Python-Sonic is a simple Python interface for Sonic Pi, which is a real great music software created by Sam Aaron (http://sonic-pi.net).
At the moment Python-Sonic works with Sonic Pi. It is planned, that it will work with Supercollider, too.
This version supports Sonic Pi versions > 4 when OSC run-code security was added.
If you like it, use it. If you have some suggestions, tell me ([email protected]).
- First you need Python 3 (https://www.python.org, ) - Python 3.5 should work, because it’s the development environment
- Then Sonic Pi (https://sonic-pi.net) - That makes the sound
- Modul python-osc (https://pypi.python.org/pypi/python-osc) - Connection between Python and Sonic Pi Server
- Add the module python-sonic - simply copy the source, or try pip install python-sonic That should work.
For local development you might want to locally install using
$ pip install -e .
- You have to start Sonic Pi first before you can use it with python-sonic
- Only the notes from C2 to C6
V ersion | |
---|---|
0.2.0 | Some changes for Sonic Pi 2.11. Simpler multi-threading with decorator @in_thread. Messaging with cue and sync. |
0.3.0 | OSC Communication |
0.3.1. | Update, sort and duration of samples |
0.3.2. | Restructured |
0.4.0 | Changes communication ports and recording |
0.4.4 | Enables GUI Token |
The API python-sonic communications with Sonic Pi over UDP and two ports. One port is an internal Sonic Pi GUI port and the second is the external OSC cue port.
For >v4 a Token is needed to communicate with Sonic Pi. This token is generated randomly at runtime when Sonic-Pi is started. The Token is extracted from the sonic-pi log files. Similarly, the GUI udp port is now randomised at start and must be read from the log file too. In order to play notes (and not just send OSC cues), a connection must be made to the GUI udp port using the Token as the first argument in the message. The Token is automatically used on send.
from psonic import *
set_server_parameter('127.0.0.1', -2005799440, 30129, 4560)
These values are found from the file ~/.sonic-pi/log/spider.log
Sonic Pi Spider Server booting... The time is 2023-10-01 11:01:41 +0100 Using primary protocol: udp Detecting port numbers... Ports: {:server_port=>30129, :gui_port=>30130, :scsynth_port=>30131, :scsynth_send_port=>30131, :osc_cues_port=>4560, :tau_port=>30132, :listen_to_tau_port=>30136} Token: -2005799440 Opening UDP Server to listen to GUI on port: 30129 Spider - Pulling in modules... Spider - Starting Runtime Server ...
This can be automated by using the function set_server_parameter_from_log
from psonic import *
set_server_parameter_from_log('127.0.0.1')
set_server_parameter_from_log('127.0.0.1', "path-to-log-file")
Note if the set_server_parameter functions are not used, a default connection is created which will not work.
There is a simple example file psonic_example.py which you can run to check that things work. First open sonic-pi, then run the following:
$ python psonic_example.py
and a note should be played.
Many of the examples are inspired from the help menu in Sonic Pi.
from psonic import *
The first sound
play(70) #play MIDI note 70
Some more notes
play(72)
sleep(1)
play(75)
sleep(1)
play(79)
In more tratitional music notation
play(C5)
sleep(0.5)
play(D5)
sleep(0.5)
play(G5)
Play sharp notes like F# or dimished ones like Eb
play(Fs5)
sleep(0.5)
play(Eb5)
Play louder (parameter amp) or from a different direction (parameter pan)
play(72,amp=2)
sleep(0.5)
play(74,pan=-1) #left
Different synthesizer sounds
use_synth(SAW)
play(38)
sleep(0.25)
play(50)
sleep(0.5)
use_synth(PROPHET)
play(57)
sleep(0.25)
ADSR (Attack, Decay, Sustain and Release) Envelope
play (60, attack=0.5, decay=1, sustain_level=0.4, sustain=2, release=0.5)
sleep(4)
Play some samples
sample(AMBI_LUNAR_LAND, amp=0.5)
sample(LOOP_AMEN,pan=-1)
sleep(0.877)
sample(LOOP_AMEN,pan=1)
sample(LOOP_AMEN,rate=0.5)
sample(LOOP_AMEN,rate=1.5)
sample(LOOP_AMEN,rate=-1)#back
sample(DRUM_CYMBAL_OPEN,attack=0.01,sustain=0.3,release=0.1)
sample(LOOP_AMEN,start=0.5,finish=0.8,rate=-0.2,attack=0.3,release=1)
Play some random notes
import random
for i in range(5):
play(random.randrange(50, 100))
sleep(0.5)
for i in range(3):
play(random.choice([C5,E5,G5]))
sleep(1)
Sample slicing
from psonic import *
number_of_pieces = 8
for i in range(16):
s = random.randrange(0,number_of_pieces)/number_of_pieces #sample starts at 0.0 and finishes at 1.0
f = s + (1.0/number_of_pieces)
sample(LOOP_AMEN,beat_stretch=2,start=s,finish=f)
sleep(2.0/number_of_pieces)
An infinite loop and if
while True:
if one_in(2):
sample(DRUM_HEAVY_KICK)
sleep(0.5)
else:
sample(DRUM_CYMBAL_CLOSED)
sleep(0.25)
--------------------------------------------------------------------------- KeyboardInterrupt Traceback (most recent call last) <ipython-input-18-d8759ac2d27e> in <module>() 5 else: 6 sample(DRUM_CYMBAL_CLOSED) ----> 7 sleep(0.25) /mnt/jupyter/python-sonic/psonic.py in sleep(duration) 587 :return: 588 """ --> 589 time.sleep(duration) 590 _debug('sleep', duration) 591 KeyboardInterrupt:
If you want to hear more than one sound at a time, use Threads.
import random
from psonic import *
from threading import Thread
def bass_sound():
c = chord(E3, MAJOR7)
while True:
use_synth(PROPHET)
play(random.choice(c), release=0.6)
sleep(0.5)
def snare_sound():
while True:
sample(ELEC_SNARE)
sleep(1)
bass_thread = Thread(target=bass_sound)
snare_thread = Thread(target=snare_sound)
bass_thread.start()
snare_thread.start()
while True:
pass
--------------------------------------------------------------------------- KeyboardInterrupt Traceback (most recent call last) <ipython-input-19-5b8671a783d6> in <module> 22 23 while True: ---> 24 pass KeyboardInterrupt:
Every function bass_sound and snare_sound have its own thread. Your can hear them running.
from psonic import *
from threading import Thread, Condition
from random import choice
def random_riff(condition):
use_synth(PROPHET)
sc = scale(E3, MINOR)
while True:
s = random.choice([0.125,0.25,0.5])
with condition:
condition.wait() #Wait for message
for i in range(8):
r = random.choice([0.125, 0.25, 1, 2])
n = random.choice(sc)
co = random.randint(30,100)
play(n, release = r, cutoff = co)
sleep(s)
def drums(condition):
while True:
with condition:
condition.notifyAll() #Message to threads
for i in range(16):
r = random.randrange(1,10)
sample(DRUM_BASS_HARD, rate=r)
sleep(0.125)
condition = Condition()
random_riff_thread = Thread(name='consumer1', target=random_riff, args=(condition,))
drums_thread = Thread(name='producer', target=drums, args=(condition,))
random_riff_thread.start()
drums_thread.start()
input("Press Enter to continue...")
Press Enter to continue...
''
To synchronize the thread, so that they play a note at the same time, you can use Condition. One function sends a message with condition.notifyAll the other waits until the message comes condition.wait.
More simple with decorator __@in_thread__
from psonic import *
from random import choice
tick = Message()
@in_thread
def random_riff():
use_synth(PROPHET)
sc = scale(E3, MINOR)
while True:
s = random.choice([0.125,0.25,0.5])
tick.sync()
for i in range(8):
r = random.choice([0.125, 0.25, 1, 2])
n = random.choice(sc)
co = random.randint(30,100)
play(n, release = r, cutoff = co)
sleep(s)
@in_thread
def drums():
while True:
tick.cue()
for i in range(16):
r = random.randrange(1,10)
sample(DRUM_BASS_HARD, rate=r)
sleep(0.125)
random_riff()
drums()
input("Press Enter to continue...")
Press Enter to continue...
from psonic import *
tick = Message()
@in_thread
def metronom():
while True:
tick.cue()
sleep(1)
@in_thread
def instrument():
while True:
tick.sync()
sample(DRUM_HEAVY_KICK)
metronom()
instrument()
while True:
pass
Play a list of notes
from psonic import *
play ([64, 67, 71], amp = 0.3)
sleep(1)
play ([E4, G4, B4])
sleep(1)
Play chords
play(chord(E4, MINOR))
sleep(1)
play(chord(E4, MAJOR))
sleep(1)
play(chord(E4, MINOR7))
sleep(1)
play(chord(E4, DOM7))
sleep(1)
Play arpeggios
play_pattern( chord(E4, 'm7'))
play_pattern_timed( chord(E4, 'm7'), 0.25)
play_pattern_timed(chord(E4, 'dim'), [0.25, 0.5])
Play scales
play_pattern_timed(scale(C3, MAJOR), 0.125, release = 0.1)
play_pattern_timed(scale(C3, MAJOR, num_octaves = 2), 0.125, release = 0.1)
play_pattern_timed(scale(C3, MAJOR_PENTATONIC, num_octaves = 2), 0.125, release = 0.1)
The function scale returns a list with all notes of a scale. So you can use list methodes or functions. For example to play arpeggios descending or shuffeld.
import random
from psonic import *
s = scale(C3, MAJOR)
s
[48, 50, 52, 53, 55, 57, 59, 60]
s.reverse()
play_pattern_timed(s, 0.125, release = 0.1)
random.shuffle(s)
play_pattern_timed(s, 0.125, release = 0.1)
One of the best in SONIC PI is the Live Loop. While a loop is playing music you can change it and hear the change. Let’s try it in Python, too.
from psonic import *
from threading import Thread
def my_loop():
play(60)
sleep(1)
def looper():
while True:
my_loop()
looper_thread = Thread(name='looper', target=looper)
looper_thread.start()
input("Press Enter to continue...")
Press Enter to continue...Y
'Y'
Now change the function my_loop und you can hear it.
def my_loop():
use_synth(TB303)
play (60, release= 0.3)
sleep (0.25)
def my_loop():
use_synth(TB303)
play (chord(E3, MINOR), release= 0.3)
sleep(0.5)
def my_loop():
use_synth(TB303)
sample(DRUM_BASS_HARD, rate = random.uniform(0.5, 2))
play(random.choice(chord(E3, MINOR)), release= 0.2, cutoff=random.randrange(60, 130))
sleep(0.25)
To stop the sound you have to end the kernel. In IPython with Kernel –> Restart
Now with two live loops which are synch.
from psonic import *
from threading import Thread, Condition
from random import choice
def loop_foo():
play (E4, release = 0.5)
sleep (0.5)
def loop_bar():
sample (DRUM_SNARE_SOFT)
sleep (1)
def live_loop_1(condition):
while True:
with condition:
condition.notifyAll() #Message to threads
loop_foo()
def live_loop_2(condition):
while True:
with condition:
condition.wait() #Wait for message
loop_bar()
condition = Condition()
live_thread_1 = Thread(name='producer', target=live_loop_1, args=(condition,))
live_thread_2 = Thread(name='consumer1', target=live_loop_2, args=(condition,))
live_thread_1.start()
live_thread_2.start()
input("Press Enter to continue...")
Press Enter to continue...y
'y'
def loop_foo():
play (A4, release = 0.5)
sleep (0.5)
def loop_bar():
sample (DRUM_HEAVY_KICK)
sleep (0.125)
If would be nice if we can stop the loop with a simple command. With stop event it works.
from psonic import *
from threading import Thread, Condition, Event
def loop_foo():
play (E4, release = 0.5)
sleep (0.5)
def loop_bar():
sample (DRUM_SNARE_SOFT)
sleep (1)
def live_loop_1(condition,stop_event):
while not stop_event.is_set():
with condition:
condition.notifyAll() #Message to threads
loop_foo()
def live_loop_2(condition,stop_event):
while not stop_event.is_set():
with condition:
condition.wait() #Wait for message
loop_bar()
condition = Condition()
stop_event = Event()
live_thread_1 = Thread(name='producer', target=live_loop_1, args=(condition,stop_event))
live_thread_2 = Thread(name='consumer1', target=live_loop_2, args=(condition,stop_event))
live_thread_1.start()
live_thread_2.start()
input("Press Enter to continue...")
Press Enter to continue...y
'y'
stop_event.set()
More complex live loops
sc = Ring(scale(E3, MINOR_PENTATONIC))
def loop_foo():
play (next(sc), release= 0.1)
sleep (0.125)
sc2 = Ring(scale(E3,MINOR_PENTATONIC,num_octaves=2))
def loop_bar():
use_synth(DSAW)
play (next(sc2), release= 0.25)
sleep (0.25)
Now a simple structure with four live loops
import random
from psonic import *
from threading import Thread, Condition, Event
def live_1():
pass
def live_2():
pass
def live_3():
pass
def live_4():
pass
def live_loop_1(condition,stop_event):
while not stop_event.is_set():
with condition:
condition.notifyAll() #Message to threads
live_1()
def live_loop_2(condition,stop_event):
while not stop_event.is_set():
with condition:
condition.wait() #Wait for message
live_2()
def live_loop_3(condition,stop_event):
while not stop_event.is_set():
with condition:
condition.wait() #Wait for message
live_3()
def live_loop_4(condition,stop_event):
while not stop_event.is_set():
with condition:
condition.wait() #Wait for message
live_4()
condition = Condition()
stop_event = Event()
live_thread_1 = Thread(name='producer', target=live_loop_1, args=(condition,stop_event))
live_thread_2 = Thread(name='consumer1', target=live_loop_2, args=(condition,stop_event))
live_thread_3 = Thread(name='consumer2', target=live_loop_3, args=(condition,stop_event))
live_thread_4 = Thread(name='consumer3', target=live_loop_3, args=(condition,stop_event))
live_thread_1.start()
live_thread_2.start()
live_thread_3.start()
live_thread_4.start()
input("Press Enter to continue...")
Press Enter to continue...y
'y'
After starting the loops you can change them
def live_1():
sample(BD_HAUS,amp=2)
sleep(0.5)
pass
def live_2():
#sample(AMBI_CHOIR, rate=0.4)
#sleep(1)
pass
def live_3():
use_synth(TB303)
play(E2, release=4,cutoff=120,cutoff_attack=1)
sleep(4)
def live_4():
notes = scale(E3, MINOR_PENTATONIC, num_octaves=2)
for i in range(8):
play(random.choice(notes),release=0.1,amp=1.5)
sleep(0.125)
And stop.
stop_event.set()
from psonic import *
synth(SINE, note=D4)
synth(SQUARE, note=D4)
synth(TRI, note=D4, amp=0.4)
detune = 0.7
synth(SQUARE, note = E4)
synth(SQUARE, note = E4+detune)
detune=0.1 # Amplitude shaping
synth(SQUARE, note = E2, release = 2)
synth(SQUARE, note = E2+detune, amp = 2, release = 2)
synth(GNOISE, release = 2, amp = 1, cutoff = 60)
synth(GNOISE, release = 0.5, amp = 1, cutoff = 100)
synth(NOISE, release = 0.2, amp = 1, cutoff = 90)
Using FX Not implemented yet
from psonic import *
with Fx(SLICER):
synth(PROPHET,note=E2,release=8,cutoff=80)
synth(PROPHET,note=E2+4,release=8,cutoff=80)
with Fx(SLICER, phase=0.125, probability=0.6,prob_pos=1):
synth(TB303, note=E2, cutoff_attack=8, release=8)
synth(TB303, note=E3, cutoff_attack=4, release=8)
synth(TB303, note=E4, cutoff_attack=2, release=8)
In Sonic Pi version 3 or better you can work with messages.
from psonic import *
First you need a programm in the Sonic Pi server that receives messages. You can write it in th GUI or send one with Python.
run("""live_loop :foo do
use_real_time
a, b, c = sync "/osc*/trigger/prophet"
synth :prophet, note: a, cutoff: b, sustain: c
end """)
Now send a message to Sonic Pi.
send_message('/trigger/prophet', 70, 100, 8)
stop()
With python-sonic you can record wave files.
from psonic import *
# start recording
start_recording()
play(chord(E4, MINOR))
sleep(1)
play(chord(E4, MAJOR))
sleep(1)
play(chord(E4, MINOR7))
sleep(1)
play(chord(E4, DOM7))
sleep(1)
# stop recording
stop_recording
<function psonic.psonic.stop_recording()>
# save file
save_recording('/Volumes/jupyter/python-sonic/test.wav')
from psonic import *
#Inspired by Steve Reich Clapping Music
clapping = [1, 1, 1, 0, 1, 1, 0, 1, 0, 1, 1, 0]
for i in range(13):
for j in range(4):
for k in range(12):
if clapping[k] ==1 : sample(DRUM_SNARE_SOFT,pan=-0.5)
if clapping[(i+k)%12] == 1: sample(DRUM_HEAVY_KICK,pan=0.5)
sleep (0.25)
Raspberry Pi sonic-track.py a Sonic-pi Motion Track Demo https://github.com/pageauc/sonic-track
Joe Armstrong: Connecting Erlang to the Sonic Pi http://joearms.github.io/2015/01/05/Connecting-Erlang-to-Sonic-Pi.html
Joe Armstrong: Controlling Sound with OSC Messages http://joearms.github.io/2016/01/29/Controlling-Sound-with-OSC-Messages.html