Now that the most complicated task of vocal synthesis is successfully
solved thanks to researches of
Music Technology Group and
addition of expressive nuances becomes finishing strokes for receiving
effective means of voice parts creation.
At Jasmine Music we have Performance Modeling technology (PM), which
is intended for humanizing MIDI data. PM algorithms can recognize
musical data, modify it, and also insert detailed Wheel, Modulation,
Expression and other control data. A number of algorithms serve for
imitating "singing voices" like Human Voice, Flute, Violin and others.
The presented MP3 files have been generated using the VL section of
the Yamaha MU100R tone generator. They illustrate possibilities of PM
technology to achieve bright musical expressiveness and "singing"
character even without real human vocal samples.
PM MP3 Examples
PM MIDI Examples for Vocaloid
All MIDI files in ZIP
Processed MIDI files can be used to drive a new voice generator such as
the Yamaha Vocaloid. They include detailed renderings of Wheel,
Modulation, Expression, Velocity and Duration data. You can compare PM
MIDI files with the original MIDI file. All MIDI events were generated
using the PM technology, and no event was edited manually.
We are very excited with the up coming Zero-G announcement of the Yamaha
Vocaloid at the NAMM Show (January 15-18, 2004, Anaheim, California). And
when it becomes available, we plan to announce additional specially
designed PM algorithms that you will be able to hear the full potential of
the Yamaha Vocaloid synthesis.
Please feel free to contact us via emails:
Dr. Ravi E. Dattatreya
Chairman, Jasmine Music Technology
E-mail: ravi @ jasminemusic.com
CTO, Jasmine Music Technology
AlexeiUstinov @ jasminemusic.com
Now Performance Modeling technology is an important part of JMT
products: Onyx Arranger 2.1, Style Enhancer 3.0 (and up coming SE4.0),
Stylizer 1.0, SEM2.0 MFX Plugins for Cakewalk/Sonar. The PM program module
has flexible, real-time controls.