Audio
Source Separation Requests output different audio files in the WAVE (wav) or the MP3 (mp3) format, depending on which model and output format you used.
The audio format you would like to fetch is specified beforehand and cannot be changed later on.
The following audio files can be fetched with the four-stems model:
vocalsbassdrumsother
With the six-stems model you can additionally fetch the following files:
pianoguitar
warning
The six-stems model inflicts higher quota costs than four-stems model!
Encoding of the MP3 audio files:
| MP3 Audio | |
|---|---|
| Codec | MPEG-1 Layer 3 (MP3) |
| Channels | Stereo |
| Sample Rate | 44.1 kHz |
| Bitrate | 128 kbps |
Encoding of the WAVE audio files:
| WAVE Audio | |
|---|---|
| Codec | Uncompressed 16-bit PCM audio |
| Channels | Stereo |
| Sample Rate | 44.1 kHz |
| Bitrate | 1411 kbps |
Use the /job/{job_id}/audio?stem_type=<STEM_TYPE> endpoint to get the corresponding audio file from the source separation job.