Sounds Odd #14: Novel Uses For Voice Recognition Software, by Elizabeth Holmes

SOUNDS ODD  by Elizabeth Holmes
Fun Facts on the Science of Sound

#14. Novel Uses For Voice Recognition Software

Am I the only one, or is voice over work making you
more aware of the sounds around you?

 

The work we do requires us to notice every auditory
detail. And, for many of us, this noticing
has become a habit outside of the studio as well. 

Whether I’m in an urban environment or a rural one, I find myself
noticing things that I used to tune out as a matter of course.

 

THE SOUNDS OF YOSEMITE

For example, a few months ago, I stayed overnight
at Yosemite National Park. The ranger
lecture at that time was “Sounds of Yosemite,” which is where I learned
about a fascinating owl study that Park biologists are conducting using voice
recognition software.

 

Yosemite is home to a rare breed of Great Gray Owl
that was separated from its nearest relatives (Canadian Grey Owls) 30,000 years
ago.

To date, efforts to save this
dwindling population of birds has involved capturing, banding and observing them
at close range. These techniques are both
intrusive and counter-productive, because they often disrupt the owls’ mating
cycles, putting them at even greater peril.

 

SOUND ENGINEERS RESCUE!

Fortunately last summer, Joe Medley, a PhD
candidate in ecology at UC Davis, found a better way to study the birds. 

He placed 40 data-compression digital audio recorders in their habitat, then perfected voice
recognition software that allowed scientists to identify individual “voices.”

 

The amount of data that they gathered was staggering –
nearly 50 terabytes of sound! Without a
way to sift through that soundscape of airplane engine noise, frog song,
coyotes barking and bears rummaging (not to mention critters chomping on the
expensive equipment), it would have taken seven years for researchers to
sort those sounds by ear.

 

Medley modified existing
voice recognition software with algorithms that searched for the specific
frequency and time intervals of the Great Gray Owls’ low hoot. 

The program is so sophisticated that it can
tell the difference between males, females and younger birds, as wells as
distinguish the cry of nesting females calling for food.  These details help scientists improve the
Great Gray Owls’ chances of survival.

 

CLOSER TO ‘HOLMES’

As it happened, a few
weeks after that vacation, my niece introduced me to yet another variation on
“speech recognition” software when she visited my home for a family
celebration. 

We hosted my dad at that
event as well, who (at 94 years old) still enjoys the music of his youth –
Big Band era jazz. 

 

The surprise for me was
that my niece (who is 60 years younger) liked the music, too. And she captured those tunes out of thin
air with Shazam, an app on her iPhone that told her instantly who was playing
and the name of the song. 

 

FUTURE OF VOICE CASTING?

It got me to
thinking: Is it just a matter of time
before voice seekers have similar software that identifies each of us
individually as voice actors? 

It’s
already being done with facial characteristics. Why not voices?  Something to think about when it comes to our own contributions to the
soundscapes around us.

This article was adapted from Great Gray Owls of Yosemite Studied Through
Use of Digital Audio Recording Software
, by Tracie Cone (Associated Press). 
—————————————————–
ABOUT ELIZABETH
Elizabeth Holmes is a writer, voice actor, and staff editor at VoiceOverXtra, based in Northern California. She is also editor of VoiceOverXtra’s book division, including Voice Over Legal, by voice actor / attorney Robert Sciglimpaglia.

Email: elizabeth@HolmesVoice.com
Earlier Sounds Odd Columns: http://bit.ly/SoundsOddColumns

Similar Posts

4 Comments

  1. I always learn so much from Elizabeth’s columns. Thanks, Elizabeth, for your words of wisdom, and thanks,
    John Florian, for publishing her articles.

  2. Fun article, Elizabeth! And great musings about how it all may affect voice actor casting! Imagine: A client gets a synthesized voice “sound” that they like. Then they send it to a voice-hosting place (voices.com, etc) who simply uses VRT to match the synthesized sound desired with existing demo sounds…presto! A match! And no need to bother the voice actor! Just steal the demo-sound and re-produce it with a synthesizer machine. That’s pretty scary! Just sayin’…hey, you know, soon we’ll probably put a man on the moon! =)

  3. This is amazing, Elizabeth! Thank you for sharing this ingenious use of voice recognition technology. What an interesting article!!

  4. Hi, Elizabeth. I really enjoyed your article on voice recognition. As an animal lover, it was great to learn how much care was taken in the owl research, and it appears the voice recognition software made all the difference. It will be interesting to observe the progression of voice recognition software in our industry. Keep us informed. Thanks!

Leave a Reply

Your email address will not be published. Required fields are marked *