Have you ever used a speech synthesizer in your music? I’ve used a 1970′s Texas Instruments Speak & Spell, Apple computer’s built in MacInTalk text to speech synthesizer, Magnavox OdysseyÂ² video game peripheral “The Voice”, AT&T’s “Natural Voices” for telephony, Commodore Amiga’s “Soft Voice” synthesizer and a handful of other talk making algorithms.
You can hear synthetic voices in a lot of popular music. Most recently Benni Benassi’s song Satisfaction highlights Apple’s MacInTalk speech synthesizer.
In my own music I sometimes have the synthetic voice mirror the last few words in a verse acting as a robotic backup singer. In my live version of the song “One Night in NYC” I have a synthetic female voice tell her side of the dark story. I recorded a futuristic minimal track with German producer Miro Pajic titled “Gigabytes Numbers” and the last minute of the song a male synthetic voice with a British accent rambles gibberish and well large numbers. To increase the futuristic effect on his voice we put it through a TC Electroinc’s Filtrator plug-in and then automated some delay effect times using Ableton Live’s standard Ping Pong Delay plug-in. Here’s an audio sample:
Here are a few online places you can go right now and create some synthetic voices:
- AT&T Research Labs
- Speak & Spell Online
- Oddcast TTS
- Vox-machina.net (Uses Apple’s MacInTalk technology)
What’s next? Software that creates real singing vocals of course. Yamaha’s Vocaloid software takes a stab at it but the technology really is not ready for prime time. However, I could see using Vocaloid for interesting sounds. You can jump over to Sound on Sound Magazine to hear a sample.
Keep in mind we are not “talking” about vocoding in this article. Vocoding uses a carrier signal and a real human voice and will be a subject of a different post in the future. For a current stream of updated info on Text to Speech check out the Text to Speech blog!