Automated FocusStacker

[rjlittlefield]

Raw data holds the color information with a gamma curve applied. When it is converted to tiff it becomes linear

With respect, you have this part backwards.

It is the raw data that is linear -- each photon contributes the same amount to a sensor value, no matter how many photons have been captured at that photosite (assuming no overflow).

Gamma is traditionally applied in creating a TIFF file, when the color profile is made to be sRGB, Adobe RGB, ProPhoto RGB, or similar.

However, using Helicon's raw-in-DNG-out workflow, the TIFF files created as intermediates are also linear, and in fact their color profile is identical to the camera's sensor. That is, their raw-to-TIFF conversion process does Bayer interpolation to fill in otherwise missing RGB values, but for non-interpolated values the TIFF pixel value will equal the raw sensor value. It is this use of a linear color profile that preserves full ability to do color corrections later. This is explained in different words but more detail at http://www.heliconsoft.com/raw-in-dng-out/ .

The only thing I know is if you want to have a 16bit result, dng pipeline is much more faster.

That is useful information, and I have no disagreement with this statement as long as it's understood as an observation about Helicon's implementation.

--Rik

[masquerade]

Yes you are right. sorry for that.

No gamma curve applied to the raw data. I dont know why and how I wrote it like that. If it were processed like that it wont be called raw.

[mjkzz]

As a result, I'm not at all inclined to provide a workflow in which raw conversion is done invisibly in a background process, using parameters that are at best "out of sight, out of mind".

I did investigate using RAW file as input, it is absolutely right that you can NOT just do it "out of sight, out of mind". Color grading is a complicated process and like Rik said, it depends on lighting, setup, etc to get best result.

The structure of data in a typical raw image file, one value per photosite with color implied by a mosaic Bayer filter pattern, is fundamentally incompatible with the image alignment process that is required for stacking.

I am not quite sure how "linear" camera sensors are, but I think most sensors are tuned to have linearity in certain range of light intensity. Having said that, these RAW data can be quite useful for alignment, we just have to develop algorithm based on RAW data, instead of RGB.

Just a thought.