Turning whitebalancing off during RAW conversion

[deepskystacker]

Turning whitebalancing off during RAW conversion

[perdrix52]

I definitely could make this change and it would be moderately simple to do (though it has impact way beyond the code I pointed you to). It would be part of the next release if done. Would I be correct to think that you still want the code to scale 12 and 14 bit data to 16 bit?

Darn - this would mean even MORE translation work to be done.

David

[perdrix52]

Hi Ivo, the current code doesn't use DcRaw at all, and we do our own WB processing. However I while I am prepared to consider allowing an option to bypass white balance processing, I don't understand why it's a problem. The code we use is very simple;

See: RawUtils.cpp lines 1093 to 1153

https://github.com/deepskystacker/DSS/blob/master/DeepSkyStacker/RAWUtils.cpp

So what could that be doing that messes up your code?
David

[IvoJager]

Thanks for the lighting quick response David - it is much appreciated!
Looking at the code, pre_mul[0], pre_mul[1], pre_mul[2], pre_mul[3] should all be forced to 1.0 for white balancing to not take place. Currently, it appears these are set by a copy from either C.cam_mul or C.pre_mul. I'm not quite sure where these, in turn, are set.

To recap, the problem occurs when the values in pre_mul are anything but 1.0 (as would be accomplished by the -r 1 1 1 1 parameter).
This causes the signal to be multiplied by an amount that cannot be known by the program used to post-process the stack. Because these multipliers cannot be known, it cannot be known by the program post-processing the stack how the signal and its noise component were impacted in each channel. This in turn makes the correct weights for creating a synthetic luminance dataset unknown. This in turn yields to a noisier synthetic luminance frame than needed, leading in turn to a poorer image and less effective results from algorithms such as deconvolution.

In addition, when stacking an image shot with the increasingly popular duoband filters for DSLRs, white balancing becomes rather meaningless, and the impact on synthetic luminance creation even more detrimental.

Does this make sense?

[IvoJager]

Sorry about the translation work, but you would make a lot of people very, very happy indeed! :)
Please keep the proper global scaling in place, as-is. It's just the disparity in color channel scaling (aka white balancing) we're trying to eliminate for people with our use cases. Say the word and I'll bring the good news to the StarTools forums!

[perdrix52]

This will be added to to the next release (4.2.3).

David

[IvoJager]

That's fantastic news David - thank you very much!