RVC INFERPY

Note

This project is still under development..

rvc_inferpy is a Python library for performing audio inference with RVC (Retrieval-based Voice Conversion). It provides a simple command-line interface (CLI) and can be integrated into Python projects for audio processing with customizable parameters.

Installation

You can install the package using pip:

pip install rvc-inferpy

Usage

Command Line Interface (CLI)

You can interact with rvc_inferpy through the command line. To view the available options and how to use the tool, run:

rvc-cli -h

Here’s a breakdown of the full command-line options:

usage: rvc-cli [-h] [--model_name MODEL_NAME] [--audio_path AUDIO_PATH] 
                 [--f0_change F0_CHANGE] [--f0_method F0_METHOD] 
                 [--min_pitch MIN_PITCH] [--max_pitch MAX_PITCH] 
                 [--crepe_hop_length CREPE_HOP_LENGTH] [--index_rate INDEX_RATE] 
                 [--filter_radius FILTER_RADIUS] [--rms_mix_rate RMS_MIX_RATE] 
                 [--protect PROTECT] [--split_infer] [--min_silence MIN_SILENCE] 
                 [--silence_threshold SILENCE_THRESHOLD] [--seek_step SEEK_STEP] 
                 [--keep_silence KEEP_SILENCE] [--do_formant] [--quefrency QUEFRENCY] 
                 [--timbre TIMBRE] [--f0_autotune] [--audio_format AUDIO_FORMAT] 
                 [--resample_sr RESAMPLE_SR] 

Command-Line Options:

• -h, --help: Show help message and exit.
• --model_name MODEL_NAME: Name or path of the model.
• --audio_path AUDIO_PATH: Path to the input audio file.
• --f0_change F0_CHANGE: Pitch change factor.
• --f0_method F0_METHOD: Method for F0 estimation (e.g., "crepe").
• --min_pitch MIN_PITCH: Minimum pitch value.
• --max_pitch MAX_PITCH: Maximum pitch value.
• --crepe_hop_length CREPE_HOP_LENGTH: Crepe hop length.
• --index_rate INDEX_RATE: Index rate.
• --filter_radius FILTER_RADIUS: Filter radius.
• --rms_mix_rate RMS_MIX_RATE: RMS mix rate.
• --protect PROTECT: Protect factor to avoid distortion.
• --split_infer: Enable split inference.
• --min_silence MIN_SILENCE: Minimum silence duration (in seconds).
• --silence_threshold SILENCE_THRESHOLD: Silence threshold in dB.
• --seek_step SEEK_STEP: Step size for silence detection.
• --keep_silence KEEP_SILENCE: Duration to keep silence (in seconds).
• --do_formant: Enable formant processing.
• --quefrency QUEFRENCY: Quefrency adjustment.
• --timbre TIMBRE: Timbre adjustment factor.
• --f0_autotune: Enable automatic F0 tuning.
• --audio_format AUDIO_FORMAT: Desired output audio format (e.g., "wav", "mp3").
• --resample_sr RESAMPLE_SR: Resample sample rate.

Example Command:

rvc-cli --model_name "model_name_here" --audio_path "path_to_audio.wav" --f0_change 0 --f0_method "crepe" --min_pitch 50 --max_pitch 800

As a Dependency in a Python Project

You can also use rvc_inferpy directly in your Python projects. Here's an example:

from rvc_inferpy import infer_audio

inferred_audio = infer_audio(
    MODEL_NAME="model_name_here",       # Name or path to the RVC model
    SOUND_PATH="path_to_audio.wav",     # Path to the input audio file
    F0_CHANGE=0,                        # Change in fundamental frequency
    F0_METHOD="crepe",                  # F0 extraction method ("crepe", "dio", etc.)
    MIN_PITCH=50,                       # Minimum pitch value
    MAX_PITCH=800,                      # Maximum pitch value
    CREPE_HOP_LENGTH=128,               # Hop length for Crepe
    INDEX_RATE=1.0,                     # Index rate for model inference
    FILTER_RADIUS=3,                    # Radius for smoothing filters
    RMS_MIX_RATE=0.75,                  # Mixing rate for RMS
    PROTECT=0.33,                       # Protect level to prevent overfitting
    SPLIT_INFER=True,                   # Whether to split audio for inference
    MIN_SILENCE=0.5,                    # Minimum silence duration for splitting
    SILENCE_THRESHOLD=-40,              # Silence threshold in dB
    SEEK_STEP=10,                       # Seek step in milliseconds
    KEEP_SILENCE=0.1,                   # Keep silence duration in seconds
    QUEFRENCY=0.0,                      # Cepstrum quefrency adjustment
    TIMBRE=1.0,                         # Timbre preservation level
    F0_AUTOTUNE=False,                  # Enable or disable F0 autotuning
    OUTPUT_FORMAT="wav"                 # Desired output format (e.g., "wav", "mp3")
)

The infer_audio function will return the processed audio object based on the provided parameters

you can also use this new inference mode

from rvc_inferpy import infernew


infer = infernew(
    model_name="model_name_here",
    sound_path="path_to_audio.wav",
    f0_change=0
)

output_audio = infer.run_inference()
print("Inference completed:", output_audio)

Download Models

you also can download your own rvc models using

from rvc_inferpy import download_rvc_model dlrvc

dlrvc("URL_HERE", "model_name")

Note

downloas models only allowed Huggingface, Google Drive, and Pixeldrain links.

Tip

Ensure that you upload your models in the models/{model_name} folder.

Terms of Use

The use of the converted voice for the following purposes is prohibited.

• Criticizing or attacking individuals.

• Advocating for or opposing specific political positions, religions, or ideologies.

• Publicly displaying strongly stimulating expressions without proper zoning.

• Selling of voice models and generated voice clips.

• Impersonation of the original owner of the voice with malicious intentions to harm/hurt others.

• Fraudulent purposes that lead to identity theft or fraudulent phone calls.

Disclaimer

I am not liable for any direct, indirect, consequential, incidental, or special damages arising out of or in any way connected with the use/misuse or inability to use this software.

Credits

• IAHispano's Applio: Base of this project.
• RVC-Project: Original RVC repository.

License

This project is licensed under the MIT License.