need advice on using flags

Hello, I've been working with UTAU for a little while, but I never end up using flags because I don't really understand very well how they're used. I've seen various lists for what different flags do, but most of the time when I use flags they just make the Utauloid sound muffled or low in volume (especially if using combinations of flags). Other than that I usually can't tell the difference between a flag being in use or not (unless it's a very obvious effect like UTAU growl or the g flag). I know not all flags work with all resamplers (I typically use fresamp14) and have different effects with different voice banks.

^That's my experience with flags. I have many questions, and I'll try my best to be clear about what I'm asking. Any one person shouldn't feel like they have to answer all of them.

From what I've seen it seems pretty typical for people to use flags for voice clarity over their whole UST. Other than that I'm not really sure how people tend to use flags. Should they be used more on a note by note basis, or more as an addition to an entire phrase or section (I imagine it might depend on the flag though)?

Are flags more useful for different voice effects or are they more to correct errors in how the UST sounds? I know they're sometimes used to make an utauloid sound more "human." Does that mean trying to change the tone of certain words within phrases or is it more just eliminating noise or possible glitches? I'm thinking it's probably a combination of both.

Last question. I've seen people write that certain flags have x effect on their utauloid, but might not have that effect with others. How specific are a flag's effects to any given voice bank? It seems like they should all generally have the same effect over many different voice banks (because that's how they should be designed as far as I know), but I can understand how one might have more of an effect on one voice bank vs a different one depending on how it was recorded and otoed.

Kidding this is the last question. A lot of flags seem to have something to do with a formant filter. I know roughly what formants are (very roughly)- what are these flags trying to accomplish? For example the flag F which "determines the strength of the formant filter."

Sorry for writing 3 pages of questions. Any help is appreciated ^^

Kiyoteru said:

A master post of all known flags was collected here. The page itself has a lot of information for quick reference, but I recommend downloading the PDF as well.
https://utaforum.net/resources/updated-flag-list.431/

Flags can do a wide variety of things, and they're very dependent on which specific resampler you're using. They can be used universally for the whole UST, on certain phrases of the song, and on individual notes, all depending on what you're trying to achieve with them.
Many flags act like EQ (such as highpass and lowpass) and aren't particularly useful, since you have more control over EQ while mixing.
I personally use Moresampler very often because it has a wide selection of unique flags. The ones I actually use are g, e, Mt, Mb, MG, and b. g is the well known "gender flag", e forces stretching instead of looping (some resamplers only stretch, those that have a looping option usually include the e flag to force stretching), Mt (Moresampler only) affects tension and can create a wide change in tone from soft to powerful, Mb is Moresampler's more natural-sounding alternative to BRE, MG is growl (which I usually turn all the way up and combine with a higher g value for an individual note), and b in Moresampler only controls the loudness of consonants. Over a whole UST I might set e, over certain phrases I might set Mt and Mb (to make the verse soft and the chorus powerful, for example) and on individual notes I might set b (to control that particular consonant).
With low-quality voicebanks I can attempt to correct the sound by using individual note flags as well. If one recording was soft and another was powerful, I might put Mt flags that "undo" the effect and try to get them sounding more neutral and cohesive.

The specificity of flags to a voicebank is really because it depends on the tone of the voice and the recording quality. You can use flags to make a normal voicebank soft and whispery, but if you use the same flags on a voicebank that's already soft and whispery you're not going to get the same results.

The formant filter is what prevents audio from sounding squeaky and chipmunky on high notes, or crazy deep on low notes. You might be adjusting the strength of the filter so that notes sound more similar to each other or less similar (since a voice would naturally have a higher tone on higher notes).

Click to expand...

Thank you, that's really helpful! I'll have to check out moresampler and try out some different flags on my old USTs.