Ever since Max Mathews taught a computer to sing "Daisy Bell" in 1961—inspiring Stanley Kubrick to recreate the performance for HAL's death scene in 2001: A Space Odyssey—humans have been intrigued by computer vocals.

In this episode of Digital Media Insider, we begin with our article "Sneaky Tricks for Speech Synthesizers" and then explore some of the amazing new software designed (or corrupted) to make your computer sing. (DMI 10-16-2006: 13 minutes 49 seconds)

Production Notes

The audio examples in this episode came from my article "Sneaky Tricks for Speech Synthesizers," Delay Lama, and some demo files I found on the web. The web files were in a variety of formats, including .mp3, .asx, and .mp2. Because my audio editor, BIAS Peak, can't open the latter two, I played them in VideoLAN VLC and streamed the output into Peak with Cycling '74 Soundflower. I then shortened the examples with cuts and fades, and exported them as AIFF files.

Next I imported the AIFFs into Ableton Live, where I arranged them around my voiceover and background music. Finally, I rendered the mix to an AIFF file and converted it to an MP3 in Peak.

I recorded the voiceover directly into Peak, this time using the new Rode Podcaster mic instead of my usual AKG D3900. The Rode's USB jack lets you plug it directly into a computer, so I didn't have to haul out the preamp and extra cabling I needed last time. After snipping out some plosives and false starts, I compressed and enhanced the voiceover with Izotope Ozone.

Next, I created regions in Peak where I wanted the voiceover segments to start, and exported those regions as individual AIFF files.

From VLC to Peak via Soundflower transfer

Because my audio editor, BIAS Peak, can't import .MP2 files, I hijacked the VLC media player's output with Soundflower and routed it into Peak. For programs without an Audio Device menu, you can select Soundflower in System Preferences.

The background music came together in Live as well. I made the opening sound effect by splicing a compressed mouth noise onto a tone cluster I generated in Native Instruments Reaktor. The main groove is from Steinberg Xphraze. (Jim Aikin turned me on to both virtual instruments in his article "My Five Favorite Soft Synths.") The piano is from the Garritan Personal Orchestra, which I discovered when we interviewed Gary Garritan. Then there are a few percussion samples dredged from my hard drive. Altogether, the music mix took just six tracks. Effects processing was courtesy of Live's default plugins and Freeverb.

Links

O'Reilly Digital Media Articles

Sneaky Tricks for Speech Synthesizers

Surprisingly, synthetic speech can add a captivating human aspect to your music. Here are several creative (and free) ways to use speech synthesizers.

Bring Your MIDI Music to Life

This introduction to MIDI continuous controllers explains how they work and how to use them to make more expressive music.

Look Ma—Hands! Choosing and Using MIDI Controllers

If you really want to play today's wonderful software instruments, drop that mouse and grab a dedicated MIDI controller. In this MP3-enhanced tutorial, you'll hear the dramatic difference controllers make in musical expressivity, then get buying and usage tips.

Software and Sites

  • AT&T Natural Voices Demo A high-quality online speech synthesizer.

  • AudioNerdz Delay Lama A VST singing plugin that emulates a Tibetan monk.

  • Dictionaraoke Fuses speech samples from online dictionaries with karaoke backing tracks.

  • KAE Labs VocalWriter A combination synthesizer and MIDI sequencer with singing instruments as well as voices. Mac only.

  • Monolake Home page for performer and Ableton Live designer Robert Henke.

  • SoftVoice Developer of formant-based text-to-speech systems used in everything from the Mac to FL Studio.

  • Spell Catcher A terrific spelling checker with abysmal error sounds.

  • Yamaha Vocaloid An advanced singing synth for Windows. Also see the MP3 demos and free trial version at Zero-G.

David Battino is the audio editor for O’Reilly’s Digital Media site, the co-author of The Art of Digital Music, and on the steering committee for the Interactive Audio Special Interest Group (IASIG). He also writes, publishes, and performs Japanese kamishibai storycards.


Return to digitalmedia.oreilly.com