"E.V.E.C." or "Enhanced Voice Expression Control", was developed to expand the expression and emotions of the set singer. E.V.E.C. is an advanced feature that can both aid and hinder a producer, due to its complexity and is more suitable for professional or experienced users.
Currently, E.V.E.C. capabilities are only available for Japanese voicebanks in Cryptons "V4X" range.
E.V.E.C. works similar to the original VOCALOID software and is an analytic based system. Though data is recorded for usage within E.V.E.C. the resulting recorded samples are used as a mathematical equation to adapt the vocal how one pleases.
E.V.E.C. was most likely born out of the desire of Crypton's development team had to create more Append vocals with different degrees of variation of tone such as "Dark-Sweet, Warm-Sweet, Weak-Sweet", much of which suddenly became a possibility with VOCALOID4s new XSY function.
E.V.E.C. adapts the vocal into a new final result entirely which behaviours, in theory, differ to its original sound. This is similar to how XSY adapts the results of two combined vocals to produce a "new vocal" effect entirely with traits of both vocals. The user can change a certain note and make adjustments to said note, effecting how this particular note sounds out ho exactly they want it to. However, E.V.E.C. is not necessarily designed to have a single function used on the entity of a songs results; though is capable of doing so. Instead, it is more suited for use of a multitude of its functions within 1 song. The feature was also designed best to combine itself with XSY specifically. It was the first addition to any Vocaloid added that fully took advantage of VOCALOID 4's new XSY, as well as its new GWL function.
This effect is difficult to create within Vocaloid itself and would require a multitude of voicebanks; this is shown with Megpoid V4. Even then, using Voicebanks may produce a higher quality result then E.V.E.C. produces right now, they are much more limiting in their overall function providing an limited amount of tone variations within a song at best. A multitude of voicebanks used within a song will take up many tracks within Vocaloid, while E.V.E.C. can have a multitude of functions applied to a single track.
When using the function within Piapro, a note is treated as a singular entity. Piapro studio will automatically note E.V.E.C.'s use within VOCALOID4 and the software will cross reference each other automatically when the song is played.
It is possible to write the data within VOCALOID4 itself instead of using Piapro Studio, though this is more complex to do. Inputting data this way skips the need to have Piapro Studio open at all to use E.V.E.C., though it takes some getting use to and understanding to enter the data this way. When using VOCALOID4, E.V.E.C. drops a note in the VSQX file which cross-references with the E.V.E.C. function. As with any other input onto a note within the Vocaloid engine, it is added manually. Users will have to break up the note, as it cannot be applied to a single note within VOCALOID4, unlike with Piapro Studio.
If users wish to break up a note within Piapro studio, they can make it break up the note automatically via the "Separate EVEC note" function. This will make the note more closely mimic the VOCALOID4 method of application. The advantage of breaking up the note as oppose to it being singular, is there is more control over when E.V.E.C. will begin and end.
For more details see the Vocaloid Otaku forums topic on the subject.
Voice colours are the concept of changing a tone of voice or "colouring it" in a particularly way by adjusting vowel sounds to act a certain way, with it adjusting the vocals expression or even tone. The idea of "colouring" a voice was first created as a concept with Hatsune Miku Append and the different voicebanks were meant to display a "expression of colours". This is often thought of mistakenly to be soul function of the E.V.E.C. systems functionality due to the focus of E.V.E.C. being mostly centred on the colours at times.
Not all releases had the same colours;
- Megurine Luka V4X had multiple voice "colors" which coloured two "expressions" VOCALOID voicebanks of "Soft" and "Hard."
- For Kagamine Rin/Len V4X the only available "colors" are "Power" and "Soft" for the "Power" Vocaloid voicebank. This is because the colours role was swapped and instead of colouring tone to expression, expression was coloured to tone. Their "Sweet", "Warm", "Cold" and "Serious" vocals handled the tone control instead and were used in XSY to give "Power" tone changes.
- Likewise, Hatsune Miku V4X also has only "Soft" and "Power" as colours. They are used with "Original", "Soft" and "Solid" to add expression. "Dark" and "Sweet" are this package's tone controllers instead.
Voice colours adapt the concept further by allowing the user to adapt the voicebanks more then they would be able to without it. This means the user can fine-detail how a voice should be according to their desire, allowing for Solid vocals with more power to give force to a high note, or power vocals with a slight softness to them to give precise expression. Some colours are better suited for specific ranges, for example Luka V4x's "Falsetto" colour is better suited for higher notes, while her "Dark" lower.
The colours act as pre-sets and adapt the vocal to make it sound different to how it originally sounded by altering its wavelength. This saves the producer time trying to figure out how to adapt the vocal using only presets within Vocaloid itself and can be done without altering anything within Vocaloid. In theory this allows the producer to combined the impact with features such as simple parameter adjustments or XSY combination, allowing for a easy access to a large array of experimentation within the software.
The second function of E.V.E.C. is voice release. This effects the ending of the voice and allows for softer ending sounds. This allows the vocal to mimic a certain way of speaking to invoke a sweeter or sexy effect. Unlike the colours, this can have less of an overall impact to the vocal as the other features of E.V.E.C. Unlike the other features, every V4X package has this feature.
Currently, there are two modes;
Consonant Extension was assigned to Megurine Luka V4X. This allows a consonant to have more impact over vowels, making it more impacting by giving it more attack or power. Little was explained about the feature and was not as impacting as the feature that replaced it.
The feature was re-designed for Kagamine Rin/Len V4X and Hatsune Miku V4X as "Pronunciation extension". It designed to strength the effect of GWL. It works by stretching out the consonant sound. This weakens (Miku) or strengthens (Rin/Len and Miku) the vocal and can be combined with the other E.V.E.C. functions. The result of doing so is that the vocal imitates the effect of "shouting" (strength) or a much more hoarse voice (soft) with the right XSY vocal and E.V.E.C. colour/pronunciation combinations.
Notably the new feature had a lot of complaints, most of which varied per release (see each individual Vocaloid for more details). Since it acts as a separate entity to the Vocaloid engine itself so has been known to have some effects on certain features like XSY within Vocaloid. It can over-complicate its own mathematical equations at times due to having to work with Vocaloid vocals and bring out faults that are barely noticeable without it, such as sample oddities and phonetic errors.
Particular of note is that XSY between two E.V.E.C. compatible vocals (such as Megurine Luka V4X "SOFT_EVEC" and "HARD_EVEC") is not always recommended and some "colours" are less stable than others. The XSY function confuses the colour functions wavelength adaption calculations at times as well as the XSY function due to it having to adpat the vocal. This is particular with the usage of multiple vocal colour adjustments within a song. E.V.E.C. colours are also often robotic, in the case of Luka V4x, some of her colours were notably problematic with XSY usage then others.
Voice colours do not necessary have to have the entire Vocaloid sample library recorded and may only contain basic phonetic or biphonetic data, rather then the full triphonetic data. For example, Luka's V4x colours only adapt vowel sounds. The result means sometimes results are jerky or not smooth; an issue that Arsloid also displays when trying to XSY "Bright" and "Soft", since the functions only are working with half the needed sounds.
As a result, the current versions of E.V.E.C. have produced many undesirable effects, this results in LQ outcomes when the vocal can at times produce HQ results without its aid. Heavy usage of the function across a large number of tracks can produce lag and be resource consuming on low spec machines, though the design of E.V.E.C. should void the need for multiple tracks within a song. Users should be aware of this.