genshin impact ai voice
So, I’ve been messing around with this idea of getting AI to mimic the voices from Genshin Impact. It’s pretty cool, right? Let me walk you through what I did.
First, I needed to find some decent AI voice models. I did some digging, and I found a few different programs and websites. Some were free, some were paid. It took a bit of trial and error to see which ones were even remotely usable.
Finding the Data (Voice Clips)
Next up, I needed a good amount of voice lines for the characters. You can’t just have the AI spit out words without teaching it how the characters sound. So, I jumped into the game and started recording. I focused on getting a variety of lines, you know? Happy lines, angry lines, battle cries – the works. This took…a while. I also spent time cruising around YouTube to find some character voice line compilations. Saved me a bit of time, I gotta admit.
Cleaning up the Audio
Now, this is important: the audio needs to be clean. No background music, no sound effects, just the character’s voice. So, I fired up an audio editor – Audacity is free. I went through each clip, carefully snipping out any extra noise. It’s tedious, but it’s gotta be done. It’s a really really boaring job!
Training the AI (The Hard Part)
This is where things got tricky. The AI needs to “learn” the voices. I used the best model I found. I uploaded all my clean voice clips and…waited. Training takes a long time. Like, go-make-a-sandwich-and-watch-a-movie long. And sometimes, it just…doesn’t work. I had to tweak settings, try different models, and basically just keep experimenting until I got something that sounded halfway decent.
Testing and Tweaking
Once the training was done (or at least, I thought it was done), it was time to test it out. I typed in some phrases and listened to the AI try to say them in the character’s voice. Some results were hilarious, others were surprisingly good. I kept making small adjustments, feeding it more voice lines, and retraining until I was relatively happy with the outcome.
Here is the final result
- Paimon sound like Paimon finally.
- Zhongli sounds older than before, perfect!.
- I think i need more voice lines for Nahida.
It’s definitely not perfect, and it’s a constant work in progress. But it’s pretty darn fun to play around with. Who knows, maybe one day I’ll have a whole army of AI-powered Genshin characters at my disposal. That is today’s record, see you next time.
