1.03

With pyknon ex.1 , ex.2 and midi2audio I have evything to generate midi and to play and export them as music file. With pyknon NoteSeq funktion i have a usabel interface to build the farmes and so to give structure to the music.

pyknon ex.1

  1. output only as midi
  2. has many features
from pyknon.genmidi import Midi
from pyknon.music import Rest, Note, NoteSeq
import numpy as np

def angel_trans(grad):
    rad = (grad *np.pi) / 180
    return rad

a simpele major frame in which the melody can play

frame_dur = [0,2,4,5,7,9,11,12]
frame_moll = [0,2,3,5,7,8,10,12]

def roud_to_frame(frame, num):
    array = np.asarray(frame)
    listnr = (np.abs(frame - num)).argmin()
    return frame[listnr]

def mod12(n):
    return n % 12

def note_name(number):
    notes = "C C# D D# E F F# G G# A A# B C".split()
    return notes[mod12(number)]

#--main--

seq = ""
for i in range(-90,270,20):
    sin = np.sin(angel_trans(i))
    normal = (sin+1)*6
    num = roud_to_frame(frame_moll, normal)
    note = note_name(num)
    seq = seq + note + " "

print(seq)
notes1 = NoteSeq(seq)
midi = Midi(1, tempo=120)
midi.seq_notes(notes1, track=0)
midi.write("demo.mid")

C C D D# F G G# A# C C C A# G# G F D# D C

pyknon ex.2

#=== section: music generation =================================
def tune_A():
    notes1 = NoteSeq(      "C4 D E F G A B C''")   # Apostroph ' = "gestrichen" = Höhe der Oktave
    notes2 = NoteSeq("r1 r1 C4'' B' A G F E D C")   # r = rest = Pause; Zahl = Länge der Pause
    return notes1, notes2

def tune_B():
    notes1 = NoteSeq("Db4- F#8 A Bb4")   # Zahl = Länge des Tones: 1=ganz, 4=Viertel
    notes2 = NoteSeq([Note(2, dur=1/4), Note(6, dur=1/8), Note(9, dur=1/8), Note(10, dur=1/4)])
    return notes1, notes2


def generate_midi_pyknon():
    #--- choose the tune
    notes1, notes2 = tune_A()         # <<<---- select a tune <<<------

    #--- squezze ir into a MIDI framework
    m = Midi(2, tempo=120)     #
    m.seq_notes(notes1, track=0)
    m.seq_notes(notes2, track=1)

    #--- write the MIDI file
    midi_file_name =  "myMidi.mid"       # set the name of the file
    m.write(midi_file_name)
    return midi_file_name


generate_midi_pyknon()

'myMidi.mid'

pyfluidsynth

more or less an python api for fluidsynth direct sound output suports several channels with diffrent instruments (Gneral-Midi) todo: ouput as an audio file.

import time
import fluidsynth

fs = fluidsynth.Synth()

#fs.start()
fs.start(driver="alsa")

sfid = fs.sfload("/usr/share/sounds/sf3/MuseScore_General.sf3")
fs.program_select(0, sfid, 0, 33)   #(tracknr , sondfontid, ??, instrumentnr)
print(fs.channel_info(0))

fs.noteon(0, 60, 30)
fs.noteon(0, 67, 30)
fs.noteon(0, 76, 30)

time.sleep(1.0)

fs.noteon(0, 60, 30)
fs.noteon(0, 67, 30)
fs.noteon(0, 76, 30)

time.sleep(1.0)

fs.noteoff(0, 60)
fs.noteoff(0, 67)
fs.noteoff(0, 76)

time.sleep(1.0)

fs.delete()

(1, 0, 33, b'Fingered Bass')

Midi: Play and Generate audio-file (first solution)

For generating audio-files I use midi2audio.

midi2audio

midi to audio converter with FluidSynth. But only over jack sound driver in Linux.

When using play_midi in Linux with Pulseaudio
–> with pulseaudio it is nessecary to change the source code of midi2audio:
'''
replace:
def play_midi(self, midi_file):
    subprocess.call(['fluidsynth', '-i', self.sound_font, midi_file, '-r', str(self.sample_rate)])

with:
def play_midi(self, midi_file):
    subprocess.call(['fluidsynth', '-i', self.sound_font, midi_file, '-r', str(self.sample_rate), '-a', 'pulseaudio'])
'''
''

''

otherwise FluidSyth can also used manualy. See below.

from midi2audio import FluidSynth
default_soundfont = '/usr/share/sounds/sf3/MuseScore_General.sf3'
soundfont = default_soundfont

def midi_play(midi_in, soundfont= default_soundfont):
    fs = FluidSynth(soundfont)
    fs.play_midi(midi_in)  # This runs FluidSyth with Jack.


def midi_audio(midi_in, name_out = 'none', soundfont= default_soundfont):
    fs = FluidSynth(soundfont)
    if name_out == 'none' :
        name_out = midi_in.replace('.mid', '.flac')
    else:
        name_out = name_out + '.flac'

    fs.midi_to_audio(midi_in, name_out)


#midi_audio('scale.mid')
midi_play('myMidi.mid')

https://raw.githubusercontent.com/schuhva/Music-Generation/master/doc/releases/1.03/scale.flac

FluidSynth manualy

import subprocess

sound_font = '/usr/share/sounds/sf3/MuseScore_General.sf3'
midi_file = 'scale.mid'
sample_rate = 44100
subprocess.call(['fluidsynth', '-i', sound_font, midi_file, '-r', str(sample_rate), '-a', 'pulseaudio'])

0

Midi: Play and Generate audio-file (second and further used solution)

This soulution offers a better sound quality and does not requier a change of the souce code. It uses audacious and Musescore

‘audacious -p -q -H myMidi.mid’

‘mscore -o ms.flac myMidi.mid’

import subprocess
default_soundfont = '/usr/share/sounds/sf3/MuseScore_General.sf3'

def midi_play(midi_in, soundfont= default_soundfont):
    subprocess.call(['audacious', '-p', '-q', '-H', midi_in ])  # -p = play , -q = quite , -H = hide gui

def midi_audio(midi_in, name_out = 'none', soundfont= default_soundfont):
    if name_out == 'none' :
        name_out = midi_in.replace('.mid', '.flac')
    else:
        name_out = name_out + '.flac'
    subprocess.call(['mscore', '-o', name_out, midi_in]) # -o = export as

midi_audio('scale.mid')
midi_play('myMidi.mid')