Piapro Wiki
Advertisement

"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 Crypton's "V4X" range and currently only the versions of Piapro Studio that use the VOCALOID API.

Purpose[]

E.V.E.C. works similar to the original VOCALOID software and is an analytic-based system which references a simple sample database. 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. At their base core, each E.V.E.C. function is essentially a series of presets for the voicebanks they are made for, allowing a user to easily adapt a voice without having to use trial and error.

E.V.E.C. adapts the vocal into a new final result entirely which behaves, in theory, different 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. Users can change a certain note and make adjustments to said note, effecting how this particular note sounds out how 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 song's results, though is capable of doing so. Instead, it is more suited for use of a multitude of its functions within one 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 VOCALOID4'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 was shown within VOCALOID itself with Megpoid V4. Even then, using Voicebanks may produce a higher quality result than E.V.E.C. produces right now; they are much more limiting in their overall function providing a 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.

E.V.E.C. was most likely born out of the desire of Crypton's development team to create more Append vocals with different degrees of variation of tone such as "Dark-Sweet, Warm-Sweet, Weak-Sweet"[1], much of which suddenly became a possibility with VOCALOID4's new XSY function.

Note that due to VOCALOID5 having dropped the XSY function, the above no longer applies to E.V.E.C. as the function is no longer available for E.V.E.C. to make use of. As of VOCALOID5 Yamaha has introduced its own methods similar to the EVEC function. Users should be aware the "Type" and "Colour" methods used by VOCALOID5 do not produce the same results as the ones displayed by EVEC even when the functions are set up identically because of how they work. This does not make E.V.E.C. completely useless, as it can still make use of "Type" and "Colour" within VOCALOID5 as it once did with XSY. However, certain combinations can result in very peculiar results, especially certain colours among Luka's V4x options and the new feature does not work the same way as XSY did.

Usage[]

When using the function within Piapro Studio, 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. The function is accessed through the "Note Setting" window into its respective tab. There the user can select each of the E.V.E.C's individual function: Voice Color, Voice Release and Consonant/Pronunciation Extension.

It's 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 used to and understanding how 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. Users will have to break up the note, as it cannot be applied to a single note within VOCALOID4, unlike with Piapro Studio, and the respective inputs have to be added onto each note as if they were phonetic data.

If users wish to break up a note within Piapro Studio, they can make it break up the note automatically via the "Separate E.V.E.C. Note" function. This will make the note more closely mimic the VOCALOID4 method of application. The advantage of breaking up the note as opposed 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 (archive) on the subject.

Individual functions[]

Voice Color[]

Voice Color allows control over the overall tone and expression of the voice. This function is based on the concept of changing a tone of voice or "coloring it" that inspired the creation of Hatsune Miku Append and the different succeeding voicebank packs that were meant to display an "expression of colors". This is often thought of mistakenly as the soul function of the E.V.E.C. system's functionality due to how the focus of the Appends were to allow a completely different tone/expression while E.V.E.C. is mostly centered on adjusting the voice color at particular times.

Not all releases had the same Voice Colors available;

  • Megurine Luka V4X has multiple voice "colors" which color two "expressions" VOCALOID voicebanks: "Soft" and "Hard."
  • For Kagamine Rin & Len V4X, the only available colors are Power and Soft for their "Power" VOCALOID voicebanks. This is because the Voice Colors role was swapped; instead of coloring tone to expression, expression was colored to tone. Their "Sweet", "Warm", "Cold" and "Serious" voicebanks handled the tone control instead and were used in conjunction with Cross-Synthesis (XSY) to give "Power" tone changes.
  • Likewise, Hatsune Miku V4X also has only Soft and Power as Voice Colors. They are used with the "Original", "Soft" and "Solid" voicebanks to add expression. "Dark" and "Sweet" are this package's tone controllers instead.

Voice Colors adapt the concept further by allowing the user to adapt the voicebanks more than they would be able to without it. This means users 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 E.V.E.C's colors are better suited for specific ranges, for example Luka V4X's Falsetto Voice Color is better suited for higher notes, while her Dark lower.

Within the VOCALOID editor itself, the closest vocal to come close to the effect of "coloring" is ARSLOID with his extension libraries "Bright" and "Soft".

The colors act as presets 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 combine the impact with features such as simple parameter adjustments or XSY combination, allowing for an easy access to a large array of experimentation within the software.

For inputting this option into the VOCALOID4 editor, it's required to put the colored vowel after the same non-colored vowel. This is due to how the colored vowels can't be synthesized on their own, and simply act as an extension of the original vowel.

  • Example: A "Husky" colored 'ru'  should be input as [4 M][M#3] or [4 M M#3] into the VOCALOID4 Editor.

To specify, the vowel's color simply adds the respective symbol as a suffix, similar to the method used for the devoiced sonorants.

Voice Color Symbol Applies to
Whisper [※#1] Megurine Luka V4X
Soft [※#2] Megurine Luka V4X
Kagamine Rin & Len V4X
Hatsune Miku V4X
Husky [※#3] Megurine Luka V4X
Native [※#4] Megurine Luka V4X
Power1 [※#5] Megurine Luka V4X

Power2 (Luka)

Power (Rin/Len, Miku)

[※#6] Megurine Luka V4X
Kagamine Rin & Len V4X
Hatsune Miku V4X
Cute [※#+] Megurine Luka V4X
Dark [※#-] Megurine Luka V4X
Falsetto [※#F] Megurine Luka V4X

Note: Replace the ※ with the corresponding vowel.

Voice Release[]

The second function of E.V.E.C. is Voice Release. This option controls the breath amount at the end of a word. This allows the producer to produce a softer ending sound, and it can be used for mimicking a certain way of speaking that invokes 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 types of Voice Release available:

  • Breath-Short
  • Breath-Long

For inputting this option manually into the VOCALOID4 Editor, simply write the symbol corresponding to the selected Voice Release into the phonetic input.

Voice Color Symbol Applies to
Breath-Short [*#1] All the V4X E.V.E.C. voicebanks
Breath-Long [*#2] All the V4X E.V.E.C. voicebanks


Consonant/Pronunciation Extension[]

The 'Consonant Extension' and 'Pronunciation Extension' are a series of features that allow the user to affect the performance, duration and impact of the Consonant at the beginning of the notes/syllables. This function has been revised in each release and behaves differently for each of the V4X voicebanks available.

Consonant Extension was assigned to Megurine Luka V4X. This option is designed to emphasize the consonant at the beginning of the note to have more impact over vowels. Basically, it works by layering several sound samples of the consonant, effectively giving the impression of a harsher and longer sound.[2]

The feature was re-designed for Kagamine Rin & Len V4X and Hatsune Miku V4X as the Pronunciation Extension function, and it's designed to take the Growl (GWL) parameter and other options into account.

For Rin/Len, this feature is designed only for strengthening the voice and can be used for producing or imitating a "screaming" or "shouting" effect when combined with the other features available (Voice Color, GWL, XSY).[3][4] For Miku, the Pronunciation Extension has two modes available: Strong Pronunciation, similar to Rin/Len's feature; and Soft Pronunciation, which produces a breathy voice more suited for gentle and quiet expressions and can produce a hoarse voice effect when coupled with the other features and parameters.[5][6]

For inputting this option manually into VOCALOID4, the user simply requires to repeat the target consonant at the beginning of the note 2 - 4 times.

It's important to mention that abusing this function can produce a stutter- or trill-like effect in some cases. This E.V.E.C function also can be combined with the Velocity (VEL) parameter for better control of the consonant's length, impact and stress.

Issues[]

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, it 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, especially if both features are used in tandem and some "colors" are less stable than others. The XSY function confuses the Voice Color's function's wavelength adaption calculations at times as well as the XSY function due to it having to adapt the vocal (basically, as both features perform their own calculations and sample processing independently, they may clash with each other). This becomes particularly noticeable with the usage of multiple vocal color adjustments within a song. E.V.E.C. colors are also often robotic; in the case of Luka V4X, some of her colors were notably more problematic with XSY usage than others.

This issue partially stems from the fact that Voice Colors do not necessarily need 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 colors 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. Another fact to consider is the range of the Voice Color, as working outside its comfort area increases the chances of odd results (Example: Luka V4X's Falsetto color is designed for high notes and doesn't handle correctly the low range).

As a result, the current versions of E.V.E.C. can produce many undesirable effects that it may lead to 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. Although the design of E.V.E.C. should void the need for multiple tracks within a song, users should still be aware of this on larger-scaled projects.

E.V.E.C. and XSY can also only be used within the VOCALOID4 or later versions of the engine. Therefore, users who do not have the full version do not have access to XSY at all and are limited to use of E.V.E.C. only.

References[]

  1. link
  2. Sonicwire - Megurine Luka V4X
  3. link
  4. Sonicwire - Kagamine Rin & Len V4X
  5. link
  6. Sonicwire - Hatsune Miku V4X

Navigation[]

Template:Tutorial directory Template:Crypton Future Media, Inc. navbox

Advertisement