[vdr] Re: More thoughts on Matrox TV out with interlaced programmes

[Tony Houghton <h@realh.co.uk>]

In <4202741C.1070901@gmx.de>, Reinhard Nissl wrote:

> Hi,
> 
> Tony Houghton wrote:
> 
> >There's also the issue of recordings vs live TV. In theory it should be
> >possible to avoid the above problem with recordings by reading the file
> >at a rate to suit the graphics card's clock. But it looks as if a
> >cReceiver has no choice but to receive the data at the rate VDR sends
> >it. Obviously for live TV that would be determined by the broadcast, but
> >I presume for recordings VDR uses a system clock. If that's separate
> >from the clock used by the graphics card and they drift much, it would
> >get ugly. In case there is a potential problem there, would it be
> >possible for cReceiver to be extended so that a plugin can feed back to
> >VDR to moderate the rate at which data is sent if possible?
> 
> VDR doesn't control the data rate while replaying a recording. The data 
> rate is only determined by the output device. I. e. VDR will deliver the 
> data as fast as the output device consumes it.
> 
> There is just a minor issue: when the output device doesn't "eat" a PES 
> packet within 300 ms (= 3 consecutive poll timeouts of 100 ms), then VDR 
> will issue a CLEAR to have the device drop any buffered data and to 
> startover with replay.

In other words, VDR already has a sort of feedback mechanism, using a
poll function, that is separate from cReceiver?

-- 
TH * http://www.realh.co.uk