I seem to have read on these forums that Allegro can (via DirectSound or Allegro's mixer) create upto 64 sounds at once for us to use.

Whatever I try to reserve any amount of voices, a subsequent check of digi_driver->voices (from memory I think that is right), brings up a measly 4 channels on my PC. If this is the number of hardware voices on my soundcard then this is still ok as long as channels are mixed before being sent to the card. However I do not seem to have mixing capabilites as playing 5 sounds resulted in the last sound (and the highest priority one) being skipped.

I know I've gone over my sound problems before, and I'm getting sick of it as well. But am I the only one who seems to have this very low amount of sound channels? Using Allegro 4.0.3 BTW..

Does anyone have an ideas/help for this one? I seem to recall Kitty Cat had an updated directX mixer? Any chance this will solve my problem?

Thanks!!

If you havent already, use the Allegro mixer, instead of the default DX mixer, which just passes sound to DirectX, and it gets to decide how to deal with things.

As for KCs changes.. CVS has the new mixer, but not all of the new code, that requires the old DX mixer to be dropped totally (might happen in 4.2, I hope it does, might get hardware mixing with all drivers instead of just the DX one)

Ah yes. I forgot to mention, the DX mixer ignores ALL of the sound parameters you give to allegro (another reason why it needs to die).

I've got it working! I tried the play command and discovered that all the mixer's only gave me 4 channels apart from the allegro mixer which gave me 8. When I tried that in game I got all 8 channels but only at about half the volume of the directsound mixer.

I couldn't seem to reserve more voices than 8, no matter what value I used for reserve voices.

I then reread your last post Kitty Cat and added a digi_voices = 16 to my allegro.cfg and it worked! I now can have 16 voices at full volume using the directsound mixer.

However I seem to have found a problem. Reserve_voices did nothing if I used DIGI_AUTODETECT which is not mentioned in the docs. So I then tried something else, I put DIGI_CARD = AXA in the config file and used DIGI_DIRECTX(0) in the program. When I recompiled and ran my game I could tell from the sound volume and from the logfile where I print the driver name that the Allegro mixer was still being used! I reversed things and used DIGI_DIRECTAMX(0) in the program and DXA in the config file and it again used the config file settings.

This means that for me at least either those listed in the docs (DIRECT_AMX(n) etc..) are wrong, or that the driver name is ignored if the config file entry is present. This is not explained in the docs.

So, reserve voices does nothing for me no matter the driver used, the driver seems to be autodetected from the cfg file anyway and so it seems I will have to distribute the cfg file to give people more voices than 4.

I've got it working, but I'm a bit frustrated at these errors or unclear explanations in the docs.

Yes, I called reserve_voices before install_sound. But even with no allegro.cfg file in my game directory it has no effect. Perhaps because it's looking in the Allegro directory?

I had a quick look at the sound.c source and it seems to work like the docs indicate. But that is not the result I get in my programs. I use Allegro 4.0.3 as well, which should be stable. Perhaps this only appears on a few PC's?

I may be able to try 4.1.14 later. I can't do it at the moment and since you are now passed out on the floor that won't help you either.

4.1.15 will default to using the software mixer. The DX mixer creates a large gap in the expected functionality between itself and Allegro's software mixer, creating problems like what you described.

The Allegro mixer doesn't sound bad. Just make sure you set quality to 2 (which is the default, iirc), a nice frequency (44.1khz is the default here), and 16-bit stereo (both of which are defaults as well).

In 4.1.14 and earlier, however, there are a few issues.. namely that is only maxes out at 12-bit mixing, 8-bit samples don't interpolate right, and the default volume per voice is extremely low. 4.1.15 fixes these issues however.. it has 24-bit mixing, both 16- and 8-bit samples interpolate at full 24-bit precision, and the default volume is half (so a full-panned sample won't clip).

Yup, it's gone. What it was supposed to do is take a 12-bit sample (with 4-bit overflow protection), clamp it to 12-bit and resize it to a full 16-bit value to pass to the driver. The reason for that was you couldn't mix two full 16-bit samples together without risk of overflowing the 16-bit buffer, so they used 12-bit mixing. However, now it does 24-bit mixing in a 32-bit buffer, so the only way it could overflow is if you played more than 256 voices simulatiously at full volume, and Allegro doesn't allow that many (only 64).

Actually, if you're static linking, that'd be the best time to do it. Don't have to worry about tincompatible DLLs.

For better or for worse.

Most people don't care so much about sound, or simply don't know how to make sound. So they don't get around to using it much (until near the end of their projects, occasionally), and thus the chance for bug reports to crop up is significantly lower than for the more oft-used modules.

I understood it was recommended practice that static linking with a stable library be used for release versions? Bear in mind this is a Shareware Game I need to make sure everything works as well as possible.

I'm looking at publishing in one months time.

Hmm.. I'm sure people said we should static link.. Well, even for a shareware game is including a .DLL standard practise? Seems somewhat unprofessional to me for some reason..