RAW Dialog - dynamic range

[MikeG]

I'm having trouble getting the RAW Dialog/Grag/Dynamic Range - Full Range to work as it does in Levels and Color.

So, I pressed F1 and read , in the section RAW Dialog - Gray Tab, the following:



I hope that this is not a question that I'm going to wish I hadn't asked, but where's the button? Is it simply the button to access the drop down selection dialog?

If it is, I use this button to select Full Range, and even with EV set at 0.0, no Hilite Recovery, I seem to regularly get one or other end of the brightness range not fully expanded - I check by using Level and Color.

Any one else having similar problems?

Mike.

[den]

MikeG...

The 'Full Range' of the Levels&Color transform may not be the same as 'Full Range' of the RAW Dialog.

I seem to remember that the Levels&Color 'Full Range' will actually clip 0.1% for both the white and black points. Perhaps, JS/Kiril will advise.

Perhaps a better confirmation of the RAW Dialog 'Full Range' is using the Histogram Tool with the Expansion set to High. If the histogram extends from 0% tone (black point) to 100% tone (white point), then the maximum number of tones has been preserved which afterall is the object of starting with a RAW file.... the Levels&Color transform may only show something like 1% to 99% tone range for a Histogram Tool - High Expansion that shows a 0% to 100% tone range.

[den]

In my previous post, histogram/tone range comparisons where considered only in the HSV color space model.

If only HSV-V is considered to establish black/white points [tone range], it is possible that one or more of the R, G, or B channels can be clipped.

To preserve maximum image data from a RAW file, I will use the RAW Dialog - Gray tab with 'Medium' Dyn. Range... then take the conversion tiff image and set White/Black points with a RGB Brightness Curve where the histogram expansion is set to High. This ensures [initially] no RGB channel clipping unless performed to preferences in downstream post-processing workflow which is usually the case to get good tone mapped contrasts for the visual elements of the image and color balance.

Obviously, this process is individualized for a specific image and not for bulk image processing.

[ksinkel]

The dynamic range in Raw dialog is a bit more complex than the one in Level and Color because it is integrated with additional functions not present in the Levels and Color. However, it works in the same way -- detecting the .01% extremes in image brightness and adjusting the dynamic range to place these at the full black and white points. Make sure that you magnify the histogram -- the low magnification histogram may not show very low frequency brightness values.

Of course you may always decide that the extreme shadow or highlight areas are not visually significant and adjust the dynamic range beyond the black and white points to get more contrast for significant portions of the image.

Kiril

[MikeG]

Thanks, Den and Kiril,

The fog (over my brain) is starting to lift - slowly.

I'll experiment with setting the RAW Dialog Dynamic Expansion to medium, and try to get a feel for that approach.

In the meantime, I'm still not clear if the behaviour shown in the screen shots below is 'correct'.

The histogram in the Gray Tab shot is at the highest (stage 3) magnification and appears to show the image short of the black point.
Full range is selected. No hilite correction. EV at 0.0

The Levels and Color screen shot still show the image shy of the black spot. Shouldn't the image be full range at this point?

Mike.

[tomczak]

I confirm Mike's concerns - I think that in some cases the Full Rnage in RAW dialog does not work as intended.

[ksinkel]

Mike,

Are you using a camera profile? A camera profile defines the black point and can create the situation you show. Here are screen shots of the gray tab for the same image with and without a profile. Notice that the black point in the first case is at about 10%. It stays at this point even if you move the DR black point slider to the eft of the zero point. In the second case, with no profile, the black point is at 0%

With profile:

rawDR1a.jpg (12.72 KiB) Viewed 10253 times

Without profile

rawDR5a.jpg (9.54 KiB) Viewed 10257 times

Kiril

[David_B]

Hello

I'm glad I'm not the only one who finds the controls a bit baffling.

I have observed a problem that I think shows that the "camera profile" and the "gamma setting" are applied in the wrong order.

For RAW, the camera trc is a mapping for pixel values from the camera to a nominal linear workspace, and should be applied to the camera values directly.

The "gamma setting" is applying a curve for the output device. Normally, increasing "gamma setting" increases the midrange brightness.

One of my camera profiles has a negative trc with gamma of -1.4, to convert from a colour negative film, with a gamma of about 0.7, to a positive image.

If I am using that profile, increasing the "gamma setting" makes the image darker. This shows that it is applied before the conversion to a positive image.

As a result, I have to use "gamma setting" of 1, and apply the correct gamma later.

Please let me know if I've got the wrong end of the stick.

[MikeG]

Kiril,

I'm NOT using a camera profile - see screen grab below.
And the histogram does follow when I drag below the zero point on the dynamic range slider - see other scrren grab.

Mike.

[ksinkel]

David,
If you are using a profile which applies a gamma correction, you should set the gamma in the raw dialog 1.0. (A gamma of 1.0 makes no change.) Otherwise you will be applying gamma twice.

However, I would caution you that using a profile that was calculated in a different workflow from the one where it is applied is not valid analytically. It may work in a particular case, but can be a source of baffling problems in others.


Mike,

Camera profile was ony a guess. Let's go back to your original question:

"In the meantime, I'm still not clear if the behaviour shown in the screen shots below is 'correct'. ... The Levels and Color screen shot still show the image shy of the black spot. Shouldn't the image be full range at this point?"

The dynamic range slider black point and the histogram are calculated at different points in the workflow. So there can be discrepancies because later processing changes the histogram. (Hence my guess about camera profile.) Also, the dynamic range black point is based on detecting very low frequency pixels that may not show up on the histogram. As long as you are getting good control there is not really a problem.

Kiril