Andreas,

There is sure a non-linear conversion to fit 12bit sensor output into 8 bit
data being compressed (see
http://www3.elphel.com/linuxdevices/articles/AT9913651997.html ). This
non-linear conversion (defined by gamma settings in camera settings) is
designed to match the sensor output noise (mostly shot noise). The
conversion parameters are saved in Exif headers, I would recommend you to
try ImageJ plugin - it converts JP46/JP4 images to linear (32-bit TIFF),
and it is GPL-ed software with the source code you can reuse in other
projects. Here it is:

http://elphel.git.sourceforge.net/git/gitweb.cgi?p=elphel/ImageJ-Elphel;a=blob;f=JP46_Reader_camera.java

Andrey

These are illustrations from that article:
http://docs.elphel.com/linuxdevices/AT8926870933_files/elphel_images/fwc_01.png
http://docs.elphel.com/linuxdevices/AT8926870933_files/elphel_images/fwc_02.png

On Wed, Nov 16, 2011 at 9:49 AM, Andreas Bean <[email protected]> wrote:

> Andrey,
>
> After investigating the problem more deeply I found out that the reason
> for my problem with the hsv saturation is a nonlinear camera response
> against the exposure time.
>
> Is there any processing done in the camera before saving in jp46 mode?
>
> Do you have any idea why the response of the r g b pixel values is not
> linear with exposure time?
>
> Andreas
>
> ______________________________**_________________
> Support-list mailing list
> [email protected].**com <[email protected]>
> http://support.elphel.com/**mailman/listinfo/support-list_**
> support.elphel.com<http://support.elphel.com/mailman/listinfo/support-list_support.elphel.com>
>
_______________________________________________
Support-list mailing list
[email protected]
http://support.elphel.com/mailman/listinfo/support-list_support.elphel.com

Reply via email to