Exposure setting display compared to LR


I have found two items of interest which might be bugs or features:
Take a look at these two RAW files:
(source: http://www.dpreview.com/previews/canon-eos-7d-mark-ii/8)
Issue 1: differing relative brightness compared to LR
If I open them in FRV the Nikon D7100 picture displays quite a bit darken than the Canon file. Visually they look the same if I raise exposure in FRV bei +0,67 for the Nikon file.
If I open them in Lightroom 5.6 they show up with similar brightness from the start.
I wonder if that's somethign that is caused by Nikon's optimistic manipulation of ISO values.
Issue 2: differing exposure compensation display compared to LR
When I open the above files in FRV and set the exposure compensation to 0,00 the pictures seem relatively dark and when I open them up in LR the exposure is always shown as the one from FRV -1, e.g. 0,33 in FRV becomes -0,67 in LR.

This is not bug, but feature of current FRV 0.9.3

Adobe products (Lightroom, CameraRaw) applies 'hidden' exposure compensation to RAW files. See more details here: http://photographylife.com/adobes-silent-exposure-compensation.

This compensation is different for different cameras and (worse) for same camera and different ISO values.

In FastRawViewer 0.9.3 user should set 'Adobe/FRV exposure difference' in XMP settings page ( http://www.fastrawviewer.com/usermanual/program-settings/xmp ). Unfortutately, for many cameras this offset need to be changed for different ISO.

In FastRawViewer 0.9.4 (public beta expected in week or so) we solved this problem for most cameras, so automatic exposure difference between FRV and Adobe programs is calculated and used.



Alex Tutubalin/FastRawViewer team

To add a little to what Alex already said and address your first observation from a somewhat different angle:

Looking at the nef with RawDigger, the white patch of CC24 is exposed to the level of 5282 out of maximum 16383, that is log2(16383/5282), that is ~1.6EV below maximum. With cr2, the level on the same patch is 6345 out of the maximum (as defined in Makernotes tag 0x02d0, the one Adobe are using to calculate the brightness compensation) of 12279; which is ~0.95 stops below maximum. The difference is log2(16383/5282) - log2(12279/6345) = 0.68 EV, as you observed.

You can also see that the shutter speed is different between the shots, 1/50 vs. 1/40. This alone is 1/3 EV difference. Another factor is that ISO 100 on 7DMk2 is achieved through the shift of the exposure meter by close to half a stop (the base ISO is close to 130 for this camera). Effectively, from the point of view of the regular in-camera metering, the shot is 0.4 EV overexposed, which makes it brighter.

PS. BTW, we checked and it seems that the Canon file you linked to is from 7D original, Mark I, not Mark II.

Thank you very much for you quick and insightful responses. I was not aware of the hidden manipulations in Adobe software. Good to know.
Yes I linked a file from 7DI. But I took both files from the page listed above which has raw files for 4 cameras. I couldnt propose to use a 7D2 file here since that currently cant be opened in lightroom or ACR.
If you will display the adobe designed adjustments I suggest to display both values, so the user sees what is "actual" and what will show up in an adobe product. Happily there still are other raw converter on the market and we shouldn't all follow a poor decision (which I think it is, to apply exposure compensations first and then tell the user there is none).
FRV is a great tool by the way.

FastRawViewer 0.9.4-441, published yesterday supports automatic 'Adobe-FRV' offset for 400+ recent cameras.

If you set tone (contrast) curve to Adobe Linear, the rendering in FRV should be very similar to Adobe CameraRaw/Lightroom


Alex Tutubalin/FastRawViewer team

Add new comment