Accuracy of MixMeister BPM Analyzer

jonp · Post by **jonp** » Sun Jan 06, 2008 8:23 am

I used MixMeister BPM Analyzer to add BPM to most of my MP3s. This worked without a problem although I think the results are questionable. One of many examples I could use is: from Coldplay's Parachutes album MixMeister BPM Analyzer tags the song Yellow as 87 BMP and the song High Speed as 154 BMP. To me there is not nearly twice the BMP, almost no difference.

Is MixMeister BPM Analyzer any good? I was hoping to use BMP to create fast playlists for use at the gym and slow playlists for chilling out etc etc. It seems BMP is no good for this?????

Others thoughts/comments?

Big_Berny · Post by **Big_Berny** » Sun Jan 06, 2008 9:25 am

AFAIK you can't just say that 150 has double speed than 75. Because there are different beats in a song and it's difficult to detect which is the correct one. That's way two songs with the same speed can get 150 and 75 because it's just divided by 2. But what you can say: If a song has 50 and another one has 80 then you know that they have a different speed.

Don't know if I'm right though...

EDIT: It's useful for sports (like jogging) anyway because if the beats is doublespeed it's no problem. It's only a problem if the beat is for example 40% faster. Then you'll probably change your speed.

EDIT2: Here are some other tools you could test...

latinmusiclover · Post by **latinmusiclover** » Sun Jan 06, 2008 4:08 pm

I used Mixmeister BPM to give me some context for about 100,000 tracks. It was to create a useful database for dancers so that they had a fixed # instead of the subjective "Fast" or "Slow". I had done the "MPM" (measure or bars per minute) for a lot of them already so I had something to compare with. Of the tracks I compared, almost all were accurate with a beat or two, and most errors were obvious -- the BPM was usually 1/2 or double (or triple) the expected range. The following conditions seemed to confuse the software:
1. Along and different speed part at the beginning or middle of a track.
2. Medley songs in which there were different musical styles and rhythms
3. Tracks with heavily syncopated beats and broken rhythms.

bluesbeat · Post by **bluesbeat** » Sun Jan 06, 2008 7:25 pm

I suspected it was less accurate with vbr files, but I haven't gone further to prove it.

Big_Berny · Post by **Big_Berny** » Sun Jan 06, 2008 7:48 pm

Sorry, forgot the link: http://www.mmartins.com/mmartins/bpmdet ... ection.asp

jonp · Post by **jonp** » Tue Jan 08, 2008 7:21 am

Thanks for the replies. I did some more reseach and IMHO BMP is not a reliable metric to use to create 'fast' or 'slow' playlists.

I am going to give MusicIP Mixer a go and see if its ability to select 'similar' music will work.

Any feedback on MusicIP Mixer?

duckcubbie · Post by **duckcubbie** » Fri Jan 18, 2008 12:48 pm

Hi Jonp

If you ever find one that comes close to giving you "fast" and "slow" playlist let me know and I will do the same. I am looking for the same thing, especially for the gym!

martialartsguy · Post by **martialartsguy** » Fri Jan 18, 2008 1:17 pm

As far as the gym is concerned, I have a playlist based on Tempo, where Tempo is very fast and genre is either Metal or Hard (as in hard rock). I use this for jogging a lot.

I also have a playlist where Mood is Wild or Angry, and Tempo is Fast or Very Fast. All this requires a lot of tagging though, but that's the reason I tag my stuff.

I do find that MixMeister really underestimates very fast stuff - for example, Thrash Metal is over 200 BPM a lot of times, but MixMeister often cuts about 100 out of the BPM. I manually change it sometimes, but for the most part I use the Tempo classification.

Accuracy of MixMeister BPM Analyzer

Accuracy of MixMeister BPM Analyzer

Jonp if ...