Linear raw processing and the Tindemans curve tools

[Juan]

Hi,

I was trying to use the Tindemans curve tools actions applied to a
TIFF file processed by Canon Digital Photo Professional (which allows
for a linear raw processing). I expected I was going to get the most
accurate colors with this procedure, followed by the curve tools but
all I get is a very dark and saturated image in Photoshop. I had read
that linear processing was very good at preserving colors and the
curve tools preserve colors so it made sense to use them together..

Am I doing something wrong?


Thanks,
Juan Dent

[Steve Sprengel]

Looking at Tindemans Color Reproduction essay, I believe his "linear" is not the same as DPP's "linear."

http://www.21stcenturyshoebox.com/essays/color_reproduction.html#_Toc116325714

 

DPP's "linear" means proportional to the number of photos hitting the sensor.  A "regular" conversion in DPP and any other RAW converter applies a gamma curve that corrects for the human eye's non-linear response, where our eye is much more sensitive to changes at the dark end as compared to the light end, so the dark-end of the sensor data has to be brightened more than the bright end to get a realistic-looking image.  Having access to the pre-eye-gamma-corrected image data is useful if you are going to be adding images together to synthetically increase the dynamic range of the data and/or to reduce the random thermal noise.

 

From reading his essay, above, I believe Tindemans' use of "linear" means not applying any sort of toning adjustments in ACR or Lightroom, where the ACR/LR toning sliders are set to 0 and are using the linear contrast curve instead of using the 5, 50, 25 and Medium Contrast toning curve defaults.  This is one step less linear than not applying the eye-correction gamma curve.

 

So regarding color-accuracy, if you are trying to be photometrically accurate, then DPPs linear will help, but the images will look strange because the eye's response is not linear.  If you are trying to keep adjustments in brightness from changing the hue/sat of your colors then using ACR/LR in a linear mode followed by Tindemans' tools, will help, but you will still have to do some non-linear things, afterwards, to compress the wider dynamic range of the camera into the narrower display/print space.

 

If you are using 32-bit Photoshop on Windows, then his newer Tonability plug-in may be easier to use than the CurveTools action.

 

Also, it seems that Tindemans now prefers LightZone's ZoneMapper, to his own CurveTools action for Photoshop:

http://www.adobeforums.com/webx?14@@.3bc3840b/2

[Juan]

On the subject of keeping color accuracy, what about using Photoshop
CS4 with adjustment layers set to luminosity blending? I have been
trying them and the color does not seem to be affected when this
blending mode is chosen. If this is as I think, then there is no need
to use Tindeman's curve tools.

Am I correct?

Thanks,
Juan Dent

On Nov 8, 8:35 pm, "Steve Sprengel" <st...@sprengels.com> wrote:
> Looking at Tindemans Color Reproduction essay, I believe his "linear" is not the same as DPP's "linear."http://www.21stcenturyshoebox.com/essays/color_reproduction.html#_Toc...

[Steve Sprengel]

You can work in Lab mode and do what the luminosity-blend-mode adjustment layers do as far as not affecting the color, and Tindemans is aware of Lab mode, so there must be something more his tools are doing; however, for what you are trying to do, it's probably just fine to use adjustment layers.