next »

[Hamilton]

My question concerns the way sound cards (and their drivers) internally route signals by default.

I would like to be able to use any standard standard audio application like Audacity, VLC media player or an internet streaming service, as a source of audio and read the stream into my own software in real time. I would like to process it using my own algorithms, then send the output to multiple analogue output channels in real time. (Many sound cards these days are capable of producing 5.1 or 7.1 i.e. up to 8 discrete channels). BASS appears to have all the tools available for me to do this.

The thing that worries me is that the standard mode of operation of sound cards I have used, is for a 'wave' source signal to be automatically routed to two of the analogue outputs by default. This certainly is the case for the Creative Audigy 2 ZS that I have been experimenting with. It is only because of the open source Kx driver project that I have been able to disable this routing, allowing my application to read the wave stream and send a processed version to six independent outputs. For audio quality reasons, if I upgrade to the Creative X-Fi range, for example, can I be sure I will be able to do the same thing?

Or is there something in BASS that can always work around this problem?

[Ian @ un4seen]

I think the only way you would be able to intercept the audio and do what you want with it would be via a virtual soundcard driver, ie. you would set the audio application to use the your virtual soundcard, and you can then do what you want with the received data.

[Hamilton]

Thanks Ian. At first glance, wanting to apply live software processing to the Wave stream doesn't seem all that outlandish a thing to want to do. As I say, it is possible on the Creative Audigy, but only because of the open source Kx driver. The 'problem' all seems to boil down to the way the record and play sliders are configured in the Windows mixer applet i.e. record slider comes after the playback slider, for some strange non-intuitive reason (that has confused me many times over the years, I think!).

I have no experience of virtal sound cards, but I have a feeling that it isn't necessarily as good as using a real sound card, as it is not completely obvious that its sample clock would be locked to that of the card doing the playback. I would envisage that BASS would be configured to use the Virtual sound card as the recording source, and the real sound card as the playback destination. Record and playback sample clocks would not be linked. If so, the playback buffer would empty or fill up over time, as the sample clocks drifted. By using the sound card for both the Wave destination and for playback, the sample clocks are perfectly locked. Or there may be tricks that get around this..?