In last days I've thought a lot about server mixing and direct mixing approaches pro/cons.
I'll resume: server mixing pro: - cheaper (between 1.5 and 3 times, excluding the added context switch) direct mixing pro: - last minute mixing is feasible Taken for granted this resume I think to have found a solution to permit to have all the pros on one easy solution. The approach choosen is of course server based (otherwise it's not possible to get rid of repeated saturation). The server process is constantly waiting for hw interrupts (one each P frames). An interrupt happens and the hw pointer is H: the server saturates an area of the mixing buffer content and transfer it to hardware. The area saturated and transferred is from the last point done (let's call it D) to H+K*P, where K is a very small positive integer (ideally 1, but the true value is dependant by interrupt latency). Now D=H+K*P At the same time the server silences the area of mixing buffer just played (then we might also remove the extra work for silencing hw buffer in kernel space). Now we reach the interesting part: what happens if a client has just added a sample at a position X < D? Two choices: 1) it saturate and transfer to hw buffer itself (like in current pcm_dmix) 2) send a token through an interprocess communication channel with server (present in server poll list) saying 'Sorry, I'm late: please resaturate and transfer the area starting at X'. It should work... Comments? -- Abramo Bagnara mailto:[EMAIL PROTECTED] Opera Unica Phone: +39.546.656023 Via Emilia Interna, 140 48014 Castel Bolognese (RA) - Italy ------------------------------------------------------- This SF.net email is sponsored by: SlickEdit Inc. Develop an edge. The most comprehensive and flexible code editor you can use. Code faster. C/C++, C#, Java, HTML, XML, many more. FREE 30-Day Trial. www.slickedit.com/sourceforge _______________________________________________ Alsa-devel mailing list [EMAIL PROTECTED] https://lists.sourceforge.net/lists/listinfo/alsa-devel
