Subject: [linux-audio-dev] Re: Audio engine
From: Benno Senoner (sbenno_AT_gardena.net)
Date: pe syys 24 1999 - 13:03:58 EDT
On Fri, 24 Sep 1999, Juhana Sadeharju wrote:
>>
> Hmm.. I looked that greedy algorithm paper:
> Efficient Editing of Digital Sound on Disk
> Vol. 32, Number 6 pp. 394 (1984)
> Author: Curtis Abbott
> Abstract: Most professional applications of digital audio require a large
> number of channels to be available, but do not require extensive editing
> of the timing relationships among the channels. For these applications,
> digital tape technology is an appropriate media
>
> It is from 1984 but it is quite good. The author works at Lucasfilm, so,
> if the system is good for them, it is good for us. I will scan it in
> for you to read if you don't have a local technical library there.
Would be nice if you can put it online somewhere.
> They have another paper written by J.A. Moorer at about the same years.
> It gives an overview of the audio system used there and designed by Moorer.
> Moorer is founder of Sonic Solutions (Sonic Studio manuals are available
> through my Manuals page at GASP and gives a quite good view to the system)
> and is pretty well-known in audio research.
>
> They talk about low level disk usage (can a Linux hard disk driver know
> about where the disk head rotates or is it buried in deep to the hard disk
> device?) and buffer fillings.
we can't make disk rotation assumptions etc, since we must deal with regular
files, but using smart buffering, we get nice performance,
60 tracks out of a 5200RPM disk is certainly not that bad.
I don't think that the algorithm proposed in the paper would bring many
benefits, since we aren't allowed to use the disk directly.
But I' certainly worth to take a look at the paper.
Using my simple sorting algorithm you get varispeed for free.
regards,
Benno.
This archive was generated by hypermail 2b28 : pe maalis 10 2000 - 07:27:12 EST