many folks on this list were interested in computer software that could simulate cross aproximations to the human singing voice.
here is a peice of code that is due to be released shortly and it's a project im keeping an eye on myself. i can imagine what mr. cylob's tracks would sound like after he has his way with this. muwahaha
http://www.yamaha.co.jp/news/2003/03022602.html
for non jappy ppl and those who enjoy babble fish's inane babble
Tip of inquiry General one One related to reporting
---------------------------------
On personal computer real singing synthesis
* The lyric and diacritic just are inputted, the music due to the singing of the person compilation
Singing synthetic software
' #12532;#12457;#12540;#12459;#12525;#12452;#12489; cVocaloid ' development
-3 in month " #12512;#12472;#12540;#12463;#12513;#12483;#12475; " (Germany), " AES " (orchid) and the like with announcement -
' The VOCALOID ' one example of the operation picture (being something which is in the midst of developing, it does)
February 26 of 2003 day
---------------------------------
Yamaha corporation (head office: Shizuoka prefecture Hamamatsu city Nakazawa Cho 10-1, president: Itoh Osamu two), recently the lyric and diacritic just are inputted, real the software ' VOCALOID which synthesizes the singing of the person (the #12532;#12457;#12540;#12459;#12525;#12452;#12489;) ' to develop, on March 5th - 9th in Germany & Frankfurt it is held in Holland & Amsterdam it is held with " the 114th AES convention " it displays & announces to " the #12512;#12472;#12540;#12463;#12513;#12483;#12475; " and March 22nd - 25th.
---------------------------------
Summary > of < new technology
Presently, as for the music production by the computer and the like, the music production whose quality is high with the completeness of the MIDI peripheral device which begins the XG format corresponding sound source, is widely done, also many of the musician of the professional, have produced the CD with the music production environment by such computer.
But, in regard to " the singing " of the person, it produces freely making use of the computer, there was a technical limit, accompanied extraordinary difficulty. The software and the system which synthesize singing existed so far, but to bring close to the singing of the genuine person, finely it is necessary to tune, it was most actualization impossible to synthesize the kind of singing where human voice and distinction are not attached.
On one hand, recording and processing various sounds making use of the sample which recorded vocal sound with the spread of the sound source sampler which it plays back, on music production site, also the technique which produces the music of the singing entering is adopted, but there was a weak point such that the lyric and expression are limited.
*
If this each time this corporation developed ' the VOCALOID ', inputs the melody and the lyric of the part of singing on the computer, it is the singing synthetic software which that way can produce the vocal part of music. It is the data base which was recorded from singing of the actually person, in order to synthesize making use of the " singing library ", character of original singing remains, real can obtain the synthetic sound of singing. In addition, the expression which is necessary for singing with simple operation you attach and (vibrato and pitch bending) also it is possible to do. Because of this ' the VOCALOID ' with, production of the music which used various voice qualities by being complete this singing library becomes possible. The language which it can synthesize is Japanese and English. Concerning telling others language consecutively correspondence is planned.
Furthermore ' the VOCALOID ', it operates with the personal computer which loads " the Windows " OS. In addition, planning correspondence concerning the MaccOs, you are
*
At this corporation, of the business development which centers the production of the singing library and the license provision regarding the sale of the software this ' VOCALOID ' concerning, is thought. Already, the krypton * future * media corporation which is the major company of sampling CD development and circulation (the address: The Sapporo city Chuo Ku, president: Hiroyuki Itoh), MUSIC-SHOP Best Service Kandler und Greil GmbH (address: German & Munich city, director: Ludwig Greil, Klaus Kandler), Zero-G Limited (address: The English & Cornwall state #12469;#12523;#12479;#12483;#12471;#12517;, director: The Edward Stratton) with it has reached to agreement in between concerning the framework of license provision. Furthermore, actual commercialization has planned next year beginning from during this year.
Details are as follows.
---------------------------------
< Main merit >
1. Just inputs the melody and the lyric, real synthesis of singing is possible
The melody and the lyric just are inputted, can synthesize the singing which is sung in like the on the screen of the personal computer. Furthermore the singing which is outputted becomes the Wave file, the DTM (the desktop * music) uses the software and and the like can also acquire accompaniment.
2. By the fact that the library is added synthesis of various singings is possible
Various vocal sounds (it is possible) with such as man and woman the singing library (the data base) by changing, to synthesize singing. As for these libraries, the person who really has various voice qualities can increase, by the fact that the phrase and the like of basic word is recorded.
3. With simple operation the expression attaching of singing is possible
By the fact that it operates simply with the software, the expression which is necessary for singing (vibrato, you sing and turn, the fist and the like) can appoint. Because of this, it is possible to do expression rich music production easily.
4. In addition to the Japanese, singing of English synthetic possibility
It removed from the singing of the actually person, sound the library is drawn up from the being connected of sound, but this sound the being connected of sound (the Singing Articulation), not only Japanese, you prepare also English. Because of this, it is possible to synthesize the singing of English in addition to the Japanese.
5. With personal computer for national welfare available
' The VOCALOID ', OS it operates with " Windows 2000 " and " the same XP " top (the CPU: Pentium III, above 1gHz, RAM: Above 512MB recommendation). In addition, correspondence is planned concerning the MaccOs.
---------------------------------
The < ' VOCALOID ' concerning the singing synthetic system >
' The VOCALOID ', research and development of signal processing in the frequency domain was adopted synthetic method " frequency domain song articulation connected method of the singing which this corporation which long time goes back and forth developed individually (the Frequency-domain Singing Articulation Splicing and Shaping)".
With this synthetic method, the song articulation which beforehand is necessary for song from singing of the actual human (fragment and vibrato et cetera of sound such as Singing Articulation = syllable, it decided beforehand, the collection " of expression of voice") to gather, after doing the conversion processing to the frequency domain, it converts to a data base.
And it adjusts to the lyric which is inputted, it removes the data which is formed from fragment of singing, it does connection processing pitch (the musical interval) as a conversion and a lyric.
At that time, in order to process in the frequency domain, adjusting to the melody which is appointed, when it becomes easy to modify pitch, becomes simultaneously, smooth connection as a lyric of fragment of singing possible.
For example, " it bloomed, (the #12469;#12452;#12479;)" the part " of the rhinoceros " is obtained by the fact that two fragments, " #12469; " and " the eye " are connected. In the timbre of the vowel " #12450; " which is included " the #12469; " and in the timbre which is included " the eye " usually because it differs, when you connect that way, is audible this time as an offensive noise. Then, in order to solve this, synthesis (smoothing processing) doing the timbre of the connected part in the frequency domain, it can obtain smooth singing by connecting.
Figure 1) In the frequency domain, processing
In addition, also pitch when expression such as vibrato is given by processing in the frequency domain, and control of the timbre can do simply. ' The VOCALOID ' with, time change of timbre such as pitch and the vibrato which are removed from the singing of the actual human is remembered in the data base, when synthesizing by the fact that it applies that, it becomes possible to reproduce the time change of pitch and the timbre which are included in expression such as originally vibrato.
Figure 2) Processing at time of vibrato
' The VOCALOID ', musical interval * lyric * expression is constituted is inputted the library every of engine, and the singing which synthesize the " score * editor " and singing (the pronunciation data base + music expression data base) from.
As for the library of singing, the actual human by the fact that you record alongside the templet of timbre change of basic vocabulary and vibrato et cetera, it is possible to produce anew.
Figure 3) ' VOCALOID ' system configuration
---------------------------------
Was drawn up with the < VOCALOID the singing sample > which
Kimi No uwasa / solo vocal of man SoundcVq file: 909kb Morceau of smooth snowscape / chorus
SoundcVq file: 133kb Amazing Grace / synthetic sample of English
SoundcVq file: 395kb
* ' The SoundcVq ', applying high-level speech compression technology, on Internet, clear sound and force it is the software which makes the transmission of the music which overflows possible.
As for plug in of the SoundcVq player, it is possible from the sight below to download.
Bunchu trade name and the company name et cetera are trade mark or registered trade mark of this corporation and each company.
---------------------------------
Tip of inquiry General one
Related to reporting
---------------------------------
Do you Yahoo!?
Yahoo! Tax Center - forms, calculators, tips, and more