Extract speech with actors and audio

Ask for help with ScummVM problems

Moderator: ScummVM Team

Post Reply
guga2112
Posts: 2
Joined: Fri Apr 02, 2021 6:18 am

Extract speech with actors and audio

Post by guga2112 »

Hi all,
some friends and I are planning to make an Italian fan dub of DOTT, so I looked into how scummvm and its tools extract and insert custom data in the game.

I was able to swap some files in the monster files, have scummvm play my own moster.sog, it worked for a small demo.

However, it took me quite some time: I had to extract the Italian lines with scummtr, the audio files, then loop through the audio files until I found the speech lines I wanted to change, overwrite them. It's not sustainable if I want to create scripts for the different voice actors (how do I know who says each line?) and if I want to give them the original lines as an intonation reference (how do I know which audio file refers to which line without having to listen to them all?)

Is there a way to 1) extract all lines from the game also saying which actor says them, and/or 2) extract all speech lines with an indication on which audio file is associated to?

I know scummspeaks is able to insert new audio lines into a game, but that's not what I want to do, since DOTT already has the lines, and I couldn't get it to extract the current line-voice mapping from the game.

Thanks a lot in advance!
rzil
Posts: 30
Joined: Wed Aug 17, 2016 12:56 pm

Re: Extract speech with actors and audio

Post by rzil »

Hi,
You can find mapping of remastered edition audio files to SOU offset here
(these are the same offsets as original game, just that the audio files names can provide some context: actor + scene):
https://github.com/BLooperZ/remonstered ... onster.tbl
You'll also be able to extract the audio files themselves by using this project and uncomment this block: https://github.com/BLooperZ/remonstered ... rce.py#L37

this will give you offset <-> audio file mapping

you can also get offset <-> text line mapping by using scummtr or https://github.com/BLooperZ/nutcracker
each spoken line will begin with the offset of the audio file as escaped bytes.

by combining both, you get text line <-> offset <-> audio file mapping

Good luck, feel free to contact me for help, I were also using those projects for the same cause.
guga2112
Posts: 2
Joined: Fri Apr 02, 2021 6:18 am

Re: Extract speech with actors and audio

Post by guga2112 »

Thanks a lot! I was finally able to extract all audio files, separate them by actor and save the corresponding translation lines in a text file. We're ready to start!

Now... putting the newly recorded audio back in the games is another story, but we first need to record it, so we have time :D I'm trying to understand the structure of the .sog file, is there some documentation or I have to look at the source code? I think I got most of it, but there are a couple of bytes I don't understand
Post Reply