Beat~ problems. BPM/Tempo Detection
Hello,
I'm having some trouble getting the "beat~" object to work, and it appears there are very few other ways of detecting BPM/Tempo from incoming audio without using a serious amount of FFT analysis and mathematics.
If anyone could suggest how best to detect BPM, or just how to get Beat~ to work, I would be most appreciative.
The main problem seems to be that even when all arguments and parameters are specified in the object, when audio is put through the object, no data seems to be output.
Thanks,
Rich.
Just found this object....
It's not 100% accurate, but will just about do the trick.
I can't get anything to work either. The object I have transmits a synced bang,but it just fires off for every transient so it fluctuates excessively from one bpm to double the bpm. Terrible for breakbeats. All in all, it's not reliable.
I am capturing beats from streaming audio and using the low end to set BMP.
I use bonk~ to split the band and use the first 2 bands to monitor the transiients , then I Split the min and max and dial each in to get the range the primary beat is in. The output is a bang per bass-beat.
What I don't have yet is the timer to capture the time between the transients and divide over 100th of a second. So I can't tell how accurate it is atm. What I did find is that beat~ is much more cpu intensive than bonk~, so using both is eventually inefficient.
pros and cons of bonk~ it has some attributes to adjust the range and decay, it's only 32bit and a bit dated.
not sure if there is a replacement for it's function.
I am still experimenting and haven't seen a better solution yet for bonk or bmp of audio
map of the output flow:
aiff -> bonk~ --> zi.slice 2 --> split (min/max) --> bang
anyone know of a an optimized flow for bang per minute?
Hey,
There is a 64 bit version of bonk:
I'm new and have searched the forums for BPM, Onset detection, and bonk. There are some examples. I think the metro object could control how fast samples are taken.
you can use the beatseeker object from the max for live device?
this does a good job (with lots of fft stuff) at detecting tempo, but then there is the question of following tempo - what is this trying to do?
p.s. also Adam Stark's BTrack does a good job of this
https://github.com/adamstark/BTrack
and he has the external there