Thoughts on NineZero

I just bought NineZero without watching a demo, because I love Kanru/Dreamtonics products.
However, I’m a little skeptical about this one, NineZero is very atypical as Astorian, Feng Yi or Cheng Xiao can be, we can’t use them everywhere.
I had hoped for more possibilities in terms of modes, there is no “Soft” “Air” mode, so we forget the Love Songs of Three Days Grace/Linkin Park/My Darkest Days.
And where I’m really disappointed are for the modes, I had really hoped for more aggressiveness in the sound, like a “Growl” mode, “Grunt”, and why not a “Guttural” mode, “Death Growl” , “Gruik”, a very very “dirty” voice.
Of course I don’t know the interpreter of “NineZero”, I don’t know if he was able to feed the “Machine Learning” with this kind of stuff, but I’m a little disappointed, but nothing serious, it’s my fault :).
On the other hand, Kanru is moving forward and now offers us very beautiful voices, thanks to him and to Dreamtonics, thanks for your work.
I can’t wait to hear the next voices because they are getting better and better :slight_smile:
See you soon :slight_smile:

ps : sorry for my english :slight_smile:


I was just going to comment on this. The voicebank is great, but the lack of dynamics is a little bit disappointing. He can only do power vocals. There’s no soft, airy, or light mode, so he just sounds loud all the time. And I definitely agree that they could’ve done more with the “aggressive” modes. I expected him to be able to sound much stronger than he does. I really hope they keep working on him and improving the voicebank, cause he does sound good.


Kanru is on the right way :slight_smile:


it is too open and sharp … the mute is not “soft” enough … maybe the update will come soon

1 Like

It’s a fun time right now.

1 Like

I think Ninezero highlights some limitations of the current AI engine, and I’m curious to see how Dreamtonics moves foward.

For comparison, consider Megpoid V4 Complete. This is a package of Vocaloid4 voicebanks with Native, Adult, Power, Sweet, and Whisper variants, which (back before the release of Vocaloid5) could be combined with the cross-synthesis feature, but were otherwise completely separate voicebanks sold as a bundle.

While I think we can all agree that SynthV has drastically more impressive output quality than Vocaloid, there’s something to be said about the sheer flexibility of GUMI V4. She can not just sing in a wide variety of styles, but also fill the various niches without compromise because each voicebank is an entirely separate product dedicated to that style.

Vocal modes are clearly intended to mimic the Vocaloid4 concept of combining multiple voicebanks to cover a wide variety of styles, however since these are all packaged within a single product, the result is often a “jack of all trades, master of none” situation.

No matter how much you tell the engine to favor Kevin’s belt mode, the AI engine still can’t “forget” about the rest of the machine learning that isn’t belting (which includes a lot of softer singing!). This is, in my opinion, the biggest drawback of the vocal mode approach.

This was also one of the most common reactions when people first compared the Standard and AI voice databases released by AHS – the AI ones simply did not have the same ability to perform in the upper range without becoming thin and inconsistent, even when comparing the Standard version by the same voice provider (such as Saki vs Saki AI).

Vocal modes have certainly improved the situation, but Ninezero has, in my opinion, highlighted the limitation of Dreamtonics’ current approach; in order to make a voice database cater to a niche, it seems that it must also exclude anything that falls outside that niche.

Of course I could be wrong, and maybe Dreamtonics will surprise us. Some people will find Ninezero fits exactly what they want, but the fact that he can only belt is going to be very limiting for songs with any sort of stylistic progression.


I’m not sure that I agree with this assessment of 9z. Although I do need to do more tests with his softer singing to see what results I can get. I think that the disappointment people have with 9z is more to do with his tone, which is actually quite accurate to the original singer. I am actually pleasantly surprised at his realism. What I perceive, after playing with 9z, is a step forward in realism and authentic singing. The issue is that his voice is a bit niche but really good, nonetheless, in that niche. I think if Dreamtonics would release a more versatile style of rock voice ( something like the lead singer of 3 doors down) more people would be amazed at how far this tech has come. So I’ll reiterate my point. I don’t think that the issue with 9z are lack of modes or the way these modes are implemented but rather that his voice is just a niche voice, no different than the real singers. Having said that, I’m not against adding more of the types of modes mentioned here, but I don’t think that would necessarily make his voice more useful outside of the type of music that he is good for, which just isn’t going to be as wide as we would like, but I still think it’s a great voice to have an worth having. I just hope that Dreamtonics doesn’t stop with 9z when it comes to rock voices because people don’t recognize the breakthrough his voice is due to his niche tone.

1 Like

I appreciate your comments, thank you :slight_smile:

@Claire: I partially agree with your comments, there is a small limitation on the AI, but I want to clarify that compared to the first version of SynthV, Kanru did an EXCEPTIONAL job with more than 1000% on voice quality. I’m no longer afraid to release my EPs with an AI without singers, it’s very appreciable.
I have mixed feelings about the “fictitious” voice modes of the voices, but I get what you mean :slight_smile:

I’m going to listen to the demos of Gumi, I haven’t been interested in Vocaloid for a long time [Thanks Kanru :)].
Claire: Have you tested/purchased the latest Vocaloid? Is it interesting for SynthV users?
And people outside of Claire who have Vocaloid6, what do you think?

@Pderbridge: Exactly, it won’t be easy to place it in a so-called “normal” song, like with Astorian, Feng Yi or Cheng Xiao.
I clarify, I didn’t say the N9 voice was bad or anything else, I had hoped for more options, but I should have looked at the modes before buying :slight_smile:

Coming back to NineZero, I wouldn’t be surprised if it finds its place in one of my DrumnBass or DubStep compositions, let’s see :smiley:


Right after listening to some Vocaloid demos…hardtuning really isn’t for me. :cold_face: :hot_face: :cold_face: :face_with_head_bandage: :face_with_thermometer:

I haven’t used Vocaloid much personally, the comparison I made was more about the approach to development and ability for the product to cover many use cases while still excelling in niche situations. Of course Megpoid V4 Complete is from 2015 so it’s not really comparable in terms of synthesis quality or realism.

1 Like

thanks for your feedback which i feel is real helpful.

Hi peeps, I recently bought 9z and I agree with many posts on the forum that it is a generally nasal voice that seems more suited to belting all the time I spent a bit of time to try and address those perceived issues.

Nasal quality - I used melodyne to drop the formants quite a bit and also randomly detune so the sound is closer to a real vocalist.

I used the various other parameters to soften some lines to then accentuate the belt and the dynamics (to my ear) while still a work in progress are a step in the right direction.

See what you think….

1 Like

I love 9Z, but it is just that he fits perfectly the type of music I write (Hard Rock).
For songs or passages that a different approach is needed, or for a call and response where a different tone is needed, I use a different voice bank (Jun has been a great complement).
But for me, discovering Synth V and 9Z together was like a gift from heaven. I had been dreaming to have exactly what they offer for years. 9Zis exactly the type of voice I have in mind for most of my songs.

I’m playing around with gender and phonems to get a little more control over the nasal voice. But I would still like to have a voice for heavier stuff. For me it is not a perfect fit. Unless I make a hard rock song. :wink:

9Z Soundcloud

Dreamtonics could ask this guy, i think he can handle everything about rock music.

1 Like