Generate a wave file from a MIDI file with Pydub

Generate WAV from MIDI.md

Simple example of rendering a midi file with Pydub

Basically, this takes a MIDI input file (I just googled and grabbed one of Maroon 5's "Animal" – I know, I know) and generates a WAV file.

NOTE: This is the slowest midi rendering program I have ever seen!

Dependencies:

• Pydub (pydub.com)
  • pip install pydub
• mido (http://mido.readthedocs.org/en/latest/)
  • pip install mido

In this example I an rendering a MIDI file of Maroon 5's "Animals" from http://www.free-midi.org

ok, ok, ok - Here's the output:

Here is the output my program generated (converted to mp3):

output.mp3

Obvious issues:

• Tempo is hard-coded
• All instruments are rendered using a Sine wave, including drums. Would be better to use different signal generators depending which instrument is being rendered (Sine is fine for flute, square is closer to saxaphone, etc)
• This is REALLLY SLOWWW. I haven't looked at performance at all, but it took almost an hour (2014, 2.8 GHz i7, Macbook Pro) to render the wav file with the code above. LOL

midi_to_wav.py

from mido import MidiFile
from pydub import AudioSegment
from pydub.generators import Sine
from collections import defaultdict

def note_to_freq(note, concert_A=440.0):
  '''
  from wikipedia: http://en.wikipedia.org/wiki/MIDI_Tuning_Standard#Frequency_values
  '''
  return (2.0 ** ((note - 69) / 12.0)) * concert_A

mid = MidiFile("./maroon_5-animals.mid")
output = AudioSegment.silent(mid.length * 1000.0)

tempo = 100 # bpm

def ticks_to_ms(ticks):
  tick_ms = (60000.0 / tempo) / mid.ticks_per_beat
  return ticks * tick_ms
  

for track in mid.tracks:
  # position of rendering in ms
  current_pos = 0.0

  current_notes = defaultdict(dict)
  # current_notes = {
  #   channel: {
  #     note: (start_time, message)
  #   }
  # }
  
  for msg in track:
    current_pos += ticks_to_ms(msg.time)

    if msg.type == 'note_on':
      current_notes[msg.channel][msg.note] = (current_pos, msg)
    
    if msg.type == 'note_off':
      start_pos, start_msg = current_notes[msg.channel].pop(msg.note)
  
      duration = current_pos - start_pos
  
      signal_generator = Sine(note_to_freq(msg.note))
      rendered = signal_generator.to_audio_segment(duration=duration-50, volume=-20).fade_out(100).fade_in(30)

      output = output.overlay(rendered, start_pos)

output.export("animal.wav", format="wav")