Hi, I want to record the sentences unity is saying to the players and give them a shodan (system shock 2) like voice effect. Does anyone have an idea how I can achieve this with freeware software? I used audacity, but coyping/pasting small portions of the soundwave (for the stutter effect) is rather time consuming. not even talking about the 2 other voices, constantly shifting in pitch and timbre, changing tempo and delay..
A just canned voice effect was not what I was looking for though, shodan has a whole set of effects, multiple voice lines saying the same in different speeds and pitches, a stutter effect and the "main" voice pitching up and down, to name just a few. Really difficult to do with audacity, and a lot of work
A lot of work: yes it is.
Really difficult to do with audacity: not really, but see above. It is work. Just listening to some of the Shodan lines lets me hear about 9 different effects used, some of which are just manual editing like you mentioned. Doable, but time consuming. And I say this as a working voice actor/sound editor.
What you are looking to do CAN be done with freeware. Audacity can do it. What can't be done is doing it fast and/or easily.
Don't want to be the bucket of cold water, but just like the old coffee stained paper in the oven to make parchment trick, you can't really get around the process involved.
|
If you want to get really really fancy you could write a script that takes your typed text, parses it through a sentiment analysis module and then based on which sentiment it thinks it detects, selects a different voice synthesizer to pronounce it. Don't worry if the sentiment analysis sometimes gets it wrong; odd choice of voice is a feature here :)
The voice pitch is nothing I found in audacity..
Time to do some necromancy on this thread.. I think I found a quite nice way to edit this. The statement however stays.. It's a lot of work. I think I spent 2.5 hour on this file:
It's german, but basically the first text unity says to the players (the invitation to the godmind).
I recorded the speech from google cloud text to speech (german Wavenet-C, pitch change: 0, speed change: 0.85) and imported that into audacity. Then first step was duplicating the track, changing the pitch of the duplicated to -6%. Then (after mixing the tracks to one track) the rest was just changing pitch and tempo in some parts of the track, creating that up and down in pitch and tempo. For some parts I copied out single words or sentences, played them lagging behind with changed pitch and tempo again. It's very much experimenting here, what sounds good. Also experiment with the volume of the additional tracks.
After that, I merged all the tracks into one, and again duplicated it, giving the duplicate a -6% pitch. Then again: merge to one. Over that track I laid a reverb effect (again play with the values, I chose a significant smaller room size, like 20 or so..).
Then I duplicated that track 2 times, and changed the Pan for one track completely to the left, and to the right for the other track. After that I chose segments, that play only left or only right, replacing the segment with silence on the other tracks.
Now I downloaded some sound files from System Shock 2, where the nice ambience sounds come from. Laid the low hum (lowloop.wav from the ambience folder) under the track, the creepy effect (nerve1.wav - nerve7.wav from the ambience folder chained together and looped) and the babbling background voice (sh2bab_1.wav - sh2bab_6.wav from the shodan folder, again chained and looped).
Last thing was the "intro effect", which is basically a recording of the ingame scene (you can find it on youtube) and the file shodoor5.wav from the shodan folder to blend it together.
Finally a fade in at the beginning and fade out at the end, that's it.
Here you see an image of the tracks (in the second stage, so the speech track is already mixed together)