Re: [Lcms-user] [ANNOUNCE] Little CMS 2.13 released

[Marti M. <mar...@li...>]

It is now fixed in git. This bug has been there since 2.10

https://github.com/mm2/Little-CMS/commit/fdbfb7694f9d7048d53674b79ddfc38068bfdaf7 <https://github.com/mm2/Little-CMS/commit/fdbfb7694f9d7048d53674b79ddfc38068bfdaf7>

A true pity, but 2.13 is already released. You can avoid the bug on the official releases by using cmsFLAGS_NOOPTIMIZE when input grayscale is being used. I guess this is not very common as the flaw has been unnoticed by years. In 2.13 is more evident because another bug was discarding channels G and B in the gray->RGB optimized transforms, anyway that was wrong too.

Regards
Marti



> 
>> On 30 Jan 2022, at 02:48, Aaron Boxer <bo...@gm...> wrote:
>> 
>> Hi Marti,
>> 
>> Thank you for this new release! I think I have found a regression, though.
>> 
>> I maintain the Grok Jpeg 2000 codec, and we use lcms2 to apply ICC profiles after
>> decompressing from JPEG 2000. 
>> 
>> The attached file file8.jp2 is a monochrome image
>> with a grayscale ICC profile.  For decompression, the codec decompresses
>> the single channel, applies the ICC profile, and then converts it to RGB 
>> (by duplicating the mono channel)
>> 
>> file8.jp2.tif is the decompressed result with the older version of lcms2.
>> file8.jp2_new.tif is the result with lcms2 2.13. You can see that the image w/ 2.13 
>> is saturating at the top of the range.
>> 
>> If you need, I can point you to the code that applies the ICC profile.
>> 
>> Let me know if I can provide any more information.
>> Thanks again,
>> Aaron
>> 
> 
> 
> 
> _______________________________________________
> Lcms-user mailing list
> Lcm...@li...
> https://lists.sourceforge.net/lists/listinfo/lcms-user