[vdr] Re: Can vdr work without antenna? (play recordings)

[Dennis Noordsij <dennis.noordsij@movial.fi>]

> > Wasn't there a remark on the DVB list about calling SetPIDS(0,0,0,0,0)
> > periodically (interrupts video stream though) to keep things working?
>
> I've tried something similar by calling that on every channel change
> (vdr might do this on channel change anyway I think) - but it made no
> difference. Possibly worth a try to do this more often, but I don't
> think this resolves the problem. Question is - can you reproduce that
> 0.9.4 is stable over a day or so on your otherwise crashy card and if
> you can, why is that so. What do later drivers do differently that makes
> it crash. I've had no luck trying to figure this out last year for a
> couple of months.

Well analysing all driver<->firmware communications in the 0.9.4 and the 
dvb-kernel driver is possible, but the firmware itself was also different, so 
if it reacts differently or the API is very different then it gets difficult 
to see what is going on.

There was an interesting thread on the DVB list of using the old firmware with 
the newer drivers.

By the way, what seems to trigger it? If one were to make a VDR output plugin 
which always sends valid data to the DVB device, instead of letting VDR use 
the device directly, could that solve things?

I.e. when you change the channel the output plugin sends valid I-frames for as 
long as there is no valid MPEG data during the channel switch just to keep a 
valid stream going to the ARM. As in, the DVB card does not know the channel 
was changed, or a recording was started, or whatever. Even trick mode is 
handled in the output device by sending only I-frames instead of letting the 
ARM do that (if I'm correct in thinking the ARM does that at the moment :-).

Would that work?

Cheers,
Dennis


-- 
Info:
To unsubscribe send a mail to ecartis@linuxtv.org with "unsubscribe vdr" as subject.