Can you play audio directly from a python script using DawDreamer?

[music-lover99]

I'd like a python script that takes an audio file, slows it down and lowers it, then plays it directly from the python script (without having to write it to a wavfile). Is this possible? This is what I have so far:

from librosa import load
import dawdreamer as daw

SAMPLE_RATE = 44100
BUFFER_SIZE = 512

def load_audio_file(file_path, duration=None):
    sig, rate = load(file_path, duration=duration, mono=False, sr=SAMPLE_RATE)
    assert(rate == SAMPLE_RATE)
    return sig

engine = daw.RenderEngine(SAMPLE_RATE, BUFFER_SIZE)

playback_processor = engine.make_playbackwarp_processor("song", load_audio_file("C:\Users\sebastian\Desktop\slowed_and_lowered\Drake - Fair Trade (Lyrics) ft. Travis Scott.mp3"))
playback_processor.time_ratio = 1.25
playback_processor.transpose = -5.

graph = [
  (playback_processor, []),
]

engine.load_graph(graph)

---

And also, when I do try and write it to a wavfile, the file comes out playing nothing. Here's my code for when I try and write it to a wavfile:

from librosa import load
from os import chdir
import dawdreamer as daw
from moviepy.editor import AudioFileClip
import scipy.io

def load_audio_file(file_path, duration=None):
    sig, rate = load(file_path, duration=duration, mono=False, sr=SAMPLE_RATE)
    assert(rate == SAMPLE_RATE)
    return sig

SAMPLE_RATE = 44100
BUFFER_SIZE = 512

engine = daw.RenderEngine(SAMPLE_RATE, BUFFER_SIZE)

playback_processor = engine.make_playbackwarp_processor("song", load_audio_file("C:\Users\sebastian\Desktop\slowed_and_lowered\Drake - Fair Trade (Lyrics) ft. Travis Scott.mp3"))
playback_processor.time_ratio = 1.25
playback_processor.transpose = -5.

graph = [
  (playback_processor, []),
]

engine.load_graph(graph)

durationOfClip = AudioFileClip("C:\Users\sebastian\Desktop\slowed_and_lowered\Drake - Fair Trade (Lyrics) ft. Travis Scott.mp3")
durationOfClip = durationOfClip.duration
engine.render(durationOfClip)

audio = engine.get_audio()

scipy.io.wavfile.write('C:\Users\sebastian\Desktop\slowed_and_lowered\output_file.wav', SAMPLE_RATE, audio.transpose())

[DBraun]

Hi, thanks for finding this bug! The issue is that loop_end needs to be set to a high value (but also not too high, which is another bug aspect). This adjustment is working for me

from librosa import load
from scipy.io import wavfile
import dawdreamer as daw

SAMPLE_RATE = 44100
buffer_size = 512
time_ratio = 1.25
song_path = "C:/Users/sebastian/Desktop/slowed_and_lowered/Drake - Fair Trade (Lyrics) ft. Travis Scott.mp3"

def load_audio_file(file_path, duration=None):
    sig, rate = load(file_path, duration=duration, mono=False, sr=SAMPLE_RATE)
    assert(rate == SAMPLE_RATE)
    return sig

engine = daw.RenderEngine(SAMPLE_RATE, buffer_size)
input_audio = load_audio_file(song_path)

playback_processor = engine.make_playbackwarp_processor("song", input_audio)
playback_processor.time_ratio = time_ratio
playback_processor.transpose = -5.
playback_processor.loop_end = 50000.  # this is necessary due to a bug

graph = [
  (playback_processor, []),
]

engine.load_graph(graph)

output_duration = (input_audio.shape[1]/SAMPLE_RATE)*time_ratio

engine.render(output_duration)

output_audio = playback_processor.get_audio()

wavfile.write('slowed_down_output.wav', SAMPLE_RATE, output_audio.transpose())

If you'd like to hear things directly from Python, you can try Jupyter/IPython notebooks.

[music-lover99]

Why does the output audio sound so digital? When I slow down and transpose the file by -5 semitones using dawdreamer, it sounds very different than if I do it with another DAW.

[DBraun]

Sorry, I forgot to remove a modification to load_audio_file. I changed it above now. As far as the audio quality goes, it's using Rubber Band with some settings which can't yet be changed from Python (#59). Perhaps you can email me files which you have permission to distribute and we can discuss the DAW differences more. My email is braun AT ccrma DOT stanford DOT edu.

[music-lover99]

Alright I'll contact you there. Thank you.