Jasmine Music Technology - Home
productspurchasesupportdownloadcompany
yv enhancer demos
some thoughts before testing vocaloid,
12.10.2003
yv enhancer questionnaire
basic scientific and methodological provisions
jmt tips for using yamaha vocaloid
yv enhancer 1.0
related links
 

Performance Modeling technology can enhance
Yamaha Vocaloid!

 
  Song Name Singer Download MP3 More details
21 Make Me Feel (Bossa) Sweet Ann, Big-Al [1,35 MB] here
20 Make Me Feel (Club) Sweet Ann, Big-Al [1,35 MB] here
19 Alove Sweet Ann [0,89 MB] here
18 I Feel Good Big-Al [0,41 MB] here
17 PowerFX Big-Al [1,00 MB] here
16 Album (7 songs) LEON -- here
15 Autumn in Central Park LEON [3,59MB] here
14 Three Singers Together LEON, LOLA, MIRIAM [2,54MB] here
13 Short Romance MIRIAM, LEON [3,09MB] here
12 East LEON [2,36MB] here
11 West MIRIAM [1,39MB] here
10 Red Hair Girl LEON [3,65MB] here
9 Yest Tolko Mig LEON [2,87MB] here
8 Flying Girl LEON [2,3MB] here
7 The Phantom Of The Opera MIRIAM, LEON [1,87MB] here
6 Please Touch Me Lola LEON [1MB] here
5 Summer-time LEON [1,29MB] here
4 Rising Star LEON [2,4 MB] here
3 You ask me LEON [927 KB] here
2 Internet Angel LEON [1,06 MB] here
1 Sad Mondays LEON [2,49 Mb] here

It is a big mistake, thinking that synthesizing vocals or something equally expressive, such as solos for violin, sax, flute or guitar is possible without introducing performance nuances into the electronic score. Great synthesis technology and perfect samples are just the material which requires control.

How can that be done? The usual way is editing controllers in the Piano-Roll, which is very tiresome; then again, one should know very well exactly how to edit these controllers. We can record parts on a MIDI-keyboard and then control them with Wheels and Joysticks, but the capabilities of these controllers are limited, and mastered not by everyone.

There is another way - Performance Modeling technology, when the program calculates performance nuances analyzing the electronic score. Of course, even the intelligent algorithm, a "robot", is very distant from real live musician, but it allows controlling the performance. The user, like an orchestra conductor, manages the manner of performance, controlling the use of various "Performance nuances", at the same time being free from the immediate editing of every controller.

We have accumulated certain experience with Performance Modeling, having released our first product in 1996 by issuing a line of Style Enhancer products and developing as many as 500 intelligent PMStyles for the variety of instruments and in various manners. Now we are improving our algorithms in order to provide the users of Vocaloid and other vocal synths with specialized software for humanizing the results and saving the time in data editing.

Below you can see our first examples in controlling Vocaloid (Zero-G LEON library). Keep in mind the following comments to have the objective picture.

  1. This is just our "trial balloon", our first attempt at Vocaloid testing, before YV Enhancer and special PMStyles are developed.
  2. Each example is just one of the possible variants of adding the "Performance nuances". In YV Enhancer you will be able to select not only the manner, but also a set of "Performance nuances" and the depth of their influence.
  3. Most of the "Performance nuances" are provided with the exaggerated depth, to make it more noticeable.
  4. All the "Performance nuances" were calculated in a fraction of a second. (To completely appreciate the job done by the PMStyle, Vocaloid users can analyze the controllers and "expressiveness elements" introduced by a style in Vocaloid *.mid file.)
  5. For each example Vocaloid synthesis parameters are given, which were calculated by PMStyle. As you can see, the examples show the sequence of introducing the parameters in the process of our mastering Vocaloid control.

Updates History