> -----Original Message----- > From: video4linux-list-admin@xxxxxxxxxx > [mailto:video4linux-list-admin@xxxxxxxxxx] On Behalf Of Rob Nielsen > Sent: Friday, 1 August 2003 8:09 AM > To: video4linux-list@xxxxxxxxxx > Subject: Re: SAA7134 Sound disappeared > > > Christian Hack wrote: > > > > Sounds like an OK fix in the interim. I am teaching myself > Linux device > > drivers at the moment. Are you (still) getting the > variations in sound > > volume though? That is starting to annoy me almost as much > as the mute > > problem. > > > It appears it might have fixed that too. I've just gone through the > recordings since it happened and they seem to be of > comparable volume. > There are differences but that's probably just the different > channels - > I can't find anything as dramatic as it was before. > Mine's definitely still got the multiple volume levels bug. However, this _could_ be due to something external to the driver setting the volume incorrectly. The way I currently understand it, saa7134 is a V4L2 driver and Myth only supports V4L(1). Thus the V4L2 saa7134 driver is working through some sort of compatibility layer (module is v4l2-compat I guess) and that _might_ be where things go screwy I have hardcoded a volume level of "2" into my driver which seems like a reasonable volume and when it normally works 2 is what I see in the debug logs. I had it go very quiet second go after hard coding mute off (according to debug messages it was a volume level of -14). Even after restarting live Tvmore than 10 times in Myth it remained on -14 and very quiet. In the latest snapshot of 24072003 line 1061 of saa7134-video.c: case V4L2_CID_AUDIO_VOLUME: dev->ctl_volume = c->value; saa7134_tvaudio_setvolume(dev,dev->ctl_volume); break; I simply changed to case V4L2_CID_AUDIO_VOLUME: dev->ctl_volume = 2; saa7134_tvaudio_setvolume(dev,dev->ctl_volume); break; And that seems to work OK. Of course it's not a proper fix, but without any doco, I think this is the best I can do for now. I use Myth's output for volume control rather than the recording input which should always remain constant. Christian