Luminous Landscape Forum
Equipment & Techniques => Digital Cameras & Shooting Techniques => Topic started by: rogoldboy on June 15, 2015, 09:52:38 pm
-
Hello All:
I am fairly well versed with Histograms and ETTR.
My basic understanding of ISO invariance is that you can have an exposure set with ISO 3200, or you can use the same shutter speed and aperture but reduce the ISO to 100 and then push it 5 stops in Camera RAW. Each image, made differently, will have roughly the same exposure value, with the +5 Camera RAW image exhibiting better noise qualities.
My question is “How does this then affect the use of ETTR?” ???
Thanks for any and all insight,
rogoldboy
-
It depends on how the ISO gain is implemented in the camera. As a really general principle, analogue gain via in-camera iso is going to be more accurate (i.e. better signal to noise ratio) than using post-processing digital gain. But there's other variables that come into play. I don't profess to understand it well (as I haven't bothered to look into it yet) but there is the concept of an ISO-less camera - or at the least a range of iso's that don't follow the above principle. The others with more technical knowledge will post with the details. As far as I understand it, some cameras with the Sony sensors - like the Nikon D800/810, Sony A, and some of the Pentax's have ISO-less performance. That means, in the case of my camera the D810, that there is little to no benefit to raising ISO over about 800 ISO. That is, you can get the same effect from a digital push as you can from in-camera ISO gain. Why that is the case conceptually, I'd like to know. It seems like a waste of opportunity to get a better SNR from analogue gain as opposed to digital gain.
-
I am fairly well versed with Histograms and ETTR... then push it 5 stops in Camera RAW...
so are you aware about the difference between Process 2010 and Process 2013 when "pushing it" in "Camera RAW", right ?
-
Hello All:
I am fairly well versed with Histograms and ETTR.
My basic understanding of ISO invariance is that you can have an exposure set with ISO 3200, or you can use the same shutter speed and aperture but reduce the ISO to 100 and then push it 5 stops in Camera RAW. Each image, made differently, will have roughly the same exposure value, with the +5 Camera RAW image exhibiting better noise qualities.
My question is “How does this then affect the use of ETTR?” ???
Thanks for any and all insight,
rogoldboy
As far as the term ISO-less is concerned, this is my understanding :-
Sensor Read noise can be introduced prior to the ISO amplifier or after the amplifier in the A/D converter circuitry. The noise introduced before the amp is amplified when the gain is increased as a result of an increase in ISO setting but the A/D noise is not.
With Canon sensors, it appears that at base ISO, A/D converter noise is a significant contributor to the total noise in the final digital signal. As ISO is increased, the noise introduced before the amplifier is increasingly amplified until at some point, it is the dominating contributor and the contribution from the A/D is insignificant. Once you reach this point, the sensor becomes virtually ISO-less as it makes little difference whether you increase the “exposure” digitally or in the amplifier. Before this point however, increasing gain digitally increases both sources of noise and hence is less effective than increasing ISO gain.
On the other hand, the latest Sony/Nikon sensors seem to have very little A/D noise (presumably because of the use of integrated column A/D’s.) So for these sensors, the noise introduced before the amplifier is always the dominant noise for any ISO. These cameras are therefore virtually ISO-less for the entire ISO range as it makes little difference whether the ‘exposure” is adjusted digitally or in the column amplifiers.
ETTR is really a separate issue and is more about making use of the full dynamic range of the camera. It may involve more exposure than is necessary for a given scene which in turn allows the exposure (and also the noise) to be wound down in pp.
Dave
-
My question is “How does this then affect the use of ETTR?” ???
Looking at ETTR as a principle for exposure, it doesn't.
ETTR is about pushing exposure as high as possible without actually clipping to maintain maximum amount of image information and this principle holds no matter what the ISO. As you increase the ISO you could argue that ETTR becomes even more important because it would lead to less visible noise than if the histogram were more central.
What I suspect you meant to ask is 'given that ETTR is meant to maximise image information how does increasing ISO affect the information you collect with ETTR': higher ISO will always lead to smearing of detail and increase in noise even if it is not immediately visible at the chosen viewing size/medium.
-
My basic understanding of ISO invariance is that you can have an exposure set with ISO 3200, or you can use the same shutter speed and aperture but reduce the ISO to 100 and then push it 5 stops in Camera RAW. Each image, made differently, will have roughly the same exposure value, with the +5 Camera RAW image exhibiting better noise qualities.
My question is “How does this then affect the use of ETTR?” ???
Well, for maximum captured IQ you would ETTR at base ISO , if possible. You would simply maximize Exposure with the camera at base ISO based on your artistic constraints (say 1/400, f/2.8 for indoor sports, ensuring that no desirable highlights are clipped). Then shoot away with abandon, forgetting about ISO. Later correct for pleasing brightness during raw conversion.
The effects of the two approaches on IQ (ISO in-camera vs push in post) are shown here (http://www.strollswithmydog.com/iq-raising-iso-vs-pushing/) for a well designed (nearly) ISOless camera.
Jack
-
ETTR is really a separate issue and is more about making use of the full dynamic range of the camera. It may involve more exposure than is necessary for a given scene which in turn allows the exposure (and also the noise) to be wound down in pp.
Correct, ETTR is basically used at the camera's native ISO, i.e. the lowest ISO that thus allows a maximum of DR. The low ISO reduces the risk of clipping at the saturation point, and the optimal exposure level will be just shy of that clipping point, based on the lightest tones that need to have accurate texture and color. The shadows will then fall where they may, with as much exposure as possible without clipping whites. This typically produces the highest S/N ratio possible for all tones in a single exposure, and offers the best basis for post-processing (with the possible exception of hue twists introduced by some profiles when changing the Raw Conversion 'Exposure' control).
So to answer the OP's question, essentially ETTR is only used at native ISO. It's all about collecting as many photons as can possibly fit in the potential wells of the sensor, without saturating highlights that matter. Collecting more photons, results in lower shot noise (at all tonalities).
With boosted ISO one just has to avoid that the amplified signal doesn't clip important highlights, but DR will be reduced due to lack of photons.
Cheers,
Bart
-
The principles still apply, though, at any ISO. What matters is exposure (i.e. amount of light hitting the sensor). So if you have the leeway, then a greater exposure is better than less exposure, whatever ISO you happen to select. Or to put it another way, you can still employ the principles of ETTR away from base ISO, particularly so in cameras that don't have ISO-less like operation.
-
The principles still apply, though, at any ISO. What matters is exposure (i.e. amount of light hitting the sensor). So if you have the leeway, then a greater exposure is better than less exposure, whatever ISO you happen to select. Or to put it another way, you can still employ the principles of ETTR away from base ISO, particularly so in cameras that don't have ISO-less like operation.
Exactly and as you pointed out first, in terms of ISO/ETTR, depends on the camera. Example of a Canon 5DMII:
(http://digitaldog.net/files/100vs800iso.jpg)
ISO 100, more noise than ISO 800.
-
Yeah, that's a marked difference. I just wish I understood the technical aspects of why that no longer applies with some sensors. Actually, on second thoughts, I don't really need to know. I'm happy that there are a resourceful bunch of geeks here who keep us updated on what the best approach is to use for each camera/sensor. :)
I remember a series of graphs that Guillermo posted years ago for the 5D that convinced me to skip ISO200 altogether, as it showed virtually no noise improvement over using ISO400. It's all handy information.
-
Yeah, that's a marked difference. I just wish I understood the technical aspects of why that no longer applies with some sensors. Actually, on second thoughts, I don't really need to know. I'm happy that there are a resourceful bunch of geeks here who keep us updated on what the best approach is to use for each camera/sensor. :)
I remember a series of graphs that Guillermo posted years ago for the 5D that convinced me to skip ISO200 altogether, as it showed virtually no noise improvement over using ISO400. It's all handy information.
http://home.comcast.net/~nikond70/Charts/PDR.htm
-
Yeah, that's a marked difference. I just wish I understood the technical aspects of why that no longer applies with some sensors. Actually, on second thoughts, I don't really need to know.
Too late :) A clumsy attempt (http://www.strollswithmydog.com/information-transfer-iso-invariant/) at such an explanation. In a nutshell: ISOless systems are able to store into the raw data (most) all the scene information collected by the sensor's photosites in one go.
Jack
-
Exactly and as you pointed out first, in terms of ISO/ETTR, depends on the camera. Example of a Canon 5DMII:
Which demonstrates that not only the OP is getting confused by the wrong use of the ETTR concept. :(
ISO 100, more noise than ISO 800.
The ISO 100 shot received 8x less exposure than what the exposure meter recommended for ISO 800. And just because the Canon architecture uses pre- and post amplification, there can be a benefit to increase the ISO setting if you deliberately underexpose (which is not ETTR), but not nearly as much as getting the benefit from correct exposure (=ETTR).
You can prove it to yourself by shooting the same shot with 8x as long an exposure time (which apparently does not clip this crop of the image you picked), and compare it with the ISO 800 shot which then received 8x less exposure and got amplified 8x after capture.
Nothing beats real signal in the form of Photons. That's what ETTR is about.
So the comparison should have been between an ISO 100 shot optimally exposed (ETTR), compared to an underexposed shot boosted by ISO (which on Canons can benefit from cranking up the ISO, other cameras benefit less due to different architecture but may already be better at low noise underexposure).
Cheers,
Bart
-
Which demonstrates that not only the OP is getting confused by the wrong use of the ETTR concept. :(
Why would he be getting confused? This thread and the original question are about ETTR on ISO invariant systems, per the pretty title. The 5DMII is not ISO invariant as the relative images clearly show.
Jack
-
Why would he be getting confused? This thread and his question are about ETTR on ISO invariant systems, per the pretty clear title.
Hi Jack,
The confusion is clear when high (or not native) ISO is mentioned in the same sentence with ETTR. They are completely different things. High(er) ISO is all about underexposure, e.g. to achieve a shutterspeed that stops camera shake or subject motion that would have occured at an ETTR exposure level. ETTR on the other hand is all about not underexposing.
There is a reason that the Dynamic range of e.g. the 5D Mark III that Andrew used for a comparison is highest at native ISO (http://www.dxomark.com/Cameras/Compare/Side-by-side/Canon-EOS-5D-Mark-III___795), but you know better than most who have not done the math that more photons are hard to beat when signal to noise is concerned. Photon shot noise statistics (Poisson noise) produce a better increase of S/N with ETTR, than with underexposure and an ISO boost of the gain (as in Canons) or with a Raw converter's Exposure control. It's physics, which as usual is hard to beat (if at all possible in a single exposure).
Cheers,
Bart
-
Why would he be getting confused? This thread and the original question are about ETTR on ISO invariant systems, per the pretty title.
Unlike Bart, I will not speak for the OP as he has. But I agree with you. The image and the results speak for themselves, Bart speaks for others. Enough said ::)
-
Unlike Bart, I will not speak for the OP as he has. But I agree with you. The image and the results speak for themselves, Bart speaks for others. Enough said ::)
Correction, I speak for myself. If you disagree with my explanation of the effects of Photon shot noise statistics on image quality, feel free to proof me wrong, or ignore me. I'm fine with either.
Cheers,
Bart
-
Correction, I speak for myself.
This wasn't you:
Which demonstrates that not only the OP is getting confused by the wrong use of the ETTR concept.
Let's examine what the OP stated in his first post:
am fairly well versed with Histograms and ETTR.
Sure seems you're placing confusion upon him and speaking for more than yourself sir.
-
I have seen many discussions on ETTR on many different sites and part of the problem is that people add their own twist on what it actually means. In his original 2003 article and the update in 2011, Michael Reichmann shows images shot at ISO200 and ISO 400 and does not even refer to base ISO yet now some people are claiming it is an important part of the definition.
I totally agree that if you are taking the trouble to ETTR to maximise informatoin then why not go the whole hog and use base ISO, but that is NOT the same as saying base ISO is a critical part of ETTR.
Interestingly, in this web page
http://photo.net/digital-darkroom-forum/00Vy86
Jeff Schewe contributes the following:
ETTR has NOTHING TO DO with the amount of levels in the shadows...it has nothing to do with the base ISO (vs altered ISO) is has EVERYTHING to do with the fact that more photons equals less perceptible noise (which means a better signal to noise ratio).
...
More photons=better noise
That's the basis of ETTR, has nothing to do with "shadow bits.
You should also test out the fact that HTTR can also actually benefit when "modest" increases of ISO are used...with today's cameras you can prolly go upwards of ISO 800 and STILL take advantage of ETTR if the scene contrast range is below the sensor...
-
Sure seems you're placing confusion upon him and speaking for more than yourself sir.
I'm not a native English speaker, but I believe that "getting confused" is not the same as "he is confused", or "he is becoming confused" and certainly not the same as "he says he is confused". But feel free to correct my interpretation of the English language, I'm eager to learn.
Getting confused can IMHO mean that other people are (in my estimation) presenting confusing information (e.g. by only telling part of the story), which might lead to confusion. Not necessarily to the OP, but there are others reading these fora who would like to learn.
Cheers,
Bart
-
I totally agree that if you are taking the trouble to ETTR to maximise informatoin then why not go the whole hog and use base ISO, but that is NOT the same as saying base ISO is a critical part of ETTR.
I hear you spidermike, but in this thread the context is ISO invariant cameras. For a given exposure they maximise scene information captured at base ISO. So with an ISOless camera you should attempt to ETTR at base ISO.
Jack
-
I'm not a native English speaker, but I believe that "getting confused" is not the same as "he is confused", or "he is becoming confused" and certainly not the same as "he says he is confused". But feel free to correct my interpretation of the English language, I'm eager to learn.
You don't have to be a native of the English language to see that the first person to use the word 'confused' was you. In relationship to the OP who isn't confused but asking for clarification.
-
You don't have to be a native of the English language to see that the first person to use the word 'confused' was you. In relationship to the OP who isn't confused but asking for clarification.
And that makes me speak for others? Now you've got me confused ...
Cheers,
Bart
-
And that makes me speak for others? Now you've got me confused ...
Yes that much is clear.
Do I really have to connect the dots for you? You brought up the OP and the term about him being confused when he isn't confused and by suggesting he is, you're doing him a disservice.
-
So with an ISOless camera you should attempt to ETTR at base ISO.
but first again you also need to know what your converter (along with camera profiles, specifically in ACR/LR case) is doing behind the scenes with exposure
-
Yes that much is clear.
Do I really have to connect the dots for you? You brought up the OP and the term about him being confused when he isn't confused and by suggesting he is, you're doing him a disservice.
There you go again. I never suggested that. What I did suggest is that there were half truths being told, which might confuse anybody (and I do know the distinction between anybody and everybody, before you start an OT attack about that).
A half truth like ISO 800 being better than ISO 100, without pointing out the non-ETTR exposure and the non-ISO-less camera used as an example. Oh wait, it was you who brought up that example. Could that explain the diversion tactic by suggesting that I pretended to speak for the OP? Never mind, it is clear to me.
Others have tried to pull off such an 'example' at other occasions, and I've responded to those as well with factual information. So don't take it personal, unless you want to, feel free to do so. We can then discuss what concept you were trying to get across with that example, and how it was relevant to the OPs question which you seem very concerned about. Looking forward to an on-topic discussion.
Cheers,
Bart
-
but first again you also need to know what your converter (along with camera profiles, specifically in ACR/LR case) is doing behind the scenes with exposure
Indeed, there is a risk of Hue twists getting in the way in the ACR/LR case. Others have reported concerns about Capture One as well, but I've not experienced that myself. Boosting even to three or four stops over an underexposure doesn't change the colorbalance a bit, maybe some isolated colors I have missed?
Cheers,
Bart
-
There you go again. I never suggested that.
You indeed didn't suggest, you said so.
Even Jack asked you rather clearly: Why would he be getting confused?
So I'm not alone.
We're getting OT like the other discussion you have posted within today (Adobe conspiracy theories). I don't have the time to address your assertions.
-
but first again you also need to know what your converter (along with camera profiles, specifically in ACR/LR case) is doing behind the scenes with exposure
If by that you mean that you should not be using LR for anything but a quick conversion, I agree :) On the other hand I am just referring to capturing the best image quality possible. How you render it to be most pleasing is another matter.
Jack
-
If by that you mean that you should not be using LR for anything but a quick conversion, I agree :) On the other hand I am just referring to capturing the best image quality possible. How you render it to be most pleasing is another matter.
Jack
what I am saying is that for example ACR/LR not only might have exposure correction dependent twists in some camera profiles, but they might apply different hidden exposure corrections (not to mention again that process 2012 exposure correction is not linear in ACR/LR) based on the nominal ISO (different ISO = different corrections) for the same camera model... so any ETTR application shall take into consideration the tools that deal with raw file afterwards before you decide how you are going to expose - like the ISO it is not part of the exposure, but it is a part of your decision how to expose (before it starts) because it affects the processing when it (exposure) ends.
-
I just had another read of Michael's article of 2011 on ETTR and it re- inforced my view that you only get benefit from ETTR if the dynamic range of the scene is less than that of the camera. In this situation, ETTR results in over exposure which is then reduced in pp thus reducing the noise. it is sort of equivalent to shooting with a lower ISO than base ISO.
You will therefore get most benefit from ETTR if you shoot at base ISO where the dynamic range is highest.
An ISO less camera will generally have better dynamic range at base ISO than one that isn't ISO less, for a given pixel size. Thus it should be more likely that you can get some benefit from ETTR with an ISO less camera at low ISO.
Dave
-
what I am saying is that for example ACR/LR not only might have exposure correction dependent twists in some camera profiles, but they might apply different hidden exposure corrections (not to mention again that process 2012 exposure correction is not linear in ACR/LR) based on the nominal ISO (different ISO = different corrections) for the same camera model... so any ETTR application shall take into consideration the tools that deal with raw file afterwards before you decide how you are going to expose - like the ISO it is not part of the exposure, but it is a part of your decision how to expose (before it starts) because it affects the processing when it (exposure) ends.
I hear what you are saying, but what we are talking about here is really just 'brightening' by linear multiplication before applying all the other adjustments - pretty basic stuff, as described in the earlier link (http://www.strollswithmydog.com/iq-raising-iso-vs-pushing/): most well behaved converters do this as a matter of course (I am familiar with Nikon NX converters, RT, DCRAW and ACR). The only converter I know that insists on introducing all sorts of secret camera specific 'twists' is Adobe's, presumably in order to keep things as simple as possible for its users. Nothing wrong with that. But until they fix it perhaps it should not be used when attempting to achieve advanced performance.
If this sounds harsh, consider that properly implemented ETTR is the only 'correct' exposure when the objective is maximum IQ. Many cameras are able to approximately meter this ideal exposure in-camera today (e.g. Nikon's Matrix/Highlight-Weighted metering with Auto ADL). ACR/LR on the other hand incorrectly treat those perfectly exposed images as underexposed/overexposed, possibly introducing weird 'twists' while rendering them. A perfect case of the software not keeping up with advances in the hardware.
Jack
-
You indeed didn't suggest, you said so.
Even Jack asked you rather clearly: Why would he be getting confused?
To which I gave a technical answer, and a link to an external source, to explain my view.
So I'm not alone.
?? I'll let Jack speak for himself if he feels inclined.
I do not need people to agree with me, although they might, if my explanation makes sense to them, or they may add another viewpoint that makes sense to me. One can only learn and benefit from such an exchange of views, or at worst agree to disagree.
We're getting OT like the other discussion you have posted within today (Adobe conspiracy theories).
Ah, that's another allegation you came up with (http://forum.luminous-landscape.com/index.php?topic=101244.msg830223#msg830223), in reaction to a direct remark I made to another poster with a smiley. So it is that personal for you, that remarks I make to other people in other threads affect your reaction in this thread about a different subject? Amazing.
I don't have the time to address your assertions.
Nobody is forcing you (or me) to respond to anything, assertions or not, so no problem if you do, no problem if you don't. When I'm asked a direct question (especially when it helps a discussion to progress) I'll try to answer, if time allows. So without further questions, we'll leave it at that then.
Cheers,
Bart
-
what I am saying is that for example ACR/LR not only might have exposure correction dependent twists in some camera profiles, but they might apply different hidden exposure corrections (not to mention again that process 2012 exposure correction is not linear in ACR/LR) based on the nominal ISO (different ISO = different corrections) for the same camera model... so any ETTR application shall take into consideration the tools that deal with raw file afterwards before you decide how you are going to expose - like the ISO it is not part of the exposure, but it is a part of your decision how to expose (before it starts) because it affects the processing when it (exposure) ends.
This may be a bit too much of a diversion from the thread topic, but I feel it is better addressed here than starting another thread:
May I ask for some clarification?
Are you saying that the (non-linear) behavior of exposure correction in LR/ACR varies depending on the ISO value used for the exposure (as determined from the exif data)? Alternatively is the behavior dependent on the image values themselves (as influenced by exposure as well as ISO)?
In either case, how does this information influence your decisions prior to exposure?
I can see how it might influence one's choice of the camera profile used in that raw conversion software, or even the choice of raw conversion software itself.
-
Are you saying that the (non-linear) behavior of exposure correction in LR/ACR
it is what process 2012 brings... so exposure correction for +5 stops in ACR/LR is not exactly the same as dialing EV+5 (including increasing ISO by 5 stops) in camera and feeding that to ACR/LR... the closer to clipping you are the more compression exposure correction in PV2012 does... granted with "ETTR" you most probably will be going backwards (pulling, not pushing), that helps (twists if present in profile still matter)
varies depending on the ISO
hidden exposure correction done by ACR/LR code might be different for different nominal ISOs, that part depends on ISO only (but then see above about non linear nature of exposure correction in PV2012) - for example with Fuji raw files.
-
what I am saying is that for example ACR/LR not only might have exposure correction dependent twists in some camera profiles, but they might apply different hidden exposure corrections (not to mention again that process 2012 exposure correction is not linear in ACR/LR) based on the nominal ISO (different ISO = different corrections) for the same camera model... so any ETTR application shall take into consideration the tools that deal with raw file afterwards before you decide how you are going to expose - like the ISO it is not part of the exposure, but it is a part of your decision how to expose (before it starts) because it affects the processing when it (exposure) ends.
Before piling on to Adobe, one should recognize that there are ways to deal with the Adobe hue twists. Sandy McGuffog (http://dcptool.sourceforge.net/Hue%20Twists.html) is the expert here and the referenced link discusses hue twists and how to remove them with his dcpTool. The DNG BaselineExposure offset that Adobe uses is not really hidden. It is in the EXIF of the DNG file and can be read with EXIF readers. Rawdigger can also read the offset. For my Nikon D800e, the BaselineExposure is +0.35 EV. One can get a good measure of the offset by exposing a uniform white or gray card according to the light meter and determine the raw saturation. This offset does complicate the evaluation of the status of the raw channels with the ACR histogram, but RawDigger is the proper tool for that purpose. FastRawViewer also works well for this.
Highlight recovery with PV2012 is nonlinear--it does roll off the highlights, but this is usually what you want with clipped highlights. If you need more linearity, you can fall back to PV2010.
Bill
-
Jack: "The effects of the two approaches on IQ (ISO in-camera vs push in post) are shown here for a well designed (nearly) ISOless camera."
thanks for that link, it is making for interesting reading.
-
So with an ISOless camera you should attempt to ETTR at base ISO.
Jack
Now I'm getting a bit confused, Jack. ;)
Surely with all cameras, whether they are ISO invariant or not, one should attempt to apply ETTR at base ISO for best results, if the required shutter speed and aperture allow it.
Likewise, with all cameras, whether they are ISO invariant or not, one needs to apply the ETTR process regardless of the ISO setting chosen, in order to get the best results, the best results being the lowest noise in shadows and mid-tones, without blowing wanted highlights.
For example, if I'm using a D7000 or D7200, which I would say are pretty close to being ISO invariant cameras, and I decide to use ISO 800, not only because I need a fast shutter speed, or because the light is not good, but also because I want to clearly see what I've shot on the camera's LCD screen, and/or share the images with others in the field, then I still need to apply the ETTR process at ISO 800 in order to achieve the lowest noise and the best review image.
In other words, any exposure which does not push the histogram to the right, whatever the chosen ISO, will result in more image noise.
Where's the confusion? The fact that one might not need to use a high ISO with an ISOless camera, if one doesn't feel the need to be able to clearly view the shots taken on the camera's LCD screen, does not mean that the ETTR principle does not apply if one does choose to raise ISO, rather than underexpose at base ISO.
However, that is not to deny that the ISOless camera has clear advantages. But that is a separate issue.
-
Before piling on to Adobe, one should recognize that there are ways to deal with the Adobe hue twists. Sandy McGuffog (http://dcptool.sourceforge.net/Hue%20Twists.html) is the expert here and the referenced link discusses hue twists and how to remove them with his dcpTool. The DNG BaselineExposure offset that Adobe uses is not really hidden. It is in the EXIF of the DNG file and can be read with EXIF readers. Rawdigger can also read the offset. For my Nikon D800e, the BaselineExposure is +0.35 EV. One can get a good measure of the offset by exposing a uniform white or gray card according to the light meter and determine the raw saturation. This offset does complicate the evaluation of the status of the raw channels with the ACR histogram, but RawDigger is the proper tool for that purpose. FastRawViewer also works well for this.
Bill
Hi Bill,
Indeed, the hue twists are easy to avoid. In fact most profiles (dcp) that one creates with the dng profile editor are "untwisted". See my remarks on this site (https://www.hansvaneijsden.com/colorchecker-perfect-skin-colors/).
Interested in your comments about Baseline Exposure Offset being visible in the exif data of the dng file. When I look, I too see a value of +0.35 for my D800. However I am a bit curious about how that value got there and how it is used by the raw converters. When I look at a NEF file (same image) that value is not found in the exif. Or, at least I cannot find it.
-
Before piling on to Adobe
nobody does - however it is worth reminding what it is... because people tend to talk about what ETTR brings to raw data, but forget that raw data will be dealt with with a raw converter and there you have its own share of nuances....
The DNG BaselineExposure offset that Adobe uses is not really hidden. It is in the EXIF of the DNG file and can be read with EXIF readers.
yes, but that means you need 1) convert to DNG 2) know to check for it
If you need more linearity, you can fall back to PV2010.
and then you might end up w/o other useful features that PV2012 brings
---
again this was not against ACR/LR but to remind the OP that his "then push it 5 stops in Camera RAW" must be with fine print attached
-
However I am a bit curious about how that value got there and how it is used by the raw converters.
as it was noted it is hardcoded in ACR/LR code... other converters, if you are talking about NEF files, either do not have any hidden expo-corrections (like RPP for example) or deal with that by increasing brightness through some tone curves/commands in tone curve files (C1) or LUTs in whatever camera profiles they use or may be even have something like Adobe does... you need to test your specific converter by comparing with known tools that don't do anything hidden
-
the dng profile editor
now with DCamProf from Torger this tools is simply obsolete for dcp profile creation, may be only useful as a visual profile editor...
-
Which demonstrates that not only the OP is getting confused by the wrong use of the ETTR concept. :(
The ISO 100 shot received 8x less exposure than what the exposure meter recommended for ISO 800. And just because the Canon architecture uses pre- and post amplification, there can be a benefit to increase the ISO setting if you deliberately underexpose (which is not ETTR), but not nearly as much as getting the benefit from correct exposure (=ETTR).
You can prove it to yourself by shooting the same shot with 8x as long an exposure time (which apparently does not clip this crop of the image you picked), and compare it with the ISO 800 shot which then received 8x less exposure and got amplified 8x after capture.
Nothing beats real signal in the form of Photons. That's what ETTR is about.
I'm sure others have already addressed this, but you contradict yourself here. You say that the ISO100 image received 8x less exposure, therefore he doesn't understand ETTR. But at the end here you say that ETTR is all about real signal in the form of photons. Which is exactly what AR and I have said. Your final pronouncement is correct. The former is not. Increasing ISO doesn't increase "real signal in the form of photons".
-
I'm sure others have already addressed this, but you contradict yourself here. You say that the ISO100 image received 8x less exposure, therefore he doesn't understand ETTR. But at the end here you say that ETTR is all about real signal in the form of photons. Which is exactly what AR and I have said. Your final pronouncement is correct. The former is not. Increasing ISO doesn't increase "real signal in the form of photons".
First, I don't interpret Andrew's post as an example of ETTR. The number of photons captured is the same in both images. It is simply a demonstration of a situation where a predetermined aperture and shutter speed result in an underexposed image at base ISO.** In such a case one has two options:
(1) Boost the ISO or
(2) Amplify the signal in post processing.
It is clear that, for the camera used for this demo, the first option results in a much less noisy image.
I think Bart's objection stems from the fact that Andrew's post could hardly be considered an example of how to employ ETTR, and I am guessing that AR himself would not describe it as such. (However, I should hasten to add, I am not empowered to speak for either of these gentlemen.)
**In fact the image is underexposed at ANY ISO. I do believe that, for the purpose of such discussions, one should confine the use of the term "exposure" to mean the amount of light (photons) reaching the sensor; thus influenced only by aperture and shutter speed. ISO then is simply an (in-camera) electronic amplification of the signal generated by exposure.
-
Andrew's post wasn't supposed to be an example of ETTR. Anyone who thinks that needs to upgrade their reading firmware. ;)
-
I'm sure others have already addressed this, but you contradict yourself here. You say that the ISO100 image received 8x less exposure, ...
Hi Bernie,
The confusing part of the example is that both shots received the same exposure, check the exposure time and the shutterspeed. I said that it received 8x less exposure than the exposure meter would have suggested (if it had been metered at ISO 100), but it was instead exposed as if ISO 800 was used for metering.
That means that if we assume that the ISO 800 shot did not result in clipping (histogram moved to the right end without introduction of clipping), the ISO 100 shot therefore received 8x less exposure than an ETTR shot would have. Therefore it doesn't demonstrate ETTR. What it does demonstrate is something completely different, namely ISO variant behavior. That behavior is typical mainly for Canon cameras but not for most others, like Sony sensor based ones from Sony and Nikon.
But at the end here you say that ETTR is all about real signal in the form of photons. Which is exactly what AR and I have said. Your final pronouncement is correct. The former is not. Increasing ISO doesn't increase "real signal in the form of photons".
Indeed it doesn't, although increasing the ISO on the ISO variant Canon example does help to improve the S/N ratio of an underexposed non-ETTR shot. But I also said that using more photons for the exposure is always a better approach, provided that we have the luxury of absence of camera or subject motion that needs to be frozen with a short exposure time. When we must prioritize shutter speed to get the shot, we will not be using ETTR. ETTR is about optimizing signal quality, and collecting as many photons as possible without clipping important highlight detail is the best way. Base ISO will then allow to maximize the DR.
Cheers,
Bart
-
Now I'm getting a bit confused, Jack. ;)
Surely with all cameras, whether they are ISO invariant or not, one should attempt to apply ETTR at base ISO for best results, if the required shutter speed and aperture allow it.
Likewise, with all cameras, whether they are ISO invariant or not, one needs to apply the ETTR process regardless of the ISO setting chosen, in order to get the best results, the best results being the lowest noise in shadows and mid-tones, without blowing wanted highlights.
For example, if I'm using a D7000 or D7200, which I would say are pretty close to being ISO invariant cameras, and I decide to use ISO 800, not only because I need a fast shutter speed, or because the light is not good, but also because I want to clearly see what I've shot on the camera's LCD screen, and/or share the images with others in the field, then I still need to apply the ETTR process at ISO 800 in order to achieve the lowest noise and the best review image.
In other words, any exposure which does not push the histogram to the right, whatever the chosen ISO, will result in more image noise.
Where's the confusion? The fact that one might not need to use a high ISO with an ISOless camera, if one doesn't feel the need to be able to clearly view the shots taken on the camera's LCD screen, does not mean that the ETTR principle does not apply if one does choose to raise ISO, rather than underexpose at base ISO.
However, that is not to deny that the ISOless camera has clear advantages. But that is a separate issue.
Well put Ray. I would add that ETTRing an ISOless camera at base ISO the IQ maximizing photographer is guaranteed the best IQ possible from the capture. Not necessarily so as ISO is raised.
Jack
-
Andrew's post wasn't supposed to be an example of ETTR. Anyone who thinks that needs to upgrade their reading firmware. ;)
Hi Bernie,
I interpreted Andrew's post (http://forum.luminous-landscape.com/index.php?topic=101242.msg830204#msg830204) as agreeing with you that at a higher ISO one can still use ETTR. That is what I disagree with, and that the example used an ISO variant camera was not what the OP seemed to ask.
At best, when you boost the ISO, the histogram is the result of amplifying the underexposed signal for a more normal looking thumbnail or OOC JPEG. It is not EXPOSING to the right (ETTR), but more post-exposure AMPLIFYING (the resulting histogram) to the right. There is a large difference in image quality.
Image quality can only be optimized by using more photons for the EXPOSURE, and to avoid clipping for the maximum possible exposure one needs to us the native ISO. Boosting ISO on an ISO invariant camera will normally just lead to underexposure, which will then be somewhat compensated by boosting the amplification during Raw conversion and postprocessing. The histogram that results is shifted to the right, but I do not consider that EXPOSING to the right.
Hope that clarifies what I intended to say.
Cheers,
Bart
-
In other words, any exposure which does not push the histogram to the right, whatever the chosen ISO, will result in more image noise.
Hi Ray,
I agree, with emphasis added to the word exposure. That's the gist of it, IMHO.
Cheers,
Bart
-
@Bart... No one is suggesting that bumping up ISO is an example of ETTR. All that we are saying is that you can still employ ETTR (particularly in a non-ISOless sensor) away from base ISO. Someone earlier (I think it might have been you) was asserting that you have to operate at base ISO to get the benefits of ETTR. That's simply not true, as I and others have explained. Andrew posted that example to show that increasing ISO doesn't necessarily result in a worse SNR as compared to doing it later in post processing. This is related to what the OP was asking.
-
Someone earlier (I think it might have been you) was asserting that you have to operate at base ISO to get the benefits of ETTR. That's simply not true, as I and others have explained. Andrew posted that example to show that increasing ISO doesn't necessarily result in a worse SNR as compared to doing it later in post processing. This is related to what the OP was asking.
Bernie, if it was me I was misunderstood because I never said that one has to operate at base ISO to get the benefits of ETTR. My first post said this instead, in the context of this thread's title, ISOless cameras:
Well, for maximum captured IQ you would ETTR at base ISO , if possible.
I am still asserting that :) Of course if there are reasons why you cannot do that feel free to ETTR at a higher ISO with the same exposure. Just be aware that with an ISOless camera it is less likely that you will capture the best IQ possible.
Jack
-
Some early digital cameras had a base ISO of 50. Now for nearly all DSLRs base ISO is ISO100, or even 160 or 200.
Does this mean it is impossible to practice ETTR with modern digital cameras?
::)
-
As an aside, ETTR is the gift of a story that keeps on giving! It's years since this idea kicked off and it's still causing contention all these years later. :D
-
@Bart... No one is suggesting that bumping up ISO is an example of ETTR. All that we are saying is that you can still employ ETTR (particularly in a non-ISOless sensor) away from base ISO.
Sorry, but now I'm becoming confused (speaking for myself again) about what you mean exactly, maybe we mean the same thing but are not saying it precisely enough? I (still) do not see how one can ETTR, away from base ISO. All that happens with an ISO boost in an ISO invariant camera is an amplification of the (resulting) histogram to the right, and that is also with a tonecurve/gamma applied. So we need to under-expose to avoid the histogram from climbing the right wall. If that's what you mean, then I understand, but I wouldn't call it "Expose to the right" but rather "Amplify to the right" after underexposure(!).
Cheers,
Bart
-
I can't really make it any simpler. ETTR is about exposure - that is, amount of photons hitting the sensor. What ISO you shoot at has zero affect on this. Only your shutter speed and aperture affect exposure. I think the reason you are getting confused is that you are assuming that an image taken at higher ISO will be clipped (i.e. the ADC will be fully saturated). But that's clearly not necessarily the case, as we all know. If there is leeway to increase exposure at ISO 800, then you can apply the principles of ETTR. You can do it at ISO 4 billion if you want. ISO has no effect on the principles of ETTR. The only effect it has is that larger ISO's have reduced dynamic range, and may also saturate the ADC.
-
If that's what you mean, then I understand, but I wouldn't call it "Expose to the right" but rather "Amplify to the right" after underexposure(!).
Cheers,
Bart
Now we're getting into semantics, Bart. The word 'right' in the phrase ETTR refers to the 'right' of the histogram. The histogram, as seen on the camera's LCD screen, does not distinguish between the different states of 'full-well highlights' (ETTR at base ISO) and the amplification of 'less-than-full-well highlights' resulting from ETTR at a higher-than-base ISO.
If one feels the need to distinguish between these two states with regard to optimum exposure, then one needs two different acronyms. ETTR at base ISO could become EFFWH (Expose For Full Well Highlights), and ETTR at higher-than-base ISO could become ATTR (Amplify To The Right). ;D
-
interpreted Andrew's post (http://forum.luminous-landscape.com/index.php?topic=101242.msg830204#msg830204) as agreeing with you that at a higher ISO one can still use ETTR. That is what I disagree with, and that the example used an ISO variant camera was not what the OP seemed to ask.
I'll post the facts, you Gentlemen can argue, discuss, come to conclusions.
I provided two captures.
Capture 1 was shot at ISO 100 with a f-stop and shutter recommendation from an Incident meter (Minolta Flash Meter III). It is what many would suggest is 'correct' exposure.
Capture 2 was shot at ISO 800 with the same settings as capture 1. Identical in every respect expect I changed the ISO setting from 100 to 800!
Capture 2 shows less noise than capture 1. This is a very magnified view of the image but it is clear that the capture with the higher ISO has less noise, better S/N.
I provided this example to dismiss two commonly heard urban legions. The first is higher ISO always produces more noise. That's clearly not the case here!
The other reason I posted was to dismiss the idea that ETTR must be conducted at base ISO. If that's true, explain the effect upon the noise at ISO 800.
Now you guys can argue, discuss and come to some conclusions in terms of semantics about what Exposure means, what ETTR means, or the benefits expected from ETTR (I'll suggest the expected benefits are exactly what is seen in capture 2 and not capture 1 despite the external meter told me capture 1 was the 'correct exposure'.
Gentlemen, start your engines....
-
I can't really make it any simpler. ETTR is about exposure - that is, amount of photons hitting the sensor.
Check.
What ISO you shoot at has zero affect on this.
Check.
Only your shutter speed and aperture affect exposure.
Check.
I think the reason you are getting confused is that you are assuming that an image taken at higher ISO will be clipped (i.e. the ADC will be fully saturated).
Well, there lies the rub. Try an ETTR at base ISO. Now boost the ISO, what happens?
We did 'expose to the right', and the histogram will be clipped at the right (because it was then amplified).
But that's clearly not necessarily the case, as we all know.
Not necessarily, but for that to not happen we must first underexpose. Calling that exposing to the right confuses (me), because you have to underexpose to get there, and its hurting image quality. What should have been done is reducing the ISO, i.e. the amplification, and retain optimal quality. Wasn't that the goal of ETTR, optimizing quality by exposure? Or is it just to fill out the histogram by changing the amplification? I'd say the former, you seem to say the latter.
Can we agree on that? Because if that's not what is confusing the real issue of quality, then what is?
If there is leeway to increase exposure at ISO 800, then you can apply the principles of ETTR. You can do it at ISO 4 billion if you want. ISO has no effect on the principles of ETTR. The only effect it has is that larger ISO's have reduced dynamic range, and may also saturate the ADC.
Yes, because ISO has nothing to do with exposure as such, it's just like yanking the volume knob, but the band is not playing any louder nor is their music getting any better (if they don't play loud enough to begin with, people at the back won't be able to hear them, so then amplifying the volume would cure that, but band still isn't playing any louder). If it has nothing to do with exposure, then why call it Expose to the right?
Again, is that what we're dealing with? Although the exposure doesn't change, you call boosting the amplification by increasing the ISO an Exposure boost?
Cheers,
Bart
-
I'll post the facts, you Gentlemen can argue, discuss, come to conclusions.
I provided two captures.
Capture 1 was shot at ISO 100 with a f-stop and shutter recommendation from an Incident meter (Minolta Flash Meter III). It is what many would suggest is 'correct' exposure.
Capture 2 was shot at ISO 800 with the same settings as capture 1. Identical in every respect expect I changed the ISO setting from 100 to 800!
Capture 2 shows less noise than capture 1. This is a very magnified view of the image but it is clear that the capture with the higher ISO has less noise, better S/N.
Hi Andrew, thanks for taking the time to explain. Assuming these were Raw captures, the brightness of the converted ISO 800 shot should have been much brighter, so I assume that the brightness was made equal in Lightroom with a.o. the Exposure slider. Is that correct?
I provided this example to dismiss two commonly heard urban legions. The first is higher ISO always produces more noise. That's clearly not the case here!
Correct, although the Canon is different than an ISO invariant sensor design.
The other reason I posted was to dismiss the idea that ETTR must be conducted at base ISO. If that's true, explain the effect upon the noise at ISO 800.
Sure. Since the actual exposure was identical, either the ISO 100 shot needs to be brightened (which will only show the shot-noise with a large percentage of read noise added), or the ISO 800 shot darkened. By increasing the ISO to 800, in the Canon specifically, there is an amplification added to the captured signal before reading it out. Amplifying the signal and then adding read noise is different from adding read noise and then amplifying that. The order of things matters, in Canons.
People like Jim Kasson have written computer applications to allow sensor simulatons that can be fed different pre- and post-amplification amounts. That could help to quantify the differences, but I'm not sure if he'd like to get caught up in this thread, and if he has the actual parameters needed for a Canon EOS 5D Mark III (he has done more simulations with Sony sensor models, like used in Sonys and Nikons).
So boosting the ISO in Canons (to something like ISO 800-1600 for different models) will improve/reduce the total noise that amplification plus read noise produces. That is particularly helpful for under-exposed shots, e.g. because a shorter exposure time was required to freeze motion. But again, this is not how ISO invariant (non-Canon) cameras respond.
So as useful the demonstration is, IMHO it's not about ETTR on ISO invariant cameras.
Cheers,
Bart
-
@Bart... Ok, I see the confusion, and it's probably my fault. I've been talking about the concept of exposing to the rightest. Not necessarily at the very rightmost point. I went down that path as it more accurately represents the scenario of the OP. Along the way I've carried that into a more specific discussion about ETTR. But as you say, ETTR is more specifically about exposing right up to the clipping point. So if you manage that at base ISO, then there is obviously nothing to be gained (and a lot to be lost) by raising ISO.
Apologies for the mixup.
-
@Bart... Ok, I see the confusion, and it's probably my fault. I've been talking about the concept of exposing to the rightest. Not necessarily at the very rightmost point. I went down that path as it more accurately represents the scenario of the OP. Along the way I've carried that into a more specific discussion about ETTR. But as you say, ETTR is more specifically about exposing right up to the clipping point. So if you manage that at base ISO, then there is obviously nothing to be gained (and a lot to be lost) by raising ISO.
Apologies for the mixup.
Hi Bernie, no problem. Glad it got clarified (at least for us two ;))
And, more importantly, we were both Right in a way. Oops ;)
Unfortunately we need to get specific about the semantics at times, to avoid talking at cross purposes.
Cheers,
Bart
-
Hi Andrew, thanks for taking the time to explain. Assuming these were Raw captures, the brightness of the converted ISO 800 shot should have been much brighter, so I assume that the brightness was made equal in Lightroom with a.o. the Exposure slider. Is that correct?
To use Michael’s terms, there were, like all ETTR work 'normalized' so yes.
Sure. Since the actual exposure was identical, either the ISO 100 shot needs to be brightened (which will only show the shot-noise with a large percentage of read noise added), or the ISO 800 shot darkened.
The ISO 100 wasn’t touched at all. It is after all, according to the meter, the correct exposure.
By increasing the ISO to 800, in the Canon specifically, there is an amplification added to the captured signal before reading it out. Amplifying the signal and then adding read noise is different from adding read noise and then amplifying that. The order of things matters, in Canons.
That's not pertinent to the overall, fixed comments found here and elsewhere about ISO and ETTR.
So as useful the demonstration is, IMHO it's not about ETTR on ISO invariant cameras.
I disagree.
-
Surely with all cameras, whether they are ISO invariant or not, one should attempt to apply ETTR at base ISO for best results, if the required shutter speed and aperture allow it.
Likewise, with all cameras, whether they are ISO invariant or not, one needs to apply the ETTR process regardless of the ISO setting chosen, in order to get the best results, the best results being the lowest noise in shadows and mid-tones, without blowing wanted highlights.
For example, if I'm using a D7000 or D7200, which I would say are pretty close to being ISO invariant cameras, and I decide to use ISO 800, not only because I need a fast shutter speed, or because the light is not good, but also because I want to clearly see what I've shot on the camera's LCD screen, and/or share the images with others in the field, then I still need to apply the ETTR process at ISO 800 in order to achieve the lowest noise and the best review image.
In other words, any exposure which does not push the histogram to the right, whatever the chosen ISO, will result in more image noise.
Where's the confusion? The fact that one might not need to use a high ISO with an ISOless camera, if one doesn't feel the need to be able to clearly view the shots taken on the camera's LCD screen, does not mean that the ETTR principle does not apply if one does choose to raise ISO, rather than underexpose at base ISO.
However, that is not to deny that the ISOless camera has clear advantages. But that is a separate issue
Well stated Ray, but the critical restraints in this situation are the required f/stop and shutter speed. If one is shooting an event in limited light and requires a shutter speed of 1/320 sec to freeze motion and an aperture of f/8 for depth of field, one sets the camera with those parameters. Once this is done, exposure (in lux seconds) is fixed. If the histogram is to the left at base ISO, one could increase the ISO and move the histogram to the right, but this does not change the number of photons collected, but it would brighten the LCD preview. With a non-isoless camera, such as Andrew's Canon, use of a higher ISO will decrease read noise and improve the image. The decrease in read noise with that camera levels out at around ISO 1600, and a further increase in ISO will not improve the read noise, but will decrease highlight headroom and risk blown highlights. Interested readers should refer to Emil's post (http://theory.uchicago.edu/~ejm/pix/20d/tests/noise/noise-p3.html#ETTR).
If the histogram is still to the left at ISO 1600, one could increase the ISO further and move the histogram to the right and get a brighter LCD preview at the risk of blown highlights, but there would otherwise be no improvement in image quality. True ETTR would be to increase the exposure (shutter speed, f/stop), but this is prevented by the chosen restraints on shutter speed and aperture.
With an iso-less camera, one could increase the ISO sufficiently to give a readable LCD preview but not further in order to preserve highlight headroom. I agree with Bart that ETTR is best applied at base ISO. At higher ISO one often can not, because of the above mentioned constraints, move the histogram to the right by increasing exposure. ETTR is a useful concept, but it places undue emphasis on the appearance of the histogram.
Giving maximum exposure would be a better term.
Bill
-
Now we're getting into semantics, Bart. The word 'right' in the phrase ETTR refers to the 'right' of the histogram. The histogram, as seen on the camera's LCD screen, does not distinguish between the different states of 'full-well highlights' (ETTR at base ISO) and the amplification of 'less-than-full-well highlights' resulting from ETTR at a higher-than-base ISO.
That's right correct.
If one feels the need to distinguish between these two states with regard to optimum exposure, then one needs two different acronyms. ETTR at base ISO could become EFFWH (Expose For Full Well Highlights), and ETTR at higher-than-base ISO could become ATTR (Amplify To The Right). ;D
Well, assuming we want to optimize image quality (why else shoot Raw instead of OOC JPEG), it might help to distinguish between the two. Not sure if Expose to the right (ETTR) and Amplify (or adjust) to the right (ATTR) are going to help much, because they still need the clarification about the dilemma between shooting constraints an technical image quality.
Cheers,
Bart
-
ETTR is a useful concept, but it places undue emphasis on the appearance of the histogram.
Giving maximum exposure would be a better term.
Bill, I fully agree with that (with your entire post in fact).
Cheers,
Bart
-
Well, assuming we want to optimize image quality (why else shoot Raw instead of OOC JPEG), it might help to distinguish between the two. Not sure if Expose to the right (ETTR) and Amplify (or adjust) to the right (ATTR) are going to help much, because they still need the clarification about the dilemma between shooting constraints an technical image quality.
Ah, I think we're reaching critical agreement!
When I simply wrote to you in my last post that I disagree, it was based on what you just wrote.
What are the benefits one expects from using an ETTR approach? I think we all agree less noise due to better S/N. The adjustment of the ISO on a Canon system provided that goal and result. So now it's semantics but that's important and I've found that such back and forth discussions here on LuLa, thanks to the level of the audience often accomplishes the goal of having well defined terms as well an explanation of the whys. Then we can attempt to perhaps persuade others to do the same.
Exposure. Is it simply the interaction of shutter and aperture? Is ETTR solely achieved by controlling those two items in an attempt to produce the best quality data? Where does ISO fit? If we agree on that terminology, what do we call the process whereby altering the ISO as I did with a Canon camera produces a similar effect? ATTR? Perhaps although I'm not certain anything is happening to the right of something else.
There are a lot of Canon cameras out there which presumably can produce less noise by tweaking the ISO as I did. So I think we should address this audience. Maybe we don't want to call it ETTR per se. But we need to come up with something to illustrate the effect.
One item my test provided that I'd hope we can all agree upon. Stating that 'raising ISO always produces more noise' isn't correct. Now what do we call this and how do we explain the results in as simple a sentence or two as possible?
-
what do we call the process whereby altering the ISO as I did with a Canon camera produces a similar effect?
The problem arises because people are introducing a variable that was never there in the original articles?
Why not simply define ETTR as pushing the histogram* to the right (without blowing highlights) as a method of reducing noise and maximising information. How you do that (be it pushing ISO or longer shutter speed etc) is dependent on the constraints of that particular shoot.
The unmentioned corollary is that its effect is maximised at base ISO, but then again any photographer venturing into ETTR already knows that and applies whether using ETTR or not.
* yes, I now that the aim is not to push the histogram but the histogram is the best tool we have to know we are achieving what we are setting out to do.
-
The problem arises because people are introducing a variable that was never there in the original articles?
Yes (if original you mean Michael's way back when).
Why not simply define ETTR as pushing the histogram* to the right (without blowing highlights) as a method of reducing noise and maximising information. How you do that (be it pushing ISO or longer shutter speed etc) is dependent on the constraints of that particular shoot.
I have no issues with that as long as others are OK with the E for Exposure lumping ISO for certain camera systems into that mix.
The unmentioned corollary is that its effect is maximised at base ISO, but then again any photographer venturing into ETTR already knows that and applies whether using ETTR or not.
Thus far, I'm OK with that. But someone else might have a reason to sway my opinion which is the reason these discussions are useful.
-
Ah, I think we're reaching critical agreement!
When I simply wrote to you in my last post that I disagree, it was based on what you just wrote.
What are the benefits one expects from using an ETTR approach? I think we all agree less noise due to better S/N. The adjustment of the ISO on a Canon system provided that goal and result. So now it's semantics but that's important and I've found that such back and forth discussions here on LuLa, thanks to the level of the audience often accomplishes the goal of having well defined terms as well an explanation of the whys. Then we can attempt to perhaps persuade others to do the same.
Yes, fully agree.
Exposure. Is it simply the interaction of shutter and aperture? Is ETTR solely achieved by controlling those two items in an attempt to produce the best quality data? Where does ISO fit? If we agree on that terminology, what do we call the process whereby altering the ISO as I did with a Canon camera produces a similar effect? ATTR? Perhaps although I'm not certain anything is happening to the right of something else.
This seems to be the major reason for most of the discussions. With the exception for Canons, changing the ISO doesn't change much besides the brightness of the LCD/Thumbnail/preview/histogram after in camera Raw conversion, White balance, and tonemapping.
What does change the Raw image quality is caused by the actual exposure level (the number of photons that achieve a certain level of exposure, due to the combination Aperture and shutterspeed). More photons is better, as long as we can achieve that within the creative (aperture/DOF) and unshaken (shutterspeed/subject motion) constraints.
There are a lot of Canon cameras out there which presumably can produce less noise by tweaking the ISO as I did. So I think we should address this audience. Maybe we don't want to call it ETTR per se. But we need to come up with something to illustrate the effect.
It's not easy without getting very technical, but it is ISO variant read-noise.
One item my test provided that I'd hope we can all agree upon. Stating that 'raising ISO always produces more noise' isn't correct. Now what do we call this and how do we explain the results in as simple a sentence or two as possible?
The leading principle is that more photons will improve the shot noise statistics more than the read noise, hence an improved S/N ratio is the result. If the actual photon exposure cannot be increased, due to constraints e.g. for reduction of camera shake or subject motion, then some cameras (e.g. Canons) can benefit from boosting the ISO to 800 - 1600 ISO which improves only the read noise characteristics.
Cheers,
Bart
-
In my (and others, going by what Andrew said) opinion, the principle ethos of ETTR is getting the best SNR. If that includes using ISO on non-ISOless sensors, then that seems like a reasonable addition to the general concept. But thinking more about Andrew's example shots earlier I can't really work out how that happened. Wouldn't boosting the ISO just boost both the signal and the noise at the 100ISO image by the same factor? Which would leave the SNR the same. In fact you'd expect the SNR to fall, as more noise would be introduced by the amplifier electronics. What's going on there?
Edit: A possible answer to my own question... Does this mean that the majority of electronic noise is introduced into the image data in the ADC? That would mean that boosting the ISO in non-ISOless cameras would boost the signal in isolation from the majority of the electronic noise.
-
Edit: A possible answer to my own question... Does this mean that the majority of electronic noise is introduced into the image data in the ADC? That would mean that boosting the ISO in non-ISOless cameras would boost the signal in isolation from the majority of the electronic noise.
Hi Bernie, could I suggest you have a look at my comments in Reply 3 of this thread. That's the sort of thing I was talking about and my thoughts are based on what I have read in Emil Martinec's article referenced by Bill Janes above (reply 62).
Dave
-
Oh yeah, cheers for that. That makes sense.
-
I provided two captures.
Capture 1 was shot at ISO 100 with a f-stop and shutter recommendation from an Incident meter (Minolta Flash Meter III). It is what many would suggest is 'correct' exposure.
Capture 2 was shot at ISO 800 with the same settings as capture 1. Identical in every respect expect I changed the ISO setting from 100 to 800!
Capture 2 shows less noise than capture 1. This is a very magnified view of the image but it is clear that the capture with the higher ISO has less noise, better S/N.
This is an interesting exercise, but it would be informative to see the entire images, not crops. A danger of ETTR is blown highlights. The ISO 800 image received 3 stops more exposure than indicated by the meter. This would lift the mid-tones by 3 stops. Since pure white (zone 9) is 2.5 stops over mid-gray, this means that tones above mid-gray would be clipped. The mid-tones have a better SNR, but clipped highlights would have a SNR of zero.
Bill
-
This is an interesting exercise, but it would be informative to see the entire images, not crops. A danger of ETTR is blown highlights. The ISO 800 image received 3 stops more exposure than indicated by the meter. This would lift the mid-tones by 3 stops. Since pure white (zone 9) is 2.5 stops over mid-gray, this means that tones above mid-gray would be clipped. The mid-tones have a better SNR, but clipped highlights would have a SNR of zero.
Hi Bill,
That was one of my concerns as well, since Andrew said that the ISO 100 shot was (according to the exposure meter) correctly exposed. An ISO 800 setting with the same exposure would then place the mid tones at the clipping level, and leave brighter tones totally clipped.
But it also shows (a bit OT) that for ISO variant Canons in this case, ISO can be used for "amplification" or ISO bracketing. However, nothing beats a real (Exposure) bracketing sequence if HDR tonemapping is part of the workflow, because collecting more Photons are the quickest route to S/N ratio improvement (at all tone levels). Without that, single exposures benefit from optimizing (maximizing, without clipping of important highlights) the exposure level, by using ETTR if the other constraints allow.
In general that means that low contrast scenes will be exposed 'hotter' than an incident exposure meter would suggest, and darker in a high contrast scene (to protect highlights). The easiest way to quickly achieve that is to (spot) measure the important highlights (e.g. spotmeter the white clouds), and use a +EV of approx 2.5-3 stops. That would take care of scene contrast automatically.
Cheers,
Bart
-
In my (and others, going by what Andrew said) opinion, the principle ethos of ETTR is getting the best SNR. If that includes using ISO on non-ISOless sensors, then that seems like a reasonable addition to the general concept.
I think so. This raises an interesting concept in my mind which may need correction or explanation based on what others write.
Bart provided a good analogy, that raising ISO is akin to raising the volume on an amplifier rather than the musicians themselves playing louder. So consider the ISO 800 image I provided. Had I used the meters recommendation, I'd have stopped down from the ISO 100 image and with less photon's as a result, more noise. Now what if I'm shooting and I'm at the limit of either the shutter speed of aperture I want using the ISO 100 recommendation so ETTR isn't going to work. I just raise the ISO on the camera leaving the 'exposure' the same, I end up with less noise. That could be a useful technique no?
Also interesting is that when I tested ETTR on this camera, I found I could not go more than 1.5 stops to the right without clipping delicate highlights. 2 stops was blown out. Yet using the amplifier as Bart describes it, I was able to go up three stops. Interesting.
-
The ISO 800 image received 3 stops more exposure than indicated by the meter.
Did it? I'm only asking if this is a fair use of the term exposure as discussed previously.
I'll look at the full image and provide a sample, it's possible I was close to or just at highlight clipping but what about my previous text where 1.5 stops over the meter using exposure (aperture/shutter) was the limit? Is it possible using ISO this way provide more 'headroom' to ETTR if that's even a fair term to use in this example? Or that ISO used this way is not fully correlated with f-stops? Don't know.
Initially the example was made to dismiss the blatant statement that higher ISO always results in more noise. Unless I'm missing something, that statement isn't true.
-
Did it? I'm only asking if this is a fair use of the term exposure as discussed previously.
I'll look at the full image and provide a sample, it's possible I was close to or just at highlight clipping but what about my previous text where 1.5 stops over the meter using exposure (aperture/shutter) was the limit? Is it possible using ISO this way provide more 'headroom' to ETTR if that's even a fair term to use in this example? Or that ISO used this way is not fully correlated with f-stops? Don't know.
According to Doug Kerr (http://dougkerr.net/Pumpkin/articles/Exposure_calibration.pdf), who quotes Canon sources, if one photographs a scene of constant luminance (i.e. a gray or white card) with a Canon camera, the rendered gray scale image should have a K value of 55%, which corresponds to a relative luminance of 0.173. Mid-gray is 0.18. A K of 55% corresponds to about 113, 113, 113 in sRGB.
A relative luminance of 0.173 is 2.53 stops below 100% [log2(1/0.173)]. Accordingly, 3 stops over the camera light meter reading would clip scene luminances from slightly below mid-gray and above. According to Doug, Canon does not allow the "famous 1/2 stop cushion" for highlight headroom. You could test your setup by photographing a white card and checking rendered values in Photoshop. The sRGB value should be 113, 113, 113.
Bill
-
You could test your setup by photographing a white card and checking rendered values in Photoshop. The sRGB value should be 113, 113, 113.
need to spell raw converter parameters too...
-
This is an interesting exercise, but it would be informative to see the entire images, not crops.
Here's a screen capture of the entire image also showing the somewhat radical settings used in Develop. Clipping indicators are on in Histogram, no clipping. RGB value over brightest area of dogs head reads 98%.
What's interesting (to me) is that while the 'exposure' was boosted 3 stops via ISO, note that the Exposure slider is set to -1.55 stops which as I pointed out earlier is just about the limit of the ETTR testing on this camera done in the past. I can't explain why there is this disconnect between the +3 ISO and -1.5 Exposure slider among the other sliders as set.
(http://digitaldog.net/files/DogISO800.jpg)
-
I can't explain why there is this disconnect
when dealing with Adobe converters do use gutted .dcp profiles (linear curve, hidden expocorrections compensated, post exposure corrections LUTs removed) + Process 2010 + brightness 50, contrast 25, curve : Medium Contrast, output to the baddest gamut-wise colorspace... something like this... at least in ACR.
-
when dealing with Adobe converters do use gutted .dcp profiles (linear curve, hidden expocorrections compensated, post exposure corrections LUTs removed) + Process 2010 + brightness 50, contrast 25, curve : Medium Contrast, output to the baddest gamut-wise colorspace... something like this... at least in ACR.
Correct. In fact, LR's Process 2012 conversions are useful in practice for conversions of images with (potential) clipping issues. However, for analysis of Raw data it is less suited because there is indeed too much going on (it's not obvious what is real data related and what is constructed) under the hood of LR to draw any meaningful technical conclusions. It can reconstruct detail from (partially) clipped highlights pretty good, but that is extrapolated data. So when the clipping indicator indicates nothing, there may still be clipped data. Useful for rescuing highlights, not so much for analysis, or for learning how to improve one's technique.
To inspect the really Raw data one needs a tool like Rawdigger. It will tell if there is clipping when the Raw (linear gamma, no WB) histogram collects increasing amounts of data in the highest bin of the histogram.
Cheers,
Bart
-
Years ago, Emil Martinec posted the following where he equated ETTR with Maximizing Exposure
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; on many cameras (those with CCD sensors, and the newer Sony Exmor sensors), there is little or no advantage to raising the ISO, which compromises point (3), even though leaving the ISO at a low value may leave the histogram "to the left" for your chosen exposure; 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 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.
-
Note that ISO is not part of exposure; on many cameras (those with CCD sensors, and the newer Sony Exmor sensors), there is little or no advantage to raising the ISO,
Yet that doesn't appear to be the case with Canon systems.
-
Yet that doesn't appear to be the case with Canon systems.
I hope you have, by now, finished reading the quote, where this is discussed.
-
I hope you have, by now, finished reading the quote, where this is discussed.
The two parts of the quote that contradict each other, yes.
-
The two parts of the quote that contradict each other, yes.
ISO is not part of exposure. It is, on certain sensors, a part of maximizing exposure considerations of aperture and shutter speed. Thereby allowing faster shutter or smaller aperture with a less than linear reduction in dynamic range and/or increase in noise.
-
To inspect the really Raw data one needs a tool like Rawdigger.
absolutely, but if smb still want to use Adobe converters this is the easiest set of parameters, then the said individual might go deeper and try may be altering ColorMatrix'es inside .dcp to make proper use of UniWB (I didn't do that myself though - as I have both RD anf FRV)
-
ISO is not part of exposure. It is, on certain sensors, a part of maximizing exposure considerations of aperture and shutter speed.
OK, that's clear. It is and it isn't.
-
OK, that's clear. It is and it isn't.
Come on Andrew. ISO has nothing to do with the amount of light reaching the sensor.
It is also known that increasing ISO will increase noise and lessen dynamic range. The amount of the change is dependent of the amplification characteristics. Knowing how your camera reacts is important to the image you want to create....which we often call, improperly, exposure.
-
Come on Andrew. ISO has nothing to do with the amount of light reaching the sensor.
I agree IF we accept that exposure is solely the relationship of aperture and shutter speed. I think that's probably true but have an open mind.
It is also known that increasing ISO will increase noise and lessen dynamic range.
Expect when it doesn't. In the example I provided, the increased ISO reduced the noise.
The amount of the change is dependent of the amplification characteristics. Knowing how your camera reacts is important to the image you want to create....which we often call, improperly, exposure.
No argument, I fully agree. But there's something going on, something 'good' when the Canon is set to a higher ISO with the same 'exposure' the meter recommended. I don't think that is something to be dismissed. The image of the dog with the higher ISO has less visible noise.
-
ISO is not part of exposure. It is, on certain sensors, a part of maximizing exposure considerations of aperture and shutter speed. Thereby allowing faster shutter or smaller aperture with a less than linear reduction in dynamic range and/or increase in noise.
The usual definition of exposure is given in this Wikipedia article (https://en.wikipedia.org/wiki/Exposure_(photography)):
"In photography, exposure is the amount of light per unit area (the image plane illuminance times the exposure time) reaching a photographic film or electronic image sensor, as determined by shutter speed, lens aperture and scene luminance. Exposure is measured in lux seconds, and can be computed from exposure value (EV) and scene luminance in a specified region."
Changing the ISO setting on the camera has no effect on the number of lux-seconds falling on the sensor. ACR and other raw converters use the term exposure loosely and this "exposure" adjustment only changes the apparent brightness of the rendered image by applying a scaling factor (if one is lucky and there are no hue twists or other untoward effects). Likewise, some photographers regard increasing the ISO setting on the camera as changing exposure, but this is not scientifically correct.
Bill
-
Here's a screen capture of the entire image also showing the somewhat radical settings used in Develop. Clipping indicators are on in Histogram, no clipping. RGB value over brightest area of dogs head reads 98%.
What's interesting (to me) is that while the 'exposure' was boosted 3 stops via ISO, note that the Exposure slider is set to -1.55 stops which as I pointed out earlier is just about the limit of the ETTR testing on this camera done in the past. I can't explain why there is this disconnect between the +3 ISO and -1.5 Exposure slider among the other sliders as set.
(http://digitaldog.net/files/DogISO800.jpg)
At 98%, the whites have at least one channel clipped and LR has recovered. Check under RawDigger or PV2010.
+3 EC on what exposure and exposure mode. More than likely, due to the amount of white in the image, it probably should have been underexposed with no EC. At +3, it only need -1.55 to set exposure properly.
-
The usual definition of exposure is given in this Wikipedia article (https://en.wikipedia.org/wiki/Exposure_(photography)):
"In photography, exposure is the amount of light per unit area (the image plane illuminance times the exposure time) reaching a photographic film or electronic image sensor, as determined by shutter speed, lens aperture and scene luminance. Exposure is measured in lux seconds, and can be computed from exposure value (EV) and scene luminance in a specified region."
I'm simply wondering if the usual definition is sufficient. I'm fine accepting it, I'm fine if, as I've seen in the past with the usual smart people here, it needs some updating.
Case in point, I think you may recall the long decision here about color numbers and colors we can see. Not the same. This was nicely hashed out here. Yet go to Wikipedia and they say:
The RGB color model is implemented in different ways, depending on the capabilities of the system used. By far the most common general-used incarnation as of 2006 is the 24-bit implementation, with 8 bits, or 256 discrete levels of color per channel. Any color space based on such a 24-bit RGB model is thus limited to a range of 256×256×256 ≈ 16.7 million colors.
Can we see 16.7 million colors? If we have a color number that represents something we can't see, is it a color?
I think this group can possibly refine, if necessary, ETTR, exposure and ISO. Maybe not. If the consensus is, exposure is solely an attribute that doesn't include ISO, I think it useful to come up with some language that describes what happened and why with the ISO 800 image that has less noise than the ISO 100 image. If we agree to separate ISO, the E in ETTR doesn't apply here. But that doesn’t mean we should ignore it so what do we call it?
-
At 98%, the whites have at least one channel clipped and LR has recovered. Check under RawDigger or PV2010.
That is entirely possible, but not really pertinent to the effect seen. Suppose I altered the exposure or ISO just a tad? Do you suspect the noise would then match that of the ISO 100 capture or it would appear less noisy? I suspect the later. So I think the fact that one channel might be clipped is minutia when the question I'm asking about is the effect of using a higher ISO with the same exposure and it's effect on the noise and what do we call this, how do we describe it to others who have stated "higher ISO always produces more noise"?
-
Exactly and as you pointed out first, in terms of ISO/ETTR, depends on the camera. Example of a Canon 5DMII:
(http://digitaldog.net/files/100vs800iso.jpg)
ISO 100, more noise than ISO 800.
You cannot check ISO noise by changing ISO, then adjusting in post. You must reduce the exposure one stop for each stop increase in ISO.
At 800 with the 5D3 you would not see much change. However, noise vs 100 would increase slightly and dynamic range would decease. While the Sony Exmore sensors have almost a linear decrease the 5d3 is fairly flat thru 1600 and a deeper angle to 3200, then goes roughly linear.
-
You cannot check ISO noise by changing ISO, then adjusting in post. You must reduce the exposure one stop for each stop increase in ISO.
At 800 with the 5D3 you would not see much change. However, noise vs 100 would increase slightly and dynamic range would decease. While the Sony Exmore sensors have almost a linear decrease the 5d3 is fairly flat thru 1600 and a deeper angle to 3200, then goes roughly linear.
I don't understand. You have two captures I provided from a 5DMII. There's a very visible difference in the noise no? And it appears DR. One looks better to me. It's the capture set at ISO 800.
-
I don't understand. You have two captures I provided from a 5DMII. There's a very visible difference in the noise no? And it appears DR. One looks better to me. It's the capture set at ISO 800.
You took them at the same exposure (aperture/shutter) and changed ISO, the adjusted in post.....which you cannot do to compare ISO.
-
The attached shows the change (reduction) in dynamic range by ISO for the D800 and 5D3. Notice how the D800 reduce approximately on a linear basis, while the 5D3 is flat in the early part.
The noise change (increase) follows these curves in the (logical) inverse....that is, the D800 increases on a linear basis, where the 5D3 increase only a little at first, the goes linear around 1600-3200
-
You took them at the same exposure (aperture/shutter) and changed ISO, the adjusted in post.....which you cannot do to compare ISO.
Let's try this again. All I want is:
A. One good sentence or paragraph to describe what happened with the ISO 800 image that resulted in less noise.
B. Agreement on what to call this 'technique' other than ETTR since it's presumably not a part of exposure.
C. Agreement on whether what I see by upping the ISO is beneficial (it appears to me that it is).
-
Let's try this again. All I want is:
A. One good sentence or paragraph to describe what happened with the ISO 800 image that resulted in less noise.
B. Agreement on what to call this 'technique' other than ETTR since it's presumably not a part of exposure.
C. Agreement on whether what I see by upping the ISO is beneficial (it appears to me that it is).
How did you get the 100 ISO and 800 ISO images to look the same? Underexpose the 100, the increase "exposure" in post?
-
How did you get the 100 ISO and 800 ISO images to look the same? Underexpose the 100, the increase "exposure" in post?
See below, already asked and answered.
-
See below, already asked and answered.
Where???
-
Where???
Reply #61 on: June 17, 2015, 10:33:13 AM »
-
Reply #61 on: June 17, 2015, 10:33:13 AM »
Found it
I'll post the facts, you Gentlemen can argue, discuss, come to conclusions.
I provided two captures.
Capture 1 was shot at ISO 100 with a f-stop and shutter recommendation from an Incident meter (Minolta Flash Meter III). It is what many would suggest is 'correct' exposure.
Capture 2 was shot at ISO 800 with the same settings as capture 1. Identical in every respect expect I changed the ISO setting from 100 to 800!
Capture 2 shows less noise than capture 1. This is a very magnified view of the image but it is clear that the capture with the higher ISO has less noise, better S/N.
I provided this example to dismiss two commonly heard urban legions. The first is higher ISO always produces more noise. That's clearly not the case here!
The other reason I posted was to dismiss the idea that ETTR must be conducted at base ISO. If that's true, explain the effect upon the noise at ISO 800.
Now you guys can argue, discuss and come to some conclusions in terms of semantics about what Exposure means, what ETTR means, or the benefits expected from ETTR (I'll suggest the expected benefits are exactly what is seen in capture 2 and not capture 1 despite the external meter told me capture 1 was the 'correct exposure'.
You did not answer how, with 3 stops difference, you got the images to look as if they had the same exposure....what did you do in post?
-
Found it
Wrong post. 61, not 57.
-
Wrong post. 61, not 57.
What changes were made to the 800?
-
need to spell raw converter parameters too...
That's right. I was referring to the sRGB image rendered by the camera into sRGB. ACR/LR are not good tools for evaluating the raw file levels because of the BaselineExposure offset, automatic highlight recovery (with PV2012), and tone curve. Even better, one should check the raw levels with RawDigger. The nominal exposure should give sensor saturation of ~18% for Canon, which does not allow 0.5 EV highlight headroom. Nikons give ~12% saturation, allowing 0.5 EV headroom for the highlights.
Bill
-
when dealing with Adobe converters do use gutted .dcp profiles (linear curve, hidden expocorrections compensated, post exposure corrections LUTs removed) + Process 2010 + brightness 50, contrast 25, curve : Medium Contrast, output to the baddest gamut-wise colorspace... something like this... at least in ACR.
In my experiments with evaluating the raw file with ACR, I get best results with PV2010, linear tone curve, and the sliders on the basic panel all set to zero. I use the Adobe Standard camera profile. These are the settings (http://www.color.org/scene-referred.xalter) recommended by the ICC for obtaining scene referred data.
Bill
-
I use the Adobe Standard camera profile
and I don't (for the reasons outlined above about expocorrection, etc) - hence I use different parameters...
These are the settings recommended by the ICC for obtaining scene referred data.
recommended when CS3 was the current version, that will be ~2007 ? before Adobe introduced what we have now in their DCP profiles now I think...
-
and I don't (for the reasons outlined above about expocorrection, etc) - hence I use different parameters...
recommended when CS3 was the current version, that will be ~2007 ? before Adobe introduced what we have now in their DCP profiles now I think...
I forgot to add that with my D800e, one must use a negative exposure correction of -0.35 EV. I should probably use a profile that definitely lacks a twist. Since I am dealing with a Stouffer step wedge and other neutral targets, color is not that important and my results with ACR 9.0 and PV2010 with a linear tone curve as outlined in the ICC post and with Adobe Standard correlate quite well with DCRaw conversions with a linear tone curve and with readings from RawDigger. What ACR profile do you recommend? If you render into ProPhotoRGB you have to deal with a gamma encoded space and this complicates checking for linearity. However, a log-log plot will allow a linear plot with gamma encoded data. A nice feature of ACR 9.x is that it permits rendering directly into Linear RIMM.
As to your reference to 2007 and CS3, my testing with current software shows that one can still get excellent results. PV2010 with the default parameters that you quote will not give a linear response. Have you plotted your data with these parameters? I would be interested in seeing a scatter plot showing exposure and rendered pixel values over a wide range, say from near clipping of highlights to -6 or -7 EV. Shown below are scatter plots showing ACR renderings into ProPhotoRGB using Adobe Standard. PV2010 with the defualt tone curve using medium contrast in the point curve and brightness and contrast values of 50 and 25 does not produce a linear curve. PV2010 with the point curve set to linear and the sliders on the basic panel does work well.
Regards,
Bill
-
Let's try this again. All I want is:
A. One good sentence or paragraph to describe what happened with the ISO 800 image that resulted in less noise.
Cameras that are not ISOless/invariant are not able to transfer to the raw data the full dynamic range collected at the photosites during the given exposure in one go, because the electronics have a smaller DR than the sensor (say 12 vs 14 stops).
(http://i1.wp.com/www.strollswithmydog.com/wordpress/wp-content/uploads/Tone-TransferISO100-ISOful.png)
By changing ISO with the same exposure you are choosing what portion of the scene information available at the photosites to store in the raw data. If you raise ISO, you are choosing to transfer shadow information to the raw data at the expense of highlights:
(http://i2.wp.com/www.strollswithmydog.com/wordpress/wp-content/uploads/Tone-TransferISO400-ISOful.png)
B. Agreement on what to call this 'technique' other than ETTR since it's presumably not a part of exposure.
Given the electronics' more limited DR, raising ISO at a given Exposure favors tones in the shadows for storage in the raw data at the expense of highlight headroom. Call it 'Sensor Tone Range Optimization' or similar.
C. Agreement on whether what I see by upping the ISO is beneficial (it appears to me that it is).
Sure it is, it is part of tone selection and optimization by the photographer with an ISOful camera in the field.
On the other hand an ISOless camera does not require such worries: after maximizing exposure just stick with base ISO because it is able to store most (all) information available at the photosites into the raw data in one go.
(http://i2.wp.com/www.strollswithmydog.com/wordpress/wp-content/uploads/ToneTransferISOless100.png)
Raising ISO in the ISOless case simply throws away highlight headroom with no corresponding benefit in the shadows. Images from articles here (http://www.strollswithmydog.com/information-transfer-iso-invariant/)and here (http://www.strollswithmydog.com/information-transfer-non-iso-invariant-case/):
Jack
PS How can I make in-line images smaller?
-
All I want is:
A. One good sentence or paragraph to describe what happened with the ISO 800 image that resulted in less noise.
Here's what I wrote earlier while answering my own similar question (in relation to the non-ISOless 5DIII). Others confirmed this is right:
A possible answer to my own question... Does this mean that the majority of electronic noise is introduced into the image data in the ADC? [YES] That would mean that boosting the ISO in non-ISOless cameras would boost the signal in isolation from the majority of the electronic noise.
-
Something to be aware of is that when going to in-camera JPEG, many cameras apply noise reduction. That noise reductions can vary by ISO (or by analysis of the image), so higher ISO, more noise reduction. I don't know whether a 5DIII does that or not - I don't pay much attention is in-camera JPEGs in what I do. But that would be my first guess as an explanation for the difference in posted images.
Sandy
-
He developed it from raw.
-
I forgot to add that with my D800e, one must use a negative exposure correction of -0.35 EV. I should probably use a profile that definitely lacks a twist. Since I am dealing with a Stouffer step wedge and other neutral targets, color is not that important and my results with ACR 9.0 and PV2010 with a linear tone curve as outlined in the ICC post and with Adobe Standard correlate quite well with DCRaw conversions with a linear tone curve and with readings from RawDigger. What ACR profile do you recommend? If you render into ProPhotoRGB you have to deal with a gamma encoded space and this complicates checking for linearity. However, a log-log plot will allow a linear plot with gamma encoded data. A nice feature of ACR 9.x is that it permits rendering directly into Linear RIMM.
As to your reference to 2007 and CS3, my testing with current software shows that one can still get excellent results. PV2010 with the default parameters that you quote will not give a linear response. Have you plotted your data with these parameters? I would be interested in seeing a scatter plot showing exposure and rendered pixel values over a wide range, say from near clipping of highlights to -6 or -7 EV. Shown below are scatter plots showing ACR renderings into ProPhotoRGB using Adobe Standard. PV2010 with the defualt tone curve using medium contrast in the point curve and brightness and contrast values of 50 and 25 does not produce a linear curve. PV2010 with the point curve set to linear and the sliders on the basic panel does work well.
Regards,
Bill
certainly you need to use the output colorspace that suits you - if you want to use (which you can) the widest gamut and linear gamma RGB colorspace for your purposes as you noted ACR will allow to select one - just supply the icc/icm - you know that... if your task is to imitate rawdigger more precisely by all means use whatever it takes - but do suggest to remove everything except the matrix parts (to make sure that there are no twists and brightness corrections through LUTs), insert a specific linear tone curve in the profile just in case (in case the profile do not have the curve ACR will use its/yours own settings, but I rather make sure that the curve is in the profile too), and make sure to use BaseLineExposureOffset compensating BaseLineExposure - you know all that too...
my use is different - I mostly use the settings to see how custom DCP profiles fit the measured targets, for example
(http://s22.postimg.org/vtnxincsv/ACR.jpg)
this shows a decent fit to a measured target with L* from measurements = 96.89, 81.26, 66.68, 50.64, 35.79, 20.50 / readouts in ACR with custom made profile and settings that I used are = 97, 81, 67, 51, 36, 20.
and RawDigger shows that clipping in raw, green channel, is around 1/6 (or rather closer to even 1/3EV if we use the center/top of the "peak", instead of the very right edge of it in histogram for a selected area on a white patch till clipping for Sony A7 that I think is 15860 or so) EV from what we have in the whitest patch - which corresponds good enough with Adobe's visual display of their "histogramm" here
PS: damn Adobe for not supplying Lab readouts with 0.1 precision - just like Iridient did overkill it with 0.01 precision, Adobe still managed to spoil the goods with just integer values for Lab readouts...
-
He developed it from raw.
I didn't see any raws posted.......
-
Cameras that are not ISOless/invariant are not able to transfer to the raw data the full dynamic range collected at the photosites during the given exposure in one go, because the electronics have a smaller DR than the sensor (say 12 vs 14 stops).
Thanks Jack, let me look this over, digest it etc.
PS How can I make in-line images smaller?
I think the size is great, easy for us older farts to see.
-
You cannot check ISO noise by changing ISO, then adjusting in post. You must reduce the exposure one stop for each stop increase in ISO.
Completely incorrect approach. To evaluate the effect in SNR of a variable (ISO), that variable must be isolated from any other (exposure in this case). This is a basic scientific principle.
Changing ISO and exposure at the same time (the usual procedure when taking pictures) leads to confusing and misleading results, e.g. the "more noise the higher the ISO" legend, which is wrong. Visible noise increases when exposure is reduced, not when ISO is pushed.
In some sensors (Canon type), pushing ISO increases SNR. In some others (Sony type, the so called isoless sensors), pushing ISO has little effect in improving SNR.
Whether ETTR can be achieved by pushing ISO (when available exposure doesn't suffice) or not, is just a semantic discusion. The true is that for a given exposure, some sensors improve their SNR (the only goal of ETTR) the higher the ISO:
Canon 350D at constant exposure:
(http://www.guillermoluijk.com/article/iso/versus.jpg)
Regards
-
Changing ISO and exposure at the same time (the usual procedure when taking pictures) leads to confusing and misleading results, e.g. the "more noise the higher the ISO" legend, which is wrong. Visible noise increases when exposure is reduced, not when ISO is pushed.
Excellent, thanks. I was hoping you'd find your way into the discussion. Your illustration just provided was the original example I recall seeing that lead to my attempt testing ISO on my Canon. Whether ETTR can be achieved by pushing ISO (when available exposure doesn't suffice) or not, is just a semantic discusion.
Agreed although it would be great if we could come to some common agreement about the language. It might not be possible. ;)
-
In some sensors (Canon type), pushing ISO increases SNR. In some others (Sony type, the so called isoless sensors), pushing ISO has little effect in improving SNR.
Whether ETTR can be achieved by pushing ISO (when available exposure doesn't suffice) or not, is just a semantic discusion. The true is that for a given exposure, some sensors improve their SNR (the only goal of ETTR) the higher the ISO:
Canon 350D at constant exposure:
And assuming the ISO 100 was correctly exposed, now the (cropped out) highlights are totally clipped. By underexposing the ISO 100 shot 16x (so we create 4 stops headroom for the subsequent amplification), we then and only then can improve the S/N ratio on some cameras without clipping highlights, like the Canons, by increasing the ISO setting 16x. However, bracketing exposures creates much better image quality for such stationary subjects, and just a bit more work in postprocessing.
Cheers,
Bart
-
And assuming the ISO 100 was correctly exposed, now the (cropped out) highlights are totally clipped.
Not the case with my or presumably Guillermo's images.
By underexposing the ISO 100 shot 16x (so we create 4 stops headroom for the subsequent amplification), we then and only then can improve the S/N ratio on some cameras without clipping highlights, like the Canons, by increasing the ISO setting 16x.
Again not the case with my or presumably Guillermo's images. My ISO 100 image is correctly exposed as recommended by the incident meter.
However, bracketing exposures creates much better image quality for such stationary subjects, and just a bit more work in postprocessing.
I doubt anyone would or could disagree with that technique.
-
Not the case with my or presumably Guillermo's images. Again not the case with my or presumably Guillermo's images. My ISO 100 image is correctly exposed as recommended by the incident meter.
Then either the image was underexposed (you can check with RawDigger how far it was from clipping), depending on how your exposure meter was used, or the output clipping was masked by Lightroom's (default) highlight restoration.
Cheers,
Bart
-
Then either the image was underexposed (you can check with RawDigger how far it was from clipping), depending on how your exposure meter was used, or the output clipping was masked by Lightroom's (default) highlight restoration.
I agree, it's not ETTR and thus 'under exposed' for optimal data. The meter wasn't telling me about ideal exposure for raw, that's part of my point.
In terms of the 'disconnect' between LR and true clipping, just how much are we talking about, 1 stop, 1/2 stop, 1/4 stop?
-
Achieving ETTR through ISO only makes sense when exposure doesn't suffice to make it. In the real world this happens hundreds of times (low light conditions, large DOF requirement, quick action,...).
Regards
-
Achieving ETTR through ISO only makes sense when exposure doesn't suffice to make it. In the real world this happens hundreds of times (low light conditions, large DOF requirement, quick action,...).
Indeed it does!
-
In terms of the 'disconnect' between LR and true clipping, just how much are we talking about, 1 stop, 1/2 stop, 1/4 stop?
Depends on the Raw and the camera it was taken with, but theoretically can be up to 5 stops. One would need to take a finely bracketed series to find out, and compare Rawdigger's real Raw data, with the point where LR would start showing the clipping indicator. Possibly also for each ISO that one wants to know it for.
Shooting a newspaper should give some (paper) structure in highlights, and some dark greys that will become mid grey. A better target would be a stepwedge with 0.10 density increments like a Stouffer 2110, because plotting the RGB values will show how the roll-off is near the clipping point, and the steps will show as spikes on the histogram, narrow spikes for low noise, wider spikes/humps for high noise, if surface reflection are avoided.
Cheers,
Bart
-
Depends on the Raw and the camera it was taken with, but theoretically can be up to 5 stops.
5 stops! I don't think my Canon would come close to that, in fact I know it can't. LR can recover, rebuild, whatever you want to call it, 5 stops (darn, 1 stop) from true raw clipping?
I'd have guessed, in my case, nor more than 1/2 stop. Perhaps less.
-
5 stops! I don't think my Canon would come close to that, in fact I know it can't. LR can recover, rebuild, whatever you want to call it, 5 stops (darn, 1 stop) from true raw clipping?
I'd have guessed, in my case, nor more than 1/2 stop. Perhaps less.
Using RawDigger (or Fast Raw Viewer) will take the guesswork out of it ...
Cheers,
Bart
-
Using RawDigger (or Fast Raw Viewer) will take the guesswork out of it ...
I will but what camera system did you test with the above that was 5 stops over? I want one.
-
I will but what camera system did you test with the above that was 5 stops over? I want one.
Lightroom (process 2012) will allow a 5 stop ''Exposure" correction. When the clipping indicators will show up depends on the subject colors, white balance, and level of under exposure, which we don't know at this moment. The choice of profile may also play a role.
Cheers,
Bart
-
Lightroom (process 2012) will allow a 5 stop ''Exposure" correction. When the clipping indicators will show up depends on the subject colors, white balance, and level of under exposure, which we don't know at this moment. The choice of profile may also play a role.
Using Rawdigger, I'm seeing about 1/2 stop between the two products. I was able to expose 1.5 stops over the meter before clipping specular white (BableColor tile) in LR using PV2010, RD showed clipping. With the 1 stop over meter, RD shows no clipping. The 2 stop over meter is blown out in with PV2010 however with 2012 it isn't. But close (-2.30 on exposure slider).
What's rather interesting is the 1.5 over, that really is over according to RD looks great with the LR recover and has less noise than the 1.0 over. Go figure. I've also thought exposure and development are hand in hand. Yes, at 1.5 over meter, RD says I'm clipped. But the image? Looks better.
-
Lightroom (process 2012) will allow a 5 stop ''Exposure" correction.
But have you found a capture or raw that can utilize this much correction, that's the question.
-
Lightroom (process 2012) will allow a 5 stop ''Exposure" correction.
you can always add extra stops through .dcp profile - BaseineExposureOffset tag
-
you can always add extra stops through .dcp profile - BaseineExposureOffset tag
Begging the question, how far can you go ETTR+PV and get good results?
-
But have you found a capture or raw that can utilize this much correction, that's the question.
I tend to expose to the right, so I have no (severe over-exposure) examples of regular shots that need such levels of correction at hand unless I start digging in my HDR shots. But then, we're rather talking about deliberate underexposure in this thread, and I will have an even harder time looking for those, especially in a 8x or 16x scenario. The lowest exposure in an HDR sequence is typically the ETTR exposure, the rest in longer to collect more photons in the shadows. When in doubt for regular shots, I bracket exposures and only keep the ETTR exposure (choice determined by RawDigger's closeness of the highlights to the clipping level).
Cheers,
Bart
-
I tend to expose to the right, so I have no (severe over-exposure) examples of regular shots that need such levels of correction at hand unless I start digging in my HDR shots.
Ditto. For real casual shooting, I have exposure compensation plus 3/4 to 1 stop. If I have the time, I'll tweak or bracket.
But getting back to something maybe new, can we, should we ETTR+D meaning exposure + development. I'm from the old analog B&W and color film darkroom days. I spent my share of time trying all kinds of exotic developers for TechPan. I wouldn’t think of considering exposure and ISO without knowing about developing the neg. I'm wondering why that shouldn't be true today and if so, just the differences between the raw converters and their controls should be part of the picture. Going full circle, if a Canon shooter and in one of the common situations Guillermo mentioned, also adjust ISO.
-
Begging the question, how far can you go ETTR+PV and get good results?
I don't know - I am just pointing to a technical possibility to overcome UI limitations in ACR...
-
I don't know - I am just pointing to a technical possibility to overcome UI limitations in ACR...
One man's limitations could be another man's Technidol! ::)
-
I agree, it's not ETTR and thus 'under exposed' for optimal data. The meter wasn't telling me about ideal exposure for raw, that's part of my point.
In terms of the 'disconnect' between LR and true clipping, just how much are we talking about, 1 stop, 1/2 stop, 1/4 stop?
http://mulita.com/blog/?p=3358 George Jardin on LR headroom
-
Begging the question, how far can you go ETTR+PV and get good results?
Sensors are pretty well linear devices, so you can go as low as you can stand in terms of SNR if your converter is well behaved without penalty (it turns out that LR is not but others are). If a sensor has a 14 stop acceptable SNR (the engineering definition) then by recovering 5 stops you've essentially got 9 (14-5) to display. Most papers are doing great when they show 7 stops of DR so you've got some latitude to play with in PP. Heck most cameras from just a few years ago were lucky to capture 9 stops of DR...
Jack
-
Most papers are doing great when they show 7 stops of DR so you've got some latitude to play with in PP. Heck most cameras from just a few years ago were lucky to capture 9 stops of DR...
Jack
Jack,
I hope you are not implying that there is no need for more than 7 stops of DR in the original image if one intends to print.
Fortuitously, that DR limitation of the print seems to match pretty closely the 'static' DR limitation of the human eye.
The following explanation is from Wikipedia.
"The retina has a static contrast ratio of around 100:1 (about 6.5 f-stops). As soon as the eye moves (saccades) it re-adjusts its exposure both chemically and geometrically by adjusting the iris which regulates the size of the pupil."
When we process an image of a contrasty scene for monitor display or for printing, we compress the DR rather than reduce it. In other words, we raise the shadows in post processing in order to mimic the effect of the dilating pupil of the eye that occurred when we viewed the real scene before we photographed it. This is why I consider the DR capabilities of a camera very important.
-
I didn't see any raws posted.......
He posted a screen shot of his development settings in Lightrooom.
-
5 stops! I don't think my Canon would come close to that, in fact I know it can't. LR can recover, rebuild, whatever you want to call it, 5 stops (darn, 1 stop) from true raw clipping?
I'd have guessed, in my case, nor more than 1/2 stop. Perhaps less.
I think you and Bart are looking at different scenarios. If I expose a Stouffer wedge so that step 1 (0.05 OD) is at clipping and then increase exposure further I can recover perhaps 1 stop. However, if I block off the brightest 15 steps (5 stops), I can expose so that step 15 is just short of clipping and then normalize the image with the raw converter to reproduce the appearance of those steps.
Bill
-
If a sensor has a 14 stop acceptable SNR (the engineering definition) then by recovering 5 stops you've essentially got 9 (14-5) to display. Most papers are doing great when they show 7 stops of DR so you've got some latitude to play with in PP. Heck most cameras from just a few years ago were lucky to capture 9 stops of DR...
Jack,
I hope you are not implying that there is no need for more than 7 stops of DR in the original image if one intends to print.
Your prayers have been answered Ray, I was not implying that :) Merely pointing out that a 5 stop push at base ISO, resulting in about a 9 stop engineering DR with current Exmor cameras, may still provide quite a usable image, depending on uses. Here for instance is a controlled 5 stop push on a D810 by Jim Kasson (http://blog.kasson.com/?p=6574).
Note however that these captures are meant to show relative ISOlessness and they are not necessarily exposed to the right per this thread's context, potentially making the base ISO image 5 stops 'underexposed'. In order to properly ETTR this (near) ISOless camera in static tripod conditions one would want to bump exposure up 5 stops, producing the best IQ possible at ISO 64 (and presumably blowing all higher ISO captures).
Jack
-
Sensors are pretty well linear devices, so you can go as low as you can stand in terms of SNR if your converter is well behaved without penalty (it turns out that LR is not but others are). If a sensor has a 14 stop acceptable SNR (the engineering definition) then by recovering 5 stops you've essentially got 9 (14-5) to display. Most papers are doing great when they show 7 stops of DR so you've got some latitude to play with in PP. Heck most cameras from just a few years ago were lucky to capture 9 stops of DR...
Jack
Some time ago I measured the DR of my monitor and printing paper (just a regular cheap one with a cheap Epson printer), under regular indoor lighting conditions. I would have liked to measure a projector as well but I have none.
The resulting orders of magnitude were:
Monitor: 6,7 stops
(http://www.guillermoluijk.com/article/mpdrange/monitor.jpg)
(http://www.guillermoluijk.com/article/mpdrange/monitor.gif)
Paper: 4,3 stops
(http://www.guillermoluijk.com/article/mpdrange/papel.jpg)
(http://www.guillermoluijk.com/article/mpdrange/papel.gif)
Ambient light has a great influence in received DR. Specially a monitor will benefit from a big improvement when all the lights are turned off (deep blacks).
Regards
-
Some time ago I measured the DR of my monitor and printing paper (just a regular cheap one with a cheap Epson printer), under regular indoor lighting conditions.
Hi Guillermo,
I think you were measuring how the camera translates subject contrast to ADUs (Analog Digital Units, or DNs )
There is a difference between:
1. Subject contrast (as measured e.g. with an Exposure meter, Densitometer, or Spectrometer),
2. Camera system contrast (includes/combines optical transmission characteristics, and sensor characteristics) into an MTF response, and
3. Sensor characteristics (relatively easy to isolate by studying the Raw data).
Raw conversion effects are then a whole other can of additional worms.
The ETTR benefits/issues/opportunities are with 3. sensor characteristics. Combining 1 and 2 may make it difficult to see which element is responsible for what exactly. That's why focusing on 3. may help to see the effect on 2. better.
Cheers,
Bart
-
In fact I used the camera sensor as a high precision tri-band exposure meter. White balancing the RAW data over the white patch helped to find out that the blacks on my (non-calibrated) monitor were a bit more bluish than its whites. The opposite happened in the print test, the ink was a bit more reddish than the blank paper.
This is something an exposure meter can't provide. The RGB sensor's CFA permitted to get one step closer to a genuine spectral analysis.
Regards
-
In fact I used the camera sensor as a high precision tri-band exposure meter.
But what factor did the lens (and mirrorbox + filterstack with IR and OLPF filters) and interaction with the area sampling photosites of the sensor play? By adding a lens into the MTF equation, the whole system response changes the isolated sensor response. Imaging a white and a black patch in the same image will cause veiling glare in the lens, which differs between lenses, so one would need to calibrate that influence out.
I'm not saying it is not useful, it is useful in the end, but it clouds a clear view on the ETTR effects which are purely sensor and supporting electronic circuits related. It's a bit similar to the covering up of clipping or other tonecurve adjustments that Raw processors may introduce. It's better to cut them out of the equation if one wants to understand the sensor's role in isolation, and just study the Raw data (before demosaicing).
Cheers,
Bart
-
The influence of the lens system in the relative photon count, i.e. in the measured dynamic range between both patches (the brightest and darkest colour the devices under test can produce), is null since both patches are measured in a single capture and linearly processed. Those RAW histograms (just white balanced) are proportional to the number of photons impacting the observer's eyes.
Regards
-
He posted a screen shot of his development settings in Lightrooom.
Lightroom does a whole host of processing behind the scenes to try to make images look better - "content aware processing", etc. If the raws could be posted, it might be possible to do more than speculate. BTW, the screenshot appears to show the file as DNG, so apparently not the original raws anyway. Unless Canon have suddenly found religion on the subject of DNG :)
-
Lightroom does a whole host of processing behind the scenes to try to make images look better - "content aware processing", etc. If the raws could be posted, it might be possible to do more than speculate. BTW, the screenshot appears to show the file as DNG, so apparently not the original raws anyway. Unless Canon have suddenly found religion on the subject of DNG :)
I found the so called 'religion' of DNG, so what? The data is as raw as the original, it makes zero difference in the results obtained in my tests. What would be gained by having those DNG's? The results of the ISO 800 speak for themselves and as to why, it's been explained. That LR does processing behind the scenes is also IMHO not pertinent unless the development of raws isn't a fact in all this, I believe it is. The facts about actual raw clipping values obtained from RawDigger was also provided. And as Bernie pointed out, the development settings you missed were provided. So I'm at a loss as to what else is necessary for you to continue this discussion about ETTR and ETTR+D.
-
I found the so called 'religion' of DNG, so what? The data is as raw as the original, it makes zero difference in the results obtained in my tests. What would be gained by having those DNG's? The results of the ISO 800 speak for themselves and as to why, it's been explained. That LR does processing behind the scenes is also IMHO not pertinent unless the development of raws isn't a fact in all this, I believe it is. The facts about actual raw clipping values obtained from RawDigger was also provided. And as Bernie pointed out, the development settings you missed were provided. So I'm at a loss as to what else is necessary for you to continue this discussion about ETTR and ETTR+D.
The difference are most probably post processing, either in-camera or in a raw processor; I was curious to either confirm or deny that hypothesis. That's only possible with the original files. To me, it would seem better to take a look at the actual files rather than speculate for 8 pages. But that's me. I'm weird that way - I like hard data. Mea culpa. 8)
-
Some time ago I measured the DR of my monitor and printing paper (just a regular cheap one with a cheap Epson printer), under regular indoor lighting conditions.
Cool.
-
The difference are most probably post processing, either in-camera or in a raw processor; I was curious to either confirm or deny that hypothesis. That's only possible with the original files. To me, it would seem better to take a look at the actual files rather than speculate for 8 pages.
I agree, it's the absence of concrete data that needlessly/purposely drags the 'discussion' on to no resolve, with only some hints at files being (not) clipped, or exposed to the right (but cropped so they cannot be verified visually) or not (anybody's guess) without measurement data.
I'll see f I can find the time to fill in that gap, and produce some examples, although I can only produce Canon files (ISO variant and thus not really on topic). So maybe I'll just simulate the data (and take the critique that it's not based on a known Raw format for granted).
Cheers,
Bart
-
The influence of the lens system in the relative photon count, i.e. in the measured dynamic range between both patches (the brightest and darkest colour the devices under test can produce), is null since both patches are measured in a single capture and linearly processed.
Hi Guillermo,
When both white and black patches are measured/photographed together, then veiling glare (with a variable amount depending on the lens coatings and barrel design) will reduce contrast. Potentially, although you seem to have tried to avoid it, additional surface reflections of the patches will also reduce the contrast.
It's better to fill the frame with a single uniform brightness, and record (with a calibrated Photon Transfer Curve) the number of Photons (or if the same gain is used then the ADU's will suffice) that were actually recorded (can be derived from the standard deviation and the ADC gain). Then change the exposure time and measure the same.
Care must be taken to exclude Vignetting and uneven illumination from skewing the Raw counts of the same photosite positions, because they increase the apparent noise level. It's even more accurate if two identical frames are used per measurement, which allows to subtract them in quadrature to eliminate non-random noise, which leaves only photon shot noise.
Cheers,
Bart
-
Lightroom does a whole host of processing behind the scenes to try to make images look better - "content aware processing", etc. If the raws could be posted, it might be possible to do more than speculate. BTW, the screenshot appears to show the file as DNG, so apparently not the original raws anyway. Unless Canon have suddenly found religion on the subject of DNG :)
This has nothing to do with me pointing out your "jpg" answer was not applicable. Your "jpg" answer is still not applicable.
-
This has nothing to do with me pointing out your "jpg" answer was not applicable. Your "jpg" answer is still not applicable.
Yes, the absence of posted raws or any screenshots that showed CR2s fooled me. My bad. But post processing is post processing. Whether its in-camera or in a raw processor on a PC doesn't make much difference. Other than if we had the raws it might be possible to work out what's happening whereas with in-camera processing we couldn't.
-
Hi Guillermo,
When both white and black patches are measured/photographed together, then veiling glare (with a variable amount depending on the lens coatings and barrel design) will reduce contrast. Potentially, although you seem to have tried to avoid it, additional surface reflections of the patches will also reduce the contrast.
It's better to fill the frame with a single uniform brightness, and record (with a calibrated Photon Transfer Curve) the number of Photons (or if the same gain is used then the ADU's will suffice) that were actually recorded (can be derived from the standard deviation and the ADC gain). Then change the exposure time and measure the same.
Care must be taken to exclude Vignetting and uneven illumination from skewing the Raw counts of the same photosite positions, because they increase the apparent noise level. It's even more accurate if two identical frames are used per measurement, which allows to subtract them in quadrature to eliminate non-random noise, which leaves only photon shot noise.
Ok I see now what you meant. Veiling glare was very far from being a problem here. With indoor indirect lighting nothing to worry.
Regarding the noise/vignetting issue, the analysis was done over a centred crop with a quite tele lens to avoid any vignetting. And to minimise the influence of noise I did a HDR composite 3 or 4 stops apart. The ultra-narrow RAW histograms are a proof of this.
Regards
-
Ok I see now what you meant. Veiling glare was very far from being a problem here. With indoor indirect lighting nothing to worry.
Regarding the noise/vignetting issue, the analysis was done over a centred crop with a quite tele lens to avoid any vignetting. And to minimise the influence of noise I did a HDR composite 3 or 4 stops apart. The ultra-narrow RAW histograms are a proof of this.
Thanks for the additional information.
Cheers,
Bart
-
But post processing is post processing. Whether its in-camera or in a raw processor on a PC doesn't make much difference.
Is having to normalize the ETTR capture in a raw converter post processing and some new revelation to you? Is that what you mean by post processing? As Bernie stated, not that it's necessary if you follow the examples and text provided, this has nothing to do with JPEGs or post processing unless you feel that ETTR captures should have zero processing parameters applied to them in the raw processor. That would seem a rather silly approach.
-
Is having to normalize the ETTR capture in a raw converter post processing and some new revelation to you? Is that what you mean by post processing? As Bernie stated, not that it's necessary if you follow the examples and text provided, this has nothing to do with JPEGs or post processing unless you feel that ETTR captures should have zero processing parameters applied to them in the raw processor. That would seem a rather silly approach.
Would you mind having that discussion with the Nikon executives? They are so 'out of camera' centric in their approach, functionally that should be built into the camera is overlooked.
-
As far as the term ISO-less is concerned, this is my understanding :-
Sensor Read noise can be introduced prior to the ISO amplifier or after the amplifier in the A/D converter circuitry. The noise introduced before the amp is amplified when the gain is increased as a result of an increase in ISO setting but the A/D noise is not.
With Canon sensors, it appears that at base ISO, A/D converter noise is a significant contributor to the total noise in the final digital signal. As ISO is increased, the noise introduced before the amplifier is increasingly amplified until at some point, it is the dominating contributor and the contribution from the A/D is insignificant. Once you reach this point, the sensor becomes virtually ISO-less as it makes little difference whether you increase the “exposure” digitally or in the amplifier. Before this point however, increasing gain digitally increases both sources of noise and hence is less effective than increasing ISO gain.
On the other hand, the latest Sony/Nikon sensors seem to have very little A/D noise (presumably because of the use of integrated column A/D’s.) So for these sensors, the noise introduced before the amplifier is always the dominant noise for any ISO. These cameras are therefore virtually ISO-less for the entire ISO range as it makes little difference whether the ‘exposure” is adjusted digitally or in the column amplifiers.
ETTR is really a separate issue and is more about making use of the full dynamic range of the camera. It may involve more exposure than is necessary for a given scene which in turn allows the exposure (and also the noise) to be wound down in pp.
These principles can be demonstrated with a few calculations. The main sources of noise in a digital capture with scenes of relatively normal luminance are shot noise and read noise. As Roger Clark demonstrates in his post (http://www.clarkvision.com/articles/evaluation-1d2/) on determining noise and signal to noise ratios (SNR) of a digital sensor, the characteristics of the sensor can be modeled reasonably well by shot noise and read noise. At relatively normal exposures, dark current is not significant. Shot noise is the square root of the number of photons captured, and read noise for a given ISO is constant. One can determine the total noise by adding the shot noise and read noise in quadrature (one adds the squares of the noise sources and takes the square root of the sum).
The table below uses Roger’s data for the Canon 1D Mark II to calculate SNRs for various exposures with this camera. At base ISO the read noise is relatively high at 16.1 e- (electrons) and it drops to 4.04 e- at ISO 800. The full well of the sensor is 53,000 e-. If one exposes to the right at base ISO, the brightest f/stop is represented by 53,000 e- and this number is halved for each progressively darker f/stop as shown in the left most portion of the table. The total noise is calculated by adding the shot noise and read noise in quadrature, and the signal to noise ratio (SNR) is the quotient of the number of electrons (signal) and the total noise.
Above base ISO, each doubling of the ISO halves the number of electrons at highlight clipping, so 6625 e- are collected by an ETTR exposure at ISO 800. The calculations for ISO 800 are shown in the middle portion of the table. The SNRs at all levels are superior for ETTR exposures at base ISO as compared to ISO 800. In a situation where shutter speed and aperture considerations restrain exposure to 6625 electrons, one can use ISO 800 rather than base ISO to reduce the read noise. The SNRs for this exposure are shown for ISO 800 and ISO 100. In the brighter f/stops of the exposure where shot noise is predominant, the SNR is only slightly worse at ISO 100, but in the shadows where read noise predominates, the SNR is better at ISO 800.
I don’t have data for Andrew’s 5D Mark II, but if one can extrapolate from data for the 1D Mark II, it is apparent that he would obtain the best SNRs by fully exposing at base ISO. With reduced exposure at 1/60 sec and f/5.6, the SNRs in the highlights are not much different, but the SNRs in the deep shadows are markedly better. The mid-tones (18%) would be represented by 1192 e-, somewhere between steps 2 and 3, and the SNR is only marginally better at ISO 800. With an ISO-less sensor, the read noise would be constant and there would be little difference between ISO 100 and ISO 800 for a given exposure.
The minimal SNR that gives acceptable image quality is somewhat a personal decision. Emil Martinec has a graphic demonstrating relatively low SNRs (Figure 19 in his post (http://theory.uchicago.edu/~ejm/pix/20d/tests/noise/noise-p3.html#bitdepth)). A SNR of 8 in the shadows is acceptable to me, and a SNR of 4 is marginal.
Bill
-
Good work, Bill.
To relate this to ISO invariant cameras, there would be no difference in read noise between ISO 800 and 100 for them.
Jack
-
These principles can be demonstrated with a few calculations. The main sources of noise in a digital capture with scenes of relatively normal luminance are shot noise and read noise. As Roger Clark demonstrates in his post (http://www.clarkvision.com/articles/evaluation-1d2/) on determining noise and signal to noise ratios (SNR) of a digital sensor, the characteristics of the sensor can be modeled reasonably well by shot noise and read noise. At relatively normal exposures, dark current is not significant. Shot noise is the square root of the number of photons captured, and read noise for a given ISO is constant. One can determine the total noise by adding the shot noise and read noise in quadrature (one adds the squares of the noise sources and takes the square root of the sum).
The table below uses Roger’s data for the Canon 1D Mark II to calculate SNRs for various exposures with this camera. At base ISO the read noise is relatively high at 16.1 e- (electrons) and it drops to 4.04 e- at ISO 800. The full well of the sensor is 53,000 e-. If one exposes to the right at base ISO, the brightest f/stop is represented by 53,000 e- and this number is halved for each progressively darker f/stop as shown in the left most portion of the table. The total noise is calculated by adding the shot noise and read noise in quadrature, and the signal to noise ratio (SNR) is the quotient of the number of electrons (signal) and the total noise.
Above base ISO, each doubling of the ISO halves the number of electrons at highlight clipping, so 6625 e- are collected by an ETTR exposure at ISO 800. The calculations for ISO 800 are shown in the middle portion of the table. The SNRs at all levels are superior for ETTR exposures at base ISO as compared to ISO 800. In a situation where shutter speed and aperture considerations restrain exposure to 6625 electrons, one can use ISO 800 rather than base ISO to reduce the read noise. The SNRs for this exposure are shown for ISO 800 and ISO 100. In the brighter f/stops of the exposure where shot noise is predominant, the SNR is only slightly worse at ISO 100, but in the shadows where read noise predominates, the SNR is better at ISO 800.
I don’t have data for Andrew’s 5D Mark II, but if one can extrapolate from data for the 1D Mark II, it is apparent that he would obtain the best SNRs by fully exposing at base ISO. With reduced exposure at 1/60 sec and f/5.6, the SNRs in the highlights are not much different, but the SNRs in the deep shadows are markedly better. The mid-tones (18%) would be represented by 1192 e-, somewhere between steps 2 and 3, and the SNR is only marginally better at ISO 800. With an ISO-less sensor, the read noise would be constant and there would be little difference between ISO 100 and ISO 800 for a given exposure.
The minimal SNR that gives acceptable image quality is somewhat a personal decision. Emil Martinec has a graphic demonstrating relatively low SNRs (Figure 19 in his post (http://theory.uchicago.edu/~ejm/pix/20d/tests/noise/noise-p3.html#bitdepth)). A SNR of 8 in the shadows is acceptable to me, and a SNR of 4 is marginal.
Bill
Thanks, Bill.....
Supports what Emil was saying in his statement on Maximizing Exposure but with data to explain it.
It is important to remember that Andrew's ISO 800 exposure is only better, as you show, because he maximized exposure (ETTR). Had he also maximized exposure at ISO 100, by lowering aperture/shutter (left section of your chart) he would have gotten the best image of the three. ISO 800 only gives a better image than ISO 100 because on ETTR.
THANKS AGAIN....
-
Yes thanks for that Bill, very interesting.
One other minor point I would raise with these considerations is this - how closely does the clipping point of the amp or A/D converter match the full well capacity of the sensor ? Presumably it's a bit higher which means that when ISO is raised it is possible to get a bit more exposure without clipping. But I think this is probably a minor issue.
Dave
-
One other minor point I would raise with these considerations is this - how closely does the clipping point of the amp or A/D converter match the full well capacity of the sensor ? Presumably it's a bit higher which means that when ISO is raised it is possible to get a bit more exposure without clipping. But I think this is probably a minor issue.
I am not sure I understand this. If Exposure is fixed, it is fixed independently of ISO. In an ISOless camera that's it. In an ISOful camera raising ISO from base may bring out the shadows while clipping the highlights stop for stop, correct? See this (http://forum.luminous-landscape.com/index.php?topic=101242.msg831158#msg831158) example.
Jack
-
I am not sure I understand this. If Exposure is fixed, it is fixed independently of ISO. In an ISOless camera that's it. In an ISOful camera raising ISO from base may bring out the shadows while clipping the highlights stop for stop, correct? See this (http://forum.luminous-landscape.com/index.php?topic=101242.msg831158#msg831158) example.
Jack
Jack, as is often the case, I probably didn't explain what I was trying to say very well.
In the comparisons being made in this thread, the assumption is made that if a certain exposure just causes clipping at base ISO, then if the ISO is increased by say 8 times, then the exposure will have to be reduced by 8 times to still just avoid clipping. What I'm saying is that this is not necessarily quite true. If we assume that at base ISO, the onset of clipping corresponds to the sensels reaching full well capacity, then as ISO is increased this will at some point no longer be the case and the clipping point will be determined by the overload point of the amplifier or A/D converter input. I am assuming that this will occur at a slightly higher voltage than that reached at the A/D input at base ISO with the sensel at full well capacity. So if the ISO is raised by 8 times, the exposure might only have to be reduced by say 7.8 times to just reach the clipping point.
Does that make any more sense ?
Dave
-
One other minor point I would raise with these considerations is this - how closely does the clipping point of the amp or A/D converter match the full well capacity of the sensor ? Presumably it's a bit higher which means that when ISO is raised it is possible to get a bit more exposure without clipping. But I think this is probably a minor issue.
The clipping point of the ADC at base ISO is usually set somewhat below the FWC of the sensor, because the sensor usually goes a little nonlinear before it goes really nonlinear. So the situation you raised won't normally happen.
That said, when I and many others calculate the FWC from experimenting with the camera, the number that we report is that corresponding to full scale at base ISO. The real FWC is probably higher.
Jim
-
The clipping point of the ADC at base ISO is usually set somewhat below the FWC of the sensor, because the sensor usually goes a little nonlinear before it goes really nonlinear. So the situation you raised won't normally happen.
That said, when I and many others calculate the FWC from experimenting with the camera, the number that we report is that corresponding to full scale at base ISO. The real FWC is probably higher.
Jim
Ah that makes sense, thanks Jim.
I think the reason I had it in my head that the A/D clipping point was a bit higher than the FWC was that quite some time ago I found that I could get a higher maximum digital raw value out of my old Canon600D (Rebel) at ISO200 (and above) than at ISO100. However that was older technology and of course my measurements could have been flawed.
Dave
-
The clipping point of the ADC at base ISO is usually set somewhat below the FWC of the sensor, because the sensor usually goes a little nonlinear before it goes really nonlinear.
Jim, when the RAW saturation level is below the end of the ADC scale (2^n-1 being n the number of bits), don't you think this could mean the FWC is below the clipping point of the ADC?.
Typical Canon 5D RAW histogram, the clipping point 3692 is below the right end of the ADC theoretical scale (4095):
(http://www.guillermoluijk.com/tutorial/satlevel/hist5d.gif)
Regards
-
For what it's worth, some green channel clipping values I've observed on the two cameras I've owned
Canon 600D
ISO100 : 12,000
ISO200-1600 : 13,700
Nikon D610
ISO100-800 : 15,780
ISO1600 : 16,273
ISO2000-3200 : 16,383
I suspect the values for ISO above 800 for the D610 relate to the use of "digital ISO".
Dave
-
For what it's worth, some green channel clipping values I've observed on the two cameras I've owned
Canon 600D
ISO100 : 12,000
ISO200-1600 : 13,700
Nikon D610
ISO100-800 : 15,780
ISO1600 : 16,273
ISO2000-3200 : 16,383
I suspect the values for ISO above 800 for the D610 relate to the use of "digital ISO".
FWIW, here are clipping values for the D800e at base ISO obtained by gross overexposure as shown by Rawdigger. The red and blue clip at 16383, presumably due to clipping in the ADC. The green clips at 15768 in both channels. If this is due to well saturation, I would expect some variation due to pixel response non-uniformity. Perhaps Jim or others can comment.
Bill
-
Jim, when the RAW saturation level is below the end of the ADC scale (2^n-1 being n the number of bits), don't you think this could mean the FWC is below the clipping point of the ADC?.
Typical Canon 5D RAW histogram, the clipping point 3692 is below the right end of the ADC theoretical scale (4095):
(http://www.guillermoluijk.com/tutorial/satlevel/hist5d.gif)
Regards
Maybe I spoke too fast. I know nothing about Canons, mainly Leicas, Nikons, and Sonys. I do note that Nikon full scale is sometimes not as high as you'd think, but there is no sign of saturation before clipping, so I think that real sensor saturation must be higher. With the Nikons, WB prescaling is part of the problem.
Jack, want to weigh in on this?
Jim
-
Thanks, Bill.....
Supports what Emil was saying in his statement on Maximizing Exposure but with data to explain it.
It is important to remember that Andrew's ISO 800 exposure is only better, as you show, because he maximized exposure (ETTR). Had he also maximized exposure at ISO 100, by lowering aperture/shutter (left section of your chart) he would have gotten the best image of the three. ISO 800 only gives a better image than ISO 100 because on ETTR.
THANKS AGAIN....
This is wrong. Andrew's images were the SAME exposure. By boosting ISO in his non-ISOless camera he increased signal to noise ratio. NOT exposure.
-
This is wrong. Andrew's images were the SAME exposure. By boosting ISO in his non-ISOless camera he increased signal to noise ratio. NOT exposure.
You have defined what ETTR does, as I said in my post. Andrew seemed to be asking why 800 ISO had lower noise than 100 ISO. It's not, if both images are normalized to either "standard" exposure or "maximized" exposure. It only appears to be if the 100 is a standard exposure and the 800 is maximized....as shown in Bill's chart.
-
I don't really understand what you are trying to convey. We already knew why Andrew's image's are the way they are. I'm just saying that you are not correct to say one was a higher exposure than the other. They were the same exposure and the noise (actually SNR) difference is the variable that has been isolated.
-
Maybe I spoke too fast. I know nothing about Canons, mainly Leicas, Nikons, and Sonys. I do note that Nikon full scale is sometimes not as high as you'd think, but there is no sign of saturation before clipping, so I think that real sensor saturation must be higher. With the Nikons, WB prescaling is part of the problem.
Jack, want to weigh in on this?
Jim, I am also not versed in Canon.
However the Nikon numbers are typically explained by black level subtraction. For instance most pre-2015 Nikons clipped blacks before writing ADUs into the raw data. At 14-bits at base ISO the black level is typically around 600 ADUs, so the maximum number possible out of an unscaled 14 bit ADC becomes 16383-600 = 15783 or so, what can be seen in the green channels. As you know the B and R channels are pre-conditioned after black level subtraction (gotta figure out why) so they can be scaled back to (or near) the top value. The black level sometimes changes with the ISO (I think for the D610 it lowers to around 100 ADUs a couple ISO stops up). This should explain Bill and David's numbers.
So your thinking is correct, as far as I understand Nikon cameras: the gain/bias to the ADC at base ISO is chosen so that it saturates before full well capacity is reached. And as long as the histogram looks like it hits a brick wall I believe that is always the case, independently of camera make, including Guillermo's Canon.
Jack
PS The top level of the ADC can be an arbitrary level, chosen so as not to show the non linear part of the curve before FWC. I believe Canon does this in several of its cameras. The value at which the curve becomes non-linear may change based on amplification (ISO).
-
This is wrong. Andrew's images were the SAME exposure. By boosting ISO in his non-ISOless camera he increased signal to noise ratio. NOT exposure.
Exactly! I thought this was a simple demo of just that fact.
-
Maybe I spoke too fast. I know nothing about Canons, mainly Leicas, Nikons, and Sonys. I do note that Nikon full scale is sometimes not as high as you'd think, but there is no sign of saturation before clipping, so I think that real sensor saturation must be higher.
Hi Jim,
That's what I've understood as well, usually (always is a big word) the real sensor saturation is higher than the ADC saturation which produces the ADU or DN we find in Raw files. This is to benefit from the virtually linear portion of the Photon Transfer Curve, which makes color math a lot easier. I seem to remember a post on DPreview, not sure if it was for Eric Fossum, that the real sensor saturation can be something like 40% higher than the ADC saturation suggests. Here (http://www.sensorgen.info/) is a nice compilation of DxO originated ADC/Raw file Saturation data for various cameras, expressed in e- or actually converted Photons. When we divide the ADC photon saturation by the number we can find for clipped highlights in a Raw file, we'll get the conversion gain.
It also demonstrates that Andrews EOS 5D2 (http://www.sensorgen.info/CanonEOS-5D-Mark-II.html) has a Read noise that keeps dropping as the ISO is boosted, all the way to 1600 (I think the ISO 3200 value is a glitch), but at the same time the saturation level reduces faster, thus lowering the total DR. As I've said before, and Bill already showed in his calculation example, nothing beats real photons if one is to maximize the S/N of a capture. However, if one approaches the read noise level by underexposure (due to constraints like avoiding camera shake or subject motion) of the brightest relevant scene data, the benefit of the reduced read noise at boosted ISO settings becomes relatively important. But this is specific for ISO variant cameras, like the Canons, and not what the OP was asking about.
With the Nikons, WB prescaling is part of the problem.
It complicates understanding what the real exposure level was, so one needs to reverse engineer that from the photon statistics for the Red and Blue color planes, compared to the Green color planes.
I believe that Nikon at least dropped the read noise clipping for the D810, so that becomes easier to accurately evaluate, and maybe that also hints at the multiplication for R and B.
Cheers,
Bart
-
Here (http://www.sensorgen.info/) is a nice compilation of DxO originated ADC/Raw file Saturation data for various cameras, expressed in e- or actually converted Photons. When we divide the ADC photon saturation by the number we can find for clipped highlights in a Raw file, we'll get the conversion gain.
Bart,
Thanks to the link for the Sensogren data for the 5DMII. Using that data, I redid my calculations for that camera. The read noise at base ISO for this camera is considerably higher than for the 1D Mark II that I used in my original calculations, and the read noise at ISO 800 is only slightly higher. The improvement in SNR by going to ISO 800 with constant exposure is even more marked.
Bill
-
Here's a screen capture of the entire image also showing the somewhat radical settings used in Develop. Clipping indicators are on in Histogram, no clipping. RGB value over brightest area of dogs head reads 98%.
What's interesting (to me) is that while the 'exposure' was boosted 3 stops via ISO, note that the Exposure slider is set to -1.55 stops which as I pointed out earlier is just about the limit of the ETTR testing on this camera done in the past. I can't explain why there is this disconnect between the +3 ISO and -1.5 Exposure slider among the other sliders as set.
(http://digitaldog.net/files/DogISO800.jpg)
The lighting in this scene is highly directional, and I question of the incident light reading gave the best exposure for the highlights of the dog's head, which are at 98% in your adjusted rendering. One could get different readings according to the direction in which the integrating sphere of the incident light meter is pointed. A proper exposure at ISO 100 would have these highlights just short of clipping in the raw file. It would be of interest to look at the raw file in Rawdigger to determine if the exposure was really optimum. It would be helpful if you would post the raw files so we could draw our own conclusions. If the highlights in the ISO 100 exposure are just short of clipping, they would be severely clipped if one used the same exposure (f/stop, shutter speed) at ISO 800. Highlight recovery could reduce clipping in the rendered image, but 3 stops of recovery is doubtful.
Bill
-
The lighting in this scene is highly directional, and I question of the incident light reading gave the best exposure for the highlights of the dog's head, which are at 98% in your adjusted rendering. One could get different readings according to the direction in which the integrating sphere of the incident light meter is pointed. A proper exposure at ISO 100 would have these highlights just short of clipping in the raw file. It would be of interest to look at the raw file in Rawdigger to determine if the exposure was really optimum. It would be helpful if you would post the raw files so we could draw our own conclusions. If the highlights in the ISO 100 exposure are just short of clipping, they would be severely clipped if one used the same exposure (f/stop, shutter speed) at ISO 800. Highlight recovery could reduce clipping in the rendered image, but 3 stops of recovery is doubtful.
Bill
I have found that the optimal exposure (no clipped highlights) can be obtained by spot metering the brightest significant highlight and adding 3 to 3 1/3 stops.
This will normally result in LR having the brightest highlight in the mid to low 80%. Anything above that, in the current PV, will have clipped and recovered channel(s). (This is very well explain in the George Jardin article I posted earlier in this thread)
This was arrived at using RawDigger for tests. As I remember, Eric Chan uses 3 1/3 for his 5D3.
-
Jack and Bart, thanks for your explanations of clipping level. All is becoming clear !
Dave
-
I have found that the optimal exposure (no clipped highlights) can be obtained by spot metering the brightest significant highlight and adding 3 to 3 1/3 stops.
...
This was arrived at using RawDigger for tests. As I remember, Eric Chan uses 3 1/3 for his 5D3.
Keep in mind that this value is camera specific because Sensitivity (ISO) in its current incarnation is unrelated to raw data.
Jack
-
Keep in mind that this value is camera specific because Sensitivity (ISO) in its current incarnation is unrelated to raw data.
Jack
I am aware and agree. That's why I mentioned the 5d3. Also, many cameras do not have spot meters, which means they would need a different approach to determine the exposure for the significant highlight.
EDIT: Did you mean camera model specific or individual camera specific? also, not sure what you mean by "unrelated to raw data"
-
I don't really understand what you are trying to convey. We already knew why Andrew's image's are the way they are. I'm just saying that you are not correct to say one was a higher exposure than the other. They were the same exposure and the noise (actually SNR) difference is the variable that has been isolated.
I agree that I did not make clear what I was trying to say. We are generally in agreement.
-
I have found that the optimal exposure (no clipped highlights) can be obtained by spot metering the brightest significant highlight and adding 3 to 3 1/3 stops.
This will normally result in LR having the brightest highlight in the mid to low 80%. Anything above that, in the current PV, will have clipped and recovered channel(s). (This is very well explain in the George Jardin article I posted earlier in this thread)
This was arrived at using RawDigger for tests. As I remember, Eric Chan uses 3 1/3 for his 5D3.
Agreed, I can confirm that I also measured 3 1/3rd for my 1Ds Mark III, unless there is a lot of noise (e.g. very high operating temp), then there may be a very tiny bit, a few pixels, of the shot-noise tail getting clipped. Also extremely saturated (flower) colors may clip one channel of the Raw data. So bracketing is still useful for that last bit of capture quality.
Cheers,
Bart
-
EDIT: Did you mean camera model specific or individual camera specific? also, not sure what you mean by "unrelated to raw data"
Camera model, and in fact it seems that individual manufacturers are fairly consistent across their own camera classes of the same generation.
As to raw data, unfortunately the latest standards that regulate metering and ISO do not concern themselves with raw data. Madness, I know. This is the formula that ties them together (from wikipedia (https://en.wikipedia.org/?title=Light_meter))
(https://upload.wikimedia.org/math/b/c/e/bce15c13b5f6af0d4d477a4d6b279120.png)
For a given spot luminance (L) from the scene the (reflected light) meter in the camera will serve up values for f-number (N) and exposure time (t) that satisfy the equation. K is a manufacturer specific constant (apparently 12.5 for CaNikon) and S is, you guessed it, ISO. But as we know the new ISO non-standard (ISO 12232:2006 (https://en.wikipedia.org/wiki/Film_speed#The_ISO_12232:2006_standard)) allows camera manufacturers to report ISOs that they individually believe will produce pleasing brightness in the OOC jpeg. And they nowadays mostly take advantage of this possibility, reported as REI in the exif - check it out in one of your files. After a whole bunch of subjective, highly non linear transformations. And they call it a 'standard'. Ugh. That's why DxO is forced to determine their own 'Measured ISOs' (Ssat), which instead are related to the raw data linearly: Hsat = 78 / Ssat.
In other words, the spot meter may provide perfectly calibrated Exposure values for 'middle gray' at ISO 100 - but we have no idea where the manufacturer will place it in the raw data. We only know that that manufacturer will be pleased by its camera's OOC image. Enter RawDigger :(
Jack
-
Camera model, and in fact it seems that individual manufacturers are fairly consistent across their own camera classes of the same generation.
As to raw data, unfortunately the latest standards that regulate metering and ISO do not concern themselves with raw data. Madness, I know. This is the formula that ties them together (from wikipedia (https://en.wikipedia.org/?title=Light_meter))
(https://upload.wikimedia.org/math/b/c/e/bce15c13b5f6af0d4d477a4d6b279120.png)
For a given spot luminance (L) from the scene the (reflected light) spot meter in the camera will serve up values for f-number (N) and exposure time (t) that satisfy the equation. K is a manufacturer specific constant (apparently 12.5 for CaNikon) and S is, you guessed it, ISO. But as we know the new ISO non-standard (ISO 12232:2006 (https://en.wikipedia.org/wiki/Film_speed#The_ISO_12232:2006_standard)) allows camera manufacturers to report ISOs that they individually believe will produce pleasing brightness in the OOC jpeg. And they nowadays mostly take advantage of this possibility, reported as REI in the exif - check it out in one of your files. After a whole bunch of subjective, highly non linear transformations. And they call it a 'standard'. Ugh. That's why DxO is forced to determine their own 'Measured ISOs' (Ssat), which instead are[/b ]related to the raw data linearly: Hsat = 78 / Ssat.
In other words, the spot meter may provide perfectly calibrated Exposure values for 'middle gray' at ISO 100 - but we have no idea where the manufacturer will place it in the raw data. We only know that that manufacturer will be pleased by its camera's OOC image :(
Jack
When I talk about a spot meter, I am talking about one in the camera, not a handheld one.
I then "calibrate" the camera, using RAwDigger to determine how much +EC to get the significant highlight "middle grey" to be just below clipping.
If you were talking about a handheld meter, I agree....you have a problem. The meter calibration will not necessarily be calibrated to the metering of the camera. Chuck Gardner, in. One of his tutorials, covered how to get the meter and camera in sync. http://super.nova.org/DPR/SpotmeterCompensation/
-
Camera model, and in fact it seems that individual manufacturers are fairly consistent across their own camera classes of the same generation.
As to raw data, unfortunately the latest standards that regulate metering and ISO do not concern themselves with raw data. Madness, I know. This is the formula that ties them together (from wikipedia (https://en.wikipedia.org/?title=Light_meter))
(https://upload.wikimedia.org/math/b/c/e/bce15c13b5f6af0d4d477a4d6b279120.png)
For a given spot luminance (L) from the scene the (reflected light) meter in the camera will serve up values for f-number (N) and exposure time (t) that satisfy the equation. K is a manufacturer specific constant (apparently 12.5 for CaNikon) and S is, you guessed it, ISO. But as we know the new ISO non-standard (ISO 12232:2006 (https://en.wikipedia.org/wiki/Film_speed#The_ISO_12232:2006_standard)) allows camera manufacturers to report ISOs that they individually believe will produce pleasing brightness in the OOC jpeg. And they nowadays mostly take advantage of this possibility, reported as REI in the exif - check it out in one of your files. After a whole bunch of subjective, highly non linear transformations. And they call it a 'standard'. Ugh. That's why DxO is forced to determine their own 'Measured ISOs' (Ssat), which instead are related to the raw data linearly: Hsat = 78 / Ssat.
In other words, the spot meter may provide perfectly calibrated Exposure values for 'middle gray' at ISO 100 - but we have no idea where the manufacturer will place it in the raw data. We only know that that manufacturer will be pleased by its camera's OOC image. Enter RawDigger :(
Jack
It is not as complicated as it would first appear. Japanese camera manufactures are required to use the SOS or REI calibration for the sensor. Since REI (recommended exposure index) allows an arbitrary value, the camera makers can use the old saturation standard which allows 0.5 EV of highlight headroom and places middle gray at 12.7% saturation. Things have not changed that much since Thom Hogan (http://www.bythom.com/graycards.htm) discussed the subject of gray cards in 2003. Doug Kerr (http://dougkerr.net/Pumpkin/articles/Exposure_calibration.pdf) discusses exposure meter calibration and points out that Canon uses ISO2721 with a k value of 12.5 so that the camera meter reading will agree with a hand held meter using a k of 12.5 as employed by Sekonic as well as Nikon. Minolta, Pentax, and Kenko (Wikipedia (https://en.wikipedia.org/?title=Light_meter)) use a k of 14 (different by about 1/6 EV). According to Doug's Canon source, Canon rates its sensors such that an exposure according to the meter reading will result in a saturation of 17.3%. They do not allow the 0.5 EV cushion for the highlights.
My Nikon D800e places the exposure of a uniformly illuminated surface (white or gray card) exposed according to the built in meter at 14% saturation as measured by Rawdigger. This is close to the saturation based value of 12.7% and is 2.84 stops below clipping.
Bill
-
When I talk about a spot meter, I am talking about one in the camera, not a handheld one.
So am I :)
It is not as complicated as it would first appear. Japanese camera manufactures are required to use the SOS or REI calibration for the sensor. Since REI (recommended exposure index) allows an arbitrary value, the camera makers can use the old saturation standard which allows 0.5 EV of highlight headroom and places middle gray at 12.7% saturation. Things have not changed that much since Thom Hogan (http://www.bythom.com/graycards.htm) discussed the subject of gray cards in 2003. Doug Kerr (http://dougkerr.net/Pumpkin/articles/Exposure_calibration.pdf) discusses exposure meter calibration and points out that Canon uses ISO2721 with a k value of 12.5 so that the camera meter reading will agree with a hand held meter using a k of 12.5 as employed by Sekonic as well as Nikon. Minolta, Pentax, and Kenko (Wikipedia (https://en.wikipedia.org/?title=Light_meter)) use a k of 14 (different by about 1/6 EV). According to Doug's Canon source, Canon rates its sensors such that an exposure according to the meter reading will result in a saturation of 17.3%. They do not allow the 0.5 EV cushion for the highlights.
My Nikon D800e places the exposure of a uniformly illuminated surface (white or gray card) exposed according to the built in meter at 14% saturation as measured by Rawdigger. This is close to the saturation based value of 12.7% and is 2.84 stops below clipping.
Bill
Hi Bill,
If it only were so simple ;-) Unfortunately the 2006 incarnation of the standard for ISO made things worse than they were originally.
REI is arbitrary and SOS is based on the rendered OOC sRGB jpeg, not raw data. Thom's article is woefully inaccurate and Kerr's out of date.
If you have any other Nikon camera from an earlier or later generation, test them out. Both my D90 and D610 put spot metered middle gray below 10% of saturation in the raw data and, with 14 stops of eDR in the D610, I am glad about that.
Jack
-
How much effect does that have on comparing cameras? For example, if camera A is touted as great for low light because it is usable at ISO 12,800 could this be because it is actually equivalent to camera B at 6,400?
-
So am I :)
OK.....so then I am confused. How does this affect "calibrating" the camera...using RawDigger....to know the EC to add to spot metered highlight to get optimized exposure....I.e. ETTR?
-
OK.....so then I am confused. How does this affect "calibrating" the camera...using RawDigger....to know the EC to add to spot metered highlight to get optimized exposure....I.e. ETTR?
This excellent article (http://www.rawdigger.com/howtouse/calibrate-exposure-meter-to-improve-dynamic-range) by the Rawdigger staff explains the process in some detail. The important steps are to determine the clipping point of the green channels and the sensor saturation according to the camera light meter reading. To complete the process, note the relationship of raw clipping to the camera histogram and blinkies. Once understood, the process is not that complicated.
One fine point that Iliah points out is that clipping is somewhat of a gradual process. It begins when the right tail of the bell shaped histogram (shot noise distribution) begins to clip and is complete when the left tail reaches clipping. To include 95% of the data, one can add two standard deviiations to the mean.
Bill
-
One fine point that Iliah points out is that clipping is somewhat of a gradual process. It begins when the right tail of the bell shaped histogram (shot noise distribution) begins to clip and is complete when the left tail reaches clipping. To include 95% of the data, one can add two standard deviiations to the mean.
That would be true if the shot noise were Gaussian. As per my previous discussion with Erik on this thread, I usually elide the fact that it's actually Poisson. I haven't run the numbers, and I'm sure the difference is small near fullscale where you're doing this test, but, since I brought it up before, I thought I'd point it out as possibly being of academic interest.
The difference is not small near zero, and I would like to publicly thank Jack Hogan for pointing that out and for suggesting fixes to some of my Matlab code that fixed some earlier errors.
Jim
-
How much effect does that have on comparing cameras? For example, if camera A is touted as great for low light because it is usable at ISO 12,800 could this be because it is actually equivalent to camera B at 6,400?
Hi,
It's anybody's guess what that means, until one calibrates for actual Raw file data levels.
Really, it's very simple, it can also be done with Fast RAW Viewer (FRV) automatically if one doesn't have RawDigger or similar. With FRV, just set the preferences to "ETTR style Autoexposure (shift histogram to the right)", and set the "Saturate up to" option to 0.0% pixels. Now shoot a uniform surface (well lit gray card, or as I do shoot through a piece of opal glass or white translucent perspex flush with the filter threads of the lens) with the camera on Aperture priority metering, and set that aperture to something like f/8 (or narrower).
FRV will tell you after Auto Exposure (Shift A), how many EVs the Raw exposure levels are removed from clipping the Raw data. To reduce the Photon shot-noise tail from clipping, you can also use an exposure correction (e.g. +2EV) on the Aperture priority, and add that to the remaining over exposure latitude.
I have a series of Flat-frames for several of my lenses on my EOS 1Ds Mark III, and I've exposed them at +2 1/3rd EV through a sheet of opal glass. At ISO 100 FRV gives me the following remaining latitudes; +0.82 EV, +1.08 EV, +1.09 EV, +1.02 EV, +1.02 EV, +0.98 EV, etc.
Depending on the camera used, at higher ISOs you may get other values, just try it for your camera.
The variation is due to random noise, slight variations in exposure time due to daylight changing after metering and before exposing, and irregularities in aperture blades and shutter curtains closing to the same opening. But on average it shows a 1 stop latitude in addition to the 2 1/3rd EV I pushed the exposure over the measured exposure level. So on average a total of 3 1/3rd EV above medium gray measurement. That was already what Rawdigger had told me, but now it's calculated for me.
Cheers,
Bart
-
That would be true if the shot noise were Gaussian. As per my previous discussion with Erik on this thread, I usually elide the fact that it's actually Poisson. I haven't run the numbers, and I'm sure the difference is small near fullscale where you're doing this test, but, since I brought it up before, I thought I'd point it out as possibly being of academic interest.
The difference is not small near zero, and I would like to publicly thank Jack Hogan for pointing that out and for suggesting fixes to some of my Matlab code that fixed some earlier errors.
Jim,
That is true, but with a 14 bit file at highlight clipping the λ is well over 1000, and the normal distribution is a good approximation.
Wikipedia (https://en.wikipedia.org/?title=Poisson_distribution):
"For sufficiently large values of λ, (say λ>1000), the normal distribution with mean λ and variance λ (standard deviation sqrt[λ]) is an excellent approximation to the Poisson distribution. If λ is greater than about 10, then the normal distribution is a good approximation if an appropriate continuity correction is performed, i.e., P(X ≤ x), where x is a non-negative integer, is replaced by P(X ≤ x + 0.5).
FPoisson(x;λ) ~Fnormal(x;λ,σ2=λ).
Another fine point is that the normal distribution is continuous whereas the ADUs of the raw file are discrete integers. For low values of λ, one can apply the continuity correction as above. For read noise with current Sony sensors, λ is less than 10, but I don't know how much differences this would make for practical work.
Bill
-
How much effect does that have on comparing cameras? For example, if camera A is touted as great for low light because it is usable at ISO 12,800 could this be because it is actually equivalent to camera B at 6,400?
Indeed, hence the value of DxO's Measured ISOs and related jabs at, say, Olympus :-)
-
Jim,
That is true, but with a 14 bit file at highlight clipping the λ is well over 1000, and the normal distribution is a good approximation.
Wikipedia (https://en.wikipedia.org/?title=Poisson_distribution):
"For sufficiently large values of λ, (say λ>1000), the normal distribution with mean λ and variance λ (standard deviation sqrt[λ]) is an excellent approximation to the Poisson distribution. If λ is greater than about 10, then the normal distribution is a good approximation if an appropriate continuity correction is performed, i.e., P(X ≤ x), where x is a non-negative integer, is replaced by P(X ≤ x + 0.5).
FPoisson(x;λ) ~Fnormal(x;λ,σ2=λ).
Another fine point is that the normal distribution is continuous whereas the ADUs of the raw file are discrete integers. For low values of λ, one can apply the continuity correction as above. For read noise with current Sony sensors, λ is less than 10, but I don't know how much differences this would make for practical work.
Bill
Well put, Bill. Imo it makes a difference when wishing to be precise in the deep shadows with today's sub 2e- read noise cameras at base ISO (e.g. the new crop of Exmors coming on line now). Aside from a different looking histogram it makes a big difference with small signals because poisson has no negative values, while gauss does. It's amazing how precisely physics fits our models when we use the right formulas. (http://www.dpreview.com/forums/post/56012887)
Jack
-
Well put, Bill. Imo it makes a difference when wishing to be precise in the deep shadows with today's sub 2e- read noise cameras at base ISO (e.g. the new crop of Exmors coming on line now). Aside from a different looking histogram it makes a big difference with small signals because poisson has no negative values, while gauss does. It's amazing how precisely physics fits our models when we use the right formulas. (http://www.dpreview.com/forums/post/56012887)
Jack
Jack and Jim K,
How would these considerations affect my modeling of the Canon sensors using the model proposed by Roger Clark? I don't have matab and am not sufficiently versed in the fine points of probability distributions to carry out this analysis, but perhaps you guys can help. These considerations could well affect the determination of engineering DR, but would be less critical when one is dealing with practical photographic DR.
Bill
-
Jack and Jim K,
How would these considerations affect my modeling of the Canon sensors using the model proposed by Roger Clark? I don't have matab and am not sufficiently versed in the fine points of probability distributions to carry out this analysis, but perhaps you guys can help. These considerations could well affect the determination of engineering DR, but would be less critical when one is dealing with practical photographic DR.
Hi Bill,
I don't think it will affect the DR assessment like you used much, unless one uses such an extreme under-exposure that we get into those low e- counts that will (occasionally) start showing a larger difference between Poisson and Gaussian distributions. The real issue is that the Poisson statistics at low counts will develop a bias/asymmetry compared to a Gaussian, and they can be quite different from observation to observation. So one would need a large number of observations to make sense of it anyway.
The read-noise is more likely to have a Gaussian distribution, and the exposure a more Poisson distribution particularly when very very low exposures are in play. But for developing a physics model like Jim and Jack did one would need to use the correct distribution model to avoid wrong conclusions over time (multiple observations).
Attached are the Probability Density Functions of the Poisson and the Normal distribution, at a mean of 5, 10 and 20.
Cheers,
Bart
-
Hi Bill,
I don't think it will affect the DR assessment like you used much, unless one uses such an extreme under-exposure that we get into those low e- counts that will (occasionally) start showing a larger difference between Poisson and Gaussian distributions. The real issue is that the Poisson statistics at low counts will develop a bias/asymmetry compared to a Gaussian, and they can be quite different from observation to observation. So one would need a large number of observations to make sense of it anyway.
The read-noise is more likely to have a Gaussian distribution, and the exposure a more Poisson distribution particularly when very very low exposures are in play. But for developing a physics model like Jim and Jack did one would need to use the correct distribution model to avoid wrong conclusions over time (multiple observations).
Attached are the Probability Density Functions of the Poisson and the Normal distribution, at a mean of 5, 10 and 20.
Cheers,
Bart
Agreed Bart. I wish you had shown the histogram for 1 as well :-)
-
Agreed Bart. I wish you had shown the histogram for 1 as well :-)
Here it is ;)
Means for 1, 5, 10, and 20.
Cheers,
Bart
-
So am I :)
Hi Bill,
If it only were so simple ;-) Unfortunately the 2006 incarnation of the standard for ISO made things worse than they were originally.
REI is arbitrary and SOS is based on the rendered OOC sRGB jpeg, not raw data. Thom's article is woefully inaccurate and Kerr's out of date.
If you have any other Nikon camera from an earlier or later generation, test them out. Both my D90 and D610 put spot metered middle gray below 10% of saturation in the raw data and, with 14 stops of eDR in the D610, I am glad about that.
Jack,
I tested out my D200 and D3 and repeated the tests on the D800e.
Saturations at the metered exposure for the green channels of the D200, D3 and D800e were 10.9%, 10.6%, and 10.3% respectively. Illumination was from a bank of 4800K solux bulbs. My previous test on the D800 was with sunlight. Perhaps the spectral response of the meter accounts for the slightly lower sat with the D800e as compared to my previous result. Adjustments for highlight spot meter adjustments were +3 EV for all the cameras. The indicated meter exposures were 1/25 sec, f/8 @ ISO 100 for the D200 and D800e. The metered exposure was 1/40 sec, f/8 @ ISO 200 for the D3 (equivalent to 1/20 sec @ f/8 for ISO 100). My Pentax Digital Spotmeter gave a reading of 1/20 sec, f/8 for ISO 100.
Cheers,
Bill
-
Jack,
I tested out my D200 and D3 and repeated the tests on the D800e.
Saturations at the metered exposure for the green channels of the D200, D3 and D800e were 10.9%, 10.6%, and 10.3% respectively. Illumination was from a bank of 4800K solux bulbs. My previous test on the D800 was with sunlight. Perhaps the spectral response of the meter accounts for the slightly lower sat with the D800e as compared to my previous result. Adjustments for highlight spot meter adjustments were +3 EV for all the cameras. The indicated meter exposures were 1/25 sec, f/8 @ ISO 100 for the D200 and D800e. The metered exposure was 1/40 sec, f/8 @ ISO 200 for the D3 (equivalent to 1/20 sec @ f/8 for ISO 100). My Pentax Digital Spotmeter gave a reading of 1/20 sec, f/8 for ISO 100.
Spot on, Bill.
-
Here it is ;)
Means for 1, 5, 10, and 20.
Cheers,
Bart
Excellent, thank you Bart! Note how the 1 gaussian histogram, if shown in its entirety, would show values to,say, -3 while poisson stops at zero. This acn make quite a difference when measuring signals of the order of magnitude of 1 LSB or below, like when studying quantization effects or determining DR.
Jack