I found out that auto pitch tuning is great but causes havoc in case you start to copy phrases around. I ended up with the same sentence at 3 different locations in the song that all sounded totally different and strange.
The cause was the pitch deviation curve, that I never edited, but was generated automatically whenever I typed a new word. If you copy and past notes around and forget to copy the curves, it sounds strange, but if you start moving sentences around and forget to move the curves (because I have no recollection of ever editing them) things become very strange.
I suggest to simplify things to always copy/move the pitch deviation curves with the notes. Maybe a special mode?
A proper manual that explains these kind of pitfalls would also be good.
P.S. wanted to use documentation tag, but that one did not exist.
When running Auto Pitch Tuning it is generally a good idea to only select the phrase you are running it on at the time. This way you don’t end up introducing changes to different parts of the project.
If you have multiple instances of the same phrase I would recommend using Groups. By grouping the notes you can place the same sequence at multiple points in the arrangement and they remain linked, so changes to one will affect the others.
Thanks Claire, very helpful, but probably not for my workflow.
As I am composing songs directly within my daw and synthesizer V, I am never sure if a phrase will be reused or not. So, grouping is not a natural choice.
Usually, this is done on the daw level, but the integration of synthesizer v with daw is still not that far.
I now understand that for this one song I must have somehow disabled instant mode and that caused all the issues.
I must admit that I did not watch all the tutorial videos because I wanted to start using the tool. I watched them now and was surprised that I did not find any thing about instant mode or auto pitch tuning.
Definitely needs better documentation. Even in Japanese would be a start, we all could use google translate to get a somehow understandable English version.
Can anybody explain instant mode? How to turn it one again? I noticed that after switching it off, all the pitch deviation curves stay even if you turn the mode one again. Can’t wrap my head around that at the moment.
Instant Mode was introduced after the video series was created. It’s one of the main reasons that I personally haven’t made video tutorials; they’re a pain to maintain and keep up to date (which is exactly why Dreamtonics needs to step up their user documentation…). (Synthesizer V Studio 1.6.0 Update | Dreamtonics株式会社)
When Instant Mode is enabled it runs auto pitch tuning on the notes while you work, and gives access to AI pitch retakes. Timbre AI retakes do not rely on Instant Mode.
These Instant Mode pitch deviations are done separately from the Pitch Deviation parameter panel, so that they are separate from the user’s manual inputs.
When Instant Mode is toggled off, all these “invisible” pitch deviations will be applied to the parameter panel additively with any existing manual changes, so the resulting pitch curve will remain the same.
When you re-enable Instant Mode it does not clear the parameter panel because that would overwrite your manual changes.
Currently there is no way to turn Instant Mode off without keeping the pitch deviations, except by pressing “Undo” immediately after turning it on, or by removing all parameter nodes afterward (with ctrl+A followed by delete in the parameter panel).
I’ve written more here but I think that at least covers your points:
Thanks claire, very helpfull informaton you referenced and wrote.
However, not allowing us to switch back to Instant Mode is a problem in case you accidentally turned it off.
There is a way to to solve this elegantly. All Dreamtonics need to do is subtract the automatic curve back from the manual curve. This should leave only the values that the user added/subtracted.
Yes, this is how it already works. The user’s manual changes are not cleared, but the curves added by instant mode get moved back to the “invisible” layer.
There is some very minor loss of precision due to some minor smoothing of the curves. Deleting the residual points in the parameters panel near zero will usually resolve this (but also some user might want to keep the smoother curve).
you could try grouping and moving to the selected parts of the song that you want, then UNgroup so that you can make the unique changes that you want, for example to each chorus at least this way your initial changes will be the starting point for the parameters. and will not mess up other parts of your song.