Essentially, foobar2000 seems to be reverting back to 16-bit output data format whenever I restart it and randomly when I change between source type (stream, file, etc.) or even different sets of tracks (eg, RATM Renegades and Daft Punk Discovery albums ripped to identical settings). I can tell this because the Audigy series hates dealing with stereo audio less than 24-bit when CMSS is enabled, where the audio sounds like 8-bit quality. It does not show that it has reverted in the preference settings, but choosing a different bitage then back to 24-bit (or greater) noticably repairs the sound.

Although obviously nobody was too concerned with this, I think I've got it down to poor drivers (typical from Creative apparently...). I still have to set the bittage to 24-bit+ when CMSS is enabled, but with the latest beta drivers (was previously on latest production drivers) it doesn't randomly screw up the sound quality or volume.

Just wanted to resolve this incase anybody found it on Google or something.

I'd guess people were concerned, just either unable to reproduce or completely lost as to whatever might be going on (or both at once).

Thanks for reporting and finding a way to remedy this to some extent.

Urgh. Okay, don't take my advice (the newest drivers are still broken). New suggestions if you are having these issues:

1. Resample all your audio output down to 48 kHz or lower, with 16-bit or lower resolution (optimally 48/16), or
2. Use kernel streaming (which has disappeared in 0.9?) to bypass Creative's hacks into how Windows handles sound. CMSS1 won't work, but you'll be able to use 96/24 without issue, or
3. Sell your card! I've heard X-Fis are good, but that could just be marketing; I am doubting Creative has improved their software.

The 'degradation' I heard earlier is perhaps a 'good thing'; I only have headphones, so perhaps the output from that socket has reduced fidelity or something for front speakers (which I presume are handled by that socket). Only thing is, the degradation isn't consistent above 48/16; every second asynchronous stream (eg. if you play through a playlist, every second track) has a different volume (and possibly sample rate and/or resolution) and breaks CMSS1.

Why use 24/96 or resampling in the first place?

http://foobar2000.org/FAQ.html#FAQ_o5

Perhaps I want to listen to high quality audio sometime? I'd rather my setup is up to the task. Ignoring foobar2000, things like DVD Audio crop up, and need to be dealt with. I also have deluded myself into thinking the Audigy 2 performs best at 24/96, and that SNESAmp's 24-bit samples at 96kHz actually sound better.

I tried a kernel streaming plug-in with Winamp, and that fixed the troubles as well. I've come to learn that the blame for this mess is perhaps better aimed at KMixer?

You want high quality audio. Fair enough.

But forcing 24/96 output does not guarantee high quality audio. Theoretically, it can only be worse if your audio is not already 24/96.

So don't force 24/96 output if you want high quality audio.

Besides, delusions, by definition, are not worth having. And statements such as "it sounds better to me" need to be backed up with proof at this forum, in order to be taken seriously.

You are misunderstanding what I am trying to say. Of course forcing 24-bit output from 16-bit sources isn't going to make anything sound better, but that is not the issue. Also, I'm not trying to suggest to anyone that what I am saying is going to make all Audigy users have better sound; I'm just remedy some issues (ie. obvious sound degradation) that people might have (this problem in general isn't exclusive to me, but initially I thought it was something wrong with foobar2000 [whereas, I presume, others have been more wise]).

Edit: Extra clarafication: I have these issues with true 24/96 pieces of audio as well. It is nothing to do with foobar2000's resampling.