Googol GLaDOS 2.0 Voice Tutorial

TL;DR - How to emulate the voices of GLaDOS et al.

If you have come here for the animation of Googol GLaDOS, I encourage you to watch it before you continue.

If you’re here to see how the voices were made…

LETS GET STARTED.1


So we see below are the following 6 AI voice assistants in the room in Googol GLaDOS 2.0.

Click on the name to find out were they are from. (In order of appearance from the movie):

GLaDOS - Portal/Portal 2 (Video game)

Watson - IBM

Siri - Apple

Android - Google (Using Google Now)

CNTRL PANEL - RFK Trilogy (My own creation)

Macintoshiba - Apple Macintosh Plus + Toshiba.

(get it? ;P please don’t sue me Toshiba & Apple! D:)

At this point you should have a brief idea of who the characters are and where they come from.

Below are the tools and methods used to create the voices of the characters, click to find how it was made.

GLaDOS

Watson

Siri

Android

CNTRL PANEL

Macintoshiba

Glossary:

TTS - Text To Speech

GLaDOS2

Tools:

Mac:

Operating System TTS Software
Mac OS X 10.7+ Samantha US English* Celemony Melodyne, Adobe Audition

Windows:

Operating System TTS Software
Windows 7 Crystal 16k2 Celemony Melodyne, Adobe Audition

Steps:

  1. Type in text into a TTS software, and save the file.
  2. Use Celemony Melodyne to ‘gladosify’ the file. (look at the voice resources)
  3. Export it as an audio file.
  4. Edit the audio in your favourite audio editor**.

Voice Resources:

Celemony Melodyne

Melodyne Tutorial

TTS generator (choose Crystal 16k)

Adobe Audition

Follow these steps and you should be able to recreate GLaDOS’s voice.

*(Mac only.)

**Or you can use your own.

Watson2

Tools:

Mac:

Operating System TTS Software
Mac OS X 10.7+ Tom US English* Adobe Audition

Windows:

Operating System TTS Software
Windows 7 Mike 16k2 Adobe Audition

What I did:

  1. Type in text into a TTS software, and save the file.
  2. Edit the audio in your favourite audio editor**.
  3. Edited voice using effects (reverb + mastering + chorus)

Voice Resources:

TTS generator (choose Mike 16k)

Adobe Audition

Follow these steps and you should be able to recreate Watson’s voice.

*(Mac only.)

**Or you can use your own.

Siri

Tools:

iOS:

Operating System TTS Software Hardware
iOS 5+ Siri MobileTerminal, Adobe Audition, Audacity Two way audio jack.
Requirements:

iPhone 4S and above

iPad 4 and above

iPod Touch 5 and above

MUST BE JAILBROKEN.

Steps:

  1. Go to Cydia and search speak then install package.
  2. Install MobileTerminal.
  3. Launch MobileTerminal.

  4. Type the speak command in your iDevice:

    speak hello world, I can speak.

    At this point you should hear Siri speaking, This means you can record the voice using your favourite audio editor*.

  5. Plug in a two way audio jack.

  6. Record using your favourite audio editor* (I used Audacity for this one.).

  7. Edit the audio in your favourite audio editor*.

Voice Resources:

Audacity

Adobe Audition

MobileTerminal

Follow these steps and you should be able to recreate Siri’s voice.

*Or you can use your own.

Android

Tools:

Android:

Operating System TTS Software Hardware
Android ICS + Windows 7 Android US Male Google TTS, Android TTS software, Adobe Audition, Audacity Two way audio jack.

Steps:

  1. Download a TTS application from the Google Play Store.
  2. Type in text into the TTS software.
  3. Plug in a two way audio jack. (If you cannot save the file.)
  4. Record using your favourite audio editor* (I used Audacity for this one.).
  5. Edit the audio in your favourite audio editor*.

Voice Resources:

Audacity

Adobe Audition

Follow these steps and you should be able to recreate Android’s voice.

*Or you can use your own.

CNTRL PANEL

Tools:

Windows:

Operating System TTS Software
Windows 7 FL Studio TTS (VST effect) FL Studio, Adobe Audition

Steps:

  1. Open FL Studio.
  2. Type in FL Studio TTS VST (pre-built by FL Studio)
  3. Choose the voice which sounds like a distorted martian.
  4. Type in the text you want.
  5. Save voice in FL Studio.
  6. Edit the audio in your favourite audio editor*.

Voice Resources:

Audacity

Adobe Audition

FL Studio

Follow these steps and you should be able to recreate CNTRL PANEL’s voice.

*Or you can use your own.

Macintoshiba

Tools:

Mac:

Operating System TTS Software
Mac OS X 10.7+ Fred* Adobe Audition

Steps:

  1. Type in text into a TTS software, and save the file.
  2. Edit the audio in your favourite audio editor**.

Voice Resources:

Audacity

Adobe Audition

Follow these steps and you should be able to recreate Macintoshiba’s voice.

*(Mac only.)

**Or you can use your own.

Where I said (TTS software) you may have to search for one depending on your OS, but there are various TTS software out there for Android, Mac, Windows and iOS so just search on Google!

If you have any questions, just ask.

Have fun!

  1. This is an updated tutorial from the gist and coderwall versions. The original version is archived at the wayback machine ↩

  2. It used to be the case that this voice didn’t work because the original URL to the voice generator was dead, I have added an alternative with a minor caveat: You’ll need to find a way to save the generated voice which is out of the scope of this tutorial (Hint: uses Flash). (Use the Mac voices instead if you don’t want to bother using the alternative voice generator.) ↩

Written by on


UP NEXT... Full Time Education vs Spare time