Forums > MaxMSP

One question, need your help

May 9, 2007 | 1:36 pm

I am trying to find a way to be able to do this: import a field recording or other sound in a software editor programme (or other, maybe max/msp?) and try to find its changes in some values.
E.g i am recording some wind: there is some changes in the sound’s pitch, distortion or any other. How can i capture that and say that in this sound the changes in its pitch is from lets say 0.3 to 0.8 ?
My need is no specific for pitch or wind, any sound of the nature would be useful because i am preparing a project and i might need various elements… Has anybody did something similar with max/msp?
Any help would be warm appreciated.
Thanx in advance


May 9, 2007 | 1:52 pm


May 9, 2007 | 7:07 pm

well, thanx for the reply.
you r right that wind is almost pitchless or pitchfull.
my interest in not to capture the wind’s pitch. it was an example.
i would like to find a patcher (or build one) which will work as an audio analyzer in some recorded sounds.
for example, in the wind i could take its ratio or somehting else.
i would like to have some data considering the audio characteristics of the file. the more characteristics the better for me.
i hope its more clear now, so any tipe are warm appreciated

thanx again


May 9, 2007 | 7:27 pm

There’s frame~ (turns fft into lists), centroid~ (spectral centroid),
analyzer~ (has lots of different analyses), wavelet~ (wavelet
analysis)… um, I know there’s more but that’s a good start. You can
use xerox~ to see how noisy a sound is. There’s of course, peakamp~
for volume. Plenty more, I’m sure.

Thanks,
Keith

On 5/9/07, sadmac

wrote:
>
> well, thanx for the reply.
> you r right that wind is almost pitchless or pitchfull.
> my interest in not to capture the wind’s pitch. it was an example.
> i would like to find a patcher (or build one) which will work as an audio analyzer in some recorded sounds.
> for example, in the wind i could take its ratio or somehting else.
> i would like to have some data considering the audio characteristics of the file. the more characteristics the better for me.
> i hope its more clear now, so any tipe are warm appreciated
>
> thanx again
>


May 10, 2007 | 7:20 am

Thank you very much.
That is really helpfull. If anybody has any futher help post it up pls.
thanx a lot

Quote: keithmanlove wrote on Wed, 09 May 2007 22:27
—————————————————-
> There’s frame~ (turns fft into lists), centroid~ (spectral centroid),
> analyzer~ (has lots of different analyses), wavelet~ (wavelet
> analysis)… um, I know there’s more but that’s a good start. You can
> use xerox~ to see how noisy a sound is. There’s of course, peakamp~
> for volume. Plenty more, I’m sure.
>
> Thanks,
> Keith
>
> On 5/9/07, sadmac

wrote:
> >
> > well, thanx for the reply.
> > you r right that wind is almost pitchless or pitchfull.
> > my interest in not to capture the wind’s pitch. it was an example.
> > i would like to find a patcher (or build one) which will work as an audio analyzer in some recorded sounds.
> > for example, in the wind i could take its ratio or somehting else.
> > i would like to have some data considering the audio characteristics of the file. the more characteristics the better for me.
> > i hope its more clear now, so any tipe are warm appreciated
> >
> > thanx again
> >
>
—————————————————-


May 10, 2007 | 10:41 pm

everything what is not releated to _power and _frequency
requires a lot of knowledge about audio DSP and programming
in general.

to analyze in a recording what you called "distortion"
for example would probably require you to develop a
groundbraking new algorithm based on neuronal networks
in combination with fft or wavelet transforms.

but that does not mean you are lost; there is a lot you
can analyze about frequency and power in a signal.

find out how to analyze power history of an audio signal
and look into filters like [fffb~] or [svf~]. (you have
to learn that yourself)
a combination of a filterbank and power history can already
make a powerful control for video (or other audio) signals.

-110


May 11, 2007 | 7:22 am

thank you.
can u recommend me some documendation since i am a newbie here?
i do have the will to read and learn but i don’t know the sources.
so pls if possible let me know.
thanx again

Quote: Roman Thilenius wrote on Fri, 11 May 2007 01:41
—————————————————-
>
> everything what is not releated to _power and _frequency
> requires a lot of knowledge about audio DSP and programming
> in general.
>
> to analyze in a recording what you called "distortion"
> for example would probably require you to develop a
> groundbraking new algorithm based on neuronal networks
> in combination with fft or wavelet transforms.
>
>
> but that does not mean you are lost; there is a lot you
> can analyze about frequency and power in a signal.
>
> find out how to analyze power history of an audio signal
> and look into filters like [fffb~] or [svf~]. (you have
> to learn that yourself)
> a combination of a filterbank and power history can already
> make a powerful control for video (or other audio) signals.
>
>
> -110

>
>
>
>
—————————————————-


May 11, 2007 | 8:15 am

sadmac schrieb:
> i would like to have some data considering the audio characteristics
> of the file. the more characteristics the better for me. i hope its
> more clear now, so any tipe are warm appreciated

I would do a fft based analysis with pfft~. Use vectral~ for envelope
following of the bins, that should give you plenty of parameters to
digest…

Another approach is a filterbank like fffb~ with envelope following…

Stefan


Stefan Tiedje————x——-
–_____———–|————–
–(_|_ —-|—–|—–()——-
– _|_)—-|—–()————–
———-()——–www.ccmix.com


Viewing 8 posts - 1 through 8 (of 8 total)