Re: [Banshee-List] Crossfading and BPM detection

[Dominik Schnitzer <dominik schnitzer ofai at>]

2009/2/18 Neil Loknath <neil loknath gmail com>:
> Crossfading would be a great feature to have!  Especially when playing songs
> with similar tempos and tonalities.  The transitions could sound really
> nice.  I'm looking forward to trying out the BPM detection.  Will build from
> trunk later tonight.

Hi Neil,

Correct BPM detection is very difficult -- the errors a BPM detector
makes are often quite fatal.

But just try it out yourself: Listen to music, start tapping the beat
untill it fits, now double (or half) your tapping-speed -- your
tapping-beat will still fit to the music piece. Thats the typical
error a BPM detector makes - e.g. 180 BPM instead of 90 or 90 BPM
instead of 180 - These fatal errors make it very hard to integrate
tempo in a music similarity algorithm - that means usually results get
worse...

Apart from that a nicely working tempo descriptor could be used to
maybe postfilter the playlist or rerank the top 10(?) playlist items.

> From what I have read, Mirage's similarity detection is based on timbre and
> totally ignores tempo.  Therefore, it might be worth while to collaborate
> with the Mirage project and work on some new features.

definitely! we're looking forward for some new approaches! I didn't
know about the camelot wheel, that looks interesting to play with.

>> I'd be interested to know more about tonality. Can it be determined
>> automatically ?

Hm. you can try guessing the tonality by using this approach:
http://ismir2004.ismir.net/proceedings/p018-page-92-paper164.pdf
but they've only tested it with classical music. I'm sure they
published something better on another ismir conference. I can ask a
colleage of mine tomorrow if you're interested.

greetings,
domi.

-- 
DI Dominik Schnitzer
Austrian Research Inst. for Artificial Intelligence
Freyung 6/6, A 1010 Vienna, Austria, Europe
Fax: +43-1-5336112-77(Fax)