[Bug 759624] videoconvert: Odd behaviour between videoconvert and v4l2src

[GStreamer (GNOME Bugzilla)]

https://bugzilla.gnome.org/show_bug.cgi?id=759624

--- Comment #1 from Dimitrios Katsaros <patcherwork at gmail.com> ---
-- CONTINUING FROM ABOVE --

gst-launch-1.0 -v v4l2src device=/dev/video3 !
'video/x-raw,width=1280,height=720,format=RGB,colorimetry=(string)1:4:7:1' !
videoconvert ! video/x-raw,format=I420  ! videoconvert ! ximagesink

Running the same pipelines with videotest src would not break the results in
the first case. Comparing the two pipelines I was able to notice that the
colorimetry was being fixed to "1:4:7:1", whereas with videotestsrc the
colorimetry was not set. After even further investigation I was able to get
this debug info for the above pipelines:

For the bad one:
00:00.511107821  1342       0x743e30 DEBUG        video-converter
video-converter.c:1501:chain_scale: 230400 <> 230400
0:00:00.511127805  1342       0x743e30 DEBUG        video-converter
video-converter.c:1545:chain_convert: matrix 4 -> 4 (1)
0:00:00.511146296  1342       0x743e30 DEBUG        video-converter
video-converter.c:1547:chain_convert: bits 8 -> 8 (1)
0:00:00.511168155  1342       0x743e30 DEBUG        video-converter
video-converter.c:1549:chain_convert: primaries 1 -> 4 (1)


For the good one:

0:00:00.588955810  1348       0x743e30 DEBUG        video-converter
video-converter.c:1545:chain_convert: matrix 4 -> 3 (0)
0:00:00.588978150  1348       0x743e30 DEBUG        video-converter
video-converter.c:1547:chain_convert: bits 8 -> 8 (1)
0:00:00.588996163  1348       0x743e30 DEBUG        video-converter
video-converter.c:1549:chain_convert: primaries 1 -> 1 (1)
0:00:00.589018669  1348       0x743e30 DEBUG        video-converter
video-converter.c:1593:chain_convert: to RGB matrix

It seems to verify that the colorimetries are different for different frame
sizes. At this point the subject is outside of my knowledge so I am not even
aware if this is a problem of v4l2, v4l2src or videoconvert. Anyone know what
is going on here?

-- 
You are receiving this mail because:
You are the QA Contact for the bug.
You are the assignee for the bug.