Experiment: Use TTS Programm Balabolka and some Wizardry

Here’s a little patch that has syllables on keys…for each Octave
C -> “This”
E -> “Is”
F -> “Psy”
A -> “Trance”
A# -> “Whis…”
B -> “…pers”

This-Is-Psy-Trance-Whispers.vital (345.6 KB)

Please don’t expect the word to be vey clearly pronounced or perfectly understandable … for background “whispers” it’s kind of close enough :wink:

I did some Wizardry …
Started using Balabolka as Text-2-Speach programm
Then I extracted all the syllables with waveozaur AND made them all exactly 4096 samples long (I think you can do the math … if you have for instance a syllables which is 7736 samples long, how do you get the target sample rate in order to resample it to be 4096 samples long)

Then I concatenated all 4096 long syllables into one wav file
Then I imported it into Vital as Wavetable … Attention: Windows-Size must be set to 4096 manually for both key frames!

Why is this different to Vitals Text-2-WT? Because Vital has no notion of words or syllables. It kind of “abuses” a wavetable to recreate a sample. With my method I can control which syllables is used because each occupies exactly one wavetable slot.