Re: [linux-audio-dev] Re: Cached Audio Patent

New Message Reply About this list Date view Thread view Subject view Author view Other groups

Subject: Re: [linux-audio-dev] Re: Cached Audio Patent
From: David Olofson (david_AT_gardena.net)
Date: ma helmi  28 2000 - 21:01:05 EST


On Mon, 28 Feb 2000, Juhana Sadeharju wrote:
> Another feature could be equally important: the beginnings should stay
> in the memory even the audiofile is playing --- somebody might retrigger
> the same audiofile. This we could circumvent by keeping the same file
> multiple times in memory (but wastes memory).

Wastes so much memory that this becomes virtually useless, I'm
afraid... Imagine a direct-from-disk sampler, playing a 1 GB grand
piano. Now, to get more than some ridiculous 10-20 voice polyphony,
quite a lot of buffering is needed per voice, to cut seek overhead
when streaming the data during playback. This is a lot of memory
already! Multiply this with the number of different samples (splits
and multisamples), and you'll get figures that invalidate the whole
idea of direct-from-disk sample playback. I certainly hope this won't
be required...

> Some random hints:
>
> I checked Amiga manuals but I don't have access to Amiga 3000 manuals
> which had a hard drive by default. Anyone?

Whan are you looking for? (Amiga 3000, 25 MHz, 50 MB HD somewhere
under a pile of computer hardware...)

> Sound Forge manual tells something about preloading, details anyone?

Hmm... Yes. AFAIK, you can configure SF3.0 and later (IIRC) to load
and prepare on NoteOn, and then start playback on NoteOff. Cheating,
that is... Don't know if this has changed. (I have a 4.0 XP that
came with Layla, but no manual... :-( Could dig some in the help
files, though.)

> I looked also at radio jingle players without success.

Hmm... One guy where I work is voicing for a radio station, and he's
also quite into the Windows based broadcast software they use. That
application plays and mixes mp3s in real time, and it's quite
responsive even with Win98 and a normal sound card. (It works best
with the vendor's own special audio cards with their own DSP power.)

Further, it has a number of slots for jingles, sound FX etc. These
can, AFAIK, be played instantly at any time, without warning the
application beforehand. I'm not sure if they're preloaded in their
entirety, or if an algorithm of the kind we're looking for is used.

> I think we should forget this and just do what we want.

Yep...

//David

.- M u C o S --------------------------------. .- David Olofson ------.
| A Free/Open Multimedia | | Audio Hacker |
| Plugin and Integration Standard | | Linux Advocate |
`------------> http://www.linuxdj.com/mucos -' | Open Source Advocate |
.- A u d i a l i t y ------------------------. | Singer |
| Rock Solid Low Latency Signal Processing | | Songwriter |
`---> http://www.angelfire.com/or/audiality -' `-> david_AT_linuxdj.com -'


New Message Reply About this list Date view Thread view Subject view Author view Other groups

This archive was generated by hypermail 2b28 : pe maalis 10 2000 - 07:23:28 EST