❓FAQ

Find the most frequently asked questions here.

Index

• General Questions

• Questions about Voice Clone Guides

• Questions about Model Creation Guides

• Questions about Other Guides-> Questions about Real-Time Voice Changer:

Most of the issues stem from Python:

1. Access the Python download page.

2. Click on the "Download Python" button.

3. Once the download is complete, run the installation file and follow the instructions to install Python.

4. During the installation process, make sure to check the boxes next to "Add Python to PATH" and "Install IDLE."

Also, never use paths with spaces or special characters:

❌ D:\Test 1\AI or D:\Test 1\The AI ✅ D:\Test-1\AI or D:\Test1\AI

General Questions

What are "epochs"?

It reflects the number of passes of the algorithm during the training phase. One can define an epoch as the number of passes of a training dataset by an algorithm. One pass is equivalent to a round trip. The number of epochs can reach several thousands, as the procedure repeats indefinitely until the model's error rate is sufficiently reduced. To put it simply, if the dataset is of good quality, the more epochs there are, the better the quality will be.

What is "pitch"?

"Pitch" in music refers to whether a sound is high (like a bird's song) or low (like an engine noise). It adds interest to the music and helps create melodies.

To make the music sound higher, you increase the pitch, and to make it sound lower, you decrease it. It's like adjusting the notes in a song to convey different emotions and moods.

What is "Feature Retrieval"?

In essence, it's about accentuation.

To delve further: Feature retrieval typically refers to the extraction of specific characteristics from a vocal signal or audio sample. These characteristics can include information such as timbre, rhythm, tempo, volume, and so on. In the context of RVC GUI, the "feature retrieval" option could be used to extract specific features from the vocal signal and potentially modify them, thus emphasizing the accent or other aspects of the vocal signal.

What is the best f0Detector?

The choice of the "f0Detector" model depends on how you plan to use it, whether it's for singing, speaking, rapping, etc. Here are recommendations for different use cases:

• RMVPE: It offers excellent quality and high performance, suitable for all purposes.

• Harvest: Suitable for basic conversations and rap with lower pitches.

• Dio: Suitable for basic conversations and rap with medium/high pitches.

• Crepe / Crepe-full: Recommended for speaking and singing with various pitches.

• Crepe-tiny: A faster and less resource-intensive version of the Crepe model, ideal for many use cases.

Select the model based on your specific needs to achieve the best possible results with voice conversion.

Questions about Voice Clone Guides

FileNotFoundError

This error occurs because you haven't added the path to the Applio-RVC-Fork/runtime/scripts directory to the Windows PATH variable. To do this, you need to modify the environment variables and add the path to the Applio-RVC-Fork/runtime/scripts directory at the top of the list.

Follow these steps:

1. Access advanced system settings:

   • In the Windows search bar, type "Advanced system settings." The System Properties window will open.

   
2. Environment Variables:

   • In the System Properties window, click on the "Environment Variables..." button at the bottom right.
3. Modify user or system variables (based on your preference):

   • In the Environment Variables window, you'll see two sections: "User variables" and "System variables."

   • If you want the modification to be available only for your user account, modify the "User variables." If you want the modification to be applied globally for all users, modify the "System variables." Note that modifying system variables typically requires administrator privileges.

   
4. Modify the PATH variable:

   • In the appropriate section (User or System), locate the "Path" variable in the list of variables and select it.

   

   • Click the "Edit..." button to make changes to the selected variable.
5. Add the path to the Applio-RVC-Fork/runtime/scripts directory:

   • In the Edit Environment Variable window, click "New" to add a new entry.
   • Enter the full path to the Applio-RVC-Fork/runtime/scripts directory. Make sure to separate multiple entries with semicolons (;) if there are already entries.
   • For example, if the path is C:\Program Files\Applio-RVC-Fork\runtime\scripts, you would add this path to the list.

6. Move the added path to the top of the list:

   • It's important to place the newly added path at the top of the list to give it priority over other conflicting paths.

   • Use the "Move Up" button to move the selected path to the top of the list. Repeat this until it's at the top.
7. Apply the changes and close:

   • After adding and moving the path, click "OK" in all open windows to save your changes and close them.
8. Reopen the command prompt:

   • To ensure that the changes take effect, you may need to reopen the go-applio.bat file that was open before making these modifications.

The PATH variable will now include the path to the Applio-RVC-Fork/runtime/scripts directory at the top of the list, and you should no longer encounter the "file not found" error when running Applio!

What settings should I use if the voice is rather low or rather high?

It depends on the sound; you have to experiment and adjust according to the accents, the voices, the breaths, etc1. Take your time, experiment and listen progressively to find the right settings.

What are the recommended GPUs for voice cloning?

• NVIDIA GeForce RTX 30xx Series (e.g., RTX 3090, RTX 3080)

• NVIDIA GeForce RTX 20xx Series (e.g., RTX 2080 Ti, RTX 2070)

• NVIDIA Quadro Series (for professional performance)

• NVIDIA Tesla Series (for intensive AI workloads)

• NVIDIA A100 (for intensive AI tasks and extreme performance)

• AMD Radeon Instinct Series (for professional tasks)

However, it's important to note that voice cloning performance also depends on the software you are using. Some voice cloning software can take advantage of specific features of certain graphics cards or other hardware accelerators. Before purchasing a graphics card for voice cloning, check the specific recommendations of the software you are using.

I still get "No module named" errors. What should I do?

First:

pip install -r Requirements.txt

Otherwise: If you continue to encounter "No module named" errors, it means Python cannot find certain libraries. Here's how to resolve it:

1. Identify the name of the missing module in the error message.

2. Execute the following command to install the missing module (replace "module name" with the actual module name):

   pip install "module name"

3. Repeat this step for each missing module until all errors disappear.

If nothing works, consider deleting everything and reinstalling.

Rather GPU or CPU

GPU: Faster

CPU: Slower

Questions about Model Creation Guides

When I press CTRL + B in Audacity, nothing happens.

Go to "Edit" > "Preferences" > "Keyboard" > Search for: "Place a label at the selection" and change the key.

What is a "dataset"?

A dataset is a structured collection of data that is typically organized into tables, lists, files, or similar formats. These data can be of various types, such as numbers, text, images, videos, etc. When creating voices, the data would be audio.

What should I do at the "Enter the path of the audio file to be processed" step (there is a default example in the correct format)?

At this step, you should provide the full path to the audio file you want to process from your local computer. You can do this by copying the file's path. Ensure that the audio file is located in a directory accessible from your computer because the program will access the file from this path for processing. If you encounter any errors, please share them so we can assist you in resolving them.

Rather GPU or CPU

GPU: Faster

CPU: Slower

I still get "No module named" errors. What should I do?

First:

pip install -r Requirements.txt

Otherwise: If you continue to encounter "No module named" errors, it means Python cannot find certain libraries. Here's how to resolve it:

1. Identify the name of the missing module in the error message.

2. Execute the following command to install the missing module (replace "module name" with the actual module name):

   pip install "module name"

3. Repeat this step for each missing module until all errors disappear.

If nothing works, consider deleting everything and reinstalling.

Questions about Other Guides

Rather GPU or CPU

GPU: Faster

CPU: Slower

Questions about Real-Time Voice Changer:

What settings should I use if the voice is rather low or rather high?

It depends on the sound; you have to experiment and adjust according to the accents, the voices, the breaths, etc1. Take your time, experiment and listen progressively to find the right settings.

I get the error message: "No such file or directory."

This error message typically indicates that the specified file or directory does not exist in the path you provided. Please ensure that the file path is correct and that the file or directory you are trying to access actually exists. If the file is located in a different location, please specify the correct path.

You can't hear me when I have the Voice Changer Realtime enabled.

In the Voice Changer software:

1. Select your main microphone for input.

2. Choose the audio device "VB-Cable: [MME] CABLE Input (VB-Audio Virtual Cable)" for output.

3. Use the monitor to listen to the output audio. Select your headset.

4. Make sure your headset is correctly configured as the default output device in your system settings.

In other software (e.g., Discord):

1. For input, select "CABLE Output."

2. For output, choose your headset.

These settings should enable you to use the Voice Changer Realtime with your microphone and hear the modified audio through your headset. Make sure all your software settings match these configurations for a seamless experience.

I'm experiencing stuttering/horrible voice quality with the Voice Changer Realtime.

First, use VB-Audio Cable, as it's simpler than Voicemeeter. This guide will focus solely on VB-Audio. You can download it from this link: https://vb-audio.com/Cable/ Download it for either Windows or Mac, depending on your platform. Once downloaded, run "VBCABLE_Setup" as an administrator. If nothing opens, do the same with "VBCABLE_Setup_x64" (also as an administrator). When the application opens, install it and restart your PC.

After the restart, open RVC. In the "AUDIO" category at the bottom left, under "output," select "CABLE Input (VB-Audio Virtual Cable)." In Discord, under "Voice & Video," set the input device to "CABLE Output (VB-Audio Virtual)." Try to see if you can hear yourself; it's likely that the issue persists.

For the last step, go to your "Task Manager," then to "Details," and find "audiodg.exe." Right-click on "audiodg.exe," choose "Set priority," and select "High." Finally, right-click on "audiodg.exe" again, choose "Set affinity," and check the box for "All processors," but only select "CPU 2."

That should solve your problem!