Description
Downloaded all of Season 1 of Aggretsuko as mkv's and used HandBrake to change the audio from Japanese to English and changing the mixdown to 5.1 channels and exporting it as a mp4 (since the editing software wouldn't take mkv's). Then, I got a editing software that had NLE and converted the mp4's as 6 separate channels for audio. I muted the other channels besides the one that had the audio with people talking, so there wouldn't be any background music or noise, eliminating the use of using a background noise remover. After exporting all episodes background-free, I edited the mov's so when Fenneko only speaks and used the final product as my dataset. TL;DR: got season 1 of Aggretsuko as mkv's and getting clean audio from it. Trained with rmvpe Please credit if used HuggingFace: <
Comments
Nice!
I am interested in the process in which you isolated the audio without using UVR, is there like a youtube tutorial about it?

Wow! This is so good! Congratulations! Now we have the English version of Fenneko, I just published the model of the Italian one hours ago- KEKW.

I usually just followed these two videos and used Davinci Resolve as my editing software, since in the 1st video, the person uses Adobe Premiere, so I had to find a free alternative. https://youtu.be/CxRuINwJI0A https://youtu.be/6O59lnnJ-Po

Yep! I'm glad on how it came out! I was also on thinking on making a Retsuko Model and a Haida model as well. I started on Fenneko since I really like her and her voice.
Thanks!

Well, good luck if you do them!

I prefer Haida...but yes, she is a very good character.

Okay. I'll see if I can get a model of his voice soon.

Good luck again! 🙂
Add a comment
Samples
Pitch
More to explore
Loading more