Re: MP3 Player Shufflin' Along

On Fri, 16 May 2008 05:46:52 +0100, John Hatpin wrote:

>Hactar wrote:
>
>>Pseudo-random algorithms are hard to do well.
>
>While that's true, it should be reasonably possible for a modern MP3
>play to play songs so much at random that should satisfy anyone. Yeah,
>lots of "should" in there, and this is the real world, so maybe it
>doesn't always happen.
>
>It's not as if the thing needs to fool a mathematician in a kind of
>Turing test - all we're asking for here is a roughly even distribution
>coupled with testing to see if a track's already been played
>"recently" and a memory for "recently" long enough to convince the
>user.

I wrote my own MP3 player program (in VB), and it allowed the user to
select a programme period (I had it set to 4 hours), with no track
being repeated within that period except in one special cir***stance.
It would also, on occasion, decide to play up to five consecutive
tracks by an artist (my setting, 10%), or an entire album (5%).  This
last is when a track could possibly be repeated within the programme
period.  How likely these special events were to occur was selected by
the user.

I found it very agreeable like that.

For a while I also had an auto-DJ telling me what was being played,
but in the end I couldn't be bothered recording the announcement
files.

-- 

Peter

I'm an alien
email: groups at asylum dot nildram dot co dot uk
Well, since the gene itself is entirely hypothetical, we can make that
part up.
- Sean Houtman