Pages: 1 2 [3] 4   Go Down

Author Topic: Is expose to the right ever wrong?  (Read 21157 times)

craigwashburn

  • Full Member
  • ***
  • Offline Offline
  • Posts: 116
Re: Is expose to the right ever wrong?
« Reply #40 on: December 26, 2011, 11:09:23 am »

I think that ETTR should be defined as "increase exposure until you have objectional highlight clipping, or objectional motion blur or objectional lack of DOF". In other words: expose hot, but not too hot and not if it makes the image look bad. Then motion blur should not be a problem either?

-h

This is how I think of it:

In circumstances where:
- There is no undesired motion, from either camera or subject, that a longer exposure would exacerbate. 
- There is no localized veiling glare
- Highlights are controlled
- You're using an ugly-noise camera
- You want to spend time correcting the exposure later

ETTR may provide a slight benefit in numbers.  Even then, I have doubts as to whether this is visible on screen or in print.   

In the real world of photography and not lab tests, some or all of these criteria are not met.  Plus, you'd need a stout tripod, mirror lockup and a remote release to eliminate all vibration to eliminate any pixel-level blurring.  Obviously, with a lot of subject matter this just isn't practical.



Logged

bjanes

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 3387
Re: Is expose to the right ever wrong?
« Reply #41 on: December 26, 2011, 11:13:37 am »

As I said before, in laboratory-like conditions, ETTR may have a modest benefit in the numbers, but in the real world, this bears out differently.   

With your wedge on a light table, you have just about replicated a standard way of measuring global veiling glare in a lens.  McCann does something similar to show that veiling glare limits information available in an image - this actually furthers my point a bit, that ETTR has no benefits and in general shooting, increases chances of a subpar image w/ respect to motion blur and glare.

Again, you are making assertions without any data to back them up. My experiment does show that increasing the camera exposure (ETTR) has no effect on veiling glare when the image is normalized to what would be obtained with an ETTL exposure, contrary to your assertion that it increases glare even when one normalizes the ETTR exposure. Your comments about motion blur are correct, but if one is shooting from a tripod and photographing a static scene, motion blur would not be a factor.

Regards,

Bill
Logged

craigwashburn

  • Full Member
  • ***
  • Offline Offline
  • Posts: 116
Re: Is expose to the right ever wrong?
« Reply #42 on: December 26, 2011, 11:37:55 am »

Again, you are making assertions without any data to back them up. My experiment does show that increasing the camera exposure (ETTR) has no effect on veiling glare when the image is normalized to what would be obtained with an ETTL exposure, contrary to your assertion that it increases glare even when one normalizes the ETTR exposure. Your comments about motion blur are correct, but if one is shooting from a tripod and photographing a static scene, motion blur would not be a factor.

Regards,

Bill

Hi Bill,

Veiling glare limits information.  It puts values into pixels which are unrelated to the scene.  You might think of it as a type of noise that originates from the optical path and not the sensor.  McCann demonstrates this, so does the Stanford paper.  Normalizing does not add information that does not exist.  Shooting at a proper exposure won't either in a linear glare situation, non-linear it might improve things.

The purpose of ETTR is to improve your SN ratio (with dubious results), but veiling glare puts a limit to this.  So your ETTR isn't accomplishing its purpose, and in the real world, increases the chance of noticeable degradation from the other factors I mentioned. 





Logged

bjanes

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 3387
Re: Is expose to the right ever wrong?
« Reply #43 on: December 26, 2011, 11:38:29 am »

I think some of the quibble here is over semantics, or rather precision of language.  There is a distinction between 'encoded levels' and 'information'.  Having more levels does not necessarily mean more information.  What photographers who employ ETTR understand intuitively is that by increasing exposure they get smoother tonal gradations, and that is indeed because there is more information in the image due to higher S/N.  There are indeed more distinguishable levels because that number is governed by the S/N; and this number of distinguishable levels, it should be noted, is distinct from the number of encoded levels which is the number of raw values encompassing any given exposure zone in the image.  One can play all sorts of tricks with the latter without changing the S/N and thus the image quality, as Guillermo's example shows.

DXO does quantify the number of distinguishable levels in an image with their Tonal Range measurement. They state "The standard deviation of noise can be viewed as the smallest difference between two distinguishable gray levels."

The tonal range for the Pentax K5 (an ISO less camera) is shown along with an Excel spreadsheet showing the actual number of levels. With an ISO-less sensor, the tonal range with increasing ISO can represent exposure to the left, since each doubling of ISO halves the number of photons collected by the sensor and the read noise does not change. One may not be using the full scale of the ADC with higher ISO, but this does not make much difference, since the quantization is limited by noise and not the number of ADC steps.

Regards,

Bill
Logged

Mark D Segal

  • Contributor
  • Sr. Member
  • *
  • Offline Offline
  • Posts: 12512
    • http://www.markdsegal.com
Re: Is expose to the right ever wrong?
« Reply #44 on: December 26, 2011, 11:39:03 am »

This is how I think of it:

In circumstances where:
- There is no undesired motion, from either camera or subject, that a longer exposure would exacerbate. 
- There is no localized veiling glare
- Highlights are controlled
- You're using an ugly-noise camera
- You want to spend time correcting the exposure later

ETTR may provide a slight benefit in numbers.  Even then, I have doubts as to whether this is visible on screen or in print.   

In the real world of photography and not lab tests, some or all of these criteria are not met.  Plus, you'd need a stout tripod, mirror lockup and a remote release to eliminate all vibration to eliminate any pixel-level blurring.  Obviously, with a lot of subject matter this just isn't practical.


I have no doubt whatsoever. Properly used, the technique works fine and helps produce cleaner images. As for the "conditions" you specify:


 - Bill Janes has amply demonstrated that "veiling glare" is a red-herring.
 - It goes without saying that one would select a high enough shutter speed to elimimate undesired motion blur. Exposure is always about compromises, so don't try invalidating a generally valid proposition by confusing it with the issue of compromises often needed in any exposure situation.
 - Controlling highlights is part of ETTR technique, not an obstacle to it. Another red-herring.
 - All cameras have a background level of noise that becomes more apparent the lower the exposure. This is still an inescapable fact of digital imaging regardless of the camera and all the technical progress to date.
 - "Spending time correcting the exposure later" is also a red-herring, because any one who shoots raw is editing the images in post-capture software anyhow.


It's been an interesting discussion, but I think this argument has run its course.
Logged
Mark D Segal (formerly MarkDS)
Author: "Scanning Workflows with SilverFast 8....."

bjanes

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 3387
Re: Is expose to the right ever wrong?
« Reply #45 on: December 26, 2011, 11:47:26 am »

Hi Bill,

Veiling glare limits information.  It puts values into pixels which are unrelated to the scene.  You might think of it as a type of noise that originates from the optical path and not the sensor.  McCann demonstrates this, so does the Stanford paper.  Normalizing does not add information that does not exist.  Shooting at a proper exposure won't either in a linear glare situation, non-linear it might improve things.

The purpose of ETTR is to improve your SN ratio (with dubious results), but veiling glare puts a limit to this.  So your ETTR isn't accomplishing its purpose, and in the real world, increases the chance of noticeable degradation from the other factors I mentioned. 

Veiling glare affects predominantly the shadows. The ETTR will improve SNR in the highlights as shown in my previous post. Overall DR will be limited by the effect of veiling glare diluting the shadows, but the ETTR will still improve SNR in the highlights. You have yet to demonstrate that the veiling glare is non-linear. For my shooting, I will continue to use ETTR, taking care not to blow important highlights. Veiling glare does not enter into the equation.

Regards,

Bill
Logged

bjanes

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 3387
Re: Is expose to the right ever wrong?
« Reply #46 on: December 26, 2011, 11:48:37 am »

It's been an interesting discussion, but I think this argument has run its course.

+1
Logged

craigwashburn

  • Full Member
  • ***
  • Offline Offline
  • Posts: 116
Re: Is expose to the right ever wrong?
« Reply #47 on: December 26, 2011, 12:12:56 pm »

I have no doubt whatsoever. Properly used, the technique works fine and helps produce cleaner images. As for the "conditions" you specify:


 - Bill Janes has amply demonstrated that "veiling glare" is a red-herring.
 - It goes without saying that one would select a high enough shutter speed to elimimate undesired motion blur. Exposure is always about compromises, so don't try invalidating a generally valid proposition by confusing it with the issue of compromises often needed in any exposure situation.
 - Controlling highlights is part of ETTR technique, not an obstacle to it. Another red-herring.
 - All cameras have a background level of noise that becomes more apparent the lower the exposure. This is still an inescapable fact of digital imaging regardless of the camera and all the technical progress to date.
 - "Spending time correcting the exposure later" is also a red-herring, because any one who shoots raw is editing the images in post-capture software anyhow.


It's been an interesting discussion, but I think this argument has run its course.

Hi Mark, I agree that it's run its course, but if veiling glare is a red herring then why are there research papers devoted to its effects and limitations?  Have you read and understood the two papers linked to here?  IMO, the loss of information negates ETTR benefits.

I believe ETTR has a very limited usefulness and is not 'generally valid' today for anyone interested in real world photography.  Especially when the discussion has no alternative but to turn to various charts and graphs that do not represent real world ink hitting paper.


Logged

Guillermo Luijk

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 2005
    • http://www.guillermoluijk.com
Re: Is expose to the right ever wrong?
« Reply #48 on: December 26, 2011, 12:18:13 pm »

Veiling glare limits information.

Yes, veiling glare limits information, BUT...

ETTR DOESN'T INCREASE THE EFFECT OF VEILING GLARE

craigwashburn

  • Full Member
  • ***
  • Offline Offline
  • Posts: 116
Re: Is expose to the right ever wrong?
« Reply #49 on: December 26, 2011, 12:28:46 pm »

Yes, veiling glare limits information, BUT...

ETTR DOESN'T INCREASE THE EFFECT OF VEILING GLARE


Lol, nice.  That's debatable.  However, it seems few in this thread have read and understood the papers and other things presented here.  Who knew this was such a touchy subject?

Enjoy the holiday folks :)

Logged

Mark D Segal

  • Contributor
  • Sr. Member
  • *
  • Offline Offline
  • Posts: 12512
    • http://www.markdsegal.com
Re: Is expose to the right ever wrong?
« Reply #50 on: December 26, 2011, 12:31:57 pm »

Hi Mark, I agree that it's run its course, but if veiling glare is a red herring then why are there research papers devoted to its effects and limitations?  Have you read and understood the two papers linked to here?  IMO, the loss of information negates ETTR benefits.

I believe ETTR has a very limited usefulness and is not 'generally valid' today for anyone interested in real world photography.  Especially when the discussion has no alternative but to turn to various charts and graphs that do not represent real world ink hitting paper.




All kinds of research papers are devoted to all kinds of stuff and that in itself proves nothing. You've had two highly respected researchers on this forum telling you that veiled glare is a non-issue in respect of ETTR. As for real world ink hitting paper, I think my Epson 4900 and previous incarnations qualify. I've been making high quality inkjet prints from a range of top-of-the-line cameras for years, and if I tell you I can perceive noise differences in those prints as a function of exposure - and we're talking real-word photos, not charts - I'm not inviting you to tell me I don't know what I'm looking at, unless you've seen what I've seen yourself. This is not a "touchy subject" - it's simply s debate about facts, and not everyone here agrees with your view of the facts. That happens - no hard feelings. As I said, this has run its course.
Logged
Mark D Segal (formerly MarkDS)
Author: "Scanning Workflows with SilverFast 8....."

Guillermo Luijk

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 2005
    • http://www.guillermoluijk.com
Re: Is expose to the right ever wrong?
« Reply #51 on: December 26, 2011, 12:39:37 pm »

Lol, nice.  That's debatable.  However, it seems few in this thread have read and understood the papers and other things presented here.

I don't think any of those papers state that ETTR increases veiling glare since a sensor is just a linear photon counter. If you expose more by increasing exposure time, you'll just have higher levels defining your veiling glare areas, but not more veiling glare.

If I am wrong, I'd be pleased that you explain why, or in which precise point of any link this is explained. If you cannot do that I'd simply request you stop saying ETTR is a bad idea because of veiling glare.

All this in a friendly Christmas atmosphere  ;D

« Last Edit: December 26, 2011, 12:45:11 pm by Guillermo Luijk »
Logged

craigwashburn

  • Full Member
  • ***
  • Offline Offline
  • Posts: 116
Re: Is expose to the right ever wrong?
« Reply #52 on: December 26, 2011, 12:56:05 pm »

I don't think any of those papers state that ETTR increases veiling glare since a sensor is just a photon counter. If you expose more by increasing exposure time, you'll just have higher levels defining your veiling glare areas, but not more veiling glare.

If I am wrong, I'd be pleased that you explain why, or in which precise point of any link this is explained. If you cannot do that, I'd simply request you stop saying ETTR is a bad idea because of veiling glare.

All this in a friendly Christmas atmosphere  ;D


Hi Guillermo,

The papers suggest that some glare is not linear.  Yes, a sensor is a linear source - although with all the variables that go between photo capture and ink hitting paper, the final result might not be.  The Stanford paper for example, finds that deconvolution does not work so great.  It could be their implementation, or, some of the glare does not follow linearly before their algorithm is applied, or a combination thereof.  Some glare, for example, is behind the shutter and off the sensor, or there could be other hardware/software interactions with the data.  Remember the good ol' days of sensor blooming?

My own experience of thousands of images, which I am sad to say does not come with attendant charts and graphs :), shows that some glare is worse to deal with than others, and simply normalizing the data does not provide the same benefit as a shorter exposure time.  Not all glare follows this, but it happens often enough, usually when a bright source is against a darker background.  But, even as this case is rare, the non benefits of ETTR plus the increased risk of motion blur and such detail killing artifacts make it a poor choice as a general rule. 

Logged

Guillermo Luijk

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 2005
    • http://www.guillermoluijk.com
Re: Is expose to the right ever wrong?
« Reply #53 on: December 26, 2011, 01:02:12 pm »

My own experience of thousands of images, which I am sad to say does not come with attendant charts and graphs :), shows that some glare is worse to deal with than others, and simply normalizing the data does not provide the same benefit as a shorter exposure time.

In thousands of images you'll surely have one that demonstrates what you say. Why not use it?

craigwashburn

  • Full Member
  • ***
  • Offline Offline
  • Posts: 116
Re: Is expose to the right ever wrong?
« Reply #54 on: December 26, 2011, 01:27:45 pm »

In thousands of images you'll surely have one that demonstrates what you say. Why not use it?


Chalk me up to the latest ETTR debater too lazy to bother digging one up from the archive.  That is why until now I've tried to avoid the "you haven't seen what i've seen!" argument.  Hence the papers that show the information limitations caused by glare, and also, simply the threat of pixel level motion blur or highlight detail loss in real world photography that undoes any benefit from ETTR.

The next time I come across the non-linear issue I'll post it, but until then, feel free to focus on that rather than the other more concrete issues with ETTR. :)

But to bring this back full circle to the original post of "Is expose to the right ever wrong?"  Yes.  I believe it is wrong in most situations that a typical photographer will encounter, despite all the apparently well-meaning charts and graphs that don't tell the whole story.


Logged

Bryan Conner

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 539
    • My Flickr page
Re: Is expose to the right ever wrong?
« Reply #55 on: December 26, 2011, 01:40:35 pm »

Chalk me up to the latest ETTR debater too lazy to bother digging one up from the archive.  That is why until now I've tried to avoid the "you haven't seen what i've seen!" argument.  Hence the papers that show the information limitations caused by glare, and also, simply the threat of pixel level motion blur or highlight detail loss in real world photography that undoes any benefit from ETTR.

The next time I come across the non-linear issue I'll post it, but until then, feel free to focus on that rather than the other more concrete issues with ETTR. :)

But to bring this back full circle to the original post of "Is expose to the right ever wrong?"  Yes.  I believe it is wrong in most situations that a typical photographer will encounter, despite all the apparently well-meaning charts and graphs that don't tell the whole story.




If ETTR is wrong in most situations, then you should have no problem to take a few minutes TODAY to make a few captures to illustrate your point. This will not involve any searching at all...only picking up a camera and pushing a button.  Otherwise, your points are....well....you do not have any.
Logged

craigwashburn

  • Full Member
  • ***
  • Offline Offline
  • Posts: 116
Re: Is expose to the right ever wrong?
« Reply #56 on: December 26, 2011, 02:00:54 pm »

If ETTR is wrong in most situations, then you should have no problem to take a few minutes TODAY to make a few captures to illustrate your point. This will not involve any searching at all...only picking up a camera and pushing a button.  Otherwise, your points are....well....you do not have any.

Is it not obvious that potential motion blur and non-linear veiling glare is a significant downside to ETTR, especially when ETTR provides zero-to-few benefits in the end?  This is the entire point.

I find photography forums to be frustrating.  Please read all of the above posts and papers before commenting.
Logged

Bryan Conner

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 539
    • My Flickr page
Re: Is expose to the right ever wrong?
« Reply #57 on: December 26, 2011, 02:26:04 pm »

Is it not obvious that potential motion blur and non-linear veiling glare is a significant downside to ETTR, especially when ETTR provides zero-to-few benefits in the end?  This is the entire point.

I find photography forums to be frustrating.  Please read all of the above posts and papers before commenting.


I have read every post. I have not read every paper.  I do not need to.  I have incorporated ETTR (and tried to avoid ETFTTR) into my knowledge base and use it all of the time.  I expose as far to the right as I can while retaining wanted highlight information, avoiding non-linear veiling glare, utilizing a high enough shutter speed to prevent unwanted motion blur, a low enough iso number, and my desired f stop.  By doing so, I definitely have seen a reduction in the noise level in most of my images when compared to my images back in the days of exposing to the middle only.  ETTR does not require the shooter to use a slow shutter speed thereby causing motion blur! There is also a potential for motion blur when under-exposing.  It is a theory that the shooter can take into consideration when setting the parameters before tripping the shutter.  

It does not matter if the problem is non-linear veiling glare or motion blur.  I will set my camera settings to avoid both.  In other words, I practice ETTR, not ETFTTR (Expose Too Far To The Right).

It is ok if you feel that ETTR provides zero-to-few benefits in the end.  In my photographic world, I definitely see the benefits provided...no question.  
« Last Edit: December 26, 2011, 02:30:06 pm by Bryan Conner »
Logged

digitaldog

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 20630
  • Andrew Rodney
    • http://www.digitaldog.net/
Re: Is expose to the right ever wrong?
« Reply #58 on: December 26, 2011, 03:17:01 pm »

- It goes without saying that one would select a high enough shutter speed to elimimate undesired motion blur. Exposure is always about compromises, so don't try invalidating a generally valid proposition by confusing it with the issue of compromises often needed in any exposure situation.

It does go without saying, unless you are in a forum discussion about ETTR when some have to come up with reasons it is not going to work well. But you and I both know that we are not always in ideal conditions. No one in their right mind would invoke motion blur when not wanted just to provide more photons to the sensor! IF the conditions can be met, using ETTR, a Tripod, Mirror lock up, the best lens etc will produce the best (better) data than not using these techniques. And as image creators who understand the processes, I can’t see why we’d ignore better quality data if we can produce it. But there are all kinds of situations where I can’t use a tripod, or lock the mirror, or even ETTR to get an acceptable if not idelized data.
Logged
http://www.digitaldog.net/
Author "Color Management for Photographers".

Schewe

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 6229
    • http:www.schewephoto.com
Re: Is expose to the right ever wrong?
« Reply #59 on: December 26, 2011, 03:50:47 pm »

Plus, you'd need a stout tripod, mirror lockup and a remote release to eliminate all vibration to eliminate any pixel-level blurring. 

Which is something you need to consider doing regardless of whether or not you use ETTR...if you want max image quality, photographers need to exercise proper technique on a whole host of technical issues...ETTR is simply one factor in a range of factors to deploy. But claiming it has little or no benefit is short changing your arsenal for no good reason.
Logged
Pages: 1 2 [3] 4   Go Up