[vdr] Re: sound clicks and "block artefacts" with Nova-CI

[Andreas Vierengel <vdr@vierengel.de>]

Am Mo, 2003-08-25 um 16.45 schrieb Stefan Betermieux:
> -----BEGIN PGP SIGNED MESSAGE-----
> Hash: SHA1
> 
> Am Monday 25 August 2003 15:44 schrieben Sie:
> > >> It seems that we can limit the problem to Nova cards and 845 chipsets.
> > >> The  only usable error message is the ccError count, which is increased
> > >> when  dropouts occur and which is printed to syslog when vdr changes the
> > >> channel  (i.e.:"cTS2PES got 0 TS errors, 10 TS continuity errors"). So
> >
> > This seems to be related to my thread "what causes cTS2PES errors?" that I
> > tried to discuss on the list before. I also used a 845 chipset mainboard at
> > the time I had the problems. Since yesterdays I didn't had any time left to
> > test more with this issue. But yesterday I put the two Nova-S cards
> > mentioned in the thread in an old K7 500 MHz System and started three
> > parallel recordings without an error for three hours. Since this system is
> > MUCH slower and the antenna configuration is the same as in may I think the
> > problem is really a 845 chipset issue. I will get some 865 chipset
> > mainboards in the next weeks to do some more tests. If you can verify that
> > your old P III mainboard will work we should look deeper why the 845
> > chipset causes that trouble.

Hi, here my results:

1.
tested with P3-850MHz Asus-P3B-F i440BX chipset.
the problem did vanish a little bit.

If my nova is plugged in a slot before my nexus on the pci bus, the
problem are the same like on my P4 system, but with less distortions.
The most distortions are on Sat-1 and Pro-7, even on live-viewing in
transfer-mode.
Simply swapping the cards (nexus first, nova second), reduces the
distortions further to a minimum, but the distortions happens now on my
nexus.
This is reproducible.
It doesn't matter if the cards do interrupt sharing or not, although
sometimes it seems to me the drivers are more stable without shared
interrupt, but this is only my opinion ;)

2. I tested with my old system once more:
Asus-P2L97 with 433MHz P2/Celeron.
This system is the best ! It runs smoothly without drops, regardless
pci-ordering of the cards, latency-settings, etc. 
The disadvantage here is that my 333 GB software-raid-5 array must boot
from floppy, because the bios will simply lock up the machine if I try
from hd.
It has only 4 PCI-slots...
Furthermore, the board can only do multiword-dma 2, which is rather slow
in cuttings or resyncing the array (about 8-9 hours) especially if my
wife has cut off the power, because I did not want to go to bed :)

So it seems to me: the faster the system is, the more likely the cards
get unstable...

I've never get any error messages during a drop, neither in
/var/log/messages, nor in the logfile from vdr's stdout/stderr output.

I think that vdr is aout of the way, so we shuld probably go to DVB
mailing list, since it must be hardware/kernel/driver related.

--Andi



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