A tool for batch adjusting narration pauses in VOICEPEAK
A god-tier tool for narration reading
I have been relying on a synthetic voice software called "VOICEPEAK" for creating temporary narrations in my video production projects. With its natural speech, intuitive GUI, diverse voice qualities, and ease of specifying parameters like scripts, I consider it one of the most convenient tools in the AI narration world. Whether for client projects, independent productions, or draft audio for YouTube videos, it is a savior for busy creators because it can generate high-quality read-aloud audio just by inputting text.
An example of reading a sentence from Aozora Bunko using Female 1 from the VOICEPEAK 6-narrator set.
The voice of this Female 1 is very beautiful. I think many of you have heard it in various places.
On the other hand, when listening closely, don't many people feel uncomfortable with the default way the "pauses" are taken for punctuation marks (commas and periods)? For example, there is an oddly long pause after a period, while commas are read as if they are flowing. I have been feeling a bit frustrated every day by the gap between expectation and behavior, such as wanting periods to have a more solid break and wanting commas to be read more smoothly. I realize that this subtle "pause" is very important when you want to get closer to actual narrators or background narration.
A temporary workaround
VOICEPEAK itself does have a feature to manually adjust "pauses" individually. However, when you actually try to do it, finding dozens or hundreds of punctuation marks in the manuscript one by one and setting them for each is truly back-breaking and unrealistic.
I have been waiting for the official team to implement features like "batch pause editing," but it seems my requests haven't reached them yet.
Still, I wondered if I could manage it with some ingenuity, so I tried a brute-force method of opening the .vpp VOICEPEAK project file in a text editor (like Notepad or VS Code) and searching/replacing only the "pau" (pause) parts. However, it is still tedious to manually replace them multiple times while being careful about the unique JSON structure and avoiding mistakes. Especially when dealing with long manuscripts or multiple files, I really wanted a way to convert them safely in bulk.
VPPConvertKutoten
As a way to solve the above worries and work burden all at once, I developed a self-made tool called "VPPConvertKutoten." This is a desktop app that allows you to batch change the "pause (seconds)" for various symbols treated as commas and periods, such as punctuation marks, "・", "!", and "?", to any value with a single action.
It is built with the Python-based Flet framework and can be launched as an app, so even those who are "inexperienced in programming" or "do not use the command line" can easily introduce and use it. I have minimized the complexity of setup, and the environment construction (automatic installation of dependent packages) is completed just by clicking "setup.bat." I kept in mind a design that is friendly to field creators busy with daily work and users touching VOICEPEAK for the first time.
Also, this app is designed to convert multiple .vpp files at once. For example, even if you have a large number of project files for multiple episodes or scenes, you can specify "0.7 seconds for periods" and "0.3 seconds for commas" all at once. I have also incorporated exception handling so that it can process as correctly as possible even if there are incorrect formats or extra trailing data in the files. A safe design that "does not overwrite the original file and always saves the converted file with a backup" is also a feature. I plan to update the development background and additional features on the GitHub repository from time to time.
If you don't really understand GitHub, please download the zip file from below. If you want to use the latest version, please download it from GitHub.
I also prepared a script for Mac, but since I don't own a Mac, I haven't been able to test it. If you have a Mac, it would be helpful if you could try it out and give me feedback.
Introduction to how to use the tool
The actual usage is also very simple.
First, download the necessary files from the VPPConvertKutoten repository (app.py, setup.bat, etc.).
If Python (version 3.8 or higher) is installed on your PC, double-click "setup.bat." This will automatically finish the installation of Flet and related modules. If you need any version specification or custom setup, you can flexibly handle it from the batch arguments.
If you don't know how to install Python, let's ask ChatGPT.Once the setup is finished, execute "run_app.bat" to launch the app. A simple GUI screen will appear.
Enter your preferred duration in seconds for both "comma pauses" and "period pauses." (For example, "comma: 0.25 seconds, period: 0.7 seconds")
Simply select your .vpp files (multiple selections are supported) using the "Select and Run" button to complete the conversion! The converted files are automatically saved in the same folder as "original filename + _modified.vpp." You can also see the number of files processed and any errors displayed on the screen, giving you peace of mind.
Drag-and-drop is currently not supported, but we plan to implement it once the Flet framework itself is updated. If you have any requests, bug reports, or ideas for useful ways to use the tool, please feel free to contact us via the README or GitHub Issues, or by leaving a comment on note.
We aimed for a design that can be immediately utilized for a wide range of purposes, from beginners to heavy-duty narration users.
Summary
I hope this tool makes your narration production and audio editing workflow in VOICEPEAK significantly easier. Especially for those creating long texts or series, it eliminates the hassle of manual adjustments one by one.
If this article was helpful, please consider giving it a "like," sharing it, or supporting me. Also, if VOICEPEAK officially implements batch conversion or advanced adjustment features in the future, this tool's role may come to an end, but until then, please use it to help with your work.
I am also waiting for your requests and feedback in the comments, such as "It would be even more convenient if this feature were added!" Based on your feedback, I will continue to work on further improvements!