github tgeczy/TGSpeechBox v-270
NVSpeech Player with phoneme editor and NVDA Addon, Speech Dispatcher module version 2.7

latest releases: v-300rc2, v-300rc1, v-300b7...
one month ago

NV Speech Player v2.7 Release Notes

New Voice Quality Parameters

Two new VoicingTone V3 parameters for enhanced voice customization:

Formant Sharpness (cascadeBwScale)

  • Controls the sharpness/clarity of formant resonances
  • Lower values (0.2–0.5) create crisper, more "Eloquence-like" definition
  • Higher values (1.2–2.0) soften formants for a more muffled quality
  • Default 0.8 provides good intelligibility out of the box
  • Accessible via the new "Formant sharpness" slider (0–100) in NVDA, Linux Renderer via CLI, and phoneme editor.

Voice Tremor (tremorDepth)

  • Adds realistic elderly/shaky voice effect
  • A 5 Hz oscillator modulates pitch, open quotient, and amplitude simultaneously
  • Creates natural "wavering voice" characteristic of aging or emotional tremor
  • Combines well with jitter/shimmer for enhanced realism
  • Will override Fujisaki pitch: Try tremor with 20% to make Fujisaki more natural!
  • Range 0.0–0.4; default 0.0 (off)
  • Accessible via the new "Tremor" slider (0–100) in NVDA, Linux renderer via CLI, and phoneme editor.

Bug Fixes

  • Trajectory limiting: Fixed an issue in the frontend's trajectory limiting pass which was rounding vowels into different sounds.
  • Glides at high rates: Glides no longer disappear as easily when speaking at high speech rates
  • reverted prior logic for voiced consonant stops. It was creating clicks for some of our users.

Language Improvements

  • Hungarian: Softened the á vowel for more natural pronunciation
  • Hungarian: Fixed N incorrectly becoming ny in certain contexts

Don't miss a new TGSpeechBox release

NewReleases is sending notifications on new releases.