Re: [linux-audio-dev] Problem and Solution: time

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

Subject: Re: [linux-audio-dev] Problem and Solution: time
From: Benno Senoner (sbenno_AT_gardena.net)
Date: su tammi  09 2000 - 20:02:43 EST


On Sun, 09 Jan 2000, Paul Barton-Davis wrote:
> > with 96 kHz the maximum duration is 12 hours - ok
>
> my US$0.02...
>
> imagine an arts festival that wants to use the legendary uptime of
> Linux to run a multi-day visitor-driven audio exhibit that has a
> "history" stretching over the time of exhibit.
>
> 12hrs - not ok.
>
> --p

Paul I agree,

I'd opt for 64bit integers for absolute sample positions, but sometime we need
only 32bits for expressing relative positions,
since 64bits integers (long long) on 32bit machines are much slower than 32bit
ints , my idea would be to use a combination of both if possible.
With 64bit CPUs I see no problems to work with 64bit ints all the time.

Of couse we could use double floats too , but I think for sample accuracy
we should avoid them due to potential rounding errors:
1.999999999999999999999 could be interpreted as 1 and not as 2,
therefore causing a drift of one sample which could indroduce nasty sideeffects.

Benno.


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 - 07:23:26 EST