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

[Andreas Vierengel <vdr@vierengel.de>]

Am Mo, 2003-08-25 um 12.50 schrieb Stefan Betermieux:
> -----BEGIN PGP SIGNED MESSAGE-----
> Hash: SHA1
> 
> Am Sunday 24 August 2003 20:29 schrieben Sie:
> > Am So, 2003-08-24 um 01.43 schrieb Andreas Vierengel:
> > > Am Sa, 2003-08-23 um 16.56 schrieb Timo Helkiö:
> > > > On Sat, 2003-08-23 at 15:43, Andreas Vierengel wrote:
> > > > > Hi,
> > > > >
> > > > > I have used VDR for 3 month so far with a Nexus-s(rev 2.1) and
> > > > > nova-ci (rev 1.1) connected via a sat-multiswitch (no diseqc) without
> > > > > problems. A few days ago I bought new computer hardware:
> > > > > P4-Celeron 2.4GHz, MSI-845-GE MAX-L Mainboard, 350W Enermax
> > > > > Powersupply and a new computer case.
> > > > > My old system was an 433 MHz Celeron (PII) on an asus p2l97
> > > > > Maibnboard. storage-system: 4x 120GB Seagate IDE => software raid5
> > > > > and raid1 for booting.
> > > > >
> > > > > Sofware is/was: debian linux 3.0, vanilla-kernel 2.4.21, vanilla
> > > > > vdr-1.2.2, liunxtv-dvb-1.0.0, gcc-2.95.4
> > > > >
> > > > > My problem is:
> > > > > recording on Nova-CI ( I don't use a cam-module at the moment)
> > > > > produces sound clicks, block-artefacts and small skips on every
> > > > > channel. The recording is not usable anymore :(
> > > > > recording on my nexus-s works perfectly the same like in my old
> > > > > computer-system.
> > > > >
> > > > > I changed PCI-Slots for both cards, tested any pci-latency setting,
> > > > > tested shared/unshared interrupt combinations => no effect.
> > > > >
> > > > > On software side I only rebuilt my kernel with optimization for P4,
> > > > > recompiled DVB-drivers and VDR.
> > > > >
> > > > > The whole thing is driving me nuts, so I'm at the point to throw out
> > > > > the nova and buy another nexus...
> > > > >
> > > > > What can I do further to investigate the problem ?
> > > > > Thanks in advance for any help !!
> > > > >
> > > > > --Andi
> > > >
> > > > Check if DMA for HD is on.
> > >
> > > Thanks for the answer !
> > >
> > > It is already on for every harddisk.
> > > hdparm -d 1 -m 16 -c 1 -u 1 -X67 /dev/hd[a-d].
> > >
> > > This cannot be the problem, since my nexus-s does record perfectly.
> > > My old system was only multiword-dma 2, which is slower than UDMA3,
> > > which I use now ( I had to use UDMA3, since kernel 2.4.x does have
> > > problems with UDMA >3 under load ). My throughput (tested with hdparm)
> > > was about 13MB/s with my old System and is now about 35MB/s.
> > > My Nova did record perfectly with the old, much slower io-subsystem...
> > >
> > > --Andi
> >
> > update:
> > The problem seems not related to my nova card.
> > I found out that always the first video-device on the pci-bus is
> > working, and the second shows the problems.
> > So it could be some kind of pci-problem.
> 
> 
> At last someone with similar problems like mine ;-)
> I have a Pentium 4 Box (with Asus 845 GE Board), a Nova-s and a dxr3 for 
> TV-Out. I have tried literally everything, from changing PCI Parameters 
> (including latency between 32 and 256), swapping cards, to changing the 
> drivers. On windows it's working fine. I am using 2.6.0 test at the moment 
> with dvb drivers from the kernel, before I have tried almost every available 
> DVB driver, even on 2.4 and 2.5 kernels. I have made sure, that it is no dxr3 
> related problem by using dvbstream and mplayer to display the stream on my 
> monitor. 
> 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 please Andreas, 
> could you verify this? I am really trying to iron this out, but until now, 
> nobody could reproduce the problem. If anyone needs to reread the previous 
> threads, please look for the subjects 
> "[vdr] budget dvb + dxr3 -> small drop outs"
> "[vdr] TS Continuity Errors on Nova-S"
> 
> bye,
> 
> Stefan

hmm, today in the morning, I pulled off my Nova, and tested only with my
nexus. I recorded ARD, and in replay, I got block artefacts an sound
glitches as before...
If I put my Nova in again, my nexus works perfect, but the nova produces
the problem.. arrrgh.

afterwards I've tested another Mainboard (i845PE) chipset, with almost
the same problems.
So far I don't have your error messages in syslog, but I will double
check this this evening !

My last resort is an old PIII mainboard, which I will test tonight, to
absolutely confirm that this problem is only P4-chipset related...

I'm also eager to find out this problem, cause I invested so much time
in the whole thing, that I would give my last drop blood for VDR :)

--Andi



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