Date: Fri, 17 Mar 2000 21:33:44 +0100 (CET) From: BEWARE OF DOGS <kvindeservice@fluffy.gets.an.analprobe.dk> To: current@freebsd.org, stable@freebsd.org Subject: Soundcard support, or something Message-ID: <Pine.BSF.3.96.1000317195047.77771B-100000@fLuFFy.iNt.tElE.dK>
next in thread | raw e-mail | index | archive | help
Moin moin,
I guess that -current is now -stable, and so is -stable, but the tests
I made were on -current before it became -release, as well as on -stable
before it was usurped by -devel, or something, my brane hertz, so I'm
sending this to both -current and -stable, even though I bet both lists
are frequented by mostly the same hackerz. So tell me I should have
sent this to -multimedia instead....
Anyway, I spent last night with a stack of sound cards and a dual-boot
machine with 3.4-stable and 4.0-current as of about a week ago, more or
less, to try and figure why I failed so miserably with audio recording
last week at this time.
In short, 4.0-current (and I will bet good money on 5.0-current too)
failed pretty miserably with audio recording from all cards, while most
were successful (with some limitations) under 3.4-stable. Pilot error
should not be dismissed.
On the other hand, all cards tested were successful with playback of
pre-recorded .wav files, although the mixer interface (aumix) sometimes
left something to be desired.
Maybe this info will be helpful to someone... I used the pcm k0dez
exclusively in -current and 3.4, no Voxware was harmed for this test.
Recording was made with a somewhat-hacked version of brec-0.96 or -0.99,
with adjustments made to SHM/SEM options in the kernel config file.
First, under 3.4-stable, one card that was detected under -current was
not found -- some Vibra 16S Soundforte card with what appears to be an
on-board FM tuner.
* Secondly, again under 3.4-stable, all the soundblaster ISA cards that
I tried only recorded a mono sum signal from both left and right input
(line) channels. This included a gen-u-wine SB16, a Vibra 16, and an
AWE 64. All these were reported at boot as SB16pnp, so there is no
doubt some simple toggle to make all of them record stereo and suddenly
become useful to me for audio input.
The two Soundblaster PCI cards I tried, recorded stereo cleanly
under 3.4, matching my previous experience with them in production
machines without problems. I've just tossed in the 16 PCI now and
have not thoroughly tested the mixer; however, I can say that the
aumix program on the 128 PCI (or PCI 128? I know one of them is
backwards) has the problem: When I attempt to select the recording
source with the space bar (to switch from mic to line), aumix exits
with SOUND_MIXER_WRITE_RECSRC. When restarted, the desired change
was in fact made. (I can't remember if I see this with -current)
I tried one other card that might be supported, a Crystal CS4235 chip
on a card whose number I know not, and it was pretty much the same
under both 3.4 and -current: Playback (line) worked, the line input
was controlled as synth on the mixer, I might have fried the mic
input since I never got sound from that, but aumix fails to select
the desired source. I had to use the commandline `mixer' and use
`=rec line' -- I think aumix tries to use the equivalent of `+rec'
which, with this card, always resets the source to mic. However, I
never actually got any sound from the line input to the recorded file
no matter what I did. I think I used this card a couple years ago
under Linux for audio input.
Now, on to -current and newpcm:
No card resulted in a clean clear recording the way most/all did under
3.4. Here are the results of each:
* The SB 128 PCI recorded a sound with a continuous crackly static in the
background of audible passages, the level of this depended on the audio
level and the static was inaudible with silence.
* The PCI 16 had almost a clean sound, so I don't think I botched the
compilation of the audio recording program too much. It was not totally
clean, with a repeating burst of static every few seconds, kinda like
a vinyl record with a bunch of dirt in a stripe across the grooves.
The CS4235 card performed (or didn't) as above, with the difference that
the recorded silent file seemed to have a bit of static-like noise.
* All the SB 16 cards -- the AWE 64, the Soundforte SF16, the SB 16,
and the Vibra 16 sounded AWFUL. The recorded .wav file sounded like
a chainsaw on overdrive, with no evidence of the audio source. It's
real kewl, but somehow not quite exactly what I'm looking for.
There are a few minor notes about the performance of aumix with each
card, but they pale in comparison with the dirty audio I recorded.
I can probably get my mitts all over a Gravis Ultrasound card that
worked ... interestingly ... under 3.4 Voxware, but I never tried it
under -current newpcm.
My conclusion is that nope, I can't toss -current onto a production
machine with my at-hand selection of soundcards for recording, and
that I'll have to stay with 3.4 for any sound work for now.
The only problem irking me with 3.4 is that all the ISA Soundblaster 16
cards that I have only record a mono sum signal, which I never caught
in many weeks of production use. Argh.
Neither -current nor 3.4 had playback problems with the supported cards.
Now, I bet I should followup to this one card at a time on the multi-
media list, no?
barry bouwms,a, tle danmark internet (a la Reply-To header)
--
*** This was posted with the express permission of ***
******************************************************
** HIS HIGHNESS KAAZMANN LORD AND MASTER OF USENET **
******************************************************
********* We are simple servants of his will *********
To Unsubscribe: send mail to majordomo@FreeBSD.org
with "unsubscribe freebsd-current" in the body of the message
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?Pine.BSF.3.96.1000317195047.77771B-100000>