" A simple and easy-to-use voice transformation (voice changer) framework based on VITS. "
Configuration requirements
1 Win10 or win11 system, the old graphics card driver must be updated to the latest version
2 Graphics card 750TI or above, if you need to train audio sources, you need a graphics card above RTX, the larger the memory, the better
3 The CPU supports AVX2 instruction set, the higher the performance, the lower the inference delay , It is recommended that the 10th generation Core or the 3rd generation Ryzen or above
4 clear and clean microphones, the sound card microphone is the best
interface preview
The test results are as follows:
After the test, the requirements for the microphone are very high. The Bluetooth headset I used caused relatively loud noise and the sound was not very clear. If the sound is intermittent, you can change the delay sampling time.
1 AI voice changer does not rely on personal voice, it only needs to be clearly spoken, traditional voice changer can only change according to personal voice 2 It can
train models, load various voices, tone and voice diversification
Compared with traditional voice changers Disadvantages of the voice changer 1. The delay is too high, usually more than 1-5 seconds. Traditional voice changers
can have no delay at least . CPU and GPU are dual-occupied, and the configuration is low and cannot be used. Most of them are currently only compatible with N card 4. Compatibility is still too poor. AI voice changer is still immature, and the stability of flashback needs to be optimized.
Friends who want to try can click to follow, and the background will reply to RVC to get the download link