Expose to the right (ETTR) or underexpose?

[Lindolfi]

Advertisement (gone after registration)

I tested already.

http://www.l-camera-forum.com/leica-forum/2214324-post168.html

 

Different test, no noise data. The test I propose you to do does not only show the concept, it shows the advantage. Real world? There is no such thing. There is your world and you decide what to do in it.

[Advertisement]

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

[douglasf13]

I should indicate to Jaybob and Douglasf13 that both have good arguments, but they are used in a different context and can therefore never lead to agreement: Jaybob claims that most photographers don't work with underexposure and therefore it must be wrong (context of convention) while Douglasf13 points at a principle that may be worth looking into (context of discovery).

 

 

I've mentioned several times that this discussion should be in two parts: IQ vs. practicality.

 

In terms of IQ, as many posts in this thread have indicated, there is no reason to boost the ISO of the M9, and there may actually be a slight decrease in performance by boosting ISO in camera.

 

When I said the following to Jaybob, I had no intent to be condescending: "You managed this far because camera makers decided to call camera gain "ISO," in order for film shooters making the transition to digital easily understandable, especially since our cameras are still geared towards jpeg output."

 

The above point is something the large majority of even professional photographers misunderstand. Our digital cameras are not like having multiple speeds of film available in one camera. Our digital cameras are still like film in that they have a single, base ISO, and everything after that is just pushing your exposure, just like you push film, whether it be with the camera's internal analog/digital gain, called "ISO," or whether it be with the raw converter after the fact.

 

In terms of practicality, using ETTR with the M9 above base ISO can actually add MORE work to post processing, because you have to pull your exposure slider down in the converter most of the time.

 

Here's a practical scenario:

 

Imaging that you're shooting a scene with the M9 at ISO 800, f2, 1/60, and that gives you a histogram that is ETTR as much as possible while still retaining highlights. However, for your next shot, if you keep your exposure the same (shutter and aperture,) and lower your ISO to 400, your midtones are exposed "correctly," and your histogram is centered without being exposed to the right.

 

Now, if I bring both of these files into the raw converter, the ISO 800 file will need to have the exposure slider pulled back one stop to match the "properly" exposed ISO 400 file, and the ISO 400 file will not need to be touched. In terms of IQ, they will both be very similar, with the ISO 400 file likely being incrementally better. This is why ETTR is unnecessary, unless we're talking about base ISO, where getting more light to the sensor will improve SNR, although, as Sandy mentioned, it may have a small cost in color hue if the midtones are exposed too far right.

 

So, in terms of practicality, why did I use ETTR in the first place? All that it did was provide a file that required an extra adjustment in the raw converter, and its IQ wasn't any better.

 

 

p.s. I should mention that the above assumes we're talking about shooting raw. JPEG, which almost all cameras are surprisingly still geared towards, is a very different matter where WB, ISO, etc., will behave differently.

[sblitz]

there is however a point where you can't shoot practically slow enough to ETTR in a given situation at base iso. at which point you have to choose to boost the iso (gain?) as part of the equation, much like using sunny 16 when it is not sunny, at some point you can't open the lens further so you have to slow the shutter speed. taking in this practicality, such as shooting a sports event or president or anything that moves, one does need to adjust the iso so the speed matches the object being shot -- and then set the appropriate shutter etc to account for ettr through the histogram. in other words, everyone here is pretty much saying the same thing applied to different circumstance.

[k-hawinkler]

there is however a point where you can't shoot practically slow enough to ETTR in a given situation at base iso. at which point you have to choose to boost the iso (gain?) as part of the equation, much like using sunny 16 when it is not sunny, at some point you can't open the lens further so you have to slow the shutter speed. taking in this practicality, such as shooting a sports event or president or anything that moves, one does need to adjust the iso so the speed matches the object being shot -- and then set the appropriate shutter etc to account for ettr through the histogram. in other words, everyone here is pretty much saying the same thing applied to different circumstance.

 

 

Actually no. As Bert has shown http://www.l-camera-forum.com/leica-forum/2215544-post178.html at least for the first 3 EV underexposure, once you have chosen aperture and exposure then ISO=160 actually produces less noise than ISO=1250, particularly when using uncompressed DNG, assuming Bert's measurements are correct.

Edited October 25, 2012 by k-hawinkler

[Jaybob]

It may be usefull to repeat the experiment yourself Jaybob. It is easy. Just keep all conditions the same, including settings of the RAW converter, with the exception of the Exposure slide. Open in photoshop en look at the histogram of a dark part of the image. Observe how the noise is higher in the high ISO photograph and see whether it works. Draw your own conclusions.

 

 

I should indicate to Jaybob and Douglasf13 that both have good arguments, but they are used in a different context and can therefore never lead to agreement: Jaybob claims that most photographers don't work with underexposure and therefore it must be wrong (context of convention) while Douglasf13 points at a principle that may be worth looking into (context of discovery).

 

 

I didn't say it was wrong, technically it appears to work on uncompressed raw files. It's just difficult to use in practice without significant measurable difference or benefit.

 

For what it's worth, your test is a direct result of my test...cut me some slack.

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!

Edited October 25, 2012 by Jaybob

[Lindolfi]

More agreement in histograms with the M9 at 160 and 1250 ISO on a MacBeth card and no combing:

 

Edited October 25, 2012 by Lindolfi

[douglasf13]

Advertisement (gone after registration)

I'm going to take a step back and attempt to explain why this single ISO technique, while probably not practically useful to most photographers, underscores why ETTR is unnecessary with the M9 above base ISO.

 

We've seen that M9 raws don't benefit from boosting ISO in-camera. However, most of us probably wouldn't be interested in shooting pictures that are so dark that you can't even review them properly on the back of the camera. This is totally understandable, and, as Jaybob has mentioned several times, it isn't all that practical. In-camera ISOs allow us to review images properly, make it easier to calculate proper exposure, and give us usable jpegs straight out of camera.

 

The point of showing this technique, in regards to this thread topic of ETTR, is to show that there is no reason to use ETTR above base ISO with the M9 raws, because you're just as well off using a lower ISO and same shutter/aperture (exposure,) with the camera's histogram planted firmly in the middle. This is more efficient, because you aren't left with overly bright raws that you have to pull in the raw converter, and you don't risk accidentally blowing highlights, and you don't risk hue twists.

 

To echo what I said in my earlier post, imagine a shooting situation where you could, say, either ETTR with ISO 800, f2 and 1/60, or you could expose "normally" at ISO 400, f2 and 1/60. Why would one shoot at ISO 800 and pull the exposure back in post, when you could just shoot with a centered histogram at ISO 400 and not need to touch it, and get as good or better results? This is essentially the crux of the ETTR above base ISO discussion.

[sblitz]

Because for the shot you are taking f4 works better. Or what if you need to shoot at 1/500? I think this discussion is very interesting and enjoyable. My only point is that the caveat with staying with base iso is when the settings aren't critical to the shot and the noise is. If settings are critical then at some level of light or lack of it one needs to push the sensor

[douglasf13]

Because for the shot you are taking f4 works better. Or what if you need to shoot at 1/500? I think this discussion is very interesting and enjoyable. My only point is that the caveat with staying with base iso is when the settings aren't critical to the shot and the noise is. If settings are critical then at some level of light or lack of it one needs to push the sensor

 

But that doesn't really address what I'm talking about, which is using ETTR above base ISO. I'd be curious for someone to describe a situation where ETTR above base ISO makes sense with M9 raws.

[dwbell]

Not one to shy away from flogging any horse, even dead ones!

 

Does the in camera amplification actually "apply" the amplification with increased ISO or does it tag the file to receive a boost in the raw convertor? (A poor comparison would be in camera jpeg sharpening)

 

What I'm getting at is if I expose a shot at 1/60 f/4 ISO 800 and there appears to be clipping of highlights, are they actually clipped, gone? Exposing the same shot at ISO200 would "look" two stops under on the LCD but would it contain information in the highlight areas that the ISO 800 shot.

[pelagia]

Here is something I am not sure about:

Our digital cameras are still like film in that they have a single, base ISO, and everything after that is just pushing your exposure, just like you push film, whether it be with the camera's internal analog/digital gain, called "ISO," or whether it be with the raw converter after the fact.

Maybe it has been discussed already, but I hesitate to agree that the higher ISO means the same as pushing the exposure, no difference, either in the camera, or in the post-processing. I suppose there are still two possibilities to consider: First, in post-processing you push the raw file. While in-the-camera, raising the gain may apply to the "pre-raw" data of the sensor (just guessing). Second, higher ISO may also mean the use of several micro-lenses combined into one cell, similarly to larger crystals of silver halide in film (just guessing again).

 

...such as shooting a sports event or president or anything that moves

So, the president moves?

[jaapv]

The first is correct. The M9 applies (some) noise reduction at higher ISO settings prior to writng the raw file specifically to correct noise that cannot be addressed in further postprocessing (Stefan Daniel at the introduction of the M9), so shooting at base ISO and pushing in postprocessing will render inferior results.

The second is not the case. The M9 sensor does not support pixel binning.

 

My take - ETTR is useful to sacrifice unimportant highlights in a controlled manner in order the preserve shadow detail and reduce shadow noise.

 

I am film-conditioned over the years to regard " get it right in the camera" as a basic mantra.

[pgk]

The M9 applies (some) noise reduction at higher ISO settings prior to writng the raw file...

The question is, I suppose, does it do any other 'pre-processing' which modifies the raw file?

[jaapv]

Like any camera, yes it does, although Leica has a tendency to keep it to a minimum.

[Lindolfi]

.....so shooting at base ISO and pushing in postprocessing will render inferior results.

 

Is this conviction based on experimental evidence?

[jaapv]

A logical conclusion from the statement referred to.

[pgk]

A logical conclusion from the statement referred to.

In terms of noise, yes. In terms of highlight blow out???

[jaapv]

Irrelevant highlights can be blown out if need be, imo. That is just a matter of exposing technique and choices. I do not buy into the credo of the digital age that highlights must be protected at all costs.

[Advertisement]

Hallo,

Look here: Leica M

[pgk]

Irrelevant highlights can be blown out if need be, imo. That is just a matter of exposing technique and choices. I do not buy into the credo of the digital age that highlights must be protected at all costs.

Depends on the subject matter entirely. Sometimes highlight detail (not specular) is important and trying to recover it can produce decidedly odd tonality.

[douglasf13]

Here is something I am not sure about:

 

Maybe it has been discussed already, but I hesitate to agree that the higher ISO means the same as pushing the exposure, no difference, either in the camera, or in the post-processing. I suppose there are still two possibilities to consider: First, in post-processing you push the raw file. While in-the-camera, raising the gain may apply to the "pre-raw" data of the sensor (just guessing). Second, higher ISO may also mean the use of several micro-lenses combined into one cell, similarly to larger crystals of silver halide in film (just guessing again).

 

There is a difference in method between pushing in-camera (ISO) and in the raw converter, but both are done after the exposure. Which method provides superior results depends on the camera and raw converter being used. With the M9 and LR4, in particular, it seems that pushing in the converter is at least as good, and possibly a little better.