Complaints about SV2

Synthesizer V Studio
1
  1. Manual Mode is MISSING — seriously this is killing my workflow i’m crying screaming begging for this to come back
  2. Expression Group option is gone (Yamine Renri can use this, not sure about other ‘Standard’ voice banks. On the plus side, Yamine Renri can inexplicably sing in english and other languages available to V1 voicebanks, but not Korean for some reason? as if Korean is exclusive to V2 voices. Expression Group is an option found in the Note tab and it’s below the alternate pronunciation options)
  3. ALT1 and AL2 => another standard voice feature. I don’t think AI voices use this. But basically you click the lil button that says “DFLT” next to the phoneme on the Note Properties panel, and it changes that specific phoneme to a alternate version. I find it interesting that Standard voices seem to have gained XLS (minus korean, which imo seems like it shouldn’t be the case as long as they’re able to add languages to VBs retroactively…) but lost other important features, like access to their ‘appends’.
  4. Pitch Deviation Panel => It functions differently. You can’t write the pitch bends generated by the AI down permanently by switching to manual mode anymore. Seems like you can do something similar by tracing the pitch to control points, but BTW the pitch control points aren’t easily visible, and also they are SUPER GLITCHY NOW…
  5. Pitch control points = useless (right now, probably just a bug and not a cut feature) bc daring to draw your own pitch curves makes the voice get all glitchy and i compared it to SV1 just to check if i was forgetting and yeah SV1 was capable of using pitches drawn on the piano roll without glitchy jumps in pitch. OH AND WHY DO V1 VOICES NOT SUPPORT PITCH CONTROL POINTS??? They had them in SV1!
  6. Srsly you get a pop up that says Version 1 voicebanks don’t support pitch control points. But if you fix your glitch notes using pitch control points via a different, Ver 2 voicebank, and then switch BACK to the version 1 voice bank… its fine T_T so stupid.
  7. It takes about 3x as long to render a loaded in project. Meanwhile on their product page they boast 300% increase in render speed… XD I think you mean 300% wait time increase. LOLLL
    image
    image729×143 21.7 KB
  8. All the AI retakes are combined to a single drop down list :frowning: i liked it a lot more when we had Pitch and Timbre in two separate columns so you could mix and match the ‘style’ with ‘pitch’
  9. I preferred the old method of downloading and checking your software version. When you click on the cloud icon to check for updates it takes a while to load and its a pop up that obscures whatever your working on. Also it seems glitchy, like it pops in fully loaded for a second and then unloads and shows a loading circle for a couple seconds before going back to what it was showing a second ago. Not to mention the process of getting my voices added to this program was confusing.
  10. Why isn’t Mouth Open/Closed a parameter on the global voice settings panel??? It seems like it should be alongside the likes of gender, tone, tension and such. I think that’s an oversight.
  11. It doesn’t make sense that Version 1 voices can’t use Korean XLS, as evidenced by Standard voices having XLS it’s obviously something handled by the editor, rather than the individual voices.
  12. Trace Pitch Into Control Points = this option literally does nothing XD unfinished-ass-program…
  13. Vocal Mode track defaults UI is annoying because you can’t just drag a straight line. it feels less precise to me. also while im complaining about this imma just say: Being able to split those into 3 different sub controls (pitch timbre and pronunciation) doesn’t seem very useful, when you consider that if you progressively control a vocal mode accross the track, there isn’t any way to control which 3rd of the vocal mode you’re actually mixing in… So if I can’t use it to change pronunciation on a small section, only the entire track
    image
    image231×671 20.2 KB
  14. No Easy Volume Correction = OK this is not just a SV2 problem, it’s also a problem in SV1, but I think because the vocal modes and parameters are stronger in SV2 it’s gotten a little more obvious. Sometimes one vocal mode is quieter than the rest, so there should be some kind of tick box or something to level out the volume automatically. It’s also a problem with the Tension parameter, lowering tension a tiny bit makes the voice too quiet to even hear. Now, raising the tension also makes the volum a lot louder too! So I think it really did get even worse.
  15. The Takes drop down list doesn’t list all of the takes if you have added more notes / split notes since you last generated takes. it’ll look like this:
    image
    And then it’s really annoying to individually change each note to the take you liked from before

I’m sure I’ll think of more complaints later but this is what I’ve dealt with so far. SV1 is still gonna be my go-to for covering vocaloid stuff until they get the SV2 editor on par with the old one (add manual mode)

1 Like
2

I am in the midst of transferring some V1 to V2 voices.

  1. I also miss the manual mode and its features.
  2. The mouth parameter seems to overreact. A bit of mouth opening seems like it is too much. For example: An ‘Ah’ sound. One should be able to make it sound similar to saying ‘ah’ for the doctor. It seems that the adjustment makes it thinner sounding? Maybe a way to reduce the resolution instead of trying to draw tiny curves?
  3. Still no area for project notes?
  4. It would be nice that one could draw in either exacting resolution or in smooth curves. Like a sort of Bezier curve between words?
  5. The volume correction is great for fine tuning individual words, but it would be extremely helpful if one could choose a word or section and increase the complete volume for that section where the existing waveform’s amplitude just increases across the word or section. It means a fine resolution and a coarse resolution.
3

I think we had it made with V1.

2 Likes
4

yeahhhhhhh im just hoping that they like. released this program a year too early and it will eventually have all the same features as SV1 :confused:
srsly like how can they boast 300% faster render speed when it’s literally the complete opposite, and just scrolling around on the piano roll and interacting with the AI can be super laggy

5

A lot of issues still need sorting out, I agree. But the performance must be relating to individual hardware configurations, rather than the software.

Here on Windows 11 and using ARA bridge mode in Cubase 14 Pro V2 is very fast to render, but having said that, I had no problems with V1.

We need to be more specific for the benefit of any Dreamtonics bods reading this in terms of hardware config.

6

You know, it’s funny… I’ve got pretty good studio monitors and Solaria II definitely sounds better than Solaria 1 from an audio perspective - it’s cleaner and in comparison you can hear some artifacts in Solaria 1 that are usually hidden by the mix. That said, when I just enter in a vocal line with lyrics, I prefer Solaria 1 every single time. She’s a little soft in her singing voice and the overall effect is very, very charming. Solaria II is more brash and strident by default. You have to do a lot of tweaking to get the charm back, and it’s too much work, so I use Solaria 1. I use SV2 for Liam and Hxvoc. Hxvoc is absolutely brilliant.

2 Likes
7

I’ve struggled quite a bit with setting universal voice settings on my tracks (as opposed to from group to group). No clue if that’s another thing the program lacks or if I’m just stupid.

8

My greatest problem is the grid visibility. Seems rather than listen to those who wanted it MORE visible, they made it LESS visible, at least to my 83-year-old eyes. It’s the kind of thing that is quite SIMPLE to program, adding a brightness slider and perhaps a color wheel or set of options.
As is, I find it almost unusable and my eyes aren’t THAT bad…

1 Like
9

Curious. I have found the grid more visible in v2 than v1 using both on a Mac. Although I agree that being able of adjust the brightness of the grid would be nice.

10

Personally, before I knew about the manual mode being removed which is pretty much a deal-breaker for me, my main issue with the product demonstration was the way it was portrayed. Like, they showed us a song using Natalie V 1 and 2, and they were talking about how much better the Synth V 2 demo sounded, and I was sitting there thinking, “No, not in the slightest. This is a pop song, and V 2 sounds like she’s singing in theater style, and V 1 sounded a lot more natural for this song.” I mean, I even noticed AI artifacts in V 2 that weren’t in the V 1 demo.
I suppose this is a matter of opinion, but I just couldn’t understand how they could listen to both and actually believe that version 2 sounded better. And honestly, I’m really not trying to attack them when I mention this. I believe they’re trying their best to make a great product. But… I really don’t know. When I hear them talking about how much better it apparently sounds even though I really don’t think so, and because of their lack of transparency and reception when it comes to the manual mode discussions, it really makes me wonder what their motives are.

11

Just shopping channel hype … Plus neither presenter the Latin American and the Japanese ‘boys’ speak well pronounced English to know !