meme / amazon-music Goto Github PK

In certain cases (it seems to be non-playlists and modern albums), Amazon will respond with an MPEG-DASH file format. This contains all of the little bits of the song which can be combined to produce the final song. VLC has support for playing and converting to .flac, .ogg, etc. as the returned response includes authentication tokens in the request URL for the audio segment.

Pseudo-code that extracts the manifest object (containing the MPEG-DASH, which can be fed into VLC):

const fs = require('fs')
// ...
const json = // Retrieve the dmls.json response
const manifest = JSON.parse(json).contentResponseList[0].manifest
fs.writeFileSync('example.mpd', manifest)

Then, either use cvlc example.mpd to stream the audio (note that the contentResponseList contains an expiry, so this .mpd will not be usable forever), or use VLC to convert the .mpd to a FLAC file:

#!/bin/bash

vlc -I dummy $1 ":sout=#transcode{vcodec=none,acodec=flac,ab=128,channels=2,samplerate=44100}:std{access=file{no-overwrite},mux=flac,dst=$2}" vlc://quit

And run the following script with: ./script.sh example.mpd example.flac, and VLC will start to download the segments and recombine to produce the DRM-free, distributable, audio file.

Options:

1. Return the MPEG-DASH file from a function specific for this, then pipe the output to a file to be played by VLC
2. Use libVLC bindings for Python (do these exist?) to convert the MPEG-DASH in memory (this would be the best solution)

The plan is to break this out into a separate project that is specifically suited for downloading albums and playlists to DRM-free FLAC files on disk (for personal use).