Date: Wed, 3 Sep 1997 08:25:03 +0200 (MET DST) From: Luigi Rizzo <luigi@labinfo.iet.unipi.it> To: kuku@gilberto.physik.RWTH-Aachen.DE (Christoph Kukulies) Cc: multimedia@FreeBSD.ORG Subject: Re: Opti 931 Message-ID: <199709030625.IAA07228@labinfo.iet.unipi.it> In-Reply-To: <199709030643.IAA08153@gil.physik.rwth-aachen.de> from "Christoph Kukulies" at Sep 3, 97 08:43:16 am
next in thread | previous in thread | raw e-mail | index | archive | help
> We have extreme cheapo soundcards around here claiming to be > full duplex. (cost about DM 45,-- == USD 24.00) > > They are built on the Opti 931 chip (www.opti.com) and > a full-duplex driver is available for Win95/NT. > > Does the new sound code support that chip (and possibly fdx?) see my previous postings on the multimedia list Subject: OPTi931 problems... To: multimedia@FreeBSD.ORG Date: Sun, 31 Aug 1997 13:38:05 +0200 (MET DST) Sender: owner-freebsd-multimedia@FreeBSD.ORG The chip appears to have at least two bugs: - full duplex operation in ulaw/alaw fails for the capture channel, which produces stereo-16bit samples; - for some reasons, the card forgets to generate some interrupts when working in full duplex. I suspect it has to do with the card generating DMA requests but _not_ counting them in certain conditions, so the system ends up with the ISA DMA controller thinking that the transfer is complete, while the card pretends to have more left. Since it appears that no other _good_ full duplex cards (i.e. based on the CS4236/37 codec) are available anymore in the local stores, and since I am the unlucky owner of several of these boards, I have implemented some fixes for the above bugs in my driver (but wait for the next snap, in a couple of days). But especially for the second one, the output will be choppy and, whatever the fix is (either in my driver or in the Win95 one), the codec will appear to have different playback and capture sample rates. The difference will be small, but it breaks the assumption that many conferencing program do about sample speeds being the same. My advice would be to get boards based on the CS4236 (and probably also the CS4237 although I do not own one) if you can, even if they cost more. The chips are much better documented, and work as described in the data sheets (which are available on line). Avoid the Soundblaster cards, as far as I know they have some stupid internal architecture so that full duplex operation requires to use 8-bit format on one channel and 16-bit format on the other one. This is painful for the driver to support, and gives much worse quality anyways. Beside the SB16 are more expensive than the CS4236. The only reason I have bought one is to be able to test my code on it (the power of being a standard....) Cheers Luigi -----------------------------+-------------------------------------- Luigi Rizzo | Dip. di Ingegneria dell'Informazione email: luigi@iet.unipi.it | Universita' di Pisa tel: +39-50-568533 | via Diotisalvi 2, 56126 PISA (Italy) fax: +39-50-568522 | http://www.iet.unipi.it/~luigi/ _____________________________|______________________________________
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?199709030625.IAA07228>