Re: AW: glitches, another story *g*

[Guido Fiala <gfiala@s.netic.de>]

Am Tue, 10 Oct 2000 schrieb Klaus Schmidinger:
>Guido Fiala wrote:
>> 
>> >He reported that he also has glitches during CPU and disk activities
>> >which are unrelated to the card. In my experience such glitches are
>> >caused by heat. Of course they would also occur during activities
>> >related to the card (recording). And, as I said before, they are much
>> >more often and stronger if you have a CI adapter connected to your
>> >card and they completely disappeared for me when I used an extra fan
>> >blowing on the card.
>> >
>> >So, if you watch TV and compile a kernel and have no glitches,
>> >fine. Then the reason for your problems lies somewhere else.
>> >Otherwise, you may have a heat problem.
>> >
>> True - think most of us have those glitches just in the recorded files,
>> but not watching TV the same time -> vdr does not take the data fast enough,
>> because it's not scheduled in realtime ???
>
>As long as VDR doesn't report any "missed data blocks" it should have
>taken all the data fast enough. VDR checks the sequence numbers of
>the data blocks and also their lengths. Also: if the glitches happen
>at different points when replaying, ths would indicate that the recorded
>data is ok, but the driver/firmware has trouble keeping picture and sound
>in sync.
>

At least for me, exactly that's happening - "missed audio/video data blocks
...".
Also - if the firmware does not write in a full buffer (but throws away the
data), you can get buffer-fulls before the sequence-chain is broken, think
exactly that happens.