Interactive Lab

Put it all together — record your voice, modify it, clone it with AI, then try to detect which is real vs AI vs modified.

If you need any help, please find an on-site Call Center Village staff member.

Read this aloud at a natural pace. It covers all major English phonemes, varied intonation, and different mouth shapes. You'll use this recording throughout the exercises below.

"The quick brown fox jumps gracefully over a lazy dog sleeping beneath the old oak tree. She sells thick, fresh seashells down by the shimmering seashore every Thursday morning. Would you kindly confirm your date of birth and the last four digits of your account number? I wasn't sure if the package arrived yesterday or if it's expected tomorrow — could you please check on that for me? Absolutely, I'd be happy to transfer you to our billing department right away. Thank you so much for your patience, and have a wonderful evening!"

Enunciate each word as clearly as you can for a more expressive voice clone.

1: Record & Analyze

Record yourself reading the script above, then visualize your voice with a spectrogram and review the audio statistics. This gives you a baseline to compare against when you start modifying and cloning your voice in the next steps.

bash

# Navigate to working directory
cd ~/callcentervillage/voice-cloning

# Record yourself reading the script above (~20 seconds)
# You can also use Audacity: Record → File → Export Audio → save as my_voice.wav
rec -r 44100 -c 1 -b 16 my_voice.wav trim 0 20

# Generate a spectrogram
sox my_voice.wav -n spectrogram -o spectrum.png

# View the spectrogram in the terminal
imgcat spectrum.png

# View audio file info and statistics
soxi my_voice.wav && sox my_voice.wav -n stat 2>&1

Satisfied with your knowledge? Continue to the next module.

Continue to Voice Agents

Loading…