ETTR TO THE FAR RIGHT
May 26, 2014 21:16:11 #
rebride wrote:
Expose for the highlights applied to slides/positve film.
For BW negative it is expose for the shadows, develop for the highlights.
My Fuji dynamic range modes for contrast control expose to the left. It stops down 1 or 2 stops (for the raw file) from the set exposure. It than holds the highlights in place and tone maps the lower end up for the jpg image.
Expose for the highlights, develop for shadows.
ETTL ???
For BW negative it is expose for the shadows, develop for the highlights.
My Fuji dynamic range modes for contrast control expose to the left. It stops down 1 or 2 stops (for the raw file) from the set exposure. It than holds the highlights in place and tone maps the lower end up for the jpg image.
Expose for the highlights, develop for shadows.
ETTL ???
Regarding b&w film, I exposed based upon a meter reading which could vary a stop or two depending upon where I pointed it. This was true of reflectance and incidental meters, built-in or handheld. I eventually added 10% to the developing time to get the negatives to print better. It was all strictly empirical. I had no use for Ansel Adams' approach because it did not lend itself readily to roll film. Thanks for pointing that out.
As for the Fuji, I guess the important thing is how you like the results regardless of how you get there. I am sure people have compared it with the conventional Canon/Nikon approach.
May 26, 2014 21:24:17 #
MW wrote:
"Avoid" may be a better term than "... (show quote)
I can live with "avoid" or shall just say "less noise"? As long as we are parsing our language so carefully, how about changing "lowest" to "lower" in keeping with your use of the comparative elsewhere in the sentence?
I have shared your other observations. As for your disclaimer, you do state the obvious and that is fine. No one here can really guarantee his or her understanding is perfect in these matters. I know one or two people may occasionally be pompous or self-indulgent or arrogant enough to think he or she can so leave them alone.
Thanks for contributing.
May 26, 2014 23:33:31 #
Michael Richman and Jeff Schewe produced an 11 hour video. In a discussion in that video Reichmann comments that the expose to the right only works at the base ISO. He said there was no point in doing it at high ISO. He started the expose to the right idea.
The original premise of this post seems flawed.
The original premise of this post seems flawed.
May 26, 2014 23:50:28 #
I thought that the ETTR technique was not motivated by noise reduction but rather to preserve shadow detail. Noise reduction was a collateral benefit. I've watched a live demo of this and it appeared to work, however the results didn't seem all that dramatic. In any case the ability to use ETTR is limited by how much if any of the highlights you are willing to blow out.
May 27, 2014 06:50:07 #
Wow. Excellent post and comment thread. Very informative. We need more of this and less..well, I'll just leave it at that.
May 27, 2014 06:51:48 #
Here are my own totally subjective observations.
On my T3i, using ETTR then pulling things down a bit in post gave me the best results in terms of noise, color and contrast.
If I do the same on my 6D, saturation and contrast suffer and I can't bring it back in post.
When using the 6D I get best results overall by keeping the histogram where it's "supposed" to be or very very slightly to the right.
I think that different sensors respond differently to over and under exposure.
On my T3i, using ETTR then pulling things down a bit in post gave me the best results in terms of noise, color and contrast.
If I do the same on my 6D, saturation and contrast suffer and I can't bring it back in post.
When using the 6D I get best results overall by keeping the histogram where it's "supposed" to be or very very slightly to the right.
I think that different sensors respond differently to over and under exposure.
May 27, 2014 09:52:37 #
For those of you who do ETTR I was wondering about hou you actually do it. Ie, do you meter and then adjust or do you keep your EV permanently set on + .3, .7, or 1.0? Or does it depend on the level of contrast in the image.
My understanding is that active d lighting has only one effect on raw images, that being to underexposed to preserve detail in highlights. Would that counter the ETTR effect, which means active d lighting should not be used with ETTR?
Thanks,
JH
My understanding is that active d lighting has only one effect on raw images, that being to underexposed to preserve detail in highlights. Would that counter the ETTR effect, which means active d lighting should not be used with ETTR?
Thanks,
JH
May 27, 2014 10:38:51 #
John Howard wrote:
For those of you who do ETTR I was wondering about hou you actually do it. Ie, do you meter and then adjust or do you keep your EV permanently set on + .3, .7, or 1.0? Or does it depend on the level of contrast in the image.
My understanding is that active d lighting has only one effect on raw images, that being to underexposed to preserve detail in highlights. Would that counter the ETTR effect, which means active d lighting should not be used with ETTR?
Thanks,
JH
My understanding is that active d lighting has only one effect on raw images, that being to underexposed to preserve detail in highlights. Would that counter the ETTR effect, which means active d lighting should not be used with ETTR?
Thanks,
JH
John, this is how I approach ETTR, right or wrong. I take the picture in an auto-exposure, shutter priority or aperture priority mode. I look at the histogram and adjust the exposure accordingly. I take a second shot, review the histogram and adjust again if necessary. For me, I want the right edge of the curve to be close to but not on the right edge of the histogram's frame. Close is close enough.
At this point, I presume I have the most information about the scene and I shoot away. I glance at the histogram every so often to make sure the lighting has not changed it. If the lighting changes, I change the exposure compensation accordingly. You do not have to fiddle with each shot, especially if you are shooting raw's with their greater latitude.
Now that I have the maximum information, I adjust the brightness in post-processing to satisfy my esthetic. You can always throw away information but you cannot create it if it is not there.
I hope this helps.
May 27, 2014 10:52:53 #
Thanks. That is sort of what I've been doing also. Some days I feel the light is changing or my angle of view is changing and I spend too much time looking at the back of the camera.
May 27, 2014 11:06:18 #
John Howard wrote:
Thanks. That is sort of what I've been doing also. Some days I feel the light is changing or my angle of view is changing and I spend too much time looking at the back of the camera.
Do what you have to and don't do what you don't have to.
May 27, 2014 13:48:33 #
amfoto1
Loc: San Jose, Calif. USA
I have used and will continue to use ETTR technique with my Canon cameras. I don't know if this applies to any other camera manufacturers, but here's what I believe is the case with Canon...
Having shot with them for a long time, I'm convinced Canon cameras tend to slightly underexpose, by design. I think this dates back to the days of slide film, when any overexposure was a real disaster. Both color neg and B&W neg films are more tolerant of slight exposure errors, slide film is not. With slide (aka "transparency" ) film the shadows are full of "data", while the highlights are the lack of any "data" on the film, up to the level of complete transparency representing pure white. The opposite is the case with negative/print films. The printing process also allows for some adjustments, while slides often are viewed directly. Slide film also tended to have a narrower dynamic range than negative films (especially higher ISO slide film... say ISO 400 or higher).
Today with digital, the dynamic range of most sensors is similar to low ISO slide film, but the distribution of data is exactly the opposite, more like negative films. The shadows are a "lack of any data", while the highlights are "full of data". Because of this, modest exposure errors to the highlight side are better tolerated than errors to the shadow side.
Yet Canon still seems to calibrate their cameras to "protect the highlights", even their DSLRs.
Using ETTR corrects for this.
When working mostly in shade or on overcast days, I use +1/3 to +2/3 Exposure Compensation (when forced to use any of the auto exposure modes, which are Av, Tv and P)... or I set my exposure +1/3 manually (when shooting in M mode). Out in full sun, I use +1/3.
I also add and subtract from the Exposure Compensation setting depending upon subject tonality... the original intent of Exposure Compensation, after all... correcting for the inherent errors of any reflective metering system. Whenever possible I use a separate, handheld incidence meter instead, which is much more accurate because it measures the brightness of light falling onto a subject rather than what's being reflected off it, so subject tonality is no longer a factor. But even then I also add +1/3 stop to my exposure. When using particularly high ISOs (3200 and up), I often will use an additional +1/3 stop.
Your test using ISO 800 with a 60D isn't high enough ISO. The 60D has pretty clean images regardless, at ISO 800. The differences are just too subtle at such a low ISO (with my 7Ds, I use 400 most of the time, and 800, even 1600 quite often... it's only 3200 and above that I consider high, noisier ISOs). Repeat the test using ISO 3200, 6400 or even higher. Then you'll see the differences more clearly.
What you will find is that whenever your original exposure forces you to increase exposure in post-processing, you will greatly increase noise along with it. If you are using a fairly high ISO (3200 and above on the 18MP Canon APS-C cameras) and if a lot of image brightening is needed, image quality will drop fast and many shots will be unusable.
On the other hand, if thanks to making an accurate or even slightly overexposed original exposure, instead you are able to leave it alone or even reduce exposure slightly in post-processing, that minimizes noise in the images. It's possible to use one or two or even three stops higher ISO with this technique, though some additional noise reduction is likely to be needed, too.
Does this work all the time? No. For one, it's important that you only slightly overexpose. Going too far "to the right" will start to seriously clip highlights. And, since it's often used with auto exposure modes, there's still a lot of room for exposure errors. Using ETTR simply improves your odds of getting a good, usable exposure with any of the auto modes. (Shooting M and locking in known-good exposures is always preferable... but simply not possible a lot of the time. In variable lighting conditions you pretty much have to use one of the auto exposure modes.)
Essentially, with digital it's "safer" to err toward the highlights, than toward the shadows. Just the opposite of slide film, back in the day.
While shooting, I check my histogram periodically. Frequently when first setting up, then every so often during a day's shoot. I look mainly to see that the LH or "shadow" end isn't "piling up", try to keep it so that the curve it's just barely or lightly touching the LH edge.... I worry less about the RH or "highlight" side bunching up or even going off the edge, though I try not to let it really go off the edge in a big way. However, there is some variation on a histogram's displayed data distribution depending upon subject tonalities that has to be taken into account.
Perhaps the easiest "test" you can do is simply make note of what you usually have to do in post-processing. Even doing all the above, way too often I still find in post-processing that I'm having to increase exposure in far more images than I am leaving "as shot" or have to reduce exposure. That tells me that, if anything, I'm probably being too conservative with my ETTR settings.
Your computer monitor must be properly calibrated - especially for brightness - for this post-processing "test" to be valid and useful. The vast majority of consumer grade LCD monitors are way, way too bright to give usable feedback. Ideally, a print should match the screen for brightness. Left uncalibrated, an overly bright LCD screen will cause you to make your images too dark and that will show up in every print you make (or online sharing, when anyone using a properly calibrated monitor views the image). An uncalibrated monitor also will make it impossible to evaluate your exposures in any meaningful way.
See the post below on another forum. It's a discussion of high ISO techniques with 7D (basically the same sensor and image processing as 60D)... It is essentially ETTR combined with post-processing NR, to make quality, usable images at ISOs most people avoid. Very good illustrations, too.
http://photography-on-the.net/forum/showthread.php?t=1079217
Having shot with them for a long time, I'm convinced Canon cameras tend to slightly underexpose, by design. I think this dates back to the days of slide film, when any overexposure was a real disaster. Both color neg and B&W neg films are more tolerant of slight exposure errors, slide film is not. With slide (aka "transparency" ) film the shadows are full of "data", while the highlights are the lack of any "data" on the film, up to the level of complete transparency representing pure white. The opposite is the case with negative/print films. The printing process also allows for some adjustments, while slides often are viewed directly. Slide film also tended to have a narrower dynamic range than negative films (especially higher ISO slide film... say ISO 400 or higher).
Today with digital, the dynamic range of most sensors is similar to low ISO slide film, but the distribution of data is exactly the opposite, more like negative films. The shadows are a "lack of any data", while the highlights are "full of data". Because of this, modest exposure errors to the highlight side are better tolerated than errors to the shadow side.
Yet Canon still seems to calibrate their cameras to "protect the highlights", even their DSLRs.
Using ETTR corrects for this.
When working mostly in shade or on overcast days, I use +1/3 to +2/3 Exposure Compensation (when forced to use any of the auto exposure modes, which are Av, Tv and P)... or I set my exposure +1/3 manually (when shooting in M mode). Out in full sun, I use +1/3.
I also add and subtract from the Exposure Compensation setting depending upon subject tonality... the original intent of Exposure Compensation, after all... correcting for the inherent errors of any reflective metering system. Whenever possible I use a separate, handheld incidence meter instead, which is much more accurate because it measures the brightness of light falling onto a subject rather than what's being reflected off it, so subject tonality is no longer a factor. But even then I also add +1/3 stop to my exposure. When using particularly high ISOs (3200 and up), I often will use an additional +1/3 stop.
Your test using ISO 800 with a 60D isn't high enough ISO. The 60D has pretty clean images regardless, at ISO 800. The differences are just too subtle at such a low ISO (with my 7Ds, I use 400 most of the time, and 800, even 1600 quite often... it's only 3200 and above that I consider high, noisier ISOs). Repeat the test using ISO 3200, 6400 or even higher. Then you'll see the differences more clearly.
What you will find is that whenever your original exposure forces you to increase exposure in post-processing, you will greatly increase noise along with it. If you are using a fairly high ISO (3200 and above on the 18MP Canon APS-C cameras) and if a lot of image brightening is needed, image quality will drop fast and many shots will be unusable.
On the other hand, if thanks to making an accurate or even slightly overexposed original exposure, instead you are able to leave it alone or even reduce exposure slightly in post-processing, that minimizes noise in the images. It's possible to use one or two or even three stops higher ISO with this technique, though some additional noise reduction is likely to be needed, too.
Does this work all the time? No. For one, it's important that you only slightly overexpose. Going too far "to the right" will start to seriously clip highlights. And, since it's often used with auto exposure modes, there's still a lot of room for exposure errors. Using ETTR simply improves your odds of getting a good, usable exposure with any of the auto modes. (Shooting M and locking in known-good exposures is always preferable... but simply not possible a lot of the time. In variable lighting conditions you pretty much have to use one of the auto exposure modes.)
Essentially, with digital it's "safer" to err toward the highlights, than toward the shadows. Just the opposite of slide film, back in the day.
While shooting, I check my histogram periodically. Frequently when first setting up, then every so often during a day's shoot. I look mainly to see that the LH or "shadow" end isn't "piling up", try to keep it so that the curve it's just barely or lightly touching the LH edge.... I worry less about the RH or "highlight" side bunching up or even going off the edge, though I try not to let it really go off the edge in a big way. However, there is some variation on a histogram's displayed data distribution depending upon subject tonalities that has to be taken into account.
Perhaps the easiest "test" you can do is simply make note of what you usually have to do in post-processing. Even doing all the above, way too often I still find in post-processing that I'm having to increase exposure in far more images than I am leaving "as shot" or have to reduce exposure. That tells me that, if anything, I'm probably being too conservative with my ETTR settings.
Your computer monitor must be properly calibrated - especially for brightness - for this post-processing "test" to be valid and useful. The vast majority of consumer grade LCD monitors are way, way too bright to give usable feedback. Ideally, a print should match the screen for brightness. Left uncalibrated, an overly bright LCD screen will cause you to make your images too dark and that will show up in every print you make (or online sharing, when anyone using a properly calibrated monitor views the image). An uncalibrated monitor also will make it impossible to evaluate your exposures in any meaningful way.
See the post below on another forum. It's a discussion of high ISO techniques with 7D (basically the same sensor and image processing as 60D)... It is essentially ETTR combined with post-processing NR, to make quality, usable images at ISOs most people avoid. Very good illustrations, too.
http://photography-on-the.net/forum/showthread.php?t=1079217
May 27, 2014 14:00:13 #
RRS
Loc: Not sure
abc1234 wrote:
CAUTION: What follows will be very technical and i... (show quote)
I would like to say that this has been one of the most informative post that I have seen in a long time. Something to read and reread and digest. Something to think about and hopefully improve our shooting technique. Thanks to all who have added to this fine post! :) :thumbup: :thumbup:
May 27, 2014 14:05:54 #
RRS wrote:
I would like to say that this has been one of the most informative post that I have seen in a long time. Something to read and reread and digest. Something to think about and hopefully improve our shooting technique. Thanks to all who have added to this fine post! :) :thumbup: :thumbup:
Thank you and thank everyone for their polite and informative responses. I have tried to keep this as factual and focused as possible.
May 27, 2014 20:23:52 #
redrocktom
Loc: Sedona
abc1234 wrote:
John, this is how I approach ETTR, right or wrong.... (show quote)
Get a nice mirrorless camera that shows the live histogram in the EVF, and you're set. You will know how much you can push your ETTR by your PP adjustment experience.
May 29, 2014 18:42:15 #
[quote=abc1234]CAUTION: What follows will be very technical and is not for everyone. This is a long, detailed post.
The idea for this post comes from a reference in another forum. Since I would like to understand this approach so I can decide to incorporate it into my exposure method, I will keep my comments as neutral as possible and invite all objective, clear and rational responses. Here is the link to the reference.
http://www.johnshawphoto.com/category/exposure/
I shot the appropriate photos as raw's according to this link and looked at them in LR.
1. I set the camera on a tripod and took a basic picture at ISO 800 (to introduce noise), aperture priority (f/7.1) and 1/30. Canon 60D, 18-200 mm lens.
2. I switched to manual exposure and increased exposure until "blinkies" started to show. This occurred at f/7.1 and 1/20.
3. I continued increasing the shutter speed in increments equivalent to 1/3 stop. This obviously made the pictures lighter and shifted the histogram to the right.
The closest I could get the histogram to the right was f/7.1 and 1/10. This is the end of the technical, factual part of the experiment.
Here are my conclusions and I invite people to comment critically, clearly and objectively.
1. Exposing beyond the right and adjusting the brightness to a desirable level do not improve noise or shadow detail.
2. The overall appearance of the standard exposed-to-the-right and adjusted exposed-beyond-the-right pictures are comparable.
XXXXXXXXXXXXXXXXXXXXX
abc1234, and others interested in this thread,
This thread has so far been long on theorization and relatively short on actual, rationally applied practice.
I've been a devotee of ETTR - following a brief, introductory period of skepticism after reading Reichmann's essay in Luminous Landscape in 2003. Then in 2004 I started working with EBTR after reading the Adobe White Paper "RAW Capture, Linear Gamma, and Exposure" by the late Bruce Fraser, and later the collaborative work of Fraser and Jeff Schewe in 2009 ( in "Real World Camera RAW) and since then have included the topic in my classes, courses, and workshops where it has been well received and won numerous convinced EBTR practitioners..
Why your lack of success, abc1234' in noise reduction with EBTR?
Haven't the foggiest, and in light of the growing acceptance of the EBTR exposure procedure based predominantly on the ease of practical elimination of noise, I can only surmise that 1.) your adherence to the requirements of the process may not have been as stringent as it ought have been, and/or that 2.) Shaw's description of the process wasn't as lucid as I had originally thought it to have been.
You stated:
"The closest I could get the histogram to the right was f/7.1 and 1/10. This is the end of the technical, factual part of the experiment."
This is only 1/3 stop more exposure than f/1.7 and 1/20 ( your putative ETTR exposure?) I don't know what that statement means in the context of determining your camera's Extra RAW-Accessible Dynamic Range (ERADR). "Closest" to what?
And nowhere can I find where you actually quantiated the ERADR of your camera.
I included a link to Shaw's article in the thread in which abc1234 was participating when I was asked by someone for other sources of info on EBTR because I recalled, after reading it some time ago, concluding "Yeah; this guy has used it and gets it!" . When reading of your problems getting satisfactory results and that you had followed Shaw's article I realized that Shaw may not have provided sufficient detail for a person new to the topic to successfully follow. I suggest you give it another try.
In order to determine your camera's Extra RAW-Accessible Dynamic Range (ERADR) proceed as follows:
As you did, start with a carefully determined ETTR exposure at an ISO of your choice. For landscapes my choice for EBTR is almost always native ISO (100 in my Sonys) or 200. But if you really want to check the effect of EBTR on high noise potential, go ahead and set ISO to 800, as you also did. Regardless of the ISO used you should experience a marked reduction of noise below the level expected with using only ETTR.
Set exposure mode to MANUAL
Set WB to anything except AUTO ( you'll be shooting RAW, fixing it is the labor of a few seconds in pp.)
Set delivered format to RAW
Set focus to MANUAL
Start with an ETTR exposure with NO tripping of the "blinkies" unless the chosen scene contains some specular reflections, in which case a slender, one-pixel thick column of highlights tripping the "blinkies" is permissable. Noting which exposure is your ETTR starting point, then make a series of nine more exposures, each with a shutter duration 1/3 stop longer than the previous one (3 full stops total beyond ETTR).
Download all the exposures into your computer and, in sequence, open each in Adobe Camera RAW (ACR), Lightroom (which is based the linear processing of ACR), or other RAW converter of your choice. Tonally normalize each by moving the Exposure slider to the left and using visual inspection until you reach an exposure in which there actually are clipped, unrecoverable highlights by visual inspection of the image. Remember, the JPEG-adjusted histogram and the in-camera JPEG-adjusted histogram and thumbnail are helpful in determining reasonable accuracy of JPEG file exposure, but it lies like a used-car dealer in regard to RAW Data exposure.
Did you discover actual highlight clipping after tonal normalization in the third exposure after ETTR? Then the camera has 2/3 stop of ERADR. Five exposures after ETTR? Then the camera has one and 1/3 stops of ERADR. Or, as I've seen in two cameras in the past two weeks, clipped highlights first appearing in the EIGHTH EXPOSURE AFTER ETTR? Thus, those two cameras each have two and 1/3 stops of ERADR .. with barely detectable luminous noise in the blacks and shadows and no detectable color noise ( inspected at 200X to 300X). Downright amazing - World records (in my limited experience!) !
So, after you've discovered your camera's amount of ERADR, what then?
What circumstances are compatible with using EBTR ? A relatively stable to motionless subject permitting longer exposures under reasonably stable illumination ( most landscapes, outdoor or studio still-lifes and portraits)...in other words whenyou aren't under rushed pressure to determine accurately your ETTR exposure. Sports photography, active children, free-living wildlife? these definitely don't leap to mind as ideal candidates for EBTR...UNLESS its overcast with little or no exprected change in E.V. Calmly determine ETTR as precicely as possible, add the pre-determined stops of ERADR, and make your EBTR exposure. Yeah, the thumbnail image in the LED will look like S--T...too bright, almost washed out with practically a written guaran-damn-tee that every highlight and mid-tone, and shadow is blown to Hell! and well they would be were you shooting JPEGs ! But fear not, oh ye of marginal faith....you are shooting RAW and that's NOT overexposed...just an appropriately exposed mass of RAW image data. Stick that sucker in your computer, open in A.C.R., tonally normalize, and revel in the perfection of exposure of the image you previsualized, or of the great variety of other versions made possible by taking greatest possible advantage of the original 12-bit or 14-bit depth image data with practically no noise to deal with! And that bit about 12- bit versus 14-bit making a difference...it holds no water in actual practivce. Do it yourself and find out! The "Bit jockeys" who spout that bushwah have obviously never actually and seriously tried EBTR...they just knew it wouldn't work...theoretically, of course...
The question sometimes arises about extrapolation of results between cameras of the same model. Until recently I had no basis for an answer, having, by bad luck of the draw, been able to test but one example of each model that was reported or passed through my hands. Recently, however, three different Nikon D800s have provided ERADR results of 2/3 stop (John Shaw), one and 1/3 stops ( mborn on UHH) and two and 1/3 stops (DLG). So yes, I strongly advise against reliance on extrapolation of ERADR data among different cameras of the same model ( which may suggest why the camera manufacturers are loathe to publicly broach the topic!).
Another exercise that may aid appreciation of the value of EBTR and use of ERADR involves realization that the width of your camera's histogram represents the dynamic range available for exposure of JPEG image files. The right half of any photosensitometric histogram represents the region with potential for capture of the brightest possible image data without clipping highlight detail. And each brighter stop provides twice the brightness of the adjacent lesser stop. In other words, the contribution of the brightest stop of the exposure has the potential of capturing at least half of the total image data of any given image...IF ... the photographer chooses to ETTR.
And, if the camera is set to deliver RAW image data, there is the option of utilizing ERADR found to the right of the "blinkies" of the JPEG-adjusted histogram.
If, for example, your camera has one full stop of ERADR, the right half of a RAW-adjusted histogram ( which no camera yet provides) could show at least half of your image data beng collected to the right of the JPEG HISTOGRAM's "blinkies" with the image data of the deep shadows being captured far to the right of the dastardly " dark side" where most noise lurks. And remember, the noise in any image depends entirely upon the zone of the histogram frame within which it was CAPTURED, not upon the region of the histogram frame to which it is redistributed during tonal normalization IF ALL TONAL NORMALIZATION IS CARRIED OUT BY LINEAR PROCESSING IN THE RAW CONVERTER, NOT BY THE NONLINEAR PROCESSING CARRIED OUT IN PHOTOSHOP.
Well, that should get you started. Let me know of any questions that might arise.
Dave in SD
The idea for this post comes from a reference in another forum. Since I would like to understand this approach so I can decide to incorporate it into my exposure method, I will keep my comments as neutral as possible and invite all objective, clear and rational responses. Here is the link to the reference.
http://www.johnshawphoto.com/category/exposure/
I shot the appropriate photos as raw's according to this link and looked at them in LR.
1. I set the camera on a tripod and took a basic picture at ISO 800 (to introduce noise), aperture priority (f/7.1) and 1/30. Canon 60D, 18-200 mm lens.
2. I switched to manual exposure and increased exposure until "blinkies" started to show. This occurred at f/7.1 and 1/20.
3. I continued increasing the shutter speed in increments equivalent to 1/3 stop. This obviously made the pictures lighter and shifted the histogram to the right.
The closest I could get the histogram to the right was f/7.1 and 1/10. This is the end of the technical, factual part of the experiment.
Here are my conclusions and I invite people to comment critically, clearly and objectively.
1. Exposing beyond the right and adjusting the brightness to a desirable level do not improve noise or shadow detail.
2. The overall appearance of the standard exposed-to-the-right and adjusted exposed-beyond-the-right pictures are comparable.
XXXXXXXXXXXXXXXXXXXXX
abc1234, and others interested in this thread,
This thread has so far been long on theorization and relatively short on actual, rationally applied practice.
I've been a devotee of ETTR - following a brief, introductory period of skepticism after reading Reichmann's essay in Luminous Landscape in 2003. Then in 2004 I started working with EBTR after reading the Adobe White Paper "RAW Capture, Linear Gamma, and Exposure" by the late Bruce Fraser, and later the collaborative work of Fraser and Jeff Schewe in 2009 ( in "Real World Camera RAW) and since then have included the topic in my classes, courses, and workshops where it has been well received and won numerous convinced EBTR practitioners..
Why your lack of success, abc1234' in noise reduction with EBTR?
Haven't the foggiest, and in light of the growing acceptance of the EBTR exposure procedure based predominantly on the ease of practical elimination of noise, I can only surmise that 1.) your adherence to the requirements of the process may not have been as stringent as it ought have been, and/or that 2.) Shaw's description of the process wasn't as lucid as I had originally thought it to have been.
You stated:
"The closest I could get the histogram to the right was f/7.1 and 1/10. This is the end of the technical, factual part of the experiment."
This is only 1/3 stop more exposure than f/1.7 and 1/20 ( your putative ETTR exposure?) I don't know what that statement means in the context of determining your camera's Extra RAW-Accessible Dynamic Range (ERADR). "Closest" to what?
And nowhere can I find where you actually quantiated the ERADR of your camera.
I included a link to Shaw's article in the thread in which abc1234 was participating when I was asked by someone for other sources of info on EBTR because I recalled, after reading it some time ago, concluding "Yeah; this guy has used it and gets it!" . When reading of your problems getting satisfactory results and that you had followed Shaw's article I realized that Shaw may not have provided sufficient detail for a person new to the topic to successfully follow. I suggest you give it another try.
In order to determine your camera's Extra RAW-Accessible Dynamic Range (ERADR) proceed as follows:
As you did, start with a carefully determined ETTR exposure at an ISO of your choice. For landscapes my choice for EBTR is almost always native ISO (100 in my Sonys) or 200. But if you really want to check the effect of EBTR on high noise potential, go ahead and set ISO to 800, as you also did. Regardless of the ISO used you should experience a marked reduction of noise below the level expected with using only ETTR.
Set exposure mode to MANUAL
Set WB to anything except AUTO ( you'll be shooting RAW, fixing it is the labor of a few seconds in pp.)
Set delivered format to RAW
Set focus to MANUAL
Start with an ETTR exposure with NO tripping of the "blinkies" unless the chosen scene contains some specular reflections, in which case a slender, one-pixel thick column of highlights tripping the "blinkies" is permissable. Noting which exposure is your ETTR starting point, then make a series of nine more exposures, each with a shutter duration 1/3 stop longer than the previous one (3 full stops total beyond ETTR).
Download all the exposures into your computer and, in sequence, open each in Adobe Camera RAW (ACR), Lightroom (which is based the linear processing of ACR), or other RAW converter of your choice. Tonally normalize each by moving the Exposure slider to the left and using visual inspection until you reach an exposure in which there actually are clipped, unrecoverable highlights by visual inspection of the image. Remember, the JPEG-adjusted histogram and the in-camera JPEG-adjusted histogram and thumbnail are helpful in determining reasonable accuracy of JPEG file exposure, but it lies like a used-car dealer in regard to RAW Data exposure.
Did you discover actual highlight clipping after tonal normalization in the third exposure after ETTR? Then the camera has 2/3 stop of ERADR. Five exposures after ETTR? Then the camera has one and 1/3 stops of ERADR. Or, as I've seen in two cameras in the past two weeks, clipped highlights first appearing in the EIGHTH EXPOSURE AFTER ETTR? Thus, those two cameras each have two and 1/3 stops of ERADR .. with barely detectable luminous noise in the blacks and shadows and no detectable color noise ( inspected at 200X to 300X). Downright amazing - World records (in my limited experience!) !
So, after you've discovered your camera's amount of ERADR, what then?
What circumstances are compatible with using EBTR ? A relatively stable to motionless subject permitting longer exposures under reasonably stable illumination ( most landscapes, outdoor or studio still-lifes and portraits)...in other words whenyou aren't under rushed pressure to determine accurately your ETTR exposure. Sports photography, active children, free-living wildlife? these definitely don't leap to mind as ideal candidates for EBTR...UNLESS its overcast with little or no exprected change in E.V. Calmly determine ETTR as precicely as possible, add the pre-determined stops of ERADR, and make your EBTR exposure. Yeah, the thumbnail image in the LED will look like S--T...too bright, almost washed out with practically a written guaran-damn-tee that every highlight and mid-tone, and shadow is blown to Hell! and well they would be were you shooting JPEGs ! But fear not, oh ye of marginal faith....you are shooting RAW and that's NOT overexposed...just an appropriately exposed mass of RAW image data. Stick that sucker in your computer, open in A.C.R., tonally normalize, and revel in the perfection of exposure of the image you previsualized, or of the great variety of other versions made possible by taking greatest possible advantage of the original 12-bit or 14-bit depth image data with practically no noise to deal with! And that bit about 12- bit versus 14-bit making a difference...it holds no water in actual practivce. Do it yourself and find out! The "Bit jockeys" who spout that bushwah have obviously never actually and seriously tried EBTR...they just knew it wouldn't work...theoretically, of course...
The question sometimes arises about extrapolation of results between cameras of the same model. Until recently I had no basis for an answer, having, by bad luck of the draw, been able to test but one example of each model that was reported or passed through my hands. Recently, however, three different Nikon D800s have provided ERADR results of 2/3 stop (John Shaw), one and 1/3 stops ( mborn on UHH) and two and 1/3 stops (DLG). So yes, I strongly advise against reliance on extrapolation of ERADR data among different cameras of the same model ( which may suggest why the camera manufacturers are loathe to publicly broach the topic!).
Another exercise that may aid appreciation of the value of EBTR and use of ERADR involves realization that the width of your camera's histogram represents the dynamic range available for exposure of JPEG image files. The right half of any photosensitometric histogram represents the region with potential for capture of the brightest possible image data without clipping highlight detail. And each brighter stop provides twice the brightness of the adjacent lesser stop. In other words, the contribution of the brightest stop of the exposure has the potential of capturing at least half of the total image data of any given image...IF ... the photographer chooses to ETTR.
And, if the camera is set to deliver RAW image data, there is the option of utilizing ERADR found to the right of the "blinkies" of the JPEG-adjusted histogram.
If, for example, your camera has one full stop of ERADR, the right half of a RAW-adjusted histogram ( which no camera yet provides) could show at least half of your image data beng collected to the right of the JPEG HISTOGRAM's "blinkies" with the image data of the deep shadows being captured far to the right of the dastardly " dark side" where most noise lurks. And remember, the noise in any image depends entirely upon the zone of the histogram frame within which it was CAPTURED, not upon the region of the histogram frame to which it is redistributed during tonal normalization IF ALL TONAL NORMALIZATION IS CARRIED OUT BY LINEAR PROCESSING IN THE RAW CONVERTER, NOT BY THE NONLINEAR PROCESSING CARRIED OUT IN PHOTOSHOP.
Well, that should get you started. Let me know of any questions that might arise.
Dave in SD
If you want to reply, then register here. Registration is free and your account is created instantly, so you can post right away.