I noticed a very strange behaviour of fmod. After some number of files played one after one, fmod suddenly stops playing. That is strange, but it gets even more strange. If I restart the application it doesn’t start playing at all. So I tried the testbed-sample and it doesn’t start playing, too. FMOD doesn’t report any error, but the sound-position does not advance and it doesn’t output any sound. Only a reboot fix this error. I tried 3.71 and 3.71 and both have this error. It makes no difference if I use MMSYSTEM or DirectSound.
An old application of mine which uses fmod 3.6.0 has not this error and plays correctly.
I’ve tested on a pc with nforce2 (soundstorm) soundchip running Windows XP.
Currently I have no idea why this could happen. I compiled the testbed with Delphi 6 and my app with freepascal.
Sorry, I meant I tried 3.74 and 3.71. Today I’ll try to compile it with 3.6, too and try it also on another PC with a soundblaster card. The newest drivers are installed on the nforce2 platform, but that doesn’t mean there can’t be any bug in them.
I tried it with 8, 16 and other Hardware-Channels, but the problem still exists. I have no idea why or when it occurs. After a more or less short time (5 to 60 minutes) it stops playing and that even in the middle of a song. I only play streams and have a maximum of 2 streams at the same time. I use the dynamic loading of fmod and so I’ve not tested the app with the 3.6 version of fmod, but all apps using 3.7x cannot play any sound after this error occured until I reboot. But applications which use fmod 3.6 works without any problem. And I don’t get any error from fmod. You only notice this bug, because you don’t hear anything and the position in the sound doesn’t advance.
This sounds like dodgy soundcard drivers. On-board audio is almost invariably broken one way or another. I’ve been running FMod quite a bit on both Win32 and Linux platform running in VMWare using on-board audio, and it’s been a bit of a painful experience, having to reset/reboot VMWare after each FMod initialization and whatnot. Using a proper PCI soundcard, even a cheap one, should help solve this somehow, at least its done near-miracles for me =P. FMod itself has been performing very well under all sorts of weird configurations here (Win32, Linux on PC, Linux on VMWare, embedded linux on diskless IFE systems, etc.) so I doubt this problem is FMod related.
- symbiote answered 14 years ago
Yes, it could be that it is a bug in the soundchip driver of nvidia. But the funny thing is that after this error happened no program which uses fmod 3.7x could play a sound but an application using fmod 3.6 could play. I changed the sourcecode to use fmod 4 and it seems to work without that error. Now I’ll try to invoke this error with the 3.74 version and look if the version using fmod 4 could play after that or if it’ll also fail.
I used fmod since 2001 or so and it worked always very good, this is the first problem I have with it and only with 3.7x
I only play streams and no samples. After one stream finished I stop and close it and load the next one. Also the error does not necessary happen during the change of the stream, but also during playing it back. So I guess my load and unload functions aren’t the source of the problem. I’m also going to test if I could reproduce the error with one of the samples like testbed.
And because the WMP could still play after this error and also fmod 3.6 but fmod 3.7x not, there seems to be a little error (or just no work-around) in fmod 3.7x in combination with nvidia soundstorm.
EDIT: I’m currently testing with the old 3.74 version and doesn’t know how to reproduce the error. I know it happens sometimes, so I guess I have to wait. I’ll let my PC on this night and report to you tomorrow. Perhaps it has something to do with some corrupt files?
Please login first to submit.