Date: Thu, 8 Jan 1998 14:15:00 +0100 (MET) From: Luigi Rizzo <luigi@labinfo.iet.unipi.it> To: multimedia@freebsd.org Subject: Please stress-tes the audio drivers! Message-ID: <199801081315.OAA05591@labinfo.iet.unipi.it>
next in thread | raw e-mail | index | archive | help
It's been some time since I don't get complaints about my audio
driver, which might mean that those using it are, on average,
satisfied with its operations, and others have given up usign it
(something I have done for years with the previous driver, so i
perfectly understand them...).
Recently, i was using the driver at high speeds (44100,16bit,stereo
-- guess where the bits come from...), and at times even in full
duplex at that speed.
While it mostly worked, I hear some background noise in presence
of other disk activity which I am not sure about the cause.
So, could you please try and report how my driver (or even guspnp,
oss, voxware3.0, ...) work at high speed, and if the use of different
drivers makes a difference on the quality of audio ?
That would help me a lot in fixing the problem(s) which could be:
- electrical interference to the analog circuitry (no fix for this);
- interference in interrupt processing or DMA activity due to IDE
activity (I am using IDE peripherals); there are some busy wait
situations in the wd/wcd/atapi stuff which could affect interrupt
processing;
- slow/inefficient interrupt processing in the audio driver;
Unfortunately I doubt the problem is in the driver: at CD speeds,
a byte is consumed every 5.67us, and a 64-byte fifo drains in 360us.
It's probably more likely that other system activities increase
the irq latency.
Thanks
Luigi
PS: other than unsupported hardware, I know of several bugs in the
driver (snd971209), on which I am currently working or have already
fixed in the upcoming snapshot of the code:
- with sb16, recording has a problem (open does not initialize the
device properly, you need to call the AIOSFMT ioctl to do it
right; vat works...);
- when you start playing from a streaming source which is not
faster than the sampling rate, you can hear some noise in the first
second or so; this is due to an annoying side effect of the
presence of the FIFO inside the audio chip.
- in one occasion, having two processes writing to the same audio
device caused a reboot (the driver is supposed to detect the
occurrence). I have already implemented a fix and have been unable
to reproduce the problem.
-----------------------------+--------------------------------------
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?199801081315.OAA05591>