AI Clone a person's voice with Piper TTS (Must provide a GPU Server) Deadline: 12 hours max
Budget / Salary$10–35
TypeFreelance project
LocationRemote
Posted1 hour ago
***** My max budget is $35 USD, SERIOUS & HONEST FREELANCERS ONLY, YOU MUST HAVE EXPERIENCE WITH AI VOICE CLONING AND PIPER TTS**
We ask that you already have access to a GPU for this project wether be on a host or your own PC.
I can send you 1 sample mp3 file that has background bed of music and dry voice. All 100 plus mp3 files are really all the same. They all are less than 5 min and they all have same voice and same bed of music. The person has given me permission to clone his voice.
1. You will use demcus to clean up background bed of music
**2. You will not use nor need all 100 + files. Just pick ones that you want less than 30 likely is enough for a good high quality voice ai clone. Your NOT converting all 100 + files. I just want to provide you enough files so you can clone ai voice. Pick as little or as many files as you want to work with.***
3. You must have used piper in past. You will use piper to ai clone - train in this project. You can bootstrap - start with the piper “ryan” preset, but the finished model must faithfully reproduce my speaker’s timbre, pacing, and articulation under normal broadcast processing.
4. You provide me all final compiled ONNX (with meta data) files at the end of the project. I def want to hear the audio of the final output before project wraps up to make sure it meets my expectations.
5. My goal is to export this voice model to piper on my radio station server. I want this voice model to play live on the air to my listeners and it will say all my scripts on the fly. You will not help me with this!!!
The END goal is a voice that sounds like the original as much as possible. I understand in piper it wont be perfect.
We ask that you already have access to a GPU for this project wether be on a host or your own PC.
I can send you 1 sample mp3 file that has background bed of music and dry voice. All 100 plus mp3 files are really all the same. They all are less than 5 min and they all have same voice and same bed of music. The person has given me permission to clone his voice.
1. You will use demcus to clean up background bed of music
**2. You will not use nor need all 100 + files. Just pick ones that you want less than 30 likely is enough for a good high quality voice ai clone. Your NOT converting all 100 + files. I just want to provide you enough files so you can clone ai voice. Pick as little or as many files as you want to work with.***
3. You must have used piper in past. You will use piper to ai clone - train in this project. You can bootstrap - start with the piper “ryan” preset, but the finished model must faithfully reproduce my speaker’s timbre, pacing, and articulation under normal broadcast processing.
4. You provide me all final compiled ONNX (with meta data) files at the end of the project. I def want to hear the audio of the final output before project wraps up to make sure it meets my expectations.
5. My goal is to export this voice model to piper on my radio station server. I want this voice model to play live on the air to my listeners and it will say all my scripts on the fly. You will not help me with this!!!
The END goal is a voice that sounds like the original as much as possible. I understand in piper it wont be perfect.
Apply on Freelancer →
Project sourced from Freelancer.com. Applications happen directly on the original platform — we never collect your data.