stereo should be enough for anyone?
Niklas Wallen <[email protected]> Tue, 9 Apr 2002 23:43:09 +0200
| Newsgroups | gmane.comp.gnu.octal.devel |
|---|---|
| Message-ID | <[email protected]> |
Hey I was playing around and changed the mono/stereo limit to OX_MAX_BUS_WIDTH (6 or 8 seems enough) channels using a simple mapping for connections. I'd like to hear your opinion about it.. lin,rin,lout,rout in machine and lbuf,rbuf in mixer became pointer arrays like samp *in[OX_MAX_BUS_WIDTH]. The mapping is stored in the input struct as an array of channel numbers. The number of channels used for a connection is determined by the number of input channels of the destination machine, so input on channel k comes from output channel input.channels[k]. To receive nothing on a channel, use the value OX_CHANNEL_NONE. Yeah, I admit a better name should be used since 'channel' belongs to another concept. Anyone with better imagination? A machine with 4 input channels for two stereo inputs could use mappings like [0,1,x,x] for one input machine and [x,x,0,1] for the other. A connection from a mono machine to a stereo machine could have a mapping of [0, 0]. Stereo to mono can only be either [0] or [1] but I'm thinking about a way to specify several channels at once. Would it be sufficient to have a 'stereo bit' to simply mix channels k and k+1? A mixer machine could always solve the more general problem of n to 1 mappings. How will panning work with this scheme? A gain for each channel? Is this interesting or are you all satisfied with stereo? (I'm gonna do it anyway, hehe) -- Niklas W