spectro-utils

Conversion from wav file to spectro image and from spectro image to wav file

install

pip install spectro-utils

how to use

get started

wav file to spectro image

import spectro_utils.audio2spectro as a2s
a2s.audio2spectro_img(audio_paths)

spectro image to wav file

import spectro_utils.spectro2audio as s2a
s2a.img2wav_file(image_paths)

more details of audio2spectro

read_audio_lazy

Reads a wav file using delayed evaluation. It saves memory compared to normal loading.

audios=a2s.read_audio_lazy(paths)

wave2spectro_lazy

Converts the waves to an spectrum using delayed evaluation.

spectros=a2s.wave2spectro_lazy(audios)

spectro2img

Save the spectro images as a image files(png or bmp).

a2s.spectro2img(spectros,image_names)

example of custom pipeline

def split_four_audio_lazy(audios):
    for audio in audios:
        n=len(audio)/4
        for i in range(4):
            start=n*i
            yield audio[start:start+n]

def custom_pipeline(paths:list[str],image_names:list[str])->None:
    g=a2s.read_audio_lazy(paths)
    g=split_four_audio_lazy(g)
    g=a2s.wave2spectro_lazy(g)
    a2s.spectro2img(g,image_names)

more details of spectro2audio

read_img_lazy

Reads a image file using delayed evaluation. It saves memory compared to normal loading.

images=s2a.read_img_lazy(paths)

img2amp_lazy

Converts the image to an power spectrum using delayed evaluation.

amplitudes=s2a.img2amp_lazy(images)

amp2wave_lazy

Converts power spectrum to sound waves using delay evaluation.

waves=s2a.amp2wave_lazy(amplitudes)

wave2wav_file

Save the sound wave as a wav file.

s2a.wave2wav_file(waves,file_names)

example of custom pipeline

def volume_up_lazy(amplitudes):
    for amp in amplitudes:
        yield amp*2000

def custom_pipeline(paths)->None:
    g=s2a.read_img_lazy(paths)
    g=s2a.img2amp_lazy(g)
    g=volume_up_lazy(g)
    g=s2a.amp2wave_lazy(g)
    return g

g=custom_pipeline(['./data/image/1.png'])
display(IPython.display.Audio(g.__next__(), rate=44100))