there are some counterarguments including the matter of color shift that can occur when overexposing. I certainly noticed some of that when I owned a 5D2 and I attempted to push exposure a half or full EV
I've read and understand Michael's argument for exposing to the right. I can't disagree in general but there are some counterarguments including the matter of color shift that can occur when overexposing. I certainly noticed some of that when I owned a 5D2 and I attempted to push exposure a half or full EV.By over exposing I assume you mean exposing more than your camera suggests, which is providing you information as though film were in the camera?
ETTR increases contrast and saturation.
I certainly noticed some of that when I owned a 5D2 and I attempted to push exposure a half or full EV.
Color shifts when you alter exposure by 1/2 stop? I would very much like to see an example of this.
I think that ETTR is one of those techniques that has a valid theoretical basis but no benefit in the real world. This of course excludes pixel-peeping!
I think that ETTR is one of those techniques that has a valid theoretical basis but no benefit in the real world. This of course excludes pixel-peeping!
I think that ETTR is one of those techniques that has a valid theoretical basis but no benefit in the real world. This of course excludes pixel-peeping!Curious what you are basing that on. Did you actually try it? Because if you do and do it right, you will see a clear difference with cleaner shadows and better detail in your shadows in prints, especially larger ones where noise becomes much more obvious.
I think that ETTR is one of those techniques that has a valid theoretical basis but no benefit in the real world. This of course excludes pixel-peeping!Back in the Dark Ages (i.e., Before Digital) I recall that many photographers said similar things about the Zone System, mainly because you had to do a little work testing your equipment and materials in order to be able to use it effectively. IMHO, the same applies to ETTR.
Thought I would add one personal IMHO that I haven't noticed in any of the EttR discussions so far.
In my mind EttR isn't just about moving the data to the right to reduce noise and increase quality. Personally I think it's the correct way to determine almost all exposures regardless of the contrast range of the scene vs. the dynamic range of the sensor.
Certainly if the scenes contrast range is less the the sensors DR, you maximize the benefits of EttR.
However, if you base your exposure on the highlights just short of clipping, you get the maximum information in your shadows, even if they clip. Often by opening up the exposure a little, you get enough shadow information to not require HDR, especially with some of today current cameras (with my IQ180, I really haven't found a time where I had to use HDR). If you let the camera choose a setting that will render a pleasing Jpeg (the goal of the camera makers) you in may block up your shadows too far.
Exposures based on the highlight clipping point is just the logical way of determine a digital exposure because the goal in digital capture is maximizing the data, not creating a pleasing on camera image.
if you base your exposure on the highlights just short of clipping, you get the maximum information in your shadows, even if they clip. Often by opening up the exposure a little, you get enough shadow information to not require HDR, especially with some of today current cameras (with my IQ180, I really haven't found a time where I had to use HDR). If you let the camera choose a setting that will render a pleasing Jpeg (the goal of the camera makers) you in may block up your shadows too far.
Exposures based on the highlight clipping point is just the logical way of determine a digital exposure because the goal in digital capture is maximizing the data, not creating a pleasing on camera image.
3) At least on some sensors, nonlinearity close to saturation may cause color shift according to some very knowledgeable people.
The final question may be how white balance messes up the exposure of each channel?
I'm still looking for a demonstration of that colour shift.
Some relevant info in a related thread:
http://www.luminous-landscape.com/forum/index.php?topic=56906.msg460680#msg460680
I see two issues:
(1) Does ETTR make any difference in the quality of the raw data;
(2) if one practices ETTR (or ETTL on cameras such as the D7000, K5, or any CCD sensor'd camera), does your raw converter of choice handle the data properly.
The answer to (1) depends on the camera, as discussed in detail in the post linked above. As for (2), I think many raw converters don't do a good job of exposure correction; Adobe has the issues with hue twist mentioned by Sandy, and Guillermo has suggested in a parallel thread (http://www.luminous-landscape.com/forum/index.php?topic=56951.0) that DPP doesn't have a true exposure control. Converters that definitely treat exposure compensation correctly are RPP and RawTherapee.
Hue twists are frequently attributed to ACR, but Eric Chan (http://www.luminous-landscape.com/forum/index.php?topic=56309.msg458159#msg458159) has stated in another thread that if one uses ETTR and normalization of the image is necessary, one should use the exposure slider first to get the tonal values in the proper range before the 3D color table is applied.I know this is the "proper" way to do it, but since the exposure slider isn't a true full linear control, I've gotten better results by setting my black point with the black slider first, then setting my white/highlight point with the exposure slider, and finally tweaking the image brightness with the brightness slider. If I go in that order I almost always get a very pleasing starting point that requires very little tweaking with the recovery and fill light sliders. Sometimes when I set my exposure slider first I have trouble getting my shadows to look right, and my highlights don't look as good after setting the black slider.
Regards,
Bill
Hue twists are frequently attributed to ACR, but Eric Chan (http://www.luminous-landscape.com/forum/index.php?topic=56309.msg458159#msg458159) has stated in another thread that if one uses ETTR and normalization of the image is necessary, one should use the exposure slider first to get the tonal values in the proper range before the 3D color table is applied.
2) The histogram is conservative, as it is not calculated from raw data. So it's possible to push exposure a bit longer, but we don't actually now how far.
I know this is the "proper" way to do it, but since the exposure slider isn't a true full linear control, I've gotten better results by setting my black point with the black slider first, then setting my white/highlight point with the exposure slider, and finally tweaking the image brightness with the brightness slider. If I go in that order I almost always get a very pleasing starting point that requires very little tweaking with the recovery and fill light sliders. Sometimes when I set my exposure slider first I have trouble getting my shadows to look right, and my highlights don't look as good after setting the black slider.
Am I crazy?
No, the Exposure stage in ACR/LR happens before the 3D LookTable, not after. Same with Blacks. So if you use these to adjust the image (to compensate for ETTR) you should get the same result. For example, compare two cases. For Case A, I take an image at a given exposure. For Case B, I take the same image, but 1 stop underexposed (e.g., halve the shutter time), then set Exposure to +1 in ACR to make up for it. Aside from shadow clipping due to the default Blacks = 5, the tones and colors in the results should be the same in both cases, even if using a lightness-dependent profile, e.g., Adobe Standard or one of the Camera Matching profiles.
We actually discourage (but do not prevent) the use of 3D tables in the HueSatMap* tags for the reasons you mention. The preferred use of the HueSatMap* tags is a 2.5D table, indexed by hue & sat only, to perform non-linear color corrections.
I am also puzzled by the discussions of the non-linearity of ACR's Exposure. ACR's Exposure in the positive direction is simply a straight multiply with a hard clip, just like digital camera exposure. In the minus direction, the only difference is that ACR tries to keep clipped whites white (i.e., we did not feel it was photographically useful to let speculars turn into gray blobs, though we still let users accomplish that with the point curve)...
I am also puzzled by the discussions of the non-linearity of ACR's Exposure. ACR's Exposure in the positive direction is simply a straight multiply with a hard clip, just like digital camera exposure.
That is why I don't understand why it appears to be a non-linearity near zero. These are the curves that model a change in Exposure into ACR with respect to the 0.0 setting (I already showed them to you in Dpreview). I just compared (input, output) pairs, I didn't any gamma calculation. And the ouput profile used was Adobe RGB, which is supposed to be a standard gamma profile (strictly applying f^(1/g) along the whole range).
So, what are those non-linearities near 0?
(http://www.guillermoluijk.com/article/acrps/exposicion_acr.gif)
It appears to me that ACR is using a non-standard gamma 2.2 for Adobe RGB, something like sRGB's gamma.
Regards
Adobe RGB has no linear segment near luminances of zero, but the spec allows a slope limit of 1/32 for 8 bit digital images in the range of 1-14. See Annex C of the Adobe RGB spec (http://www.adobe.com/digitalimag/pdfs/AdobeRGB1998.pdf). Eric Chan has confirmed that this linear segment is used in all Adobe applications.
We actually discourage (but do not prevent) the use of 3D tables in the HueSatMap* tags for the reasons you mention. The preferred use of the HueSatMap* tags is a 2.5D table, indexed by hue & sat only, to perform non-linear color corrections. This can be important with cameras with suboptimal spectral sensitivities or when photographing under difficult illuminants. Lightness-dependent corrections are not necessary because the input data is in a linear, scene-referred space. The Chart Wizard feature of the DNG Profile Editor writes into the HueSatMap* tags and writes only 2.5D tables.
Tone controls such as black point (especially), brightness, and contrast work on individual RGB channels so will cause hue shifts. I would only expect the exposure slider to behave as a true gain in the resulting output with black point set to zero, brightness and contrast neutral, and the tone curve set to linear.
I put the blame for the shifts on a combination of profile and the various controls acting in a non-linear way. But Eric seemed to be saying that ACR will be linear with those controls assuming no saturation ???
… So if you assume clipped white values are infinite, in order to give a pleasing appearance you basically have to use a curve or non-linear exposure adjustment. This is essentially what Adobe does with ACR and Lightroom exposure. The exposure slider isn't really a true "exposure" adjustment equally applied to all values at least when using negative values.
Perhaps a raw file pre-processor that multiplies raw sensel values with some value (assuming that clipping is not an issue) before importing it into a standard, nicely organized raw developer (I guess dcraw could easily be hacked into something like that).
-h
that doesn't seem to be what Eric is saying. A few years ago, I said that correcting for ETTR will cause hue shifts, and specifically that such hue shifts occur in LR/ACR. Eric seemed to be saying above that hue shifts won't occur if you adjust the "basic" exposure controls.
With below image*, and a measuring point set to the non-clipped part of the sky, minus-ACR-Exposure clearly goes along with a reduction of color saturation.
Hence, there are different scaling factors applied per RGB channel…
the one and only you need for ETTR exposure correction
With the measuring point placed as shown above, there is no indication for channel clipping in ACR.
If you open your RAW file in Rawnalyze you'll see the actual RAW histograms.
I don't think that DPP does any recovery,
however, Rawnalize does not seem to be available online anymore since the author passed away.
I don't think that DPP does any recovery,You might be able to download it from this site:
however, Rawnalize does not seem to be available online anymore since the author passed away.
Peter
--
Thanks Bill. That must be the reason for the non-linearity in the calculated curves (which of course doesn't mean any non-linearity in the way Exposure works in ACR).
So if the gamma slope in Adobe RGB's implementation is limited (i.e. Adobe RGB's gamma lifts the shadows less than a pure gamma), and we open an image encoded with a pure 2.2 gamma assigning it to Adobe RGB, we will be seeing the deep shadows a bit brighter than expected. My concern about this is because I use pure 2.2 gamma images and then assign them in Adobe RGB, but I find OK to have some shadow lifting.
Regards
So in the minus direction it is really a jump function ?
Clipped white is fixed at RGB 255 while a multiplier < 1 is applied all other values, without any smoothing ?
Hmm
Peter
--
Peter, no it is not a jump function. It is a smooth function that maps pure white to the rest of the curve below it. (If it was a step/jump function you would see horrible artifacts in your highlights and upper midtones!)
Hi Sandy, thanks for running this test and reporting the results. Emil is right, in that black subtraction (in ACR's case, via Blacks slider) is typically done per-channel so it is expected to result in color changes (including hue). Setting Blacks to 0 will turn off that behavior. Generally speaking, you'll find that a given step/error size will manifest much more strongly in blue than the other colors, because the internal working space of ACR uses the ProPhoto RGB primaries, and the blue primary is pretty far out there ...
Rawnalyze does not seem to provide any indication for Raw clipping in this mid part of the sky.
Again, for this example all other tonal controls were zeroed in ACR, which is supposed to provide a "linear rendition". Point curve linear as well. ProPhoto RGB selected for output, which to my knowledge is what the RGB readings refer to. Further, the baseline matrix profile was used to avoid any related hue twists. Also, Rawnalyze does not seem to provide any indication for Raw clipping in this mid part of the sky.
Zero'd tone controls and 'linear' tone curve don't produce a linear output in ACR; gamma is still applied according to the chosen color space. So lower tones will be amplified more than higher tones, simply because that's the way a gamma curve works. And there will be hue/saturation shifts as a result. Only the sort of test Sandy is doing, where one tries to achieve the same output density with shots taken at different ISO, should be independent of the applied gamma, which takes place after the exposure slider is applied in the processing pipeline.
Zero'd tone controls and 'linear' tone curve don't produce a linear output in ACR; gamma is still applied according to the chosen color space. So lower tones will be amplified more than higher tones, simply because that's the way a gamma curve works. And there will be hue/saturation shifts as a result.
Zero'd tone controls and 'linear' tone curve don't produce a linear output in ACR; gamma is still applied according to the chosen color space. So lower tones will be amplified more than higher tones, simply because that's the way a gamma curve works. And there will be hue/saturation shifts as a result. Only the sort of test Sandy is doing, where one tries to achieve the same output density with shots taken at different ISO, should be independent of the applied gamma, which takes place after the exposure slider is applied in the processing pipeline.
Sandy, that does sound unusual. You mentioned earlier that when Blacks, Brightness, and Contrast were all 0 (i.e., no-ops) that the hue shift wasn't there. Now it seems that with Blacks 0, but the other two at defaults, there is a hue shift (due to a blue difference). I don't have an explanation for that. But I'd be happy to study your test images (you have my email).
Eric
Perhaps ACR's highlight strategy extended beyond the strictly RAW-clipped area. In this thread (http://www.luminous-landscape.com/forum/index.php?topic=56469.msg460340#msg460340) you can see an example.
Put another way, we've found that users prefer that their raw files are tone- and color-rendered by default (i.e., have a film-like S-curve with shoulder and toe applied), rather than a straight scene-referred rendering which usually looks flat.