Vocaloid

The Human Voice

It's still the most expressive and emotionally communicative musical instrument available to mankind. UNLESS, of course, you cannot sing. Or you don't have access to a quality singer when you need one, to bring reality to your ideas or your lyrics. Either you can sing or you can't. Either you have a great singer who realizes your ideas or you don't. That is the way of things. For most creative music producers, this means that in one sense the human voice has also always been a sonic boundary - a musical frontier, through thousands of years of evolution.

Until Now!

A TRUE Revolution

When it comes to creating quality vocal tracks, Yamaha's VOCALOID singing synthesis technology literally changes everything. The future will now be very different. The last bastion of human musical expression - the singing voice - has been realistically harnessed in synthesis. Now it's time for YOU to express your music in YOUR own words. You will be able to sing your heart out - through LOLA, LEON, MIRIAM, and other VOCALOID virtual vocalists from Zero-G, thanks to the incredible new technology developed by Yamaha.

LEON and LOLA are virtual male & female soul vocalists modelled on real professional singers - two of UK's most in-demand (and expensive!) session singers. MIRIAM is modelled on the stunning voice of Miriam Stockley, the voice of the multi-million selling ADEIEMUS albums. When installed into your PC they will literally allow you to create singing of superb quality and realism. They will sing ANY words you ask them to in English - literally anything - be they beautiful lyrics or comical trivialities, Monteverdi madrigals or maniac chants. You can create vocal tracks of soulful singing in any lyrics you want. You just type in your lyrics, and synthersize. Then add expression to taste. These singers are under your total control, and the really mind-blowing thing is - they can truly sound like professional singing voices. With very little practice the results you get will completely fool your friends - they will believe they are listening to a real singer performing.

Your own backing singers, 24/7 and NO SESSION FEES.

With Vocaloid, creating your own professional sounding backing vocals is quick and easy - even for complex harmonies and arrangements. And as you develop your Vocaloid editing skills, you will find that even lead vocal lines become a reality, thanks to the arsenal of expression tools provided, which permit manipulation of the finest nuances.

Now you can communicate emotion so directly. You can add HUMANITY to ANY production in ANY style without needed a human. You can CONTROL and PERFORM your OWN private virtual professional vocalists. You can even scat. You can create stunning realism. OR... You can live in the exciting world between human and machine, where these singers can even easily perform like superhuman virtuoso androids - as true hybrids of human voice and synthesizer. The potential for innovation here is limited only by your own imagination.

WHAT WILL YOU MAKE THEM SING?

Do you simply want them to perform your lyrical ideas as an aid to your songwriting process? Or will you use their freedom from human limitations to create stunningly innovative and experimental vocal effects? They can sing over a superb pitch range, they can sing much faster than a human could if you want them to, as fast as you like in fact, and they can jump any interval yet hit every note beautifully, IF that's what YOU want them to do. And they'll NEVER tire or complain...

In addition to singing any words or combination of syllables you can imagine, LEON, LOLA & MIRIAM will spread any sustained vowel (or voiced consonant) across as many notes as you like, with perfect legato. You can select from several different natural vibrato types and drag and drop your chosen type to any note or notes and control the time-position and amount of the vibrato for superb expressiveness and realism.

And that's just the beginning! - You have control over a myriad of sound-shaping parameters, which you can bring to bear on any part of your creation. And of course, you can even change the singer (if you have purchased and installed more than one virtual singer) - perhaps get LEON or MIRIAM to sing your creation instead of LOLA - instantly!

IMAGINE WHAT THIS ALL MEANS FOR YOUR MUSIC

The potential for Yamaha's Singing Synthesis Technology is virtually unlimited. LEON & LOLA are released at the NAMM show, January 2004. They are the genesis, the "Adam and Eve" of the new era of virtual singing. A third virtual vocalist, MIRIAM will take Vocaloid to new heights. Pretty soon you are going to hear synthesized vocals in a lot of music. Don't get left out of the picture. Experience the difference Zero-G Virtual Vocalists can make for yourself!

HOW DOES IT WORK?

Few events have dramatically deiverted the course of musical history. Synthesis, MIDI, Sampling - inventions of the magnitude are very rare events indeed. Singing Synthesis is such an event. Most acoustic instruments can be simulated well using various synthesis techniques, but (until Yamaha's Vocaloid technology came along) the singing voice resisted any serious simulation attempts. Hardly surprising, since singing has an extremely wide range of articulations, timbres, and transitions between sounds. And singing can communicate words as well as melody, which means there is a double layer of meaning, unlike other instruments. The human ear is so used to hearing the voice that even the finest tonal shifts or anomalies are immediately noticed. However, Yamaha's Vocaloid is in fact a totally new vocal-synthesis technology, achieveing a much higher level of sophistication in this exciting area! The team at the Yamaha Advanced System Development Center in Japan has created software that emulates the singing voice with incredible accuracy.

Zero-G's development team have been working closely with Yamaha to create many 'vocal fonts' (virtual vocalists) for the Vocaloid system. The team starts each project by recording a professional vocalist singing literally all possible phonemes and transitions between syllables. Each transition is slightly different depending on the particular combination of phonemes, and these differences play a big part in how we understand words and whether a vocal track sounds natural or artificial. For example, the phoneme "p" sounds slightly different at the beginning of a word than it does at the end, and it affects the vowels next to it differently than say the phoneme "t".

The recordings of the professional singers are converted to the frequency domain using Fast Fourier Transform and divided into thousands of separate phonetic transitions which are processed in a unique way and then stored for use with the Vocaloid synthesis engine. Expressive tools such as vibrato, pitch bend, and attack are also derived from the real singing and stored separately.

To create a vocal track, you enter music and lyrics into the Vocaloid Editor. The melody can be entered by hand in the piano-roll style Editor or imported from a Standard MIDI File; the words are entered manually. Expressive elements can be entered via a graphic palette using drag-and-drop. Further detailed programming of expression parameters can be done graphically, for finey detailed results.

The data that you have entered in the Editor is sent to the synthesis engine, which fetches what it needs from the phonetic and expression databases to synthesize the track. To sing the word "part", for example, the software combines four elements from the phonetic database: "p" (as it sounds at the beginning of a word), "p-ar" (the transition from "p" to "ar"), "ar-t" (the transition from "ar" to "t"), and "t" (as it sounds at the end of a word). The two "ar" elements are blended together, and the resulting vowel "a" is lengthened as necessary to accommodate your chosen melody and rhythm.

Different pitches are derived by shifting the fundamental and overtones while leaving the vowel formants relatively untouched. The database elements were derived from phrases originally sung at different pitches, limiting the amount of shifting the engine needs to do (and therefore improving ultimate realism).