Well, I'll certainly follow up on the ASIO4ALL driver. Thanks for tracking
this.

I'll diff 8/14 and see what I did wrong, too.

Original Message:
-----------------
From: Art Hunkins abhunkin@uncg.edu
Date: Wed, 15 Sep 2004 15:28:29 -0400
To: csound@lists.bath.ac.uk, gogins@pipeline.com
Subject: Re: [Csnd] Followup: Realtime Csound5 Windows Audio Test


Michael,

Bingo! How silly of me to have forgotten -B and -b. This was the big
problem. Scratch all my "two-clicks-per-second" results that were reported.
Sorry about that - and thanks for getting me back on the right track.

So, some new results. I compared, on both my Gina and stereo soundcards,
your last three Windows distributions: 2004-08-14, 2004-08-16, and
2004-09-13.

I get equally good ASIO realtime sound from *both* ASIO-only (08-14 and
09-13) distributions. With the last of these, as you mention, there is the
problem of system lockup when the run is interrupted by CTRL-C, and
sometimes even when the run completes (I *believe* only with the ASIO4ALL
driver). 08-14 has neither of these problems, and I'd suggest using it to
solve the current ASIO issues.

2004-08-16 (with MME and DirectSound as well as ASIO) gives poor sound for
me with *all* drivers, and even locks up the system before play with MME.
OTOH, all cut off fine when interrupted (if they play at all!) The dastardly
poor quality is also related to the problem that the run never ends; it
*must* be terminated via CTRL-C.

I hope you can take a look at the http://www.ASIO4ALL.com ASIO driver.
Whereas it worked fine *for my stereo card* in CsoundAV (WindowsME), it
gives extremely unacceptable sound output when used with CS5, any of the
above distributions.

It is unclear to me what, if any, interaction the ASIO4ALL control panel has
with the buffer settings within Csound (-b, -B). I don't *think* any, and
I've set the control panel's settings as high as possible to permit the
greatest latency (and so the best possibility of uninterrupted sound). But -
whatever - the sound is terrible.

Since ASIO4ALL is the only ASIO driver available for my stereo soundcard
(and I suspect many others), things do not look good for stereo cards and
CS5. I'd been very hopeful following my wonderful CsoundAV (and for that
matter, CS4) experience.

And things work so well with ASIO on my Gina multi-channel card - just as
well with ASIO4ALL and "native" ASIO.

If anyone else has had experience with CS5 (Windows) and various ASIO
drivers, including ASIO4ALL, I'd appreciate your sharing it. Incidentally,
the http://www.ASIO4ALL.com website has a lot of good, user-oriented
technical info for those of you who can understand it. I think I even
understand *some* of it. (But I'm not sure.)

Art Hunkins
----- Original Message -----
From: 
To: 
Sent: Monday, September 13, 2004 11:55 AM
Subject: Re: [Csnd] Followup: Realtime Csound5 Windows Audio Test


Your .csd doesn't set -B, -b, and ksmps all to the same value, so it can't
work. This is a shortcoming in Csound 5 beta that will be fixed later.

Original Message:
-----------------
From: Art Hunkins abhunkin@uncg.edu
Date: Mon, 13 Sep 2004 11:58:21 -0400
To: csound@lists.bath.ac.uk, gogins@pipeline.com
Subject: Re: [Csnd] Followup: Realtime Csound5 Windows Audio Test


I'm stymied!

Just tried the new latest CS5 Windows version that Michael posted
(2004-09-13). Same results as before: nothing but a short click twice per
second, using my attached test file. One additional problem, however: in
contrast to the previous version, this one does not "clean up" normally at
the end. Instead, it freezes and requires a complete system reboot.

I will install this CS5 on several of my other Windows machines (98SE and
ME) to see if I get similar results.

Michael, I wonder if you have something installed on your system that I
don't have on mine, and that would make a difference in sound output with
CS5, compared to CS4 and CsoundAV?

Meanwhile, I'd appreciate anyone else running Csound in Windows to try out
my test file with Michael's most recent CS5 version, and report the results.

To run this version, you need an ASIO driver for your soundcard. If you
don't have one, but do have a WDM driver (in Win98SE or higher), you can
download the free one at http://www.ASIO4ALL.com . The attached .csd file
runs in real time (-odac0), but requires no MIDI or FLTK. It should output a
(mono) 440Hz sine wave for 30 seconds, rather than the periodic clicks I'm
getting.

Art Hunkins

----- Original Message -----
From: 
To: ; 
Sent: Sunday, September 12, 2004 4:05 PM
Subject: Re: [Csnd] Followup: Realtime Csound5 Windows Audio Test


I made a small change in InOut/pa_blocking.c which I have committed in
Csound 5. On my Windows XP Home Edition system using the M-Audio Ozone for
both MIDI input and ASIO audio output, I can perform on the keyboard with
acceptable latency and without the clicks you mentioned. Both channels
work, too, and the stereo image is correct.

There is still a problem with stopping a performance in the middle; that
leaves a hanging Csound process after which real-time audio does not start
properly again. I am working on this.

Later today, whether or not I fix this stop/restart problem, I will upload
a new file release.

This is ASIO-only, but I may also find time to experiment with putting the
other drivers back into the build.

Original Message:
-----------------
From: Art Hunkins abhunkin@uncg.edu
Date: Sun, 12 Sep 2004 13:51:05 -0400
To: csound@lists.bath.ac.uk
Subject: [Csnd] Followup: Realtime Csound5 Windows Audio Test


Michael,

I did further checking that compared the audio output from your *two* last
CS5 Windows versions - using the same test file under the same conditions.

My results were the same for your lastest version as for the previous (ASIO
only) one. All the drivers on both my stereo card and my Echo Gina
(multichannel) showed up as expected; and all performed with the
"two-clicks-per-second" audio output I previously described.

A bit puzzling was that this "click rate" was perhaps a *bit* faster for
some drivers than others - I couldn't find a reason for this. Perhaps
related to this (and perhaps not!), the playtimes for the 30" file also
varied somewhat: from a low of about 12" to a high of about 22". Again, I
couldn't make a pattern of it. Most were right at 17" however.

My personal conclusion is that at least ASIO seems to be (mal-)functioning
the same in the two versions.

I did have a problem establishing, with your most recent version, which of
the driver outputs were MME and which were DirectSound. (The ASIO were
always clearly marked.) For example, my simple stereo card marks its output
devices thus: MS Sound Mapper, Avance Sound - 41K, Primary Sound Driver -
48K, Avance Sound - 48K, and ASIO4ALL - 41K(!). In that order.

Should I assume the first two to be MME, and the second two, DirectSound? (I
was also surprised to find the ASIO4ALL driver marked 41K; it usually shows
up as 48K). Is the 48K a key that the driver is not MME?

Art Hunkins

--
Send bugs reports to csound-bugs@cs.bath.ac.uk
              (or to http://www.cs.bath.ac.uk/cgi-bin/csound )
To unsubscribe, send email to csound-unsubscribe@lists.bath.ac.uk

--------------------------------------------------------------------
mail2web - Check your email from the web at
http://mail2web.com/ .


--
Send bugs reports to csound-bugs@cs.bath.ac.uk
              (or to http://www.cs.bath.ac.uk/cgi-bin/csound )
To unsubscribe, send email to csound-unsubscribe@lists.bath.ac.uk


--------------------------------------------------------------------
mail2web - Check your email from the web at
http://mail2web.com/ .


--
Send bugs reports to csound-bugs@cs.bath.ac.uk
              (or to http://www.cs.bath.ac.uk/cgi-bin/csound )
To unsubscribe, send email to csound-unsubscribe@lists.bath.ac.uk


--------------------------------------------------------------------
mail2web - Check your email from the web at
http://mail2web.com/ .