Command line utils

Requirements

You need to install ffmpeg, ffprobe to use it.

https://ffmpeg.zeranoe.com/builds/

And then add /bin/ with ffmpeg.exe and ffprobe.exe to your environment

Help

python main.py help

Dataset creator

How to use:

python main.py create_dataset <path to all audio> <path to save> <dataset name> <pony_name> <noise level>

Path arguments:

• All path arguments must be absolute and end with \ (Windows) or / (Linux)
• Wrap your path arguments into ""

Structure inside of all audio samples:

├───samples
|   ├───S1
|   │   ├───s1e1
|   │   │   └───Source files
|   │   ├───s1e2
|   │   │   └───Source files
|   │   ├───s1e3
|   │   │   └───Source files
|   │   ├───s1e4
|   │   │   └───Source files
|   │   ├───s1e5
|   │   │   └───Source files
.....

Pony name argument:

• Be sure, that it is the same as in audio. For example you need use Twilight instead of Twilight Sparkle

Dataset name argument:

This is the name of directory for you dataset. Could be any, but if you use spaces - wrap argument into ""

Noise level argument:

• Case sensitive
• Could be Noisy, Very Noisy, Clean
• You need separate them with only comas if you use not one noise level. Example: "Noisy,Clean"
• Wrap this argument with ""

Example of usage:

python main.py create_dataset "F:\projects\pony_pre_proj\all" "F:\projects\pony_pre_proj" only_clean Twilight "Clean"

Extract only audio of one pony

How to use:

python main.py separate <path to all audio> <path to save> <name of taget folder> <pony_name>

Example of usage:

python main.py separate "F:\projects\pony_pre_proj\all" "F:\projects\pony_pre_proj" twilight_folder Twilight

All arguments of this command are the same as for dataset creation.