Granular synthesis: Difference between revisions

From XPUB & Lens-Based wiki
No edit summary
No edit summary
 
(2 intermediate revisions by the same user not shown)
Line 1: Line 1:
The technique of producing new sound waves with microsecond samples is generally termed [http://en.wikipedia.org/wiki/Granular_synthesis granular synthesis].  
The technique of producing new sound waves with microsecond samples is generally termed [http://en.wikipedia.org/wiki/Granular_synthesis granular synthesis].  


[[User:Max_Dovey | Max Dovey]] is working on manipulating audio recordings about the limits of time, exploring how the audio can be treated in a way that works at the given time scales.
[[User:Max_Dovey | Max Dovey]] is working on manipulating audio recordings about the limits of time, exploring how the audio can be treated in a way that works at the given time scales. The following [[python]] script can be used as a filter with say [[sox]] to filter the samples of an audio file and play it back in real time.


'''filtersamples.py'''
'''filtersamples.py'''
Line 39: Line 39:
<source lang="bash">
<source lang="bash">
sox input.aiff -t raw -e unsigned-integer -r 44100 -c 1 - | python filtersamples.py | play -t raw -e unsigned-integer -r 44100 -c 1 -b 16 -  
sox input.aiff -t raw -e unsigned-integer -r 44100 -c 1 - | python filtersamples.py | play -t raw -e unsigned-integer -r 44100 -c 1 -b 16 -  
</source>
== Live Audio Pipeline ==
<source lang="bash">
rec -t raw -e unsigned-integer -r 44100 -c 1 - | python filtersamples.py | play -t raw -e unsigned-integer -r 44100 -c 1 -b 16 -
</source>
</source>

Latest revision as of 17:57, 14 October 2014

The technique of producing new sound waves with microsecond samples is generally termed granular synthesis.

Max Dovey is working on manipulating audio recordings about the limits of time, exploring how the audio can be treated in a way that works at the given time scales. The following python script can be used as a filter with say sox to filter the samples of an audio file and play it back in real time.

filtersamples.py

import sys, struct 

c = 0
samples = []
numsamples = 17640 # 0.4 sec
# numsamples = 1764  # 0.040 sec == 40 ms == 0.040 * 44100 => 1764
# numsamples = 176 # 0.004 sec == 4 ms == 0.004 * 44100
# numsamples = 44 # 0.001 sec == 1 ms == 0.001 * 44100
# numsamples = 441 #0.01 sec == 1dec == 0.01 * 44100 = 441
# numsamples = 44 #0.001 sec == 1centi == 0.001 * 44100 = 44.1
# numsamples = int(4.41) #0.0001 sec == 1milli == 0.0001 * 44100 = 4.41
# numsamples = int(1) #0.000001 sec == 1micro == 0.000001 * 44100 = 0.0441

loop = 10

while True:
    sample = sys.stdin.read(2*numsamples)
    if not sample:
        break

    samples = struct.unpack("H"*numsamples, sample)

    # sys.stderr.write(str(n)+"/n")
    # o = max(35000,n)
    # take your slice and repeat
    for i in range(loop):
        out = struct.pack("H"*numsamples, *samples)
        sys.stdout.write(out)

And the command to play the audio through the python script, starts from an input audio, pipes the (raw) bytes through the python script, and finally piped back to sox's play command to play. Note the matching options to be able to in and out of a raw format.

sox input.aiff -t raw -e unsigned-integer -r 44100 -c 1 - | python filtersamples.py | play -t raw -e unsigned-integer -r 44100 -c 1 -b 16 -

Live Audio Pipeline

rec -t raw -e unsigned-integer -r 44100 -c 1 - | python filtersamples.py | play -t raw -e unsigned-integer -r 44100 -c 1 -b 16 -