>>149108103
https://vocaroo.com/1iKdkh6JigRM

https://aistudio.google.com/app/generate-speech
https://github.com/Plachtaa/seed-vc
https://github.com/chameleon-ai/vevo

>>105635242
I use Seed-VC for voice conversion. I use Minimax Text to Speech cuz it's cheap. It's $5 dollars a month for 100,000 characters and emotion control. I paid $30 for 1 million credits which equals 1 million characters that expire after 2 years. Note if you do plan to use Seed-VC, be sure to use the V1 model for fine tuning. Edit the max len in the config_dit_mel_seed_uvit_whisper_small_wavenet.yml config file to 700 (which equals 8.75 seconds) for fine tuning. Edit the dataset with audio slicer. Also note that the converted Seed-VC voice won't be able to capture the emotion unless the source input file like a Minimax output file for example is more than 16 seconds long. Seed-VC requires clean data, use BandIt Plus model via Music Source Separation Training by ZFTurbo, then clean it up with resemble enhance by resemble-ai, then clean up that file with the audacity plugin, Acon Digital DeVerberate 3.

Here's a sample.
Minimax Output File
https://vocaroo.com/1cw90SLQn2VU
Minimax Output File fed to a Seed-VC fine tune model of Merida from Brave
https://vocaroo.com/1m6QldyyUzTa