#7 quick audio degradation

[Bob]

Installation went smoothly, and program seems to function properly at first (including all graphic elements). But when I create a very simple sound patch like a sine wave, the audio starts ok, then quickly degenerates into a cyclic, choppy mess. Very strange. Once the problem manifests, it will not go away. No crash involved.

Have no clue how to troubleshoot this. I run many sorts of audio programs with this machine (e.g., Steinberg products including Cubase4), and am surprised by this behavior. I'm guessing it is some sort of driver issue.

System details:

HP pavillion laptop; 3.2 GHz P4; 1.12 GB RAM; WinXP Home

When OSW starts, it finds Midi Yoke and ASIO4ALL v2.

I can provide screenshots and whatever other necessary details. (will try to upload a screenshot now.)

Many thanks in advance!

[Bob]

Screenshot of OSW start

[Bob]

Logged In: YES
user_id=2188399
Originator: YES

I have experimented with many variations of settings with ASIO4ALL (actually, it is v2.8); still have the same problem. A simple sinewave patch at whatever frequency will start nicely, but quickly degenerates into a slow series of click noises (within 20 seconds more or less, depending on settings).

Here is the output of a little utility, for what it's worth:

A4AInfo v1.0 (c)2003-2005, Michael Tippach

============================
Start WDM Device Enumeration
============================

---------------------- 8< ----------------------
Device Path: PCI\VEN_1002&DEV_4341&SUBSYS_006B103C&REV_00
Device BUS NUMBER = 00494350
SoundMAX Integrated Digital Audio / Analog Devices, Inc. / smwdm.sys
Instance ID: 00000800
+SoundMAX Digital Audio
NumPins: 00000005
+Pin 00000000 Dataflow: 00000001
MaxChannels: 00000006
MaxBits : 00000010
MinSampleRt: 00000064
MaxSampleRt: 0000BB80
+Pin 00000002 Dataflow: 00000002
MaxChannels: 00000002
MaxBits : 00000010
MinSampleRt: 00001F40
MaxSampleRt: 0000BB80
+Pin 00000004 Dataflow: 00000001

---------------------- 8< ----------------------
Device Path: root\swenum
Device BUS NUMBER = 00000000
Plug and Play Software Device Enumerator / (Standard system devices) / swenum.sys
Instance ID: 0000087C

============================
Exit WDM Device Enumeration
============================

[Amar Chaudhary]

Logged In: YES
user_id=315707
Originator: NO

I have unfortunately been seeing this behavior lately, particularly with multi core (or hyper-threaded before that) processors. Usually, it does come and go, though - i.e., if you turn on and off audio from the AudioOutput transform, it may come back. I have much more problem with this on my MacBook than I do with my older PC laptop (2.4 GHz P4).

Also, does it change or go away when you turn up the buffer size?

These aren't solutions, just trying to figure out if it's the same behavior I have seen. It's not a driver issue per se, but probably something in how OSW accesses drivers (through PortAudio).

[Bob]

Logged In: YES
user_id=2188399
Originator: YES

Many thanks for the acknowledgment! I sure hope you will have time to discover and fix the problem because it looks like this program could be a lot of fun.

I tried using the newest ASIO4ALL 2.9(beta2), just in case that might make a difference, but noticed no change in behavior.

Any differences due to buffer settings did not seem significant enough to be noteworthy.

In all of about a dozen attempts, once it failed, it would not recover. Toggling the AudioOut transform was the first thing I tried to do in all cases. No luck.

Just in case it is relevant, the machine's more precise model # is: zv5330us

Thanks again for your attention!

[Bob]

picture of test in a track

[Bob]

Logged In: YES
user_id=2188399
Originator: YES

File Added: sineTestGraphic.JPG

[Bob]

picture of test patch

[Bob]

Logged In: YES
user_id=2188399
Originator: YES

File Added: sineTestPatch.JPG

[Bob]

Logged In: YES
user_id=2188399
Originator: YES

I uploaded a a couple more files to illustrate the problem.

sineTestPatch is a picture of the test patch I made. The gain is set up to give some attenuation from unity.

sineTestGraphic is how that same testWav looks in a track. The scale at the top is time in seconds. In this instance, the sound continues for about 13 seconds before it starts to spew at two-second intervals. When I zoom in on the chunks, they are actually pretty well-formed, just very brief.

Even though I was running everything very soft, I managed to clip the recording a bit (!). Used Windows Sound Recorder in the background--haven't done *that* in years! :>)

I tried to upload an mp3 of the test sound, but the file is too big for SourceForge. :>(

Hope this helps to define the symptoms a little better!