Topic: Castlevania Symphony of the Night Japanese fan Translation Ai PSX Dub new video

[mgfcortez]

well got me another pc so I can train models so been working on a Japanese fan translation that I found of a old Japanese text patch. I really enjoyed the script, so been cloning it. and probably got 90% of it finished just need to clone some more models to get better emotions because eleven labs can only do so much. I have already trained Alucard, Dracula, Death, ferryman, librarian, need to do a model for Richter and Maria, just need to sort the hundreds of audio files to use as datasets to train them. I Dub most the audio with eleven labs in the past few months. and eleven labs gave me such a big amount of audio like my gospel of Mark a hour and 40 minutes of Alucard it was was pretty easy to train Alucard with that. I am looking forward to finishing this script it's a dream come true to do something so cool then get to play with it. anyway my software for recording video's has been freezing the video. so recorded this with my phone on the emu. sorry if the quality suffered but wanted to go ahead and share some of the script that I already polished and fined turned in game. also recreated Alucard's memory of his mother dying as a bonus at the end I just wanted to do it because I can lol enjoy.

https://youtu.be/bXuQGZxYJXM?si=piCzuQttRXxY7Wyt

[Reinhart77]

well that was fun, thank you.

i wonder if this could be used to make an English audio of the radio drama Nocturne of Recollection?  maybe train AI on the Japanese voices for some of the original characters?

 i was having some fun with ChatGPT playing Symphony of the Night, asking it to give Richter Mode a story.  I lol-ed after Richter threw a tantrum and trashed the place because the Master Librarian refused to help him.

[mgfcortez]

thank you for replying.  nobody wants to reply after 200 plus people seen this post. so glad you said something, and glad you enjoyed it. far as the radio drama I already plan on doing it but sorry, I'm going to do it with the PSX voices. I will finish this Japanese Translation fan dub first. after all I got a few more models to train. trained Richter the other day because I wanted better emotions from him than eleven labs. but like I said eleven labs have increased my datasets to help me train better models. Like with just the 5 minutes of audio from the game wasn't really sounding good when I trained my first Alucard model. but then I used my gospel of Mark I made with Alucard reading it on eleven labs, an hour and forty minutes. it came out great, so I been scraping up all my audio for each voice I train. anyway, yeah if someone wants to try to train the Japanese voices. I wish them luck, but IDK if you can scrape together enough audio to get good models. maybe. I guess if you took the Japanese audio from both the game and the radio drama it might work out IDK. but I tried to get good audio from the PSX Dracula, but only a minute and a half wasn't enough to get good audio from eleven labs or training a model myself. so went with PSP Dracula because he had audio from 5 games. and even then, it was only hit and miss on eleven labs. got good audio only 30% of the time. and depending on what I said, truthful the opening Dracula lines were changed many times and wasted a bit of money until I got good audio. so there not 100% faithful to the Japanese fan script. but closer than anything else. it's basically the same but just expanded like with the force and power part, But anyway, thanks to the good audio from Castlevania Arcade I made for him and the script for SOTN I trained me a good Dracula DXC model.

anyway cloning voices is very cool.

[Reinhart77]

it is pretty cool playing around with that stuff, so much you can apply it to.  for the audio drama, psx voices would be good when possible, i just wasn’t sure what you’d use for the original characters who weren’t in SotN.  not sure if models trained on Japanese voices would sound good at all in English.  or maybe it’d be better to just use a different voice for those.  i do enjoy listening to this kind of stuff while driving or walking. i appreciate the time and effort it takes to train these models too. it’d be interesting to hear games that had text-only dialog as well, such as Dawn of Sorrow.  at the very least you could use Alucard’s voice in that one, ha. 

[mgfcortez]

well I actually was thinking about trying to train a voice from Castlevania curse of darkness use Isaac's voice for the male Succubus in the radio drama. far as the other voices I haven't made up my mind on yet, might try to train the Japanese voice's if I can't pull nothing else off. maybe use Alucard's voice from the PSP or Netflix show for his friend IDK. haven't finished the voices for SOTN yet. just figured snice I already training these voices. it would be nice to use them, and I wanted to do the radio drama a long time ago, I edited the first part years ago to take the Japanese voice off. now it's easier with AI. I got something called Rip X pro. that can separate music from voices. I couldn't wait to give them that 150 dollars. after all I use to do a bit of Dubbing in the past. and I took the whole Chinese movie Monkey king and edited the Chinese out and Dubbed it in English. very proud of that. took me a month of full long days to do that. now I can just load it up and strip it right off, I love it. and I want to be able to clone voices so it was a no brainer for me. anyway feel free to offer any idea's on voices for the radio show.

[Reinhart77]

Good idea using Isaac’s voice for the Inccubus. I could hear that. 

[13th Street]

Cool! I'm so glad this technology can be used for things like this. 

[Inccubus]

Neat! I'd love to see this paired up with a text edit.