After an evaluation, GNOME has moved from Bugzilla to GitLab. Learn more about GitLab.
No new issues can be reported in GNOME Bugzilla anymore.
To report an issue in a GNOME project, go to GNOME GitLab.
Do not go to GNOME Gitlab for: Bluefish, Doxygen, GnuCash, GStreamer, java-gnome, LDTP, NetworkManager, Tomboy.
Bug 710323 - terrible framerates with ximagesrc
terrible framerates with ximagesrc
Status: RESOLVED DUPLICATE of bug 693037
Product: GStreamer
Classification: Platform
Component: gst-plugins-good
1.2.0
Other Linux
: Normal normal
: git master
Assigned To: GStreamer Maintainers
GStreamer Maintainers
: 752573 (view as bug list)
Depends on:
Blocks:
 
 
Reported: 2013-10-16 20:43 UTC by Jakub Steiner
Modified: 2018-05-04 09:07 UTC
See Also:
GNOME target: ---
GNOME version: ---


Attachments
webm screencast (1.21 MB, video/webm)
2013-10-16 20:43 UTC, Jakub Steiner
Details
Debug log (no colors) (163.11 KB, application/x-xz)
2015-07-21 20:13 UTC, Francesco Frassinelli
Details

Description Jakub Steiner 2013-10-16 20:43:10 UTC
As the screen screen recording gives me 0byte files I tried to record one of my displays, set to 1920x1200 and scaled with `xrandr --output DVI-I-1 --scale 1x0.9` with the pipeline kindly provided on #gstreamer IRC channel:

gst-launch-1.0 ximagesrc show-pointer=true startx=2560 starty=0 endx=4479 endy=1080 ! video/x-raw,framerate=30/1 ! videoconvert ! videoscale ! videorate ! vp8enc threads=12 ! webmmux name=mux ! filesink location="./video.webm"

This results in a very laggy video far from being 30fps despite my workstation having 12 cores to chew on that. To use a less involving codec, I tried encoding to mjpeg to little improvement. 

gst-launch-1.0 -e ximagesrc show-pointer=true startx=2560 starty=0 endx=4479 endy=1080 ! video/x-raw,framerate=30/1 ! videoconvert ! videoscale ! videorate ! queue ! jpegenc ! qtmux ! filesink location="./video.mov"

Both files attached. This is gstreamer1-plugins-good-1.2.0-1.fc20.x86_64
Comment 1 Jakub Steiner 2013-10-16 20:43:42 UTC
Created attachment 257451 [details]
webm screencast
Comment 2 Jakub Steiner 2013-10-16 21:21:59 UTC
The mjpeg one is too big to attach -- http://jimmac.musichall.cz/stuff/video.mov
Comment 3 Edward Hervey 2013-10-17 07:21:16 UTC
Tempted to say it's a duplicate of bug #693037
Comment 4 Antonio Ospite 2013-10-28 15:54:27 UTC
(In reply to comment #0)

Jakub can you try with:

ximagesrc use-damage=false ...

if that improves things for you this issue can indeed be closed as duplicate of #693037
Comment 5 Jakub Steiner 2013-10-30 19:39:45 UTC
Unfortunately cannot test this because the same pipelines as above now generate black frames.
Comment 6 Francesco Frassinelli 2015-07-18 11:57:58 UTC
(In reply to Jakub Steiner from comment #5)
> Unfortunately cannot test this because the same pipelines as above now
> generate black frames.

That's another bug: https://bugzilla.redhat.com/show_bug.cgi?id=1136317
Use remote=true.
Comment 7 Jakub Steiner 2015-07-20 11:19:15 UTC
This slideshow of a video has been recorded with the following pipeline:

gst-launch-1.0 ximagesrc remote=true use-damage=false show-pointer=true startx=2560 starty=0 endx=4479 endy=1080 ! video/x-raw,framerate=30/1 ! videoconvert ! videoscale ! videorate ! vp8enc threads=12 ! webmmux name=mux ! filesink location="./video.webm"

http://jimmac.musichall.cz/stuff/video.webm
Comment 8 Francesco Frassinelli 2015-07-20 11:20:26 UTC
(In reply to Jakub Steiner from comment #7)
> http://jimmac.musichall.cz/stuff/video.webm

404
Comment 9 Francesco Frassinelli 2015-07-21 20:08:06 UTC
*** Bug 752573 has been marked as a duplicate of this bug. ***
Comment 10 Francesco Frassinelli 2015-07-21 20:13:16 UTC
Created attachment 307864 [details]
Debug log (no colors)

Debug log generated with this command:
$ LC_ALL=C gst-launch-1.0 -e ximagesrc ! vaapipostproc ! v4l2sink device=/dev/video1 --gst-debug="*:5" --gst-debug-no-color &> ximagesrc-slow.log

I can confirm this bug. If I use use-damage=false the CPU usage is similar, but frames doesn't seems to be dropped, no timestamp errors when reproducing.

I can confirm this bug(In reply to Edward Hervey from comment #3)
> Tempted to say it's a duplicate of bug #693037

Performance is bad and you're right probably.
Comment 11 Edward Hervey 2018-05-04 09:07:28 UTC

*** This bug has been marked as a duplicate of bug 693037 ***