ETTL? =[edit]

Refer to http://en.wikipedia.org/wiki/Talk:Exposing_to_the_left — Preceding unsigned comment added by 2.30.163.217 (talk) 22:59, 29 August 2013 (UTC)Reply[reply]

There is a common misconception that ETTL means decreasing exposure (and ETTR increasing exposure), while in fact ETTR may mean increasing or decreasing exposure, referring to some default camera metering (or maybe from a random previous exposure manually chosen in the M mode), with the aim of best digital signal capture. The latter is achieved when the exposure is maximized subject to the constraint that the (important) highlights are not blown. In the histogram this means that the right edge of the histogram is aligned with its right edge. The histogram example is technically correct but for a low-DR scene. The challenging exposure issues are with wide-DR scenes where the dynamic range exceeds the range of the in-camera histogram. Luminance histogram is shown only, JPG vs raw histogram issues are not explained, and a bunch of practically relevant trade-offs are not discussed. It'd be a good idea to add an example of processing ETTRed raw files of wide-DR scenes to fit within standard display DR, preferably in different possible ways. — Preceding unsigned comment added by 212.235.212.118 (talk) 10:53, 17 June 2015 (UTC)Reply[reply]

Criticism[edit]

The premise behind ETTR is the falicy that each bit records a different zone of f stop of light thus meaning the last bit of a 14-bit raw file records massive detail all in the highlights. This is of course an absolute absurdity to anyone who has even a basic understanding of how semiconductors and digitization works but this article makes no mention of it. — Preceding unsigned comment added by 68.150.59.159 (talk) 01:49, 15 August 2011 (UTC)Reply[reply]

You are wrong, or perhaps you have simply misunderstood. I'm trying to make the article clearer. Please let me know what you think. Regards, nagualdesign (talk) 00:19, 26 November 2011 (UTC)Reply[reply]

The common understanding of ETTR (...highest (brightest) stop uses fully half of the discrete tonal values...) is just plain wrong and utter nonsense! Truth is: When you "shift the histogram to the right", you basically increase exposure time. The longer the exposure, the less noise, it's that simple. Sensors catch light, a basic physical principle, nothing else, they have no distribution of complex tonal values or such. — Preceding unsigned comment added by Ahinterl (talk • contribs) 09:25, 8 March 2013 (UTC)Reply[reply]

Histograms[edit]

Eagle-eyed experts may have spotted that the image histograms are non-linear, whereas ETTR deals entirely with in-camera, linear histograms that are subsequently processed in a RAW converter (also linear) before being exported to Photoshop (other non-linear editors are available). A more accurate image of an ETTR histogram would look almost exactly like the 'normal' histogram but shifted to the right, without being stretched out. The same mistake is made on the Luminous Landscapes website, which I reproduced without thinking. Would anybody like me to change the image to avoid confusion? Conversely, does any body think that changing the image would be more confusing? ..I realize that I may be talking to myself here. nagualdesign (talk) 00:19, 26 November 2011 (UTC)Reply[reply]

Apologies for my impatience but I decided to swap the Photoshop-style, non-linear histograms for Canon EOS-style, linear histograms. It's more technically correct, even though it does not look like the ones on the Luminous Landscape website. nagualdesign (talk) 04:34, 26 November 2011 (UTC)Reply[reply]
I'm not sure about those histograms divided into five sections (which I presume relate to the nominal 5-stop range of the camera discussed). Surely the point of ETTR is that those stop dividers should be drawn at 16,32,64,128, illustrating that each higher stop uses twice the number space range of its predecessor? GideonReade (talk) 15:26, 1 January 2012 (UTC)Reply[reply]
The 5-segment histograms are copied off my EOS 600D. I also presumed that each line represented +/-1EV, and a quick test confirms this, yet the camera can take photos with more than 5 stops range. Go figure. (Something to do with the histogram being based on a jpg, rather than the actual raw data.) The only bit of BS are the gradients underneath, which are non-linear and inaccurate. They don't really belong there, strictly speaking, but they give a quick idea of how to read the histograms (that they run from dark to light). nagualdesign (talk) 09:24, 2 January 2012 (UTC)Reply[reply]

Misconceptions[edit]

I have amended the article as the reasons given for the disputed validity of this technique misrepresent the concept. I expect that this is due to misconceptions rather than a blatant straw man argument, so I'll attempt to explain it here and, hopefully, a better way of describing the technique may result, so that others do not have the same misconceptions.

1) "The basic principle, that the lighter tones in an image contain more information than the darker tones is not consistent with the physical realities of image sensors and their operation." Of course any tone may be considered to be a single piece of information - the tonal value (a number). What ETTR actually says is that the range of tonal values describing the lightest tones (the first 'stop') is larger than the range of tonal values that describe the darkest tones.

2) "A connection is made between the fact that a difference of 1 stop represents a doubling or halving of exposure, while the assumption is made that the same principle is applies to the image sensor and the way it records information." The sensor does not record information, it collects light. More properly, it converts light into electric charge. Double the light exposure and you double the charge (up to a limit). This charge is then digitized (converted to a number), such that a doubling of the number represents a doubling of the light exposure. In reality the accumulation of charge and the digitizing process may not precisely follow this rule due to physical limitations and the methods employed to mitigate these limitations, but the basic principle is valid. After digitization these values may be gamma corrected and then recorded (as in a JPEG), at which point we have gone beyond the scope of ETTR (which deals purely with RAW data).

3) "The image sensor will, however, (depending on its design and features) use between 8 and 16 bits of information to record the brightness of a given pixel in the image. If we assume an 8-bit sensor, and for the time being ignore the color of the image, the available brightness values are 0-255. Fully black will be recorded as 0000 0000(binary) = 0(decimal), white will be recorded as 1111 1111 (binary) = 255(decimal). In both cases we see that the image sensor uses 8 bits to record the information, irregardless of the actual value being recorded." As noted above a single tonal value (be it black, white or any shade in between) is indeed represented by a single number, however mid-gray (in this regime) would be represented not as the mid-value, 128(decimal), but by a much smaller number. This would leave many more tonal values available to describe the upper 'half' of the sensor's range of sensitivity as the lower 'half'.

4) "In ETTR, it is assumed that the part of the image with the highest light intensity uses half of the available recording values.." No it isn't. No 'part of the image' uses half of the available recording values. It's the range of tones which describe the highest stop of light sensitivity that use half of the available tonal values.

5) Inserting phrases such as "are said to", "it is assumed", etc. when describing the proposed technique, and making assertions such as "is not consistent with the physical realities.." when rebutting the technique smacks of POV.

Please comment below before making changes to undermine the article. Regards, nagualdesign (talk) 00:04, 6 March 2012 (UTC)Reply[reply]

I would suggest that the x-axis of the histogram be labeled: My initial reaction to those histograms was that they didn't correctly show the scaling of the response due to the doubling of the exposure. I am coming from a sensor/FPA POV where I would consider the photon to electron to 'digital unit' (A/D output) to be the 'linear' response so I would consider this histogram to be non-linear since each vertical bar seems to represent a one stop change (a doubling).

Considered from this POV, it seems clear that increasing the exposure or opening the aperture will increase the intensity resolution in 'digital units' throughout the intensity range while leaving much (but not all) of the noise constant thereby resulting in a better S/N ratio.

In color photography this process can have a negative effect on color fidelity and even contrast as the intensity increases due to non-linearities in the FPA. There is, for any FPA, a 'sweet spot' in the output range that minimizes this latter effect. If one were to do ETTR with this in mind, the necessarily smaller difference between a non-ETTR exposure and this sweet-spot ETTR versus a full-on ETTR exposure would limit the benefit to S/N.

In monochrome FPAs with high linearity throughout the range, there is great benefit from an ETTR approach. Keep in mind that HDRI essentially uses an ETTR-like approach to capture low-lights by allowing high-lights to blow-out, so to the extent that HDRI 'works' to capture low-light areas with a higher S/N, ETTR should be expected to 'work' as well. — Preceding unsigned comment added by Softboyled (talk • contribs) 20:16, 6 September 2012 (UTC)Reply[reply]

Reply[edit]

I edited the article to clarify the fact that this concept is in fact disputed.

Frankly, I believe that [1] proves that the concept is invalid, thus the article should simply briefly state that ETTR is a flawed concept. Out of respect for the time you have put into the article, I chose to merge the dispute into the text, to add balance. The phrases "are said to" etc were added to clarify that the research we have shows that ETTR does not work, although "it is said" that it works.

However, until research surfaces to disprove [2], I strongly believe this encyclopedic article should reflect the fact that the concept is proven wrong, although it still has some following.

Best regards, Thoger — Preceding unsigned comment added by 84.48.118.251 (talk) 08:59, 7 March 2012 (UTC)Reply[reply]

Thanks for the reply, Thoger. Unfortunately the author of Why "Expose to the Right" is just plain wrong also misrepresents ETTR in their blog. The tests that they performed were of little practical value as they used ISO gain to expose to the right. As others have pointed out in the comments section of that blog, only 2 of the examples laid out are even valid. And setting out to disprove something that the author already 'knows' to be wrong is not exactly scientific.
For a more well reasoned analysis of the technique you may wish to read Noise, Dynamic Range and Bit Depth in Digital SLRs, the first of the external links, published by Chicago University. It's author has tested the technique with some rigor and shows that it does indeed work, albeit for slightly different reasons than is often assumed. One thing it does not seem to deal with is selectively adding 'fill-light' in post production (a common technique), which a simple test confirms does indeed lead to posterization. However this simple test would be considered Original Research and is therefore inadmissable. If I could find a reputable source that has published such tests I would add it to the refs.
It is certainly a fact that ETTR is disputed - there are many people online who get frustrated even discussing the concept - hence the need to have 'both sides of the argument' represented in the article. But it is far from factual to claim that its denigrators are correct in their assertions.
If you'd like to reproduce the experiment yourself simply photograph a range of dark tones, like the shadow under your settee (a smooth gradient is best), using a normal exposure, then take another with double the Time Value. Next take the second photograph and stop it down in ACR (other RAW convertors are available), or better yet, take the first photograph and stop it up. Now open both files in Photoshop and count how many 'steps' there are in each shadow. You should see greater (or should I say worse) posterization in the 'normal' photograph. You can use a drastic curves or levels adjustment to see the banding more clearly, but do remember to treat both photographs in an identical manner and make sure that you're working in 16-bit mode. (Convert them to 8-bit and they will look nearly identical.)
Regards, nagualdesign (talk) 15:02, 7 March 2012 (UTC)Reply[reply]

References

  1. ^ http://chromasoft.blogspot.com/2009/09/why-expose-to-right-is-just-plain-wrong.html - Why "Expose to the Right" is just plain wrong
  2. ^ http://chromasoft.blogspot.com/2009/09/why-expose-to-right-is-just-plain-wrong.html - Why "Expose to the Right" is just plain wrong

Further misconceptions[edit]

I've reverted a lot of the edits made recently as there were a few misconceptions being banded about. The facts about signal-to-noise ratio were already mentioned in the opening paragraph, including the reference to Emil Martinec's work. Representing it as an 'important paper' that all but debunks Michael Reichmann and Thomas Knoll is wrong, however. That paper actually shows that ETTR works in the simple case of overexposing (within limits) and then globally stopping down an image in post-processing, compared to simply taking a normal exposure. However, it does not deal with cases in which an image is processed via dodging and burning and other techniques, which any photographer will tell you is the very reason they shoot in ETTR: Latitude.

When burning (or otherwise darkening) part of an image you are in effect discarding data, whereas when dodging (or otherwise lightening) part of an image you're often attempting to represent or reconstruct data that really isn't there, having been lost in the noise or, in the case of adding fill light to dark areas, having been poorly digitized in the first place. To put it simply, by overexposing in the first place nothing needs to be lightened in post. If you'd like to quote the section(s) of the paper that you referenced which you think demonstrates the 'misdirected reasoning' of Reichmann and Knoll, perhaps we can discuss how the article might be improved. nagualdesign 01:31, 31 December 2015 (UTC)Reply[reply]

Even more misconceptions, and ETTR for high-contrast scenes (which may require NEGATIVE exposure compensation)[edit]

There is a very common misconception that ETTR always means increasing exposure (and that there is an ETTL standing for decreasing exposure). In fact it is about adjusting exposure so that the histogram is flush right, without a spike indicating blown highlights (to be more precise, blown important highlights). Whether the exposure correction is positive or not, depends on the the initial value of the camera meter.

In the simple low-contrast context, initial metering indeed usually produces a central blob in the histogram, and ETTR requires a positive correction. This is good as an initial didactic part. But in practice, the gains of ETTR with low-contrast images may be worthwhile only with poor-quality image sensors, or when really heavy post-processing follows.

The challenge where ETTR gets really useful these days is high-contrast (high-dynamic-range - HDR) scenes, which used to be the domain of multiple-exposure HDR techniques. Specifically I mean moderately-high DR, which still fit within the DR of the better image sensors when shooting RAW, but not within the DR of current in-camera JPG engines, about 8 stops. Certain in-camera DR extensions (Nikon ADL, Sony DRO etc) may be able to process some more DR, see for instance "Dynamic range" pages of camera reviews at DPReview.com. But one is still able to do much better manually by shooting raw, exposing carefully (using ETTR with some further considerations), and processing with a wide-DR capable raw converter (for instance, Adobe Lightroom, or the free Raw Therapee).

So, I think explaining that ETTR may actually require a negative exposure correction is practically important, and I made a honest attempt to fix it. I think the standing ETTR wiki was a major disgrace to Wikipedia, because many budding photographers trying to learn advanced techniques may actually get a wrong impression here. Unfortunately, my edits were quickly undid by Nagualdesign. I fully admit my contribution lacks references etc because I'm not a skilled wikipedian and don't have much time for this. There is much more that should be added to put ETTR in the full context of practical single-exposure HDR processing. The external link "ETTR Exposed" is a very good reference, with image and jpg/raw histogram examples. I merely wanted to point out that claiming that ETTR always means increasing exposure is plain wrong, and to give a hint that ETTR is also relevant in a high-contrast context. BR, _sem_

The benefits of ETTR (greater tonal range in dark areas, greater signal-to-noise ratio, fuller use of the colour gamut and greater latitude during post-production), which is essentially a form of over-exposure, cannot be gained through under-exposure. What you are describing is called "exposing for the highlights" (to the detriment of the rest of the image), not ETTR. If the dynamic range of the scene exceeds that of the camera then ETTR isn't a viable option; instead you can either expose for the highlights, expose for the shadows, take a normal exposure or use HDR/bracketing. nagualdesign 00:55, 4 January 2016 (UTC)Reply[reply]
ETTR is nothing but exposing for the highlights, because it is defined as aligning the right (highlights) part of the histogram. It is quite true that if scene DR exceeds camera DR this goes to the detriment of the shadows range (noise and tonal range). The confusing part is that the better sensors these days have considerable sensor (raw) DR, that is about 14 stops above the noise floor (according to the DxOMark); take two or three stops away depending on your personal quality standards of the useful photographic DR, you have 11 to 12 stops available, which is much more than the camera DR of about 8 stops (which is there mainly because by what current displays are typically able to show). The point is that with such sensors, one is able to practice ETTR (expose for the highlights), with respect to the raw histogram, therefore avoid blown (important) highlights, and still get shadows clean enough for the rather heavy lifting needed in post-processing.
This is not in conflict with the basic understanding of ETTR assuming a low-contrast scene, but a simple extension of it. Not feasible with most smartphone cameras indeed, but this doesn't mean it isn't useful.
Unfortunately, even current big cameras don't support raw histograms and decent enough HDR rendering yet - until they do, understanding ETTR is a fundamental concept which lets one do this manually. ETTR with a raw histogram (well, an approximation of it) is available in Magic Lantern, an alternative firmware for Canon DSLRs (unfortunately Canon sensors are not the best out there regarding DR at base ISO, and additionally suffer from banding which is not considered in DR measurements). UniWB is an approach of modifying the default camera rendering so that the in-camera histogram becomes a good approximation of the raw histogram in the highlights range, unfortunately resulting in odd greenish JPGs. Several new cameras sport a "flat" (low contrast) rendering mode intended for capture of a wider DR (the resulting images tend to look dull and require post-processing); in this mode, the in-camera histogram is somewhat closer to the raw histogram.
The Magic Lantern firmware also implements an auto-ETTR function (I've got no personal experience with it). Certain recent Nikon DSLRs have a highlight-weighted metering option, which is essentially also auto-ETTR (however, according to a few expert reviewers it is not useful for ETTR with respect to the raw histogram, as needed for single-shot HDR use). One should expect issues with auto-HDR algorithms because the right edge of histogram content is often poorly defined (for instance with small specular highlights) and because the the metering sensors of DSLRs are low-resolution (have large "pixels"); one might expect better performance with mirrorless cameras or in live-view, where the main sensor is used for metering - but it is unrealistic to expect that the entire image sensor is used in metering, due to camera processor computing power limitations.
HDR with multiple bracketed exposures is indeed a superior approach regarding the quality of the shadows, but unfortunately limited to still scenes! BR, _sem_
Thank you for the reply, although it fails to cut to the heart of this issue. Before I explain why, I would like to note that I have read every single reference used on this article in full, plus many more, and I understand them completely. I also have Magic Lantern installed on my camera and have personally tested and verified ETTR as described by Thomas Knoll et al. There is no need to assume ignorance on my part.
Now, the whole concept of ETTR is predicated on the fact that by deliberately overexposing an image then stopping down during post-production you acquire greater tonal range in the shadows, improved SNR, fuller use of the colour gamut and greater latitude during post-production (particularly when adding fill light). It might properly have been called Deliberately Overexposing Within Limits (DOWL), as that is precisely the prescription offered by the technique. However, as this technique involves careful attention to the histogram it is more properly termed ETTR.
ETTR was primarily intended for anyone shooting in RAW who also processes their images in Photoshop/ACR. Even in 2003 (when ETTR was first described) most DSLRs had enough dynamic range to adequately capture most scenes, but heavy post-production could noticeably reduce image quality. That is still true today. ETTR wasn't intended for point and shoot cameras or iPhones. And if all you do is overexpose then stop down the whole image in post (without dodging/burning, adding fill light, etc.) you have gained virtually nothing.
When it comes to high-dynamic-range scenes (those which equal or exceed the DR of the camera) ETTR becomes redundant, as the digitization of a normally exposed image is already optimal (the right half of the RAW histogram is already in use). In such cases it may be advantageous to Expose for the Highlights by underexposing the image. However, this will result in reduced tonal range in the shadows, reduced SNR, lesser use of the colour gamut and much less latitude during post-production (particularly when adding fill light). It is therefore the antithesis of ETTR, notwithstanding the fact that you may have carefully made sure that your histogram extends as far to the right as possible without clipping. In reality, when you Expose for the Highlights you move the histogram to the left (increasing underexposure) until the highlights are no longer clipped. In photography you almost always avoid highlight clipping, but that does not imply that you are always employing ETTR.
The major problem if the article was rewritten to include Exposing for the Highlights (or "ETTL"), which is basically what you are attempting, is that all the benefits and rationale of ETTR explained in the article would no longer apply. Every test you might think of which demonstrates the efficacy of ETTR would be turned on its head if you Expose for the Highlights.
The section of the article which you have added (ETTR with high dynamic range scenes) is, frankly, prattle. Since you cannot provide robust references for your assertions, and don't seem to understand either Wikipedia or ETTR, yet persist in edit warring I will leave the article as is for now and open a Request for Comment (below). nagualdesign 03:00, 5 January 2016 (UTC)Reply[reply]
ETTR is, fundamentally, exactly the same thing as exposing for the highlights. You don't expose an arbitrary amount of EC to the right. In the original low-contrast ETTR context, you increase exposure until the highlights edge of the histogram content hits the highlights edge of the histogram, any by exactly the amount of difference from the initial camera metering. But even in this context, if you wouldn't start with a roughly centered histogram of a low-DR scene with a blob somewhere in the midtones range, but at a random previous position in the M mode instead, you would sometimes have to decrease exposure if the starting histogram was blown in the highlights, wouldn't you?
The motivation for exposing to the right (that is highlights) edge comes from the hard saturation properties of digital sensors. Film had different properties (Ansel Adams etc), blowing highlights was not so problematic because the saturation was smooth by chemical means. Digital sensors saturate abruptly; in blown areas with small gradients one can actually see edges at lines where each of the three colour channels get saturated, and weird rainbows of false colours between them. And they are nastily obvious because they are in bright areas of unprocessed images (there are odd things at the shadows edge too, which are the subject of the ETTR procedure, but normally only get visible when lifting shadows heavily or with some other heavy forms of pp). The main concept behind ETTR is the distinction between the visually desired exposure and the exposure optimal from the digital measurement point of view.
So, there is no such thing as ETTL in digital shooting. ETTR / exposure for the highlights is maximizing exposure within limits with the aim of best data capture. Increasing/decreasing exposure is the "manipulated variable" here, not some ETTR/ETTL.
ETTR was originally invented with low-DR scenes in mind. But at the time, there were no sensors with 14 stops of DR at hand. With the sensors originally considered (and today's compact camera & smartphone sensors - ETTR surely is applicable to compact cameras, I practice it often with a Canon G16, no highlights headroom but I don't really like its meter), the ETTR concept is important in order to get clean shadows from a normal-DR scene. If you got two stops using ETTR in 2003, the improved sensor gives you the same advantage today without ETTR chimping, because of the improved DR (on account of less noise and more bit-depth). Careful exposure based on ETTR is still valuable with the improved sensors, but the practical advantages are visible mainly with higher-DR scenes (these sensors can be used for many scenes which were once in the exclusive domain of multiple-exposure HDR).
The lack of raw histograms is a particular problem with ETTR in the wide-DR-scene context that blurs the understanding of the meaning of ETTR for this purpose. If raw histograms were available on every camera, using ETTR for 11-stop scenes would be quite natural. But camera user interfaces are "JPG-centric", and the raw image has the status of an optional extra. And, unfortunately, the camera image rendering engines are not catching up with sensor DR improvements so well. I know because I can render properly exposed wide-DR raw files in LightRoom much better than the camera does using its instant DR extension (Nikon ADL, Sony DRO...). Mind these extensions are not some faux-HDR, because they render with a natural-look intent and actually do capture a substantially higher input DR, than default, about 12 stops with ADL on recent Nikon full-frame DSLRs (see respective DPR camera reviews). But they don't get the full 14 stops available in raw - which is partially because they don't practice ETTR, which one can see by using RawDigger to check raw data of such images which are ETTR'd according to the camera histogram. (But this is not the only problem here; there is probably the issue of finite bitlength in processing, and also automatic HDR-style rendering to a pleasing and naturally-looking output image within the standard output DR gets increasingly more challenging with increasing scene (input) DR.).
A further refinement on top of ETTR is the notion of exposing for the highlights deemed important only, for the sake of cleaner shadows (the basic aim of ETTR). For instance, if there is the sun or sharp-edged specular highlights in the frame, it is usually okay to leave them blown (notice the limit has changed with the introduction of handy "highlights recovery" algorithms available in some raw convertes but not yet in-camera rendering engines, which let one blow some other unimportant highlights without the ugly side-effects). Then one is no longer looking at the histogram content as one chunk that needs to be flush right for basic ETTR; the exposure is increased a bit further, so that the spikes that can be associated with the unimportant highlights go into saturation.
You claim: "When it comes to high-dynamic-range scenes (those which equal or exceed the DR of the camera) ETTR becomes redundant, as the digitization of a normally exposed image is already optimal (the right half of the RAW histogram is already in use)." If the scene DR is larger than the sensor DR, there is no optimal exposure just compromises (bracketing and HDR blending is possible with still scenes). In my experience with Nikon DSLRs, with scenes over the standard ca 8 stops shown by the histogram but within the sensor DR so that a non-blown raw capture is possible, the default metering is somewhere for the midtones, but rather haphazard than optimal, prone to blowing highlights unnecessarily, with no simple and reliable way of raw-ETTR using automatic metering (no ML for us unfortunately). A couple of the latest models feature Highlights-Weighted Metering, which is reportedly prone to underexposure of raw data unfortunately (diglloyd.com) (probably made for JPG). This is why these things need to be explained to novices on daily basis, and it really doesn't help to have a Wikipedia article in a form which offers no clue. It would be really great if someone could explain things more nicely and add references, but I don't think burying head in the sand is an acceptable procedure. BR, _sem_
Please try to discuss things clearly and succinctly in a way that addresses the issues raised, rather that one person making one point and the other then making an entirely different point ad nauseum. Also, try to assume good faith, and leave out comments like, "I don't think burying head in the sand is an acceptable procedure."
Since you quoted one of the things I said I will respond to that. As I said before, "When it comes to high-dynamic-range scenes (those which equal or exceed the DR of the camera) ETTR becomes redundant, as the digitization of a normally exposed image is already optimal (the right half of the RAW histogram is already in use)." To reiterate, by 'optimal' I specifically meant that the right half of the RAW histogram is in use. This implies that full use has been made of the available dynamic range of the camera. At this point overall IQ cannot be improved using ETTR (as described by Thomas Knoll et al) and, as you mention, any artistic decisions will involve compromises. (Note also that each image in a HDRI sequence is a compromise.)
My advice would be to allow the RfC to play out before piling on additional reading material. You have made your point(s) for all to read. Maybe take a step back for now and see how the discussion develops. Sincerely, nagualdesign 23:36, 5 January 2016 (UTC)Reply[reply]
I am sorry if you took that sand as an offence; with that just I wanted to point out to you that I didn't perceive your undos and replies exactly as "good faith". If I seem to be discussing ad nauseam, this is because I feel like you're not showing interest in arguments; I was trying to explain my views better, I do not think there are entirely different points there. But it is not my intention to be confrontational, I would appreciate if we could work together to make the article better. I think it is a better idea to keep a discussion (I don't want to see this as a dispute), I would like to discuss matter we don't agree in, not to quarrel but to get things more clear. Generally, I have zero wikipedian ambitions, and I have no illusion my lines are perfect. I wouldn't care what wikipedia says on ETTR for my private needs. But I keep seeing novices who are confused over ETTR. Not the basic low-DR part, that's easy, and you've got that covered roughly. The higher DR and JPG-vs-raw histograms, that's where they are confused. Some come quoting this article, with not much clue how to use ETTR in a bit more challenging cases. I thought a better explanation here might help many. I did come in good faith.
Regarding your specific reply: "To reiterate, by 'optimal' I specifically meant that the right half of the RAW histogram is in use. This implies that full use has been made of the available dynamic range of the camera. ..." I think you didn't catch the particular situation I am addressing, maybe because you are used to a camera which has the ability to show raw histograms (with ML) but does not have a sensor with as much useful DR as some Nikons and Sonys. These other cameras are, with careful exposure and in raw, able to record scenes with quite high DR without compromise. ETTR would be directly applicable, if they were able to show the raw histogram (well, a good approximation of it). With a minor deviation from the circumstances described for the low-contrast scene: in my experience with Nikon cameras, one cannot always assume that the initial camera metering will produce an image with a "centered" raw histogram. Quote from you: "ETTR becomes redundant, as the digitization of a normally exposed image is already optimal (the right half of the RAW histogram is already in use)." To reiterate, by 'optimal' I specifically meant that the right half of the RAW histogram is in use." In my experience, in such cases, using default camera metering, the right half of the histogram may be sometimes in use but only partially, and sometimes there may already be blown highlights, despite ample unused shadows range. How would you form ETTR advice for such cases?
Reading your sentences: "ETTR was primarily intended for anyone shooting in RAW who also processes their images in Photoshop/ACR. Even in 2003 (when ETTR was first described) most DSLRs had enough dynamic range to adequately capture most scenes, but heavy post-production could noticeably reduce image quality." reminded me that some of the above issues, related to jpg-vs-raw-histogram difference, have been present all along, are not high-DR specific.
I would also like you to comment this paragraph above: "::::ETTR is, fundamentally, exactly the same thing as exposing for the highlights. You don't expose an arbitrary amount of EC to the right. In the original low-contrast ETTR context, you increase exposure until the highlights edge of the histogram content hits the highlights edge of the histogram, any by exactly the amount of difference from the initial camera metering. But even in this context, if you wouldn't start with a roughly centered histogram of a low-DR scene with a blob somewhere in the midtones range, but at a random previous position in the M mode instead, you would sometimes have to decrease exposure if the starting histogram was blown in the highlights, wouldn't you?" BR, _sem_
Some thoughts regarding metering and histograms
When using a light meter (or built-in meter) it's quite straightforward to match the mid-tones of the scene to the mid-tones of the negative. For low-key or high-key scenes you have to do a little thinking and set your exposure (or exposure compensation) lower or higher, respectively, if you want your image to match the scene.
If your camera shows a histogram you can view the range of tones in the image and ensure that there is no excessive clipping. The histogram is normally gamma-corrected in order to represent how we see; ideally, an evenly lit, normally-exposed scene will show a flat/level histogram tapering down at the ends.
For low-key or high-key scenes you would expect to see a histogram weighted to the left or right, respectively. However, it's quite common for photographers to deliberately increase exposure for low-key scenes (expose for the shadows) or decrease exposure for high-key scenes (expose for the highlights), such that the histogram sits somewhere in the middle, and correct it later in post. (High-DR scenes may also be exposed for the shadows or highlights.)
Most gamma-corrected histograms show 5 stops (of equal width), which is good for assessing the tonal range of the image, aiming for a fairly even spread. A linear histogram, on the other hand, reveals that this 'even' spread is often bunched up in the left half. If you intend to develop your digital negatives it's advantageous to make use of the entire linear range to ensure maximum IQ, in which case an in-camera, gamma-corrected histogram should be bunched up to the right, hence expose to the right.
It's worth noting that the finished image ought to have an even, gamma-corrected histogram, so having that on your camera, along with the live viewfinder, helps you to pre-visualize the finished image (linear histograms not so much). The clipping warning of a gamma-corrected luminosity histogram can be a bit oversensitive though. Personally I prefer to use a linear RGB histogram purely for the accurate highlight clipping warning, safe in the knowledge that I can always reproduce a good tonal range (normalized histogram) in Photoshop.

Regarding your final paragraph, I don't quite follow all of it to be honest, but when you say, "ETTR is, fundamentally, exactly the same thing as exposing for the highlights", I have to disagree. As I understand it exposing for the highlights implies slightly underexposing the image to ensure that highlight details are retained. nagualdesign 01:35, 7 January 2016 (UTC)Reply[reply]

Generally I quite agree with everything regarding "conventional" exposure - exposing for the midtones / highligts / shadows, high- and low-key shooting consideration. "Conventional" exposure assumes that one should expose as to achieve the desired final exposure, opposed to ETTR where the exposure when shooting is set as to maximize recorded image quality, and the desired final exposure is set in post-processing. Where we don't seem to agree is the equivalence of EFTH and ETTR. Yes, the two come from quite different premises. But I believe their outcome happens to coincide, because when maximizing exposure to the right edge of the histogram with ETTR, one also happens to expose for the highlights - just because the highlights are at the right edge of the histogram content (subtle differences do exist - EFTH is not defined so precisely and its exact effect on the right edge of the highlights is not certain, while ETTR may be merely sensitive to poorly defined histogram edge due to small highlights). It is true that by EFTH (and ETTR) one may have to sacrifice the quality of the shadows in case of higher-DR scenes, but this is an issue only when the DR of the sensor is not sufficient for the DR of the scene.
Regarding raw/jpg, linear/logarithmic/gamma-corrected histograms, certainly each has its value. Gamma-corrected is surely the most relevant for final processing. I think a logarithic raw histogram is best for judging exposure for raw ETTR, though it may be a hard nut for wide masses - well those that decide to shoot raw should be able to grasp them, but the industry doesn't seem to like embracing them. Perhaps a gamma-corrected histogram with marked points of blowing could be a compromise solution, combined with a "flat" image setting and a sufficiently wide histogram display. Currently, gamma-corrected histograms tend to fail for the pre-visualisation purpose in the higher-DR case, simply because in-camera rendering isn't good enough at the automatic HDR tonemapping part. The current in-camera DR extensions (experience with Nikon ADL) do help with better HDR tonemapping, but make things worse for judging raw data exposure because they involve heavy and scene-dependent instant post-processing that heavily affects the histogram, and they also actively affect exposure (underexpose raw data by up to a stop or two). _sem_

I'd like to offer you a little reading material, since you have offered me so much. :-) First, an Adobe whitepaper: Raw Capture, Linear Gamma, and Exposure Regarding exposing for the highlights it says, "You may be tempted to underexpose images to avoid blowing out the highlights, but if you do, you’re wasting a lot of the bits the camera can capture, and you’re running a significant risk of introducing noise in the midtones and shadows. If you underexpose in an attempt to hold highlight detail, and then find that you have to open up the shadows in the raw conversion, you have to spread those 64 levels in the darkest stop over a wider tonal range, which exaggerates noise and invites posterization. Correct exposure is at least as important with digital capture as it is with film, but in the digital realm, correct exposure means keeping the highlights as close as possible to blowing out, without actually doing so. Some photographers refer to this concept as “Expose to the Right” because you want to make sure that your highlights fall as close to the right side of the histogram as possible."

I fully agree with everything here. With the addition that the threat to the shadows and possibly midtones depends on the difference between the scene DR and the sensor DR in the raw case (or between the scene DR and the camera JPG engine DR in the JPG case). _sem_

This is precisely the sort of source that can and should be cited in Wikipedia articles. And since WP is founded on verifiability, not truth, no amount of explaining things as you understand them can make up for a lack of good references. On WP that's known as original research, which is prohibited.

I can't take credits for inventing any of this. All I know, I have learned mostly from the web, some from books, and just verified myself. I fully agree that it would be great to look up proper references, and I did supply some. But I think it is firstly important to form content in a didactically proper and reasonably accurate form. No point in tedious editing if there is no consensus on the content.
Some more relevant references here: http://www.rawdigger.com/howtouse/exposure-for-raw-or-for-jpegs http://www.rawdigger.com/howtouse/beware-histogram
I am aware of certain problematic points where decent references are probably non-existent. For instance, there is no proper documentation on camera metering algorithms that I would be aware of. Some rather anecdotal evidence is present in certain web forum threads, but these are subject to disappearing image examples etc. The in-camera DR extensions are even more of a mystery. There are rumours that ADL, DRO, HTO, ALO etc originate form Apical (http://www.apical.co.uk/assertive/camera/, http://www.dpreview.com/articles/3798759501/apical, https://www.google.si/patents/US7302110, https://www.google.si/patents/WO2002089060A2?cl=en).
Conversely, I can find numerous examples equipped with raw and jpg histograms, indicating that raw images offer a stop or two of additional "highlights headroom". But this doesn't seem to be an universal truth; my humble Canon G15 doesn't seem to have any, and if the jpg histogram seems blowm, the raw image is most likely blown to. Though, this isn't a bad thing according to the ETTR concept. Rather, it is a bad idea to waste good DR on the mostly wasted highlights headroom, especially with sensors which don't have lots of DR. _sem_

Here's another interesting article from digitalcameraworld.com inasmuch as the use of images and straightforward description explains ETTR in an easy to understand fashion. IMO, we should stick to something similar on this article. Introducing the notion that slightly underexposing is sometimes also considered ETTR is just misleading, IMO. nagualdesign 05:03, 7 January 2016 (UTC)Reply[reply]

I agree this article is nicely written, regarding the very basic things. But it fails to address the more challenging aspects of ETTR use with raw histograms. So novices after learning that bit very soon see higher-DR examples and come asking "Why is your source image so dark? How can you say this is ETTR? You cannot get that clean shadows without multiple exposures!" But it is nothing but raw-based-ETTR exposure using a camera that sports a wide-DR capable sensor, followed by four slider drags in LightRoom :) (http://www.dpreview.com/reviews/nikon-d7200/13) Can't we offer them a better clue?
BTW, Jack Hogan has proposed to include a more thorough discussion of ISO. BR, _sem_
You say you have experience with ML and the raw histograms inthere. The sample screenshots I've seen seem to have a relatively small histogram window. Is it possible to set a wider histogram in ML, for a better assessment of scene DR and more accurate ETTR?
And did you also try the ML dual-ISO feature, which allows one to trade resolution for DR (http://www.magiclantern.fm/forum/?topic=7139.0, https://rohidassanap.wordpress.com/2013/08/10/guide-to-using-dual-iso-mode-on-your-canon-5d-mark-iii/, https://www.youtube.com/watch?v=kQNR4HcfjtI)? I assume the resulting DR is comparable to what one gets with the recent Sony and Nikon sensors with a single exposure at base ISO. Notice most references to it recommend the use of raw histograms and ETTR.

RfC: Should this article cover Exposing for the Highlights

As the subtitle asks, should this article cover Exposing for the Highlights? Please read the above section before providing opinions. The IP editor appears to believe that any exposure in which the histogram reaches the right edge (without clipping) is a form of ETTR, even if the resultant image is underexposed. There are indeed (non-WP) articles online that make the same assertion, but they do not explain that underexposure negates the benefits of ETTR (in fact, having the opposite effects). Also, should the article explain that ETTR is not intended for high-DR scenes? Unfortunately I have not been able to provide good references for that, and even though it is self-evident with a little experimentation original research is not permitted on WP. nagualdesign 03:00, 5 January 2016 (UTC)Reply[reply]

Yes, a lot of different interpretations of what is being exposed to the right of what. This page should instead help to clarify the confusion by explaining how to achieve ETTR.

So for a given scene situation with Exposure in Manual mode, as long as no highlights in which the photographer wants to retain detail are clipped in the raw data:

  1. shutter speed should be the longest allowable by the photographer's constraints for camera/motion blur;
  2. aperture should be the largest possible subject to the photographer's sharpness and DOF requirements;
  3. ISO should be set in-camera to the lowest level resulting in the cleanest shadows.

1 and 2, representing manual Exposure, should be maxed out first at base ISO. If desirable highlights have been clipped in the raw data at this stage the artist may decide to compromise some blur, sharpness and/or DOF by adjusting 1 and 2 until they no longer are (just). If the camera is ISOless that's it, mission accomplished. If the camera is not ISOless and no desirable highlights have been clipped in the raw data, ISO should be increased in-camera (3) until either just before they are or the camera becomes ISOless, whichever comes first. In neither of the two situations above the histogram is necessarily all the way to the right, although it often is for obvious reasons.

This is valid for both low DR and high DR images, exposing for the highlights need not apply. I feel that having a separate section on ETTH, low contrast vs HDR is quite confusing (at least to me). Here are a couple of links with further ideas Jack Hogan (talk) 16:00, 5 January 2016 (UTC)Reply[reply]

Comment: Of course, the low-DR and high-DR cases are essentially the same, but under the assumptions that the sensor DR is sufficient for the scene DR, that a raw histogram or its good approximation is available when judging exposure, and that the camera meter defaults to a lower exposure value than ETTR. The second one is unfortunately mostly invalid, except with Magic Lantern and one Leica AFAIK. The third one is often violated, though it depends a lot on the camera and shooting mode - I find it completely irrelevant but Nagualdesign does not accept using "ETTR" for anything involving a negative exposure compensation. Also, I tried making a special section on the higher-DR case to leave the original article intact as much as possible, because this was indeed the way ETTR was first established, and to avoid conflict with Nagualdesign. BR, _sem_
@Jack Hogan: I completely agree with everything you have said here. Let me ask a direct question; Should the section ETTR with high dynamic range scenes be removed? nagualdesign 23:36, 5 January 2016 (UTC)Reply[reply]
Categories