ETTR, when and why

[SrMi]

Advertisement (gone after registration)

7 hours ago, pop said:

I don't think so. Mapping a low contrast scene to a medium capable of rendering a higher contrast is straightforward. Of course, you can improve the result by increasing the exposure if the recording medium performs better for higher levels. 

The issue becomes more relevant when the contrast of your scene exceeds the capability of your sensor. In this case it seems useful to remind photographers who are mostly accustomed to negatives on film that a digital sensor clips the highlights while film clips the shadows. Featureless white areas in an image are often a more serious defect than featureless black blobs. YMMV, of course.

I do not understand which part of this you disagree with:
"The basis of ETTR is to expose the sensor to more photons than with 'regular' exposure. More photons equal better SNR. Better SNR is preferable, regardless of sensor generation."

ETTR does not change the contrast of an image; it attempts to improve the signal's quality by optimizing the exposure (while protecting highlights!).

[Advertisement]

The Leica Q3 is available for pre-order at B&H Photo Video and Adorama

[pop]

vor 2 Minuten schrieb SrMi:

The basis of ETTR is to expose the sensor to more photons than with 'regular' exposure.

It's this part which I think applies only in some cases. Since photographers are bound to overexpose when they apply the same procedures to exposing sensor as they did to exposing negative film, the discipline called ETTR would expose the sensor to fewer photons, not more.

[pgk]

14 minutes ago, SrMi said:

"The basis of ETTR is to expose the sensor to more photons than with 'regular' exposure. More photons equal better SNR. Better SNR is preferable, regardless of sensor generation."

ETTR does not change the contrast of an image; it attempts to improve the signal's quality by optimizing the exposure (while protecting highlights!).

BUT, the problem is that this logical ideal is usurped for many reasons. It's complicated. Bottom line, as I always say, is that photography is about practical application. If you try both ETTR and 'normal' exposures then they may give very similar results, or the ETTR image may have colour artefacts. My experience is that the so called SNR 'benefits' of ETTR are so marginal that I cannot remember an instance where such an exposure has yielded a 'better' result. The only benefit of what some refer to as ETTR is in my experience, to ensure as much shadow detail as possible is recorded, but whether you can seriously call exposing for the shadows ETTR, because its digital and new(wish), I very much doubt. That's just careful exposure, and 'correct' exposure itself is all too often not easy to define but relies on the photographer's experience and the desired result.

[SrMi]

8 minutes ago, pop said:

It's this part which I think applies only in some cases. Since photographers are bound to overexpose when they apply the same procedures to exposing sensor as they did to exposing negative film, the discipline called ETTR would expose the sensor to fewer photons, not more.

Good point. The statement "The basis of ETTR is to expose the sensor to more photons than with 'regular' exposure" is not correct. ETTR can incur both positive and negative exposure compensation, as preserving highlights in digital cameras is very important. The resulting image can appear both too dark or too bright before the brightness is adjusted in the post.

[adan]

7 hours ago, pgk said:

Sorry Andy but you don't and we've had this argument before. Older Mandler lenses record less tonality not more.

Yes, we have. So I will reply with the same evidence I did before. If you have photographic evidence to the contrary, preferably with M-mount lenses, I would be happy to see it.

Same scene, shot with 28mm Elmarit-M ASPH, and Mandler 21mm Elmarit-M. Identical "same as previous" post-processing, same exposure. Two crops from a single exposure from each lens.

You tell me - which lens simultaneously retains more detail and tonal separation in the:

highlights - subtle textures and nicks in the white-painted concrete, and the color saturation of the beige foundation floor.

shadows - a large number of tonal separations: seams in the black wall, little bits of detritus along the base of the black wall, red overlays painted on the black wall, the spots and tones and colors within the shaded beige concrete pavement?

And which lens reduced the tonal separation by simultaneously blowing out the whites and muddying the shadow separation in the same exposure?

It is a reproducable and predictable effect I can demonstrate over and over, dating back to the film tests I made with contrasty, highlight-blocking, Zeiss G lenses vs. Mandler lenses in 2001 (which resulted in my dumping Zeiss/Contax for Leica-M/Mandlers).

Welcome, dear visitor! As registered member you'd see an image here…

Simply register for free here – We are always happy to welcome new members!

 

[adan]

Back to ETTR; It pays to remember that when Michael Reichmann proposed the technique in 2003, it was related to a specific technical characteristic of digital sensors and digital processing.

They are linear in their basic behavior. One photon in results in one electron out (with CMOS it is a voltage, not a physical electron, but still directly proportional to the photon flow.)

The camera then counts the electrons or measures the voltage and converts the brightnesses to digital bits.

If one exposes an image so that the brightest point in the image just reaches saturation level (a brightness value that is 255 in 8-bit files, for jpegs), then an object one stop darker in the scene will produce half the light/photons/electrons/volts/bits, and have a "truly raw" value of half of 255 (127.5 - a medium gray). And another object reflecting a stop less light would receive a value of 64-ish (a very dark gray)

In other words, the native "zone-system" of a sensor would normally result in a digitized brightness scale (rounded) of 255, 128, 64, 32, 16, 8, 4, 2, 1, 0.

And it is pretty hard to distinguish shadow tones if the available values are 1, 2, 4, and 8. And if one tries to rescue those shadows by brightening them, say, one stop, they posterize into 2 tones or 4 tones or 8 tones. That meant in the digital infancy of 2003, the 8-bit file format put a hard ceiling on usable, visible dynamic range of about 5 stops.

And 8-bit .jpgs were the dominant output format right up until 2003 and beyond, when Adobe introduced Adobe Camera Raw for Photoshop. Raw images were a tiny niche of digital photography before then. Very, very few people fiddled with the much-larger 16-bit TIFFS (if a camera even offered that file option) or the proprietary raw processors from camera makers.

It was a really big deal on dpreview when a new camera offered "raw" support in that era. It was a high-end feature. (Remember when the Digilux 2 locked up for 8 seconds between pictures to shoot in raw?)

So ETTR was a technique for "the rest of us" to get the widest tonal range out of jpgs. It still works with raw - it just is not as important now.

Edited September 25, 2020 by adan

[Jeff S]

Advertisement (gone after registration)

Ctein weighed in on ETTR years ago at TOP.

https://theonlinephotographer.typepad.com/the_online_photographer/2011/10/expose-to-the-right-is-a-bunch-of-bull.html
 

Jeff

[adan]

I especially liked the link-through to Ctein's essay on the Photo-Fetishistic League.

And the quote which I rearrange here:

"If all you can play is Three Blind Mice, who cares how perfectly your piano is tuned?"

Edited September 25, 2020 by adan

[SrMi]

30 minutes ago, Jeff S said:

Ctein weighed in on ETTR years ago at TOP.

https://theonlinephotographer.typepad.com/the_online_photographer/2011/10/expose-to-the-right-is-a-bunch-of-bull.html
 

Jeff

Like most who demonstrate the "damage" of the ETTR approach, Ctein uses a high contrast image to argue against ETTR. 

ETTR does not mean OTTR 🙂 (Overexpose to the right), it does not mean that the image must be brighter. 

ETTR is about protecting highlights as well. It is not about pushing the shadows until they are "noise-free." 

Jeff Schewe demonstrates the benefits of ETTR in Un-debunking ETTR.

Jim Kasson on Why ETTR?

DPReview on ETTR Exposed.

 

[Jeff S]

8 minutes ago, SrMi said:

 

Like most who demonstrate the "damage" of the ETTR approach, Ctein uses a high contrast image to argue against ETTR. 

ETTR does not mean OTTR 🙂 (Overexpose to the right), it does not mean that the image must be brighter. 

ETTR is about protecting highlights as well. It is not about pushing the shadows until they are "noise-free." 

Jeff Schewe demonstrates the benefits of ETTR in Un-debunking ETTR.

Jim Kasson on Why ETTR?

DPReview on ETTR Exposed.

 

If you read all the comments below the linked TOP post, Ctein responds directly to Schewe’s article and other critiques.

Don’t shoot the messenger.  Ctein is an opinionated guy.  
 

Jeff

[Jeff S]

Ctein has also written about ‘dynamic range’, and how that term is almost never used correctly, and is really exposure range (latitude).

Be sure to read both Part I and II...

https://theonlinephotographer.typepad.com/the_online_photographer/2012/09/dynamic-range-is-not-exposure-range-part-ii.html
 

Too bad he’s no longer a regular TOP contributor.
 

Jeff

 

[SrMi]

23 minutes ago, Jeff S said:

If you read all the comments below the linked TOP post, Ctein responds directly to Schewe’s article and other critiques.

Don’t shoot the messenger.  Ctein is an opinionated guy.  
 

Jeff

Thanks for the pointer. The comments list on TOP is longer than this thread :).
I do not get Ctein's response, as ETTR explicitly pays attention to the highlights, while automatic exposure does not. 
From Wiki:
In digital photography, exposing to the right (ETTR) is the technique of adjusting the exposure of an image as high as possible at base ISO (without causing unwanted saturation) to collect the maximum amount of light and thus get the optimum performance out of the digital image sensor.
By that definition, Ctein applied ETTR to his image using negative exposure compensation to avoid unwanted saturation (aka highlight clipping). The Wiki ETTR definition does not imply that one has to push the exposure.

Is ETTR worth the trouble? Jim Kasson has approximated the answer: "From a noise point of view, you can turn your full-frame camera into a micro four-thirds camera by underexposing by two stops."

Nothing wrong with MFT cameras, though :).

[SrMi]

54 minutes ago, Jeff S said:

Ctein has also written about ‘dynamic range’, and how that term is almost never used correctly, and is really exposure range (latitude).

Be sure to read both Part I and II...

https://theonlinephotographer.typepad.com/the_online_photographer/2012/09/dynamic-range-is-not-exposure-range-part-ii.html
 

Too bad he’s no longer a regular TOP contributor.
 

Jeff

 

Thanks for the link. Another windmill that did not yield :). 

Edited September 25, 2020 by SrMi

[pgk]

16 minutes ago, SrMi said:

From Wiki:
In digital photography, exposing to the right (ETTR) is the technique of adjusting the exposure of an image as high as possible at base ISO (without causing unwanted saturation) to collect the maximum amount of light and thus get the optimum performance out of the digital image sensor.

Wiki is not always accurate. Obtaining 'optimum performance' via ETTR would only be true if the output from the in-camera sensor interpretation was not 'tuned' or 'adjusted' by the manufacturer to be how they want its output to be. As far as I am aware there is no camera which simply outputs direct from the sensor so the image file is 'processed' to some extent.

[pgk]

4 hours ago, adan said:

You tell me - which lens simultaneously retains more detail and tonal separation in the:

highlights - subtle textures and nicks in the white-painted concrete, and the color saturation of the beige foundation floor.

shadows - a large number of tonal separations: seams in the black wall, little bits of detritus along the base of the black wall, red overlays painted on the black wall, the spots and tones and colors within the shaded beige concrete pavement?

The higher contrast aspheric lens. If it was. to so then manufacturers would not coat their lenses in order for the output to be low contrast and more tonality to be retained. Look at the MTF data which will show what is going on.

[Jeff S]

38 minutes ago, SrMi said:

The comments list on TOP is longer than this thread
... snip

I do not get Ctein's response, as ETTR explicitly pays attention to the highlights, while automatic exposure does not. 
The Wiki ETTR definition does not imply that one has to push the exposure.

 

As is often the case, just as Mike Johnston intends.  He recently hired two long time contributors to assist him in moderating the deluge of comments.

Ctein specifically defines ETTR (expose to the right) as pushing the exposure to the right of the histogram (first paragraph in the link).

Jeff

[SrMi]

Another relevant article:

How to Use the Full Photographical Dynamic Range of Your Camera

Excerpt:

The thing with a camera's dynamic range is that those exciting numbers become valid only when the signal reaches the maximum; that is, the exposure is the hottest possible for the given camera at a given ISO. If your whitest white is not exposed to the maximum, the top portion of the dynamic range is not being used. This immediately means that in order to use the dynamic range of the camera to its fullest and simultaneously to have as little shadow noise as possible, the whitest whites where you want to keep some texture details need to be exposed so that they are just below clipping.

 

[pgk]

7 hours ago, SrMi said:

Another relevant article:

Starts with a fallacy. You could not use a higher than base ISO if your aim is to utilise the maximum dynamic range. If ETTR works for you that is fine, but trying to convince anyone that its a solution to utilising maximum dynamic range is far too simplistic.

[Advertisement]

Hallo,

Look here: Interesting Leica gear

[pop]

vor 47 Minuten schrieb pgk:

Starts with a fallacy. You could not use a higher than base ISO if your aim is to utilise the maximum dynamic range. If ETTR works for you that is fine, but trying to convince anyone that its a solution to utilising maximum dynamic range is far too simplistic.

The article starts with the sentence "Suppose you've read somewhere that the dynamic range of your camera at a certain ISO setting is 11 stops". Base ISO certainly qualifies as "a certain ISO setting".

An overexposed part of an image will become a featureless white blob. An underexposed part of an image will be rendered with an increasing amount of noise until it's rendered as a featureless black blob. Hence, the usable dynamic range of a sensor at any ISO setting will be attained by exposing as much as possible without overexposing parts that must not be clipped. Cooling your sensor might help.

Of course, there's always HDR, but this is quite another topic.

[pgk]

1 hour ago, pop said:

The article starts with the sentence "Suppose you've read somewhere that the dynamic range of your camera at a certain ISO setting is 11 stops". Base ISO certainly qualifies as "a certain ISO setting".

Boosting ISO and ETTR are opposing constraints. Boosting ISO is, in effect, underexposing. ETTR is, in effect, overexposing. Seems to me that this is a rather contradictory technique .....