Talk:Macne Nana V4/@comment-144.138.30.68-20181004225549/@comment-53539-20181005090128

ITs not just the voice, the tempo range jumped as well.

You essentially have two types of layering; stationary and ascending. One controls how the voice sounds overall and contributes to Vocaloids like Miku sounding like Miku, the other controls how it acts when you go up and down the octaves. You don't need to know 100% what samples contribute to which but this is a lot of whats going on.

''Note; As you add sounds you get half layers caused by layers of pronunciations naturally allowing this. Things such as expression is often captured in certain sounds such as those involving blending or stressing contributing to both layer types. Larger VOCALOID voicebank languages can have more tendacy for "0.5" layers and is likely where Yohioloid gets his. In addition as the engine improves it handles sounds better so you get more "give" in a Vocaloid or more realism so range can improve as Vocaloid does. You also have quality and so fourth, all which contribute to things and allow also more range. We'll ignore this for now because these are all side-effect stuff.''

Most Vocaloids in the past have had just 2 layers of stationary sound, including Miku herself back in V2 and all 6 of her Appends, which is the bigger contributor to things then ascending layers. This is true at least for Japanese Vocaloids, and very much true well up into the late V3 era.

Layers of sound contribute to how a Vocaloid sounds in terms of realism, so the more layers you have in both directions the more, in theory, realistic the vocal is. This is not 100% always true, bare in mind Vocaloids like Miku attempt not to be realistic so the traits recorded are deliberately off. Essentially, adding new layers in theory creates a stronger vocal and allows Voclaoid to have more "give" in the voice. Large ranged Vocaloids often turn out to have more then 2 layers of both types of layers. Its likely Nana had an extra layer added which allowed her to get away with a larger range, to simply put. If you look at a new Vocaloid and see it has too many traits of V2 era ranged vocals, in due respect its likely 2 layered though this is not 100% certain.

There are also vocaloids with 1 layer... I'm not going to point out these voicebanks, because people don't like hearing it... But you can spot them if you look. Short answer? V1 vocals, but I can say this because there isn't much resistance to that idea as V1 was bad. This is why when people say V3 Kaito has a smaller range... No... He didn't. V3 has the larger range.

One thing to note about Nana is that she is based on her Reason and Garage Band vocal in Vocaloid, not the UTAU voice. The original vocal had a vague sound to it with no attempt to sound human. So Nana's ability to sound like herself isn't so solid as many other Vocaloids as she had a lot of vagueness about her from her original vocals. This is likely why she gets away with so much, it doesn't take much for Nana to sound like herself because herself was original a wide scope of sounds. Its something a lot of the UTAU fans didn't realise about Nana because they only heard her in UTAU. So with the two factors in mind, the adding of an extra layer theory and the fact Nana has a wide scope of "give" to how she sounds both stack up with each other.

I will note, there are draw backs which is why not every Vocaloid adds layers despite their advantage like with Nana.
 * Essentially, the most notable problem is that you can impact how the voice sounds and this can be a negative factor to producers. So studios can cheat by giving more voicebanks for older voices rather then extending them, which is a habit seen in regards to older V2 era vocals.  Your adding core sounds especially with stationary adding. Even between V3 and v4 Nana you can hear this impact with her slightly being higher pitched now.
 * When talking about Alter/ego and Chipspeech Plogue noted adding more layers gives you more realism, but the issue is that you loose flexibility, which is where "V1 Kaito has more range" comes from. V1 Kaito had lots of "give". You can get back by adding more layers... But...
 * Get a layer off pitch and you get the problems of Lily V2 wherein the voice has issue pitching naturally, as it messes with the maths Vocaloid itself handles and ends up with you having a Vocaloid who doesn't play well.
 * Adding more layers also adds more time in development and more costs, especially for professional singer studio time as they get paid by the hour often. This bites into profit and possibly also into price at times. This is why despite more Vocaloids having 3 layers appearing, you also have issues of there being tendency to stick to the traditional 2.

I wish people gave Nana more a try over some of the other Vocaloids because a lot of thought went into her 3 vocals and effort. Even if all this layer stuff doesn't bring home the point, she still has some of the highest quality samples among Vocaloids.