Thanks to both of you for the information! I'll work on adding it to the OP!
Resource Vocal Synth Voice Characteristics
- Thread starter Scarlet Illusion
- Start date
I think for information like this I'll try to restate some of it in a concise way and the link to the originals for more details.
Last edited:
I updated the OP by adding info about Sonika and Gumi English. Also, I made an archive of this thread on the Wayback Machine just in case something ever goes awry.
Here's the link: Vocal Synth Voice Characteristics
Here's the link: Vocal Synth Voice Characteristics
I just snatched up my copy! I can't use it on my current loan computer (it runs Windows 8, and Z Wave requires Windows 10 to use), but I'm eager to use it once I get my laptop back!
@Moderator Sorry to bother you all! If it's not much trouble, would it be possible for one of you to move this thread to "The Studio" section? I feel like this thread would fit better in that section. If not, that's all right! 
(Please and thank you!)
(Please and thank you!)
Some Renri info:
Certain plosives are explosively plosive, especially when one moves into higher notes. Particularly nasty culprits are と and か, but you may find others. I recently managed to reduce this by manually changing the attack levels in Audacity using the envelope function. Audacity's envelope tool is non-destructive, so if the changes are too much—or not enough—you can go back and redo them. (I imagine other audio editors also have non-destructive editing, but not Edison because FL Studio likes to be different.)
In regards to large pitch jumps sounding fakerooney, increasing the Duration-R value for the note before the jump, and/or increasing the Duration-L value for the note after the jump can help smooth things out.
Certain plosives are explosively plosive, especially when one moves into higher notes. Particularly nasty culprits are と and か, but you may find others. I recently managed to reduce this by manually changing the attack levels in Audacity using the envelope function. Audacity's envelope tool is non-destructive, so if the changes are too much—or not enough—you can go back and redo them. (I imagine other audio editors also have non-destructive editing, but not Edison because FL Studio likes to be different.)
In regards to large pitch jumps sounding fakerooney, increasing the Duration-R value for the note before the jump, and/or increasing the Duration-L value for the note after the jump can help smooth things out.
I'm on Windows 7 and Waves 11.0 (including Z-Wave) works without issues.
Huh. I've tried multiple times, but it just won't let me. It's not a big problem, though. This computer's super slow and buggy so I'm just gonna wait to really start using Vocaloid again until I get my laptop back, at which point I can download it without issue.
I'm beginning to think that computers worldwide are conspiring against you.
I wrote about this in status update but figured it might be a good idea to mention it here, also I hope I'm using the right terminology! Here's my "english isn't my native language" card
Kasane Teto's english vociebank often doesn't pronounce vowels at the end, for example when trying to make her say the word "go" with the "- goU" phoneme she won't pronounce the "U" at the end, it's pretty much identical to the "- go" phoneme. To achieve the correct pronunciation you have to create another note with the right vowel at the end. Keep in mind that in some cases the result will be somewhat choppy sounding.
Also try experimenting to see what works the best, I personally think "- go" with "u" creates the best result but that might not be case for every word. For instance in the word "say" ( the default "- seI" is also pretty much just "- se" ) both "seI" + "I" and "se" + "I" work fine.
Kasane Teto's english vociebank often doesn't pronounce vowels at the end, for example when trying to make her say the word "go" with the "- goU" phoneme she won't pronounce the "U" at the end, it's pretty much identical to the "- go" phoneme. To achieve the correct pronunciation you have to create another note with the right vowel at the end. Keep in mind that in some cases the result will be somewhat choppy sounding.
Also try experimenting to see what works the best, I personally think "- go" with "u" creates the best result but that might not be case for every word. For instance in the word "say" ( the default "- seI" is also pretty much just "- se" ) both "seI" + "I" and "se" + "I" work fine.
Last edited:
I have something to note about Tei ( you're the bigger Tei expert here so tell me whether you disagree with me or not 
) but I'd say she has more of a background noise problem than engine noise problem. Here comes the rambly part that you may or may not add or just add bits that you deem as helpful :
) but I'd say she has more of a background noise problem than engine noise problem. Here comes the rambly part that you may or may not add or just add bits that you deem as helpful :Moresampler makes it more apparent, imo previous resamplers turned the bg noise into engine noise so vbs like Tei and Matsu had about the same amount of it but moresampler is better at picking it up and well..not turning it into straight-up engine noise. I noticed it with Pumpking especially, it's crazy how robotic he is but with moresampler he sounds very natural just with lots of bg noise, yea I think Tei is like that too. As for Matsu..eeh I just think hes one of the voices that the UTAU engine doesn't work for that well
Last edited:
Mind if I ask which of her banks you're using? I don't notice a whole lot of noise in her 3.1.1 VCV (I think it says on her website they used noise removal for the samples on that one anyways), but I think 2.0 has a huge issue with engine or background noise being noticeable. (It's more expressive, but the amount of noise is bad.)
Totally agree about Matsu, though. But I have noticed combining Moresampler as the wavtool and Fresamp as the resampler eliminates the problem, but then causes new problems. (Matsu would probably benefit from being a SynthV or Vocaloid.)
Totally agree about Matsu, though. But I have noticed combining Moresampler as the wavtool and Fresamp as the resampler eliminates the problem, but then causes new problems. (Matsu would probably benefit from being a SynthV or Vocaloid.)
It actually is her 3.1.1 VCV, it could be an improvement from her previous version, I dunno her 3.1.1 VCV is the only vb of hers I've used so I can't tell but I very clearly hear a certain "buzz" in her vb, on some samples more than others. Maybe we should ask more people what their opinion is on this tho
I was thinking the exact same thing about Matsu, him being on Vocaloid or SynthV I mean. Also may I ask, what other probelms does that cause???
Oh, I know exactly what "buzz" you're talking about! She has it bad on any of her samples containing "e" (like "te", "me", etc.). I agree that we should have some other people voice their opinions, too.
Also, as far as the additional issues with Matsu, it kinda distorts(?) his voice and sounds similar to if you applied a light flanger when mixing vocals. Every resampler besides moresampler seems to do that to just about any voice.
Also, as far as the additional issues with Matsu, it kinda distorts(?) his voice and sounds similar to if you applied a light flanger when mixing vocals. Every resampler besides moresampler seems to do that to just about any voice.