[PD] Re: Using ov519 webcams (Eyetoy) with pdp/Gem (jpeg frames)

[martin pichlmair]

> I already tried a bit of copy and pasting in pdp (with breaking
> compression-less v4l support, as it's just for me for now), but I
> don't really find my way around in PDP yet. Some thing I encountered:
> I'm not sure, where I need to do all this, probably in the
> pdp_v4l_bang function?  Also I don't know, what kind of data pdp
> expects to get in the end, that is, what format should the image be in
> and do I need to set headers and so on. I really don't know my way
> around pdp yet. :(

you can look into pdp_llconv to find out from which formats you can 
convert to the internal pdp format (signed 16bit yvu 4:1:1 
(RIF_YVU__P411_S16)). naturally some of the conversions take longer 
than others, rgb being slowest.

you could of course do it in pdp_v4l_bang itself. the v4l object uses 
memory mapped io (i think). at least it assignes a memory portion to 
the driver to write the frame into - thats why you find no call a la 
get_frame(). thus in _bang it merely copies the frame into a packet 
(and converts it). if your driver supports that approach you can use 
pdp_v4l quite straightforward and just decompress the frame in bang - 
after it is written by the cam driver and before it is converted to 
RIF_YVU__P411_S16.

lg
martin