Speaking vs singing comparison

Pazuru

Ritsu's Renegades
Defender of Defoko
Hello!

So, some time ago, I've posted a thread stating that my UTAU, Yaemi Keisuke, didn't have emotion in her voice. Most people that have answered told me that was because I spoke instead of singing in my samples.

After a bit of training, I've tried again and made a short comparison between the two. Speaking voice comes first, then singing.



The old one has the following caracteristics:
-Tetra-pitch (4 pitches, B3, D4, G4, C5)
-No end breaths
-Speaking voice, not voice acted

The new one will have the following caracteristics (sample is just monopitch) :
-Tri-pitch powerscale (3 pitches, B3, D4, F4)
-End breaths, in and out
-Singing voice, slightly voice acted

I personally think that the speaking one had more resampler-generated noise than the singing. Also, it had this accent that I don't like very much.

What do you guys think? I'd love constructive criticism and opinions. I'm also open to new ideas!

Thank you for reading this, and have a good day!
 
Last edited:
  • Like
Reactions: Nohkara and partial

Nohkara

Pronouns: He/him
Supporter
Defender of Defoko
I think that you've improved a lot!

Maybe sound configurations (oto.ini) needs some adjustment but the voice tone itself is great.

The pronunciation was mostly ok but I suggest to practice /a/ and /o/ sounds because these sounded more like /{/ and /@/ than /a/ and /o/ (unless this is what you wanted/done on purpose for V-Key style genre or smth).
 
  • Like
Reactions: Pazuru

Pazuru

Ritsu's Renegades
Defender of Defoko
Thread starter
I think that you've improved a lot!

Maybe sound configurations (oto.ini) needs some adjustment but the voice tone itself is great.

The pronunciation was mostly ok but I suggest to practice /a/ and /o/ sounds because these sounded more like /{/ and /@/ than /a/ and /o/ (unless this is what you wanted/done on purpose for V-Key style genre or smth).
I've done a quick oto for this sample! But you're right, I'll check it!
(I was aiming for that pronunciation, but if that bugs too many people, I'll change it).

EDIT: I checked my oto, and it seems fine to me. Maybe I just didn't fit my UTAU to the UST...

Thank you for your feedback!
 
Last edited:

Pazuru

Ritsu's Renegades
Defender of Defoko
Thread starter
I think that you've improved a lot!

Maybe sound configurations (oto.ini) needs some adjustment but the voice tone itself is great.

The pronunciation was mostly ok but I suggest to practice /a/ and /o/ sounds because these sounded more like /{/ and /@/ than /a/ and /o/ (unless this is what you wanted/done on purpose for V-Key style genre or smth).
(SInce I can't edit my reply anymore, for some reason, I'm going to reply here).
After checking my oto, I've noticed that a lot of sounds weren't otoed. It strange, because I remember doing it, but I guess SetParam just decided to delete what I've done.
 

Similar threads