Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Linear 16bit output #81

Open
adrienlucca opened this issue Aug 1, 2016 · 11 comments
Open

Linear 16bit output #81

adrienlucca opened this issue Aug 1, 2016 · 11 comments

Comments

@adrienlucca
Copy link

Hi,

what is the way to go if I "simply" want a Linear 16bit output (witout gamma correction) in an untagged RGB tiff?

Thanks

@rolkar
Copy link
Contributor

rolkar commented Aug 1, 2016

I understand your confusion here. I am a bit confused myself :)
We have added diverse switches. I think this is the truth
Note - this is assuming TIFF, it does not work for DNG

In order to get tiff you shall use this switch
-tiff

Then you can use one of those switches
-unprocessed
-color none
The first one will give you the data as they are.
The second will apply denoising etc.
Giving both switches will result in the last one overiding

@erikrk
Copy link
Contributor

erikrk commented Aug 1, 2016

I think the question was about linear RGB output, i.e. after color space conversion. This has unfortunately not been implemented.

@adrienlucca
Copy link
Author

Hi,

thanks for the comments.

So what happens if I simply use the -tiff -color none switches ?
I can see that it seems linear, however:

  • What is the content of the RGB channels? Is it the camera RGB values? Or what type of RGB is it?

If I only turn on the -tiff switch, is a gamma applied on the values ? And what colorspace is used ?

Thanks

@rolkar
Copy link
Contributor

rolkar commented Aug 1, 2016

AH! Sorry! It is not so obvious what to do though. In theory, you can provide e.g. linear sRGB or linear Adobe RGB. But, no receiver will understand what that is. It simply do not exist as a known standard. Of course, that do not mean that it is not possible to output it.

@erikrk
Copy link
Contributor

erikrk commented Aug 1, 2016

With "-color none" you get data that has been corrected for white
balance and denoised, but it's still in the camera's native color space,
and linear. Without "-color" if defaults to "-color sRGB" and you get
data that has been, after white balance correction and denoising,
converted to sRGB, with gamma. As Roland just pointed out, there is no
standard for representing linear RGB, and this is why we haven’t added
the option to generate it.

On lun, 2016-08-01 at 06:21 -0700, adrienlucca wrote:

Hi,

thanks for the comments.

So what happens if I simply use the -tiff -color none switches ?
I can see that it seems linear, however:

  * What is the content of the RGB channels? Is it the camera RGB
    values? Or what type of RGB is it? 

If I only turn on the -tiff switch, is a gamma applied on the values ?
And what colorspace is used ?

Thanks


You are receiving this because you commented.
Reply to this email directly, view it on GitHub, or mute the thread.

@rolkar
Copy link
Contributor

rolkar commented Aug 1, 2016

It is a bit strange that no one seems to care. 16 bit linear color space seems rather useful too me.

@adrienlucca
Copy link
Author

Thanks,

how do I get data that has not been white balanced?

Also, is there any specified color space for the Sigma RGB ? Do we have the CIE XYZ values associated with it? Or do we know the spectral sensitivity curves they correspond to?

@rolkar
Copy link
Contributor

rolkar commented Aug 1, 2016

Good questions :) There was a discussion in the DPReview scientific forum some weeks ago regarding color spaces and RAW data. No, RAW data do not have any color space. This is, in particular, true for Foveon. The reasons for not having a color space is 1. the spectral response of the channels are not linear combinations of any known color space and 2. the illuminant will zap it anyhow. So, a perfect conversion is not possible.

The method used, for Foveon cameras, is to give conversion matrices for converting the RAW data to either XYZ (for older cameras) or RGB/D65 (for newer cameras) for different lighting conditions, including custom white balance. Then just apply them, and get as good as you can.

@rolkar
Copy link
Contributor

rolkar commented Aug 1, 2016

With the switch -unprocessed you can get non white balanced data. But, you will also get the original noise.

@rolkar
Copy link
Contributor

rolkar commented May 3, 2017

Can we call this closed?

@erikrk
Copy link
Contributor

erikrk commented May 3, 2017

I think we can close it since there is no standardized way of representing linear 16-bit data.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

No branches or pull requests

3 participants