You're looking at a specific version of this model. Jump to the model overview.
deepsbhat1984 /voice-clone-pro:9a19c963
Input schema
The fields you can use to run this model with an API. If you don’t give a value for a field its default value will be used.
| Field | Type | Default value | Description |
|---|---|---|---|
| source_url |
string
|
https://getjob-resumes.s3.us-east-1.amazonaws.com/music_app/5/song/fe77b87687094000852ee017273e15a6.mp3
|
URL of the sung song whose melody/lyrics to keep. A full "
"mixed song is fine - the vocal is separated automatically.
|
| target_url |
string
|
https://getjob-resumes.s3.us-east-1.amazonaws.com/music_app/replicate-test/deepfemale-0b30d8e1.m4a
|
URL of the reference voice clip to convert INTO (a clean "
"15-30s+ clip of the target voice; isolated singing works best).
|
| diffusion_steps |
integer
|
40
Min: 10 Max: 150 |
Quality vs speed. 40 is a good default; higher = more "
"detail, slower.
|
| grit_restore |
number
|
0
Max: 1 |
Extra DSP grit/dynamics restore on top of the model. "
"Default 0 - YingMusic already preserves grit; extra processing can "
"muddy the clean mix. Only raise for a specific flat-sounding clip.
|
| reverb_amount |
number
|
0
Max: 1 |
EXTRA reverb on top of the model's mix. Default 0 - "
"YingMusic already mixes + reverbs the accompaniment, so adding more "
"double-reverbs it ('echoing too much'). Leave 0 unless you want more.
|
| return_vocal_only |
boolean
|
False
|
DIAGNOSTIC: return ONLY the converted vocal (no music), so "
"you can verify it sounds like the target voice BEFORE judging the full "
"mix. If it doesn't sound like the person, the mixing is irrelevant - "
"the conversion/reference is the problem.
|
Output schema
The shape of the response you’ll get when you run this model with an API.
Schema
{'format': 'uri', 'title': 'Output', 'type': 'string'}