for the bump from ISO 100 to ISO 400 in the lightest tones is that turning up the ISO lessens the effect of pixel-response non-uniformity (PRNU). I don't consider this to be an effect worth worrying about in cameras with PRNUs below half a percent.

Compare the Canon results with the D800E ones:

(http://www.kasson.com/ll/D800EsimulatedCorSNR.PNG)

The simulated D800E is essentially "ISO-less" ( I don't like that term, but it has acquired a certain currency). The real D800E has more problems than the simulated one at ISOs above 1600-3200, probably because of the digital gain. Note that the Canon SNRs are slightly better than the Nikon ones for brighter tones at the higher ISOs; that's because the Nikon's full well capacity is lower. If the Nikon images were res'd down to Canon levels, this advantage would disappear.

Jim
Title: noise sources after analog amplification, before ADC
Post by: BJL on May 27, 2013, 09:35:02 PM
I've finished my Canon 5DIII simulation, and I think that the advantages of increasing the ISO that people are seeing can be explained by the post amplifications portion of the read noise ...
This fits with what I have gathered from other sources: Canon's recent sensors are unusual in having early analog amplification (on-chip) but late ADC (off-chip) and so a significant opportunity for noise to enter the signal between analog amplification and ADC. In contrast, CCD's and maybe older Canon CMOS sensors did analog amplification later, off-chip just before ADC, and most other modern CMOS sensors do both analog amplification and ADC very early, at the bottom of each column of pixels. In either case there is far less opportunity for noise to enter between analog amplification and ADC. (But the ADC itself can add some noise, quantization noise in particular.)
Title: Re: noise sources after analog amplification, before ADC
Post by: Jim Kasson on May 27, 2013, 09:43:44 PM
Canon's recent sensors are unusual in having early analog amplification (on-chip) but late ADC (off-chip) and so a significant opportunity for noise to enter the signal between analog amplification and ADC. In contrast, CCD's and maybe older Canon CMOS sensors did analog amplification later, off-chip just before ADC, and most other modern CMOS sensors do both analog amplification and ADC very early, at the bottom of each column of pixels. In either case there is far less opportunity for noise to enter between analog amplification and ADC. (But the ADC itself can add some noise, quantization noise in particular.)

That makes sense. Thanks.

Jim
Title: Re: what benefits to substantial "underexposing" through lower ISO in low light?
Post by: Jack Hogan on May 28, 2013, 03:10:33 AM
To answer, I will quote my post that you are are quoting:
"adding avoidable complications to the raw conversion process" like having to adjust the levels individually for each conversion whereas a high proportion of on-meter exposures do not need this, and "especially if you then need to construct custom profiles to avoid color shifts." So only inconvenience, and Jim Kasson has just pointed out one case where the extra PP effort can be well worth it, but when you can easily enough use the ISO that "fits" with your choice of shutter speed, aperture and the light levels, why make work for yourself with post-processing manipulations and run at least a risk of color twist at more extreme levels under-amplification? My man warning is that some people are claiming that this extra effort gives an IQ advantage that does not in fact exist in this particular situation.

Yes, potential brightness correction in post is a compromise with the potential extra DR, depending on your workflow.  I shoot Raw+Jpeg and fine tune all of my keepers, so that's not an issue for me, but I can see how it could be for others.

I think the main take-away from this approach is that it teaches photographers to decouple Exposure (photon counting, shutter speed, f/number) from ISO (processing, perceived brightness, SNR, DR).  The remaining problem to solve (the one that is causing a lot of confusion) is that only one parameter, ISO, is used to control both objective IQ related quantities like SNR/DR and subjective ones like perceived output image brightness.  It's my guess (and wish) that in a generation or so what we call ISO in-camera will be split into two components, initially hidden from the beginner: a 'gain' to control how the sensor output gets processed analogically before being written to the Raw data and a 'Perceived Lightness' target brightness for the OOC image independent of the average values of the Raw data.

Cheers,
Jack
Title: Re: ETTR vs ISO
Post by: Jack Hogan on May 28, 2013, 03:12:17 AM
I've finished my Canon 5DIII simulation, and I think that the advantages of increasing the ISO that people are seeing can be explained by the post amplification portion of the read noise, which I put at 7.5 analog-to-digital converter (ADC) counts (aka ADUs). The effect of this relatively large noise can be countered by increasing the gain of the amplifier ahead of the ADC, which is what you do when you turn up the ISO.

Good show, thanks Jim.  Makes one wonder whether to save space one should stick to 12-bit mode.
Title: Re: what benefits to substantial "underexposing" through lower ISO in low light?
Post by: FranciscoDisilvestro on May 28, 2013, 03:43:42 AM
I think the main take-away from this approach is that it teaches photographers to decouple Exposure (photon counting, shutter speed, f/number) from ISO (processing, perceived brightness, SNR, DR).  The remaining problem to solve (the one that is causing a lot of confusion) is that only one parameter, ISO, is used to control both objective IQ related quantities like SNR/DR and subjective ones like perceived output image brightness.  It's my guess (and wish) that in a generation or so what we call ISO in-camera will be split into two components, initially hidden from the beginner: a 'gain' to control how the sensor output gets processed analogically before being written to the Raw data and a 'Perceived Lightness' target brightness for the OOC image independent of the average values of the Raw data.

Cheers,
Jack

What I would like is a live view mode or EVF with two sets of "Blinkies" , one for highlights and the other for the shadows, based on actual raw values, not rendered values:

- The highlight blinkies just show saturated or blown raw values.
- The shadow blinkies are a little more complex, the idea is to be able to configure a minumun acceptable SNR ratio (for non-technicals it could be just a "quality" factor) so that any areas with a raw level that will result in lower than desired SNR will blink.

With those two blinkies, you could just play with the 'gain' until you achieve the best compromise between clipped highlight and noisy shadows. This of course for those that are confortable with RAW processing.
Title: Re: what benefits to substantial "underexposing" through lower ISO in low light?
Post by: hjulenissen on May 28, 2013, 04:37:47 AM
What I would like is a live view mode or EVF with two sets of "Blinkies" , one for highlights and the other for the shadows, based on actual raw values, not rendered values:

- The highlight blinkies just show saturated or blown raw values.
- The shadow blinkies are a little more complex, the idea is to be able to configure a minumun acceptable SNR ratio (for non-technicals it could be just a "quality" factor) so that any areas with a raw level that will result in lower than desired SNR will blink.

With those two blinkies, you could just play with the 'gain' until you achieve the best compromise between clipped highlight and noisy shadows. This of course for those that are confortable with RAW processing.

I believe that Magic Lantern (3rd party software for Canon DSLRs) provide raw-based histogram, blinkies etc.

http://www.magiclantern.fm/whats-new/95-tutorial/catexposure/125-new-fast-zebras

-h
Title: Re: what benefits to substantial "underexposing" through lower ISO in low light?
Post by: FranciscoDisilvestro on May 28, 2013, 05:17:32 AM
I believe that Magic Lantern (3rd party software for Canon DSLRs) provide raw-based histogram, blinkies etc.

http://www.magiclantern.fm/whats-new/95-tutorial/catexposure/125-new-fast-zebras

-h

It looks nice, but it seems to me it works with rendered data and not raw values. It is not clear to me if the "shadows" warning just show values (again rendered) near zero or if you could customize a SNR value.

Thanks, regards.

Edit: I found more info and you can setup the shadow threshold but I'm pretty sure it is based on rendered values.
Title: Re: ETTR vs ISO
Post by: hjulenissen on May 28, 2013, 06:59:42 AM
http://www.magiclantern.fm/forum/index.php?topic=5149.msg31247#msg31247
Quote
Alright, now you have true RAW histogram and zebras. 5D2 only.

To use, simply choose RAW in Canon menu and set histo/zebra to RGB. No new menus - it should just work.

https://bitbucket.org/hudson/magic-lantern/commits/115f297ac7c8
Quote
After some quick tests, RAW histogram and the zebras are matching exactly what I see from RawDigger

Sweet!
Quote
Black level got fixed... working on many digic IV cameras... you just have to find the bins.
Seems that it is sort of an early feature, but anyways...

Strange that it takes reverse-engineering to give us the features that we want, while the camera manufacturers seems to focus on in-camera HDR, the infamous "PictBridge" button and other features that seems to appeal more to casual users (catered for by iPhones, superzoom and m4/3?) than those interested in clunky, expensive 35mm SLR cameras.

-h
Title: Re: ETTR vs ISO
Post by: FranciscoDisilvestro on May 28, 2013, 07:14:47 AM
Great!, It looks promising, thanks for sharing.

I only wonder how much powerful and useful options we could have if only the camera makers offered a documented SDK to customize the camera firwmare.
Title: Re: ETTR vs ISO
Post by: Jim Kasson on May 28, 2013, 10:19:08 AM
Good show, thanks Jim.  Makes one wonder whether to save space one should stick to 12-bit mode.

I'll run a test.
Title: under-amplified raw through conservative ISO in low light, and usable previews
Post by: BJL on May 28, 2013, 10:38:00 AM
Yes, potential brightness correction in post is a compromise with the potential extra DR, depending on your workflow.  I shoot Raw+Jpeg and fine tune all of my keepers, so that's not an issue for me, but I can see how it could be for others.
I also shoot raw+JPEG, and I still see one inconvenience to substantially "under-amplifying" in low light --- along with a potential solution, and one that is indeed already more or less offered by some cameras.

The problem is that if I used an ISO sensitivity setting that for some shots is well below what would be needed for "on-meter" exposure, the JPEG's and previews of the raw files will sometimes be way too dark, and then the problems start before I get to fine-tuning my keepers:
it becomes necessary to make rough level adjustments on every "under-amplified" image just to get a preview that allows me to decide which shots _are_ the keepers.

One solution for me involves:
a) A preview image (the OOC JPEG, or default raw->JPEG conversion) that has been amplified to roughly correct levels.
b) Raw files in low-light situations that have lower levels from "conservative amplification": that is, the raw level histogram has a significant gap at right.
Camera makers using almost "ISO-less" sensors could accommodate both goals by keeping the analog gain lowish, thus allowing an abundant safety margin against highlight clipping by over amplification in the raw file, but tagging the raw file with information about the intended exposure level, with this used in the in-camera JPEG conversion and in default conversion of the raw files by any software that knows about the camera (which includes Adobe DNG Convertor, Lightroom etc. once they are updated for each new model of camera.)

The good news is that many camera makers already do something like this, so I think we might be over-thinking a situation that has already been addressed. I refer to cameras that use conservative analog gain at elevated ISO speed settings along with providing information about this that is used by raw convertors and in making in-camera JPEGs, thus offering both raw files with abundant highlight headroom and default JPEG versions with appropriate "viewable" brightness levels. These camera models are fairly easy to detect: they are the ones that offer substantial blown-highlight recovery from raw files, and for which the DxO measurement of "sensitivity" at elevated ISO settings is significantly less than the ISO sensitivity setting on the camera. That is, those cameras which some misguided followers of the over-broad ETTR doctrine ("the raw histogram should always be as far to the right as possible", even in low light/high ISO speed situations where read noise overwhelms ADC quantization noise) accuse of "cheating on" or "overstating" their ISO speed settings.
Title: Re: ETTR vs ISO
Post by: Jim Kasson on May 28, 2013, 10:48:13 AM
Makes one wonder whether to save space one should stick to 12-bit mode.

Here are the simulation results, with the curves being the 14-bit SNRs less the 12-bit SNRs:

(http://www.kasson.com/ll/Canon5DIII12-14bit.PNG)

I don't know quite what to make of it. It shows that 14-bits gives an improvement in SNR for shadows at low ISO, but actually hurts a tiny bit for shadows at high ISO. In the middle tones and above, it makes no difference.

Jim
Title: Re: ETTR vs ISO
Post by: jrsforums on May 28, 2013, 11:03:35 AM
Good show, thanks Jim.  Makes one wonder whether to save space one should stick to 12-bit mode.

Unlike Nikon, Canon does nt have a switch to 12bit mode.
Title: Re: ETTR vs ISO
Post by: Jim Kasson on May 28, 2013, 11:15:39 AM
The same is true of the D800E, although to a lesser extent:

(http://www.kasson.com/ll/D800E12-14bit.PNG)

Jim
Title: Re: ETTR vs ISO
Post by: Jack Hogan on May 28, 2013, 02:26:12 PM
Here are the simulation results, with the curves being the 14-bit SNRs less the 12-bit SNRs:

(http://www.kasson.com/ll/Canon5DIII12-14bit.PNG)

I don't know quite what to make of it. It shows that 14-bits gives an improvement in SNR for shadows at low ISO, but actually hurts a tiny bit for shadows at high ISO. In the middle tones and above, it makes no difference.

Hmmm, weird.  Did you just truncate the 12 bits?  What % of full scale is your Series1 at?
Title: Re: ETTR vs ISO
Post by: Jim Kasson on May 28, 2013, 02:52:37 PM
Did you just truncate the 12 bits?

Rounded the analog value, just like an ADC would be designed to do. You want the transitions to occur half an LSB to either side of the nominal.

What % of full scale is your Series1 at?

1/40,000 at  ISO 12800, and 128/40,000 at base ISO. [Clairification: that's with respect to FWC, with respect to the maximum output of the ADC, it's 128/40000 for all ISOs.] The ISO 12800 exposures are on average about 1.5 electrons/plane for the Canon, and close to one electron per plane for the Nikon.

Jim
Title: Re: ETTR vs ISO
Post by: Jack Hogan on May 29, 2013, 03:12:33 AM
Rounded the analog value, just like an ADC would be designed to do. You want the transitions to occur half an LSB to either side of the nominal.

1/40,000 at  ISO 12800, and 128/40,000 at base ISO. [Clairification: that's with respect to FWC, with respect to the maximum output of the ADC, it's 128/40000 for all ISOs.] The ISO 12800 exposures are on average about 1.5 electrons/plane for the Canon, and close to one electron per plane for the Nikon.

Jim

Intuitively one would expect virtually no difference between the two given the 8 ADU noise 'floor'.  I wonder if what we are seeing is actually quantization noise.  How large of a sample do you use?  If too small the higher bit depth could appear to produce a better apparent SNR in the deep shadows - but it might simply be better defined noise.

Jack
Title: Re: ETTR vs ISO
Post by: Jim Kasson on May 29, 2013, 10:19:15 AM
 How large of a sample do you use?

I used a 1600x1600 sample, so 800x800 = 640,000 green pixels. I used a separate sample for each data point.

Jim
Title: Re: ETTR vs ISO
Post by: Jim Kasson on May 29, 2013, 11:23:12 AM
Intuitively one would expect virtually no difference between the two given the 8 ADU noise 'floor'. 

128/40,000 at base ISO is 52 14-bit ADUs, so it's quite a bit above the noise floor.  If you're talking about the negative values at high ISOs, I agree. Upon thinking about it, I think the small SNR reduction in those regions for the 12-bit resolution is due to the ADC digitizing analog values that would have been different in the 14-bit case into the same value, thus lowering standard deviation. I can run some tests that isolate that effect and report back.

Jim
Title: Re: ETTR vs ISO
Post by: bjanes on May 29, 2013, 11:50:01 AM
Hi TJ,

I don't shoot Canon, but from what I can tell the 5DIII uses post-ADC digital gain only after ISO 12800, so no point raising ISO beyond that unless you need 'properly' bright images SOOC. Looking at Bill Claff's graph, it seems that you do get a benefit in shadow SNR by raising ISO in-camera from 100 all the way up to 12800, although the law of diminishing returns needs to be considered after about ISO 2500: raising it from 2500 to 12800 only nets you a 0.2 stop improvement in shadow SNR at the expense of 2.3 stops of reduced DR: a decision best evaluated each time based on the situation (blips in the curve above 12800 are suspicious, as the data up there is often 'cooked').

The other take-away from the curve is that if you are after best SNR/DR it appears that you are better off increasing ISO one stop at a time (as opposed to in 1/3 or 1/2 stop increments) in correspondence of the steps in the graph: IOW if you are sitting at ISO 300 and you feel the urge to increase ISO a little, there is no point going to 400 or 500.  Just leave it there with the benefit of slightly better DR until you feel the need to go to 600.

Jack

PS the 5DIII appears to have an effective QE of about 14% and PRNU of about 0.4%, plus a fairly noisy analog amplifier contributing about 8 ADU random noise throughout the range, so the last three bits never contain much information.

(http://i.imgur.com/ZO7DC2t.jpg)

Jack,

Your graph is interesting and can serve as a topic for discussion. It fits Emil's simplified read noise formula (http://theory.uchicago.edu/~ejm/pix/20d/tests/noise/noise-p2.html#read_vs_iso) R2 = (G R0)2 + R2, where R is the total read noise, G is the ISO, R1 is noise upstream of the ISO amplifier (presumably mainly the sensor itself) and R2 is the noise downstream of the amplifier (presumably with the ADC representing a major component). Roger Clark in his sensor performance analysis (http://www.clarkvision.com/articles/digital.sensor.performance.summary/#unity_gain) (Fig 8a) implicates the ADC as a main contributor of read noise for the Canon camera. The amp noise itself is a tweener in Emil's analysis.

If I interpret it correctly, your graph shows a constant sensor read noise component which is multiplied by the ISO amplifier. What you term the amp read noise could include the ADC noise which is constant and adds in quadrature to the sensor noise to give the up-sloping total read noise curve. Bill Claff (http://home.comcast.net/~NikonD70/Charts/RN_ADU.htm) states that a curved plot of the read noise expressed in ADU numbers indicates that the ADC contributes significantly to the read noise whereas a straight line plot indicates that the sensor is the main source of the read noise.

Bill's plots for the Nikon D800e are combined and shown below for ISO up to 1600. The read noise in electrons is relatively constant but the values are affected by the ADC readout and the noise in ADUs is linear indicating the source lies mainly in the sensor. The curves cross at approximately the unity gain of the sensor (ca 320), but I don't know if this is fortuitous or significant.

(http://bjanes.smugmug.com/Photography/Sensor-Analysis/D800e-Sensor/i-kX6NdR8/0/O/NikonReadNoise.png)

What do you think?

Bill

Title: Re: ETTR vs ISO
Post by: Jim Kasson on May 29, 2013, 12:16:02 PM
(http://bjanes.smugmug.com/Photography/Sensor-Analysis/D800e-Sensor/i-kX6NdR8/0/O/NikonReadNoise.png)

What do you think?

The data point at ISO 1600 looks suspicious to me. In my tests on a real-world D800E, I see the read noise referred to the sensor at ISO 1600 as just slightly worse than at ISO 800. I also don't know of a mechanism to have the read noise drop suddenly like that. Otherwise, this is generally close to what I see in my testing.

Jim
Title: Re: ETTR vs ISO
Post by: bjanes on May 29, 2013, 12:26:49 PM
The data point at ISO 1600 looks suspicious to me. In my tests on a real-world D800E, I see the read noise referred to the sensor at ISO 1600 as just slightly worse than at ISO 800. I also don't know of a mechanism to have the read noise drop suddenly like that. Otherwise, this is generally close to what I see in my testing.

Jim

Jim,

Bill does note that at ISO 1600 and above, the D800e appears to manipulate the raw data prior to writing it out to the memory card. Others have noted this by looking at the histograms in Rawdigger.

Bill
Title: Re: ETTR vs ISO
Post by: Jim Kasson on May 29, 2013, 12:55:23 PM
Bill does note that at ISO 1600 and above, the D800e appears to manipulate the raw data prior to writing it out to the memory card. Others have noted this by looking at the histograms in Rawdigger.

Good point. I have not noticed this effect in my measurements, but it might be the result of the digital gain that I've noticed the camera applying at ISO 3200 and above (leftward bit shifts of the digital information).

If that's the case, then the ADC is effectively progressively losing resolution, with possible effects on the noise statistics like I was struggling with a few posts above in another context.

Now I really want to run some tests.

Jim
Title: Re: ETTR vs ISO
Post by: Jim Kasson on May 29, 2013, 01:36:18 PM
I've run some tests on quantizing signals plus Gaussian noise at varying resolutions, both with and without the quantizing of the sensor values to integer numbers of electrons. The results are to a great extent independent of resolution, at least in the 12 to 14 bit range.

So that's not it. I wonder if the effect of decreasing resolution for shadows at high ISOs slightly improving the SNR can be seen in a real-world camera that has both 12 and 14 bit options. The effect in the simulated D800E is so slight that it would be really hard to observe with a real camera.

It's a mystery to me.

Jim
Title: Re: ETTR vs ISO
Post by: Jack Hogan on May 29, 2013, 02:58:19 PM
If I interpret it correctly, your graph shows a constant sensor read noise component which is multiplied by the ISO amplifier. What you term the amp read noise could include the ADC noise which is constant and adds in quadrature to the sensor noise to give the up-sloping total read noise curve. Bill Claff (http://home.comcast.net/~NikonD70/Charts/RN_ADU.htm) states that a curved plot of the read noise expressed in ADU numbers indicates that the ADC contributes significantly to the read noise whereas a straight line plot indicates that the sensor is the main source of the read noise.

Yes to all of the above, Bill.  The two read noise values (sensor and amp/ADC) in the simplified formula are found fitting the total read noise derivations.  [Please note that as mentioned earlier to Jim the inverse gain that is shown in my 5DIII chart is incorrect because I forgot that some Canons have severe zero offsets.  At ISO 100 it should read 5.11 instead of 4.13, with the other ISOs proportionately lower.]

Bill's plots for the Nikon D800e are combined and shown below for ISO up to 1600. The read noise in electrons is relatively constant but the values are affected by the ADC readout and the noise in ADUs is linear indicating the source lies mainly in the sensor. The curves cross at approximately the unity gain of the sensor (ca 320), but I don't know if this is fortuitous or significant.

Yes again.  Total read noise in e- and ADU are related by the inverse gain, which you can find simply by dividing the two together.  For instance for the D800e at ISO 100, from Bill's site, RN100(ADU)=1.261 and RN100(e-)=4.228, therefore igain according to Bill is 3.39 e-/ADU at this ISO.  When will the two read the same number?  At the ISO where gain is 1.  No luck involved this time :)

Quote
(http://bjanes.smugmug.com/Photography/Sensor-Analysis/D800e-Sensor/i-kX6NdR8/0/O/NikonReadNoise.png)

What do you think?

Looks good, it shows a nice, almost 'ISOless' system.  I think you may have a typo in the electrons curve at ISO 1600: it looks lower than 2.721 e-

Jack
Title: Re: ETTR vs ISO
Post by: Jack Hogan on May 29, 2013, 03:01:22 PM
Good point. I have not noticed this effect in my measurements, but it might be the result of the digital gain that I've noticed the camera applying at ISO 3200 and above (leftward bit shifts of the digital information).

As you can see in the image in Reply#80 a couple of pages back, the D800e appears to start applying ISO digitally just before ISO 1600 and thereafter.

Jack
Title: Re: ETTR vs ISO
Post by: Jim Kasson on May 29, 2013, 03:19:48 PM
As you can see in the image in Reply#80 a couple of pages back, the D800e appears to start applying ISO digitally just before ISO 1600 and thereafter.

Jack


You're right, Jack. I wrote about that a few months ago, (http://blog.kasson.com/?p=2932) but forgot the histo dropouts at 1600.  Of course, the D800 uses digital gain at all ISOs in the red and blue channels.

Jim
Title: Re: ETTR vs ISO
Post by: Jack Hogan on May 29, 2013, 04:07:37 PM
I've run some tests on quantizing signals plus Gaussian noise at varying resolutions, both with and without the quantizing of the sensor values to integer numbers of electrons. The results are to a great extent independent of resolution, at least in the 12 to 14 bit range.

OK, that makes sense, especially for as large a sample as you are using.

So that's not it. I wonder if the effect of decreasing resolution for shadows at high ISOs slightly improving the SNR can be seen in a real-world camera that has both 12 and 14 bit options. The effect in the simulated D800E is so slight that it would be really hard to observe with a real camera.

It's a mystery to me.

Jim

Can't think of anything else myself.

Jack
Title: Re: ETTR vs ISO
Post by: Jack Hogan on May 29, 2013, 04:10:56 PM

Of course, the D800 uses digital gain at all ISOs in the red and blue channels.

Jim

Yes, it's what Iliah Borg calls WB preconditioning.

Jack
Title: Re: ETTR vs ISO
Post by: Jim Kasson on May 29, 2013, 04:39:41 PM
Yes, it's what Iliah Borg calls WB preconditioning.

I don't know why any digital multiplication of unsigned integers isn't better done in the raw converter , but there it is.

Jim
Title: Re: ETTR vs ISO
Post by: Jim Kasson on May 29, 2013, 04:43:38 PM
(http://bjanes.smugmug.com/Photography/Sensor-Analysis/D800e-Sensor/i-kX6NdR8/0/O/NikonReadNoise.png)

By the way, if you use dark-field exposures to calculate read noise, you get values that are quite different from these, probably because of the D800E's truncation of negative values. I understand that Canon has a built-in offset to keep this from happening.

Jim
Title: Re: ETTR vs ISO
Post by: bjanes on May 29, 2013, 05:52:31 PM
By the way, if you use dark-field exposures to calculate read noise, you get values that are quite different from these, probably because of the D800E's truncation of negative values. I understand that Canon has a built-in offset to keep this from happening.

Jim,

Those read noises were from Bill Claff's data. He took the read noise from the optical black region whose data are not truncated. He uses his own proprietary software, but one can use RawDigger to examine the optical black area.

Here is the optical black area of the D800e at ISO 100 with the optical black area selected by hand (one could use selection by the numbers to get a bigger sample). The read noise for the green channels is 1.1 ADU. The histogram is approximately gaussian and an offset of 600 ADU has been applied to prevent truncation.

Bill

Title: Optical Black RN versus Curve Fitted RN
Post by: Jack Hogan on May 30, 2013, 05:39:37 AM
By the way, if you use dark-field exposures to calculate read noise, you get values that are quite different from these, probably because of the D800E's truncation of negative values. I understand that Canon has a built-in offset to keep this from happening.

Jim

Jim,

Those read noises were from Bill Claff's data. He took the read noise from the optical black region whose data are not truncated.

Yes I have always noticed and wondered about that myself.  Why would the optical black standard deviation values be quite a bit better than those we compute using standard curve fitting?  Perhaps they are totally unprocessed - or differently processed?  Which is more representative for our purposes?
Title: Re: Optical Black RN versus Curve Fitted RN
Post by: Jim Kasson on May 30, 2013, 10:38:52 AM
Why would the optical black standard deviation values be quite a bit better than those we compute using standard curve fitting?  Perhaps they are totally unprocessed - or differently processed?  Which is more representative for our purposes?

If the optical black values are simply offset, they should be the most direct way to measure read noise. The sample size is limited, though. I'm assuming that those sensels are representative; that's the whole point, but it's not guarenteed.

Jim
Title: Re: Optical Black RN versus Curve Fitted RN
Post by: bjanes on May 30, 2013, 11:20:02 AM
Yes I have always noticed and wondered about that myself.  Why would the optical black standard deviation values be quite a bit better than those we compute using standard curve fitting?  Perhaps they are totally unprocessed - or differently processed?  Which is more representative for our purposes?

If the optical black values are simply offset, they should be the most direct way to measure read noise. The sample size is limited, though. I'm assuming that those sensels are representative; that's the whole point, but it's not guaranteed.

Jack & Jim,

To address these concerns, I repeated my analysis with Rawdigger, selecting the entire optical black area for analysis.

This frame shows the entire raw image with the optical black on the right selected by the numbers.
(http://bjanes.smugmug.com/Photography/Sensor-Analysis/D800e-Sensor/i-kPGvnzf/0/O/ManualSelection.png)

Here is the raw histogram. 17K pixels are included for each channel, with 34K green values. The bell curve is skewed to the left. I don't know what this means.
(http://bjanes.smugmug.com/Photography/Sensor-Analysis/D800e-Sensor/i-WPSQTZ7/0/O/i_10-Sel-7378-0-14x4924.png)

Converting the selection to a sample gives the following values. The SDs for the green channels representing the read noise in ADUs is larger than before, with a mean of 1.52 ADUs.
(http://bjanes.smugmug.com/Photography/Sensor-Analysis/D800e-Sensor/i-4Hs25cT/0/O/Samples_img10.png)

My data from a previous analysis using ImagesPlus and Roger Clark's methodology are shown here. At low exposures, the graph departs from linearity as shown, due to truncation of the read noise. The minimum raw values include zero.
(http://bjanes.smugmug.com/Photography/Sensor-Analysis/D800e-Sensor/i-LVbqXD4/0/O/LowExpGraph.png)

Taking the linear portion of the data and plotting the variance of the noise against the ADU values gives the following with the gain and read noises shown. The results are actually slightly better for read noise than the Rawdigger analysis of the optical black area.
(http://bjanes.smugmug.com/Photography/Sensor-Analysis/D800e-Sensor/i-GS7R38F/1/O/varianceVSdn.png)

Your expert comments are welcome.

Correction June 9, 2013, see post (http://www.luminous-landscape.com/forum/index.php?topic=78677.msg636574#msg636574) below.

Bill



Title: Re: Optical Black RN versus Curve Fitted RN
Post by: Jack Hogan on May 30, 2013, 05:56:18 PM

Here is the raw histogram. 17K pixels are included for each channel, with 34K green values. The bell curve is skewed to the left. I don't know what this means.

If I read your heading correctly and the Exposure time is 2 seconds I guessing that could be the result of dark current/shot noise.

My data from a previous analysis using ImagesPlus and Roger Clark's methodology are shown here. At low exposures, the graph departs from linearity as shown, due to truncation of the read noise. The minimum raw values include zero.

Taking the linear portion of the data and plotting the variance of the noise against the ADU values gives the following with the gain and read noises shown. The results are actually slightly better for read noise than the Rawdigger analysis of the optical black area.
(http://bjanes.smugmug.com/Photography/Sensor-Analysis/D800e-Sensor/i-GS7R38F/1/O/varianceVSdn.png)

Your expert comments are welcome.

I assume you did not take two identical images and subracted them in order to get rid of PRNU which is proportional to signal, and that's probably why your curve above is not linear as the signal increases.  If so you may want to limit the curve fitting to data in the 0.5% to 5% range (that would be about ADU 80-800).

Jack
Title: Re: Optical Black RN versus Curve Fitted RN
Post by: bjanes on May 30, 2013, 07:06:48 PM
If I read your heading correctly and the Exposure time is 2 seconds I guessing that could be the result of dark current/shot noise.

Jack, that is a good pickup. I repeated the optical black with an exposure of 1/125 sec and the read noise matched the long method even better. Since this is from the optical black area, there is no shot noise as no light hit the sensels.

I assume you did not take two identical images and subracted them in order to get rid of PRNU which is proportional to signal, and that's probably why your curve above is not linear as the signal increases.  If so you may want to limit the curve fitting to data in the 0.5% to 5% range (that would be about ADU 80-800).

No, I took duplicate frames and subtracted them as per Roger's method. IHMO the data are about as linear as one could expect.

Bill
Title: Re: Optical Black RN versus Curve Fitted RN
Post by: Jack Hogan on May 31, 2013, 02:31:04 AM
Jack, that is a good pickup. I repeated the optical black with an exposure of 1/125 sec and the read noise matched the long method even better.

What about the shape?

Quote
Since this is from the optical black area, there is no shot noise as no light hit the sensels.

Yes, although I was referring to the black current (it comes with its own shot noise :-)

Quote
No, I took duplicate frames and subtracted them as per Roger's method. IHMO the data are about as linear as one could expect.

Bill

Yes, it looks like very good data, I am impressed.  I wonder why the slope does not change randomly, instead dicreasing monotonically as the signal increases.  Signal level has very little influence on igain k(e-/ADU) when read noise is negligible with respect to shot noise.  What would the slope be in the 0.5-5% range?

Jack
Title: Re: Optical Black RN versus Curve Fitted RN
Post by: BartvanderWolf on May 31, 2013, 03:07:59 AM
Jack, that is a good pickup. I repeated the optical black with an exposure of 1/125 sec and the read noise matched the long method even better. Since this is from the optical black area, there is no shot noise as no light hit the sensels.

Hi Bill,

That's why I prefer to use the shortest possible exposure time the camera allows, and make sure that the viewfinder is blocked as well, specifically for read noise analysis. I also prefer to use a body-cap instead of a lens and lenscap, to avoid any possible influence from the lens electronics (motor noise and wide aperture gain boost).

For camera's like those from Canon that record with the Raw data read-noise offset, that allows to use the effective image area for sampling the read noise, and for e.g. Nikon cameras that clip the read noise in the effective image area, that still allows to use the masked pixel area without significant influence from e.g. dark current, even to the edge (a normal exposure can spill some light (!) to the first or second sensels directly adjacent to the effective image sensels).

Quote
No, I took duplicate frames and subtracted them as per Roger's method. IHMO the data are about as linear as one could expect.

It's the preferred method with real sensors. That will remove yet another unknown before it can pollute the dataset and possibly lead to wrong assumptions.

Cheers,
Bart
Title: Re: Optical Black RN versus Curve Fitted RN
Post by: bjanes on May 31, 2013, 11:19:27 AM
That's why I prefer to use the shortest possible exposure time the camera allows, and make sure that the viewfinder is blocked as well, specifically for read noise analysis. I also prefer to use a body-cap instead of a lens and lenscap, to avoid any possible influence from the lens electronics (motor noise and wide aperture gain boost).

For camera's like those from Canon that record with the Raw data read-noise offset, that allows to use the effective image area for sampling the read noise, and for e.g. Nikon cameras that clip the read noise in the effective image area, that still allows to use the masked pixel area without significant influence from e.g. dark current, even to the edge (a normal exposure can spill some light (!) to the first or second sensels directly adjacent to the effective image sensels).

Bart,

Good suggestions, as usual for you. To investigate the effect of possible light spillage or blooming into the masked pixels, I took exposures at 1 second and 30 seconds of daylight sky at ISO 100 and examined the columns one by one going from the left to right of the masked areas.

The masked area is shown here.
(http://bjanes.smugmug.com/Photography/Sensor-Analysis/Long-Exposure/i-3NvF3TF/0/O/Img3_ScrCap.png)

Data for 1 second exposure
(http://bjanes.smugmug.com/Photography/Sensor-Analysis/Long-Exposure/i-Tp5P6ht/1/O/SamplesImg3.png)

Data for 30 sec exposure
(http://bjanes.smugmug.com/Photography/Sensor-Analysis/Long-Exposure/i-sCz2hVd/0/O/SamplesImg1.png)

No spillage into the masked pixels is apparent. The longer exposure does show some evidence of dark current (thermal noise), and the columns show heterogeneity for reasons that are not apparent to me.

Turning on long exposure noise reduction produces a different appearance for the masked pixels as shown.
(http://bjanes.smugmug.com/Photography/Sensor-Analysis/Long-Exposure/i-hpWMVBw/0/O/Img2_ScrCap.png)

Examination of the data demonstrates removal of thermal noise as expected, but the residual noise is even less than the read noise, and the bright pair of columns just to the right of the masked pixels is no longer present.
(http://bjanes.smugmug.com/Photography/Sensor-Analysis/Long-Exposure/i-pTwT6Sz/1/O/SamplesImg2.png)

My conclusion is that use of the masked pixels is a good method to determine read noise if one keeps the exposures less than 1 second or so, avoiding excessive thermal noise that contaminates the read noise. The result is the same as the more laborious regression method plotting the variance of noise against the ADU numbers and performing linear regression.

Regards,

Bill
Title: Re: Optical Black RN versus Curve Fitted RN
Post by: Jim Kasson on May 31, 2013, 11:33:53 AM
My conclusion is that use of the masked pixels is a good method to determine read noise if one keeps the exposures less than 1 second or so, avoiding excessive thermal noise that contaminates the read noise. The result is the same as the more laborious regression method plotting the variance of noise against the ADU numbers and performing linear regression.

Good work, Bill.

Jim
Title: Re: Optical Black RN versus Curve Fitted RN
Post by: Jack Hogan on June 01, 2013, 01:40:54 PM
My conclusion is that use of the masked pixels is a good method to determine read noise if one keeps the exposures less than 1 second or so, avoiding excessive thermal noise that contaminates the read noise. The result is the same as the more laborious regression method plotting the variance of noise against the ADU numbers and performing linear regression.

Regards,

Bill


Yes, excellent work as always Bill.  There is something that doesn't sit right with me in your ImagePlus graph, though:

(http://bjanes.smugmug.com/Photography/Sensor-Analysis/D800e-Sensor/i-GS7R38F/1/O/varianceVSdn.png)

Isn't this graph in DN, and therefore so would be the square root of the y intercept (read noise) at 4.13 DN?  Converting it to e- at a k of 3.46 e-/DN, the read noise would be 14.3 e-, which doesn't jibe with other data we have on the 'e'.   On the other hand, if one takes the values in your table and concentrates in the .5-5% range (or 84-838DN) one minimizes any residual influence of PRNU and RN in the curve, getting a better overall fit (R^2=0.9999).  These values are all referred to a 14 bit DN:

(http://i.imgur.com/Kigm9yM.jpg)

Gain (K) is 3.24 and read noise is 3.2e-, just short of 1 DN.  These values appear to me to be more accurate.  What do you think?

Jack
[EDIT] I've updated the table with a little more information.  Notes:

1. Shot Noise is calculated subtracting the newly found Read Noise from the combined Shot+Read Noise in quadrature all the way down.
2. The PRNU column (above the Shot Noise Dominated Box, where Read noise is immaterial, 1 vs 60^2) calculates PRNU subtracting the combined Shot+Read noise from Total Noise in quadrature.
3. Below the Box (where PRNU is much smaller than Shot and Read Noise) Read Noise in DN is calculated by using fitted gain k to determine shot noise from the square root of the signal in DN times k, then subracting it from the combined Shot+Read noise in quadrature and re-converting it to DN.  The average value for Read Noise so calculated is 1.08 DNs or 3.5 e-.  We are still an order of magnitude away from SNR=1,  RN is too small a component of Shot+Read, and Shot Noise and Read+Shot Noise are too close, so any small measurement variation results in quite a large variance of results with this method.  PRNU also skews the estimates, even if marginally.  I normally would not use these levels for this purpose, but it gives us an indication.
4. The gain column K is simply the ratio of Signal to Shot Noise in 1 squared.
Title: Re: ETTR vs ISO
Post by: bjanes on June 01, 2013, 03:50:56 PM
Jack,

Yes, I think you are correct about the read noise intercept. A major error on my part. I am traveling now and don't have access to my data. Will check when I get home.

Thanks,

Bill
Title: Re: ETTR vs ISO
Post by: Jack Hogan on June 01, 2013, 06:03:40 PM
Jack,

Yes, I think you are correct about the read noise intercept.

You know, thinking about it, the Variance/Signal plot seems to me to be an excellent and accurate way of determining gain k (as long as precautions are used to stay away from either end of the curve), but imho the y intercept method of determining Read Noise is very sensitive to changes in the slope: With k in hand, I think I would prefer to use the subtraction of Signal-determined shot noise in quadrature from the combined Shot+Read noise around signals of less than 5 DN.  I wonder if zero blocking might influence the statistics there, and if so whether that could be compensated for with a little math.

Jack
Title: Re: ETTR vs ISO
Post by: bjanes on June 05, 2013, 10:59:32 AM
You know, thinking about it, the Variance/Signal plot seems to me to be an excellent and accurate way of determining gain k (as long as precautions are used to stay away from either end of the curve), but imho the y intercept method of determining Read Noise is very sensitive to changes in the slope: With k in hand, I think I would prefer to use the subtraction of Signal-determined shot noise in quadrature from the combined Shot+Read noise around signals of less than 5 DN.  I wonder if zero blocking might influence the statistics there, and if so whether that could be compensated for with a little math.

Jack,

You are correct that the slope of the regression line is critical in determining read noise. Looking at my data, I see that there is non-linearity in the plot of variance vs data number at higher exposures. Since the regression uses least squares, the high DNs have a disproportionate influence on the regression line. Using the range of DNs from 2654 to 13, I get more reasonable data that agrees with the Rawdigger determined read noise and also with Bill Claff's data. When selecting the lower range for the DNs, it is important to exclude those where the read noise is clipped.

I'm not certain as how to apply your method. The variance of the subtracted duplicate images includes shot noise and read noise adding in quadrature. To determine the shot noise by subtracting the read noise in quadrature, one needs the read noise. Re-reading Emil's post (http://theory.uchicago.edu/~ejm/pix/20d/tests/noise/noise-p2.html#readandshot) on noise, I finally see how the regression line is determined:

 Thus, read and photon shot noise contribute as:

 N2 = R2 + S/g
, where N is the total noise, R is the read noise, S is the signal, and g is the gain.

Regards,

Bill
Title: Re: ETTR vs ISO
Post by: bjanes on June 05, 2013, 11:21:11 AM
Yes. Exposure settings for aperture and shutter are identical, the only difference is the ISO setting. And the differences in the amount of noise is significant as one would expect with ETTR (optimal exposure for raw).

I don't care about the Histogram! But what I see is a vast difference in noise whereby the ONLY setting change is ISO. So it's obviously affecting the degree of noise.

You saying increasing the ISO in this case is to reduce the read noise, this isn't also ETTR? Seems the number of photons collected should be the same since what he's calling exposure is fixed in both examples. The results however are clear in terms of the differences in noise and I understand that not all camera sensors respond as this Canon does.

Roger Clark has a nice post (http://www.clarkvision.com/articles/iso/) discussing ISO, exposure, and ETTR that clears up much confusion. He regards exposure in the usual sense as the number of photons collected. If one varies the ISO setting of the camera in manual mode and uses the same shutter speed and aperture, exposure is not affected. However, if one is in auto mode, changing the ISO does affect the exposure, since the shutter speed and/or the aperture are affected. He discusses ETTR for Canon cameras and concludes:

"ISO is a relative gain, varying by camera and has nothing to do with sensitivity or true exposure. ISO is simply a post sensor gain and digitization range."


True exposure is discussed in an add on post (http://www.clarkvision.com/articles/exposure/).

Regards,

Bill
Title: Re: ETTR vs ISO
Post by: Jack Hogan on June 09, 2013, 05:44:17 AM
Looking at my data, I see that there is non-linearity in the plot of variance vs data number at higher exposures. Since the regression uses least squares, the high DNs have a disproportionate influence on the regression line.

Hi Bill,

Yes, and if you look closely, you will see that there is also a non-linearity at the lowest exposures in the Read+Shot noise variance plotted, making the slope deviate from the ideal.  Hence the suggestion to not use points below about 0.5% of full scale lest the slope of the fitted line be polluted and bent downwards, resulting in wrong values for k and RN.  I believe the ideal plot needs to be of shot noise variance only versus mean signal in order to determine k.  When plotting the same data in the form of SNR vs DN [noise in this case being Read+Shot] excluding data below 0.5% of FS and above 5% (or tighter), the slope of the resulting line needs to be -3dB/decade.  If it isn't, it is polluted.

Quote
I'm not certain as how to apply your method. The variance of the subtracted duplicate images includes shot noise and read noise adding in quadrature. To determine the shot noise by subtracting the read noise in quadrature, one needs the read noise.

Yes, and we have both (RN from the curve fitting variables), so we can simply calculate shot noise by:

Shot Noise = sqrt[[read+shot)^2 - RN^2]

Cheers,
Jack