[gst-devel] Grab video

[Andrey Nechypurenko]

>> Here is the pipeline which works on my system:
>>
>> v4l2src always-copy=FALSE !
>> video/x-raw-yuv,format=(fourcc)YUY2,width=640,height=480,framerate=30/1
>> ! ffmpegcolorspace ! video/x-raw-yuv,format=(fourcc)UYVY ! TIVidResize
>> name=qos-scaler contiguousInputFrame=TRUE !
>> video/x-raw-yuv,format=(fourcc)UYVY,width=640,height=480 ! TIVidenc1
>> codecName=h264enc engineName=codecServer bitRate=320000
>> encodingPreset=2 genTimeStamps=TRUE byteStream=TRUE ! rtph264pay pt=96
>> ! appsink
>
> That's really interesting - when I run this, I get this error:
>  linux/omap3530/Resize.c:585: Resize_execute: Assertion
> `(Buffer_getPhysicalPtr(hSrcBuf) & 0x1F) == 0' failed.
> Looking closely at the source files, I see that the file in question is
> just a bunch of "not implemented" stubs.
>
> You have a slightly different package set than I do, maybe I can figure out
> what to use (although what you are using is *older* than the ones I have)

What I learned while using gstreamer is that caps are matter ( c )
Joda :-) . I was able to solve 99% of problems just by specifying the
caps manually. In particular, TI elements like TIVidenc1, would
auto-negotiate caps which are not really supported on particular
platform. That is why it is always better to try to specify all caps
as precisely as possible and after the pipeline will run, try to
remove some of them to make it more generic. I am usually starting
from analizing what formats are supported by the camera and specify
exactly the one I am sure is supported, and then go on with next
elements in pipeline. It is almost like a detective work :-) .

BTW, if you are interested, you can take a look at this project:
www.gitorious.org/veter . There is a wiki page with overview. Here you
will be able to find some more examples in the misc/*.conf files which
could be of interest for you either.

Good luck in your quest for running pipelines :-)

Andrey.