To get upsampled YCbCr JPEG, just use
TIFFSetField( tiff, TIFFTAG_JPEGCOLORMODE, JPEGCOLORMODE_RGB );
before using TIFFReadEncodedTile / TIFFReadEncodedStrip
Le 10/06/2022 à 21:29, L. E. Segovia via Tiff a écrit :
Hi again,
A quick update: it's been already reported to this list... in 2003.
https://www.asmail.be/msg0054686955.html
We have our own upsampler for lossless compressors but the JPEG
(uncompressed) strip looks like it's laid out differently. If anyone
knows how to upsample it, I'll be glad to fix it in Krita.
amyspark
On 21/05/2022 16:37, L. E. Segovia via Tiff wrote:
Hi all,
Recently one of my colleagues at Krita found an abandoned TIFF in our
source code repository that, to the best of our knowledge, had never
been checked into our test suite.
Turns out, it is a JPEG-compressed, YCbCr TIFF. I've recently added
color space profiles for YCbCr, but when loading this sample inside
our app, the output shows significant corruption as well as
misaligned strips.
I've tried the tiff2rgba and rgba2ycbcr tools to generate copies, and
can confirm that:
- Loading YCbCr TIFFs with anything but JPEG compression works.
- Loading RGBA TIFFs with any compression, including JPEG, works.
which leads me to conclude that the expected layout of YCbCr + JPEG
doesn't match what it says on the tin ("Ordering of Component
Samples", page 93 of the standard). Unfortunately, I can't verify my
suspicions with libtiff's source code itself, as in all cases, either
the data is already in RGB form or libjpeg-turbo is asked to convert
it to RGB.
Does anyone know what the layout of this kind of data is? Or,
alternatively, where I could start looking?
Best,
amyspark
--
http://www.spatialys.com
My software is free, but my time generally not.
_______________________________________________
Tiff mailing list
[email protected]
https://lists.osgeo.org/mailman/listinfo/tiff
