[gst-devel] I still don't get it : multiple sources side by side

[Jerome Alet]

Hi,

Here's a more precise description compared to my message from the other
day which remained basically unanswered.

I'd like to obtain a single 992x480 Ogg/Theora video file with the 
following visual aspect (scaling below is incorrect, but boxes' 
dimensions are) 

    +-------------------+-------------------------------------------+
    |                   |                                           |
    |                   |                                           |
    |      Box 1        |                                           |
    |                   |                                           |
    |      352x288      |                                           |
    |                   |                                           |
    |                   |                   Box 3                   |
    +-------------------+                                           |
    |                   |                 640x480                   |
    |                   |                                           |
    |      Box 2        |                                           |
    |                   |                                           |
    |      192x288      |                                           |
    |                   |                                           |
    +-------------------+-------------------------------------------+
    
As input I've got :

        - 1 DV Camera or DV Recorder (Audio and Video)
        
        - 1 V4L Device (Video only)
        
        - 1 PNG or JPEG image (possibly remote with gnomevfssrc)
        
The audio in the output will come from the DV and will be copied
unmodified to the output.

Box 1 contains the scaled down video from the 720x576 DV input.

Box 2 contains the JPEG image (or PNG), which will have to be reloaded
automatically for example every 60 seconds.

Box 3 contains the video from the V4L device (webcam), scaled to 640x480.

I've printed and read gstreamer's documentation and used gst-inspect
a bit, but I can't make something out of this.

If I understand correctly I need the following :
        
        Box 1 + Audio :
        
                - dv1394src
                - dvdemux
                - dvdec
                - queue
                - audioconvert
                - videoscale
        
        Box 2 : 
        
                - gnomevfssrc
                - jpegdec or pngdec
                - maybe decodebin, if I allow both png and jpeg
                
        Box 3 :                
        
                - v4lsrc
                - queue
                
Final output :                

                - videobox
                - theoraenc
                - vorbisenc
                - oggmux
                - filesink
                
I think I can see how to combine some of these elements to obtain
parts of the final result I want, but I really can't understand how
to combine all this stuff together.

I've read many examples but none of them so far seems to be as complex
as what I want to do. 

Any help is very welcome.

Thanks so much in advance.

bye

Jerome Alet