Tuesday, October 25, 2022

Virtual Voices for Pokémon Characters

I've always been fascinated by a certain device in the Pokémon anime that allows practically any character to speak into the device and to be heard as an entirely different character. It has its uses, notwithstanding unsavory ones in at least one instance. Apart from that, though, it has at least some entertainment purposes, not just in that world but also in the real world, the latter should the device come into existence in real life. Thankfully, through the wonders of today's technology, it may be possible to realize such a device, especially for the specific purpose of Pokémon characters, even with only throwing out conjectures.

(For the purpose of character examples, I personally refer to Ash here, but this can conceivably apply to any Pokémon character.)

Fans of Japanese pop culture today will no doubt recognize a particular series of software for the generation of virtual voices - I won't say which, but it becomes rather obvious - and with some adaptation, they could fit in a Pokémon framework. What makes them "tick" is that they are based off of actual voices, and this leads to the possibility of one for Ash being based off Rica Matsumoto's voice work, at least for the Japanese side of things. For the English side of things, this could get rather convoluted even with the natural involvement of Sarah Natochenny's voice, and a separate sophistication may be necessary - though this technology is arguably already rather sophisticated.

Outside the framework of that software series, today there are what is called text-to-speech (TTS) generators that do a pretty good job, and I'm sure most people today have had experience with them for one purpose or another. They can generate voices locally or even through the Internet (online, or by the "cloud"). For Pokémon characters like Ash, though, they would need to be really adjusted to make them sound right and true to the characters, which is a hard task but not an impossible one, since a great variety of voices in various languages also exist. This also has the potential to be more user-friendly and be closest in purpose to the device I mentioned in the beginning.

A solution that's neither of the above - and one that is comparatively low-tech to the high-tech ones above - is to keep a voice bank for the characters, which for the example of Ash would be almost every piece of dialogue that Rica and/or Sarah has ever provided. This solution isn't too versatile, but can be fairly standard. Further, it has been theorized that generation of English speech only requires 44 phonemes (the "bones" that make up speech in the language), and so if Sarah's voice bank contains all of that, then it may be a precursor to virtual voice generation. The purpose may be as it should be.

Technology today is amazing, as remarked by a certain famous guy in Pallet Town, and that seems to apply well to virtual voices, especially if they can concern something as specific as Pokémon characters like Ash. It would then be just a matter of development to find out which one works the best and universally, and in that way, it would fulfill the realized concept of that specifically interesting device from the anime that alters characters' voices. The fascination and the concept may not have been so far-out after all.

Three years ago: One-Upping Ash
Four years ago: Dilemma of Colors

No comments:

Post a Comment

Hi folks! Feel free to comment, but know that I'll be selecting only the most appropriate and relevant comments to appear. Think before you post.