speaking and singing utau difference?

AmberTonnerre

Ruko's Ruffians
Defender of Defoko
What is the vocal difference bettween an utau recording from a "singing voice" vs a "speaking voice"? For example, a lot of utau tutorials recommend that you sing your samples vs speaking them. So what's the difference in the end product? Sorry if this has already been asked I'm just curious.
 

Nohkara

Pronouns: He/him
Supporter
Defender of Defoko
What is the vocal difference bettween an utau recording from a "singing voice" vs a "speaking voice"? For example, a lot of utau tutorials recommend that you sing your samples vs speaking them. So what's the difference in the end product? Sorry if this has already been asked I'm just curious.

Vowels in singing are pronounced differently than in talking.

For example in Japanese, vowels are pronounced with more open mouth unlike in spoken Japanese. Also, vowels are a bit more relaxed.

A very common mistake that UTAU users do when they record a Japanese VB is that they pronounce /i/ and /u/ way too tightly that makes it sound more like talking than singing.

Some consonants might also sound different in singing VS talking. For example, when singing in English, D sound can be aspirated sounding almost like T. While in talking, D is not strong at all.

How vowels and consonants sound talking VS singing differs between languages, so I recommended to listen to the target language’s songs carefully and if can, try to find music or cover videos when you can see lipsynching (so you can see the shape and size of mouth when said __ phoneme).

This is what we usually mean for “sing your samples” aka. “pronounce things how you would when you sing naturally”.

I learned this from WinterdrivE. When I asked him for tips about exact thing that you asked and that how can I do the same, his answer was basically that take a part of any song and sing them a loud. Then practice singing the same line like in monotone like how you would record a regular VCV while keeping the tone of your singing.

I practiced this every day and watched some tutorial videos that taught me some singing techniques and eventually, my voice got good enough to record a VB with my natural singing-like voice.

I recommend to use a slow BGM guide that is 100 or even slower when recording with that kind of voice. It’s really difficult, if not impossible, to record in singing-like tone without practice, so I definitely recommend start from slow one.

I hope that this answered your question and find this answer helpful.
 
Last edited:

WinterdrivE

Ritsu's Renegades
Defender of Defoko
+1 to what Yuki said.

Speaking tends to sound a lot more closed and tight than singing, and the result of spoken samples is that the end result typically ends up overuniciated. Also, piggybacking off what Yuki said, every language relaxes different sounds in different ways and the pronunciation can change in different ways, even for the same/similar phonemes. (and sometimes even within a particular genre of music)

As a sort of example, here’s my speaking voice in Lyrebird imported into UTAU vs my actual UTAU:



https://clyp.it/mvnfp5dp

It’s not a great comparison since the first one is Lyrebird and not directly my spoken voice, but try to pay attention to he pronunciation itself. In particular, pay attention to L and R which tend to sound very tight and overenunciated in spoken samples vs singing samples
 

Nohkara

Pronouns: He/him
Supporter
Defender of Defoko
+1 to what Yuki said.

Speaking tends to sound a lot more closed and tight than singing, and the result of spoken samples is that the end result typically ends up overuniciated. Also, piggybacking off what Yuki said, every language relaxes different sounds in different ways and the pronunciation can change in different ways, even for the same/similar phonemes. (and sometimes even within a particular genre of music)

As a sort of example, here’s my speaking voice in Lyrebird imported into UTAU vs my actual UTAU:



https://clyp.it/mvnfp5dp

It’s not a great comparison since the first one is Lyrebird and not directly my spoken voice, but try to pay attention to he pronunciation itself. In particular, pay attention to L and R which tend to sound very tight and overenunciated in spoken samples vs singing samples


Thank you for adding some points that I forgot to mention.

Yes, it's definitely true that song genre affects a lot in pronouncing! Like, for example, Enka has extra (sometimes over to top) harsh consonants while V-Key singing style sounds like a pseudo-English accented Japanese.
 
  • Like
Reactions: AmberTonnerre

Similar threads