Feature Request: Phonemes Documented Within Synthesizer V

In prior versions of Synthesizer V, there used to be a phoneme list.

This phoneme list is no longer accessible via the UI.

It is possible to find it in the Synthesizer V Studio Pro folder clf-data, but that’s not really documentation - that’s literally data for Synthesizer V to use, not documentation.

Additionally, there’s no examples for what the additional phonemes do.

I’m familiar with the ARPABET. The only reason I discovered that phonemes like dr and tr are supported is because they appeared when English words were converted to ARPABET phonetic equivalents. However, no documentation for these phonemes from within the application.

Please add documentation for the supported phonemes back into Synthesizer V.


Yes. There are too many “hidden” features - like “br”

1 Like

YES! I am so for this! Also, can the Alternate Phonemes that you gain access to with Studio Pro PLEASE become labeled 1-xx so that I can use them in my personal user dictionary…:sob:
It’s the little things, man.

Also BTW is it just me are some alternate pronunciations(like, for the same exact letter input) DEFINETLY not reachable just by toggling the alt button? Like I keep needing to use ‘sil, cl’ or just give more space bewtween starts+ends, or change a note that is like [b r aa] to two separate notes like [b][r aa] where the [b] is taking up the same amount of space but DEFINITELY seems to change sample.
Plus I am confused on how to get parity / smooth transition between Tsurumaki’s ENG and JP voice, because I am unsure if she has any japanese R (or otherwise) in the ENG bank.


Since SynthV appears to use diphones, sometimes the way to change a phoneme is leave the the one you want alone, but change the one the precedes it.

1 Like

:thinking: Good point! I’ll have to keep it in mind… when I encounter that again.
…And yet… :flushed:
I had no clue that SVS/SVS-P was limited to resampling off of diphones, mainly because of a (possibly deceiving!) glitch/bugged-feature/‘hold-over from old arpasing VB formatting’ that i found earlier on!
It was in PREVIOUS VERSIONS of SVS-BASIC, and it seemed to get PATCHED OUT in recent updates! It allowed a free-trial / ‘Basic’ user to access alternate phonemes, just without the UI button (and interestingly; Certainly MORE alt options than whats accessible through the UI alone.)

The work-around was to append a number to the individual phoneme (ae2 aa12 r1 r3 r16 t18 for instance) and it would grab an alternate sample of that sound, IF and only if one existed of that number. They seemed to be labeling the alt-phonemes like 1…2…3… and so on, with 1 being the default that it was already using w/o a # on it (if i remember correctly!).

** SO if you got too high up in #, it would still work, however all #s beyond the max # of variants sounded the same, like it was repeating the last one. **

^^^ (((Note on above = NOT SURE THO B/C: sometimes I was PRETTY SURE i had reached the max (~10,11,12) b/c they all sound the same, but then after going up far enough say (~15, or even 18) suddenly there is just a WAY different sounding one. Glitch maybe? Or maybe around the mid-range of #s it is going through dif ‘OTOings’ of the same samples :thinking: )))

ANYWAY I just think the implications of stifled/canned/removed features/capabilities is fascinating and i wanna know how that whole thing actually worked under the hood!

What made this bug/feature INTERESTING:

  • The unintentional feature appeared to work on single phonemes, regardless of diphone
  • Sometimes it was VERY clearly using a different sound-source sample / alternate, but other times it was more close, yet a different sample, or seemingly(?) the same sample source BUT it sounds as though it is the tail end of that consonant/vowel, or the beginning (so like a different OTO-alias, relating it to UTAU for lack of better terms). Think: “Use this same ‘R’ sample, but * the envelope+cutoff* focuses on the START of the R, rather than the END of the R”…
  • There were a LOT more options than just 1,2,and 3 :roll_eyes: :unamused: Which makes me wish I could still access this! Especially since now I own Studio Pro!
  • Surprisingly, they didn’t JUST remove the access to(ADMITTEDLY GRATUITOUS) alt-pronunciations from Basic, but they are not in Studio Pro either!
  • I yearn for this feature, because as of now there is seemingly no WAY to use an Alt-Pronunciation in your user dictionary entries! I know its nit-picky, but I wanna get funky with my Dictionary! I might get a lil crazy and try to make Maki speak french for instance. Or, seeing how her VA is not american, i might find that a certain word JUST don’t sound right unless I use an alt. Yes this is rare, because luckily they seem to have done a DAMN fine job with Maki’s English pronunciation in general, and you can TYPICALLY just use a slightly dif vowel/cons to achieve a “American” or “Natural”/“Singing”/“Speaking” pronunciation of the word. But ya know, for edge cases this option would be p dope!
  • HONESTLY i just want more clear documentation on the available (SEE: POSSIBLE) phonemes for a given bank. Because I’m pretty sure I’ve stumbled upon some WEIRD undocumented phonemes that I never see used in Maki’s bank.

btw if you got this far SORRY for the really random long-winded and late as hell reply lol woops

1 Like

Hi, I just found out an interesting SynthV’s feature. Put a dot first then you can type anything even with only consonant.

It is a standard feature. After dot you are supposed to write phonemes separated by spaces. But without a documentation it is hard to find out. I saw it in someone’s video. :smiley:


As hataori already noted, it a way to rapidly enter phonemes.

It’s actually a better way to enter custom pronunciation. That’s because if you change the phonemes above the note, and then later change the text of the note, the phonemes you put up there will remain there.

If you instead use the dot notation to indicate a custom pronunciation, when you change the text, the phonemes will change as well.

1 Like