Martin McCormick wrote: > deloptes <delop...@gmail.com> writes: >> [1] http://www.tldp.org/LDP/LG/issue38/blanchard.html >> [2] http://tldp.org/HOWTO/Plug-and-Play-HOWTO-11.html >> [3] http://people.freebsd.org/~tanimura/docs/4236b.pdf > > I have had time to digest the information in those 3 > links and must say thank you for posting them. I now have a much > better understanding of why I have always had trouble getting ISA > sound cards, even the really good ones, to work under Linux > sound. As I said once before in an earlier posting, when Linux > sound works, it works extremely well so I have nothing but good > things to say about it in general. >
Yes this is true and we like and enjoy it :) > I am now amazed that the ISA cards ever worked at all for > any card other than Card 0. I also can see why there were > problems with maintaining card order after rebooting even with > udev rules in /etc/modprobe.d/alsa-base.conf. The problem > basically boils down to not having enough resources free to let > each card operate it's DMA access lines and interrupt request > line without stepping all over the other card or cards. > > On the system I was trying to use, earlier versions of > Linux allowed a PCI and ISA-style card to coexist but not totally > peacefully. If Card 0 had audio on it, Card 1 still worked but > you could hear rhythmic pops which, while not terribly loud, > shouldn't have been there at all. I suspect that it was DMA > contention as I think they did get 2 different interrupts. > You can check all of this in the kernel. Basically one would check which IRQ/DMA is free and use this/assign it to the card manually. I think it was done via driver setting, but probably kernel can take those settings from the bios. > I believe that wheezy and the newer versions of Linux > sound are right in making it harder to bend things even if the > results are not what we always want since the idea is to have > stability and predictability. I will put a USB card on here and > declare victory. > > This box is used to record 4 independent 8-bit audio > channels at 8-KHZ sampling. That is positively horrible for music > or high-quality voice but for VHF and UHF scanner radios, that is > approximately the frequency range of audio one gets so it all > sounds better than one might think. > > Maybe my ears are just shot but seriously, the radios > sound quite normal but good sound such as music or broadcast > audio is muddy and Mr. Nyquist is laughing as he spins in his > grave. > > Martin McCormick This depends on various things. regards
