Hiya

Whats the most reliable way to guess the number of beats in a loop loaded in?

I have got this working by the way by dividing the length(ms) of the loop by the metro frequency*4 and putting the result through int to get nice round beat numbers. This works to a point, except obviously when the tempos are different. I wanted Pd to work it out, without me having to parse the tempo from a filename. (i.e 128_drumloop.wav)

I'm sure this is possible - how do I do it?

thanks

the way i did it was kinda hackish, but worked most of the time.

my logic is that almost all the loops i have will be in the 100-200bpm range.

..which means that each beat is going to be between 300->600ms

soundfiler will give you the length of the file loaded in bytes, so divide this by 44.1 (or whatever samplerate you use) to get the length of the whole loop in ms.

then, divide the total time by 2...if the new time is between 300->600ms, output 2 as the number of beats...if not, divide the total time by 4..if the new time is between 300->600ms, output 4 as the number of beats...if not, divide the total time by 8...etc

this does assume 4/4 loops with 4,8,16,32... beats, but that's all i have anyway, so it works well

excellent cheers mate

what kind of automatic function do you use to output the right number?

@hardoff said:

my logic is that almost all the loops i have will be in the 100-200bpm range.
l

...this •DEFINITELY• wouldn't work with my beats. =-)
I get as close as 96, but something around 72 is ideal.

>what kind of automatic function do you use to output the right number?

i don't have access to the original patch right now (it's on the hard disk of a broken laptop..along with a lot of my old stuff...aaarrrgghh!!)

but basically, it was a long chain of [moses] objects, connected to message boxes...

[inlet] <- complete loop time in ms
|
[t b f]
|
[moses 600]
| /
[1 ( [moses 1200]
| /
[2 ( [moses 2400]
| /
[4 ( [etc..etc..]

[f ]
|
etc...

where the message boxes [1 ( , [2 (, [4 ( etc, all connect to the right inlet of the float at the bottom, and then the original [t b f] is connected to its left inlet

@hardoff said:

>what kind of automatic function do you use to output the right number?

i don't have access to the original patch right now (it's on the hard disk of a broken laptop..along with a lot of my old stuff...aaarrrgghh!!)

but basically, it was a long chain of [moses] objects, connected to message boxes...

[inlet] <- complete loop time in ms
|
[t b f]
|
[moses 600]
| /
[1 ( [moses 1200]
| /
[2 ( [moses 2400]
| /
[4 ( [etc..etc..]

[f ]
|
etc...

where the message boxes [1 ( , [2 (, [4 ( etc, all connect to the right inlet of the float at the bottom, and then the original [t b f] is connected to its left inlet

Ah, the holy moses. I never seem to end up using this humble object enough. I managed to do the above, but using [>=300] ,[<=900] , [&&] a bunch of [select], message boxes and a counting [until] loop. I think yours is far more elegant and less costly on the cpu

cheers