Hi Dmitry,
Thank you for the explanation. I have switched to
SpectrAnalyzer.asFrequencies(), as a test I am reading frames out of a
wav file with a single note (G) being played several times.
So with asFrequencies() then, is there a simple formula for determining
which
chunk is which actual frequency? For example I have these constants:
CHANNELS= 1
NUM_FREQS= 512
RATE= 44100
Now, if I analyze my data:
analyzer= sound.SpectrAnalyzer( CHANNELS, SAMPLES, NUM_FREQS )
freqs= analyzer.asFrequencies( data )
Then the len(freqs) is 2, and len(freqs[0]) for example is 512.
So the actual data in the lists in freqs is the decibel value for each
frequency correct? I am not sure how to go about determining which value
is for which frequency.. although I can see how it should be available
given the 3 constants defined above, I am just not sure how to go about it.
Sorry to be so dense :) This is the first time I have tried to create
any kind of audio application before.
Thanks,
Rob
Dmitry Borisov wrote:
asBands actually returning the bacnds like low frequencies, mdeium
frquencies and stuff.
I believe you need asFrequencies() which will split the whole interval
0-11KHz into equal chunks and calc the power of the desired audio
chunk in a particular range.
See doc for example:
http://pymedia.org//docs/pymedia.audio.sound.html#SpectrAnalyzer
For now the smallest frequency range you can get via asFrequencies()
is 44000/ ( 4* 512 )= 22Hz
Dmitry/
----- Original Message ----- From: "Robert Helmer"
<[EMAIL PROTECTED]>
To: <[email protected]>
Sent: Sunday, January 08, 2006 12:27 AM
Subject: Re: [Pymedia-users] pitch detection
Hi Dmitry,
Thank you for your reply! I have been trying out the asBands()
method, which seems to yield some result to different voice tones.
However, it seems that asBands() can only return a limited set of
possible frequencies.. I am basically doing this :
snd= sound.Input( RATE, CHANNELS, sound.AFMT_S16_LE )
snd.start()
analyzer= sound.SpectrAnalyzer( CHANNELS, SAMPLES, NUM_FREQS )
while 1:
largestFreqName= 0
largestFreqValue= 0
previousLargestFreqValue= 0
bands= analyzer.asBands( BANDS, snd.getData() )
for i in range( 0, len( bands ) ):
for j in range( 0, len( bands[i] ) ):
freqName= bands[i][j][0]
if freqName >= 1000:
continue
freqValue= bands[i][j][1]
if freqValue > largestFreqValue:
previousLargestFreqValue= largestFreqValue
largestFreqName= freqName
largestFreqValue= freqValue
print largestFreqValue
Am I understanding how this data may be used correctly? Is there some
way to do something similar with the analyzer.asFrequencies() method?
I cannot seem to make any sense out of the returned data through
experimentation :)
Thanks!
Rob Helmer
Dmitry Borisov wrote:
Hi,
It could work for you.
Basically SpectrAnalyzer is taking the fixed number of samples >512
and returns the frequencies strength. If you use much more samples
it may not work well for you though.
Sorry I never did pitch detection myself, so I cannot tell if it'll
work for you.
Dmitry/
----- Original Message ----- From: "Robert Helmer"
<[EMAIL PROTECTED]>
To: <[email protected]>
Sent: Tuesday, January 03, 2006 12:25 PM
Subject: [Pymedia-users] pitch detection
Hello,
I am working on a voice-training python application that involves
pitch detection. I have had some good results using FFT but it's a
bit too slow (I can't get good detection without a very large
sample size it seems).
Could SpectrAnalyzer be useful for pitch detection? I am a bit
confused as to how to use this class properly; the example
(sound_viz.py) doesn't work in the latest release due to what looks
like API changes. I think being able to experiment with a working
implementation would help me to be able to play around with FFT and
other algorithms (Schmitt-triggering for instance).
I took a peek at the implementation of SpectrAnalyzer in C and I am
not sure how much overlap there is between what I am trying to do
and SpectrAnalyzer; it seems quite capable but my inexperience in
digital audio processing is holding back my understanding.
I've been learning just how little I know about DSP by embarking on
this, so any suggestions would be greatly appreciated!
Thanks,
Rob Helmer
-------------------------------------------------------
This SF.net email is sponsored by: Splunk Inc. Do you grep through
log files
for problems? Stop! Download the new AJAX search engine that makes
searching your log files as easy as surfing the web. DOWNLOAD
SPLUNK!
http://ads.osdn.com/?ad_id=7637&alloc_id=16865&op=click
_______________________________________________
Pymedia-users mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/pymedia-users
-------------------------------------------------------
This SF.net email is sponsored by: Splunk Inc. Do you grep through
log files
for problems? Stop! Download the new AJAX search engine that makes
searching your log files as easy as surfing the web. DOWNLOAD SPLUNK!
http://ads.osdn.com/?ad_id=7637&alloc_id=16865&op=click
_______________________________________________
Pymedia-users mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/pymedia-users
-------------------------------------------------------
This SF.net email is sponsored by: Splunk Inc. Do you grep through
log files
for problems? Stop! Download the new AJAX search engine that makes
searching your log files as easy as surfing the web. DOWNLOAD SPLUNK!
http://ads.osdn.com/?ad_id=7637&alloc_id=16865&op=click
_______________________________________________
Pymedia-users mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/pymedia-users
-------------------------------------------------------
This SF.net email is sponsored by: Splunk Inc. Do you grep through log
files
for problems? Stop! Download the new AJAX search engine that makes
searching your log files as easy as surfing the web. DOWNLOAD SPLUNK!
http://ads.osdn.com/?ad_id=7637&alloc_id=16865&op=click
_______________________________________________
Pymedia-users mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/pymedia-users
-------------------------------------------------------
This SF.net email is sponsored by: Splunk Inc. Do you grep through log files
for problems? Stop! Download the new AJAX search engine that makes
searching your log files as easy as surfing the web. DOWNLOAD SPLUNK!
http://sel.as-us.falkag.net/sel?cmd=lnk&kid=103432&bid=230486&dat=121642
_______________________________________________
Pymedia-users mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/pymedia-users
