Leica and histograms

[Alan Goldhammer]

Plus another issue: once you have an exact linear response, how do you (we) propose to benefit from ETTR? If the response is linear, then shifting the data makes no difference. The reason that ETTR works is because it is a multiplication, not a transition.

I'm not so sure this is correct.  S/N ratio is a pronounced factor and ETTR addresses this by allowing more light in thus increasing the signal.  Noise level is presumed to be constant.

Alan

[32BT]

I'm not so sure this is correct.  S/N ratio is a pronounced factor and ETTR addresses this by allowing more light in thus increasing the signal.  Noise level is presumed to be constant.

Alan

I think we mean the same thing, but I believe it is important to note that allowing more light in is the *only* option for producing ETTR, which equates to some form of auto exposure adjustment. How many photographic situations allow an automatic increase of exposure time that will produce data that needs re-adjustment afterwards? Are there enough such situations that merit a "complete" overhaul of the exposure paradigms familiar to photographers? (Also meaning situations where you can actually capture the dynamic range).

In addition, if the limits of filling up the capturing bins is not entirely exact or has some non-linear transition point, and thus the definition of maximizing the RAW file is not strictly defined, and so on, and so forth, is it therefore a truly useful and important addition to implement for manufacturers?

Obviously, we would very much like to see clipped or questionable data in our RAW file. Mostly because we want to be able to readjust exposure to alleviate the problem if necessary, but doesn't that equate to the exact opposite: we will sacrifice S/N ratio for non-clipping RAW data?

[NikoJorj]

If the response is linear, then shifting the data makes no difference.

The response is linear, but with more and more noise towards the bottom, ultimately giving non-linearity when S/N is really too low... See eg http://www.luminous-landscape.com/forum/index.php?topic=46071.0 (in the last answers).
So the benefit comes from higher S/N, not from linearity itself as with film.

- do you want Fstops? (= gamma 2.0)
- do you want gamma 2.2 / sRGB / Lab?
- or user selectable?

Fstops definitely - it doesn't makes much sense imho to try to render the data too accurately at this stage. I don't want to check tonalities at the back of my camera, only make sure that the data captured is of good quality, so I don't mind if the preview is very crude because on a LCD in the sun it is anyway.

While i realize this is somewhat technical, it may also be important because of the logarithmic RAW histogram we propose. For Canon RAW files for example, the gamma curve is not applied to the exact zero origin, but rather to a specific lower threshold level. This may not be true for other manufacturers, but obviously has significant impact on what we will be looking at…

Good point, and that may be another reason to allow user-selectable thresholds for the shadow/highlights limit.
makes me think, these kinds of things are also discussed here http://www.luminous-landscape.com/forum/index.php?topic=57654.0

I think we mean the same thing

I do too!

How many photographic situations allow an automatic increase of exposure time that will produce data that needs re-adjustment afterwards?

Of course, there are many situations where ETTR means under-exposure relatively to the gray card, because of those fluffy clouds that have to be preserved eg. And there are also the situations where one has to choose between the Charybda of HL clipping and the Scylla of shadow noise.

In addition, if the limits of filling up the capturing bins is not entirely exact or has some non-linear transition point, and thus the definition of maximizing the RAW file is not strictly defined, and so on, and so forth, is it therefore a truly useful and important addition to implement for manufacturers?

For me, the benefice of an automated ETTR exposure would be first on high contrast scenes to avoid unwanted clipping highlight (I generally tend to jump in the arms of Scylla), and then on optimizing S/N.

[Guillermo Luijk]

All transistors, including the light sensitive ones, are non-linear. It may take a lot of very meticulous and careful electronic balancing to get the data from the linear part of the curve. The RAW data we usually see in files, is obviously not related to the response curve of the sensor considered as transistors. Probably far from it

I measured sensor response and it was highly linear over a wide range of stops and up to saturation:

I've suggested it in another thread but I think it's worth mentioning again: I don't think the histogram (linear, logarithmic, or otherwise) is very useful for determining ETTR or clipping, because it doesn't show you WHERE in your image the clipping happened. For this purpose, I think an over-the-image clipping visualization (I.e., blinking warning or similar) is much more useful because it can show exactly where the clipping occurred. And this can be meaningfully done per raw channel. (Live preview images are subsampled anyways so the fact that Bayer planes are subsampled is not an issue.).

I don't agree 100% here Eric, because when some clipping are irrelevant (reflections, lamps,...) those little areas are some stops above any other scene's area. So those allowed-to-clip zones are very easily identified on a EV histogram, just look for a big empty gap on it. Although I agree visual clipping is far better.

I proposed here: EVF/Live View for the advanced user (brainstorming) for a hypothetical EVF the following real time information being displayed:

.

[ErikKaffehr]

Hi,

I share your views!

I would say that ETTR is simply making best use of dynamic range. It's like exposing for shadows and developing for highlight in the analogue time. With negative film you always exposed for the shadows.

Best regards
Erik

The response is linear, but with more and more noise towards the bottom, ultimately giving non-linearity when S/N is really too low... See eg http://www.luminous-landscape.com/forum/index.php?topic=46071.0 (in the last answers).
So the benefit comes from higher S/N, not from linearity itself as with film.

Fstops definitely - it doesn't makes much sense imho to try to render the data too accurately at this stage. I don't want to check tonalities at the back of my camera, only make sure that the data captured is of good quality, so I don't mind if the preview is very crude because on a LCD in the sun it is anyway.

Good point, and that may be another reason to allow user-selectable thresholds for the shadow/highlights limit.
makes me think, these kinds of things are also discussed here http://www.luminous-landscape.com/forum/index.php?topic=57654.0
I do too!
Of course, there are many situations where ETTR means under-exposure relatively to the gray card, because of those fluffy clouds that have to be preserved eg. And there are also the situations where one has to choose between the Charybda of HL clipping and the Scylla of shadow noise.
For me, the benefice of an automated ETTR exposure would be first on high contrast scenes to avoid unwanted clipping highlight (I generally tend to jump in the arms of Scylla), and then on optimizing S/N.