source · discord · email | support via: gh sponsors · ko-fi · paypal · cashapp

new klattsch has a discord now. join →
phoneme sequence
syntax help
tokens (whitespace-separated)
  PHONEME              ARPABET code (AY, IH, AA, S, ...)
  PHONEME!             stressed (transient: +8 Hz lift, longer duration)
  PHONEME+N            rising pitch glide of N Hz; new pitch sticks
  PHONEME-N            falling pitch glide of N Hz; new pitch sticks
  PHONEME(+N)          transient rise: ornament only, doesn't carry forward
  PHONEME(-N)          transient fall
  ( PHONEMES )         syllable group: phonemes share one rate slot
  ,  ;  .              short / medium / long pause (100/200/300 ms)
  bN     b=N           absolute base F0 in Hz (also accepts notes:
                       bC4, bC#5, bDb3, bA-1, etc.)
  b+N    b-N           relative base F0 (running pitch shifts by N)
  b                    reset base F0 to the utterance's starting value
  rN     r=N           absolute per-phoneme rate in ms
  r+N    r-N           relative rate (positive = slower)
  r                    reset rate to opts default
  pN     p=N           insert exact N-ms pause
  sN     s=N           absolute formant scale (1.0 = male baseline,
                       1.17 = typical female, 1.3 = child)
  s+N    s-N           relative scale shift (decimal: s+0.1, s-0.05)
  s                    reset scale to opts default
  vN                   vibrato depth in Hz (peak deviation; 0 = off)
  v+N    v-N           relative depth shift
  v                    reset vibrato depth
  wN                   vibrato rate in Hz (LFO frequency, default 5)
  w+N    w-N    w      relative / reset
  hN                   breathiness / aspiration mix (0..1)
  h+N    h-N    h      relative / reset
  tN                   spectral tilt (-0.9 darker .. +0.9 brighter)
  t+N    t-N    t      relative / reset
  gN                   vocal effort (0 lax .. 1 tense; 0.5 default)
  g+N    g-N    g      relative / reset
  [base=N] etc.        verbose form, equivalent to bN, rN, pN, sN
  # rest of line       comment

f0 evolution
  stress (!) is transient: doesn't carry forward
  bare pitch deltas +N/-N are sticky: mutates the running pitch
  parenthesized deltas (+N)/(-N) are transient: this phoneme only
  bN (or bC4 etc.) is an absolute reset to that pitch
  bare b returns to the utterance's starting pitch
  examples:
    "AY+20 D IH+10 D"   D at base+20, IH at base+30, D at base+30
    "AY(+20) D IH(+10) D"  D at base, IH at base, D at base (ornaments)
    "AY+20 D IH(+10) D" D at base+20, IH peaks at +30, D back at +20

examples
  HH AH L OW                       hello, default voice
  b140 HH AH L OW                  higher voice
  bA3 HH AH L OW                   higher voice (note name)
  AY+15 D IH D                     "I did" with rise; D and IH stay raised
  AY-15 . b120 AY+10               fall, reset, then rise
  D IH D DH AE(+40) T              "did THAT" with accent ornament on AE
        
canned utterances
audition phonemes
defaults (commands override this)
120 110 1.00 0 5.0 0.00 0.00 0.50