Build your own unit selection voice (SPCC version)

Where to do this exercise

You should perform this exercise in the working directory ~/Documents/ss_2017. If you don’t see this directory, then download this zip file and unzip it in your Documents directory.

Important!

Every time you start a new shell (e.g., a new terminal window), you need to source the setup.sh script. Let’s do it now:

bash$ source setup.sh

You won’t see any output, but various environment variables have been configured for you.

Configure Festival’s audio output

To configure the audio output for Festival, there is a file in your home directory (not in Documents) called .festivalrc (the name starts with a period). It should contain these two lines of code:

(Parameter.set 'Audio_Method 'Audio_Command)
(Parameter.set 'Audio_Command "play -t raw -r 16k -e signed -b 16 -c 1 $FILE")

If you need to create this file, and to do all the script editing that comes later, use any plain text editor (but not a word processor), such as gedit

Let’s get started

Now follow the Build your own unit selection voice instructions, but with these small changes:

• Introduction – read this
• Prepare your workspace – skip
• The recording script – skip
• Make the recordings – skip – we will use one of the free ARCTIC corpora instead
• Prepare the recordings – skip
• Label the speech – do this
• Pitchmark the speech – do this
• Build the voice – do this
• Run the voice – do this

Some steps take a little time to run, but you can read the instructions whilst that happens

• do_alignment 16 minutes

If you run out of time, there is a pre-built version in ~/Documents/ss_prebuilt