Waveguide: the VL70-m's throat formant and harmonic enhancer #159

New issue

Open

opened 2026-10-06 16:24:14 +00:00 by bcox · 0 comments

bcox commented

2026-10-06 16:24:14 +00:00

Owner

bcox asked on 2026-10-06 for a ticket for each gap docs/feature-survey.md names without one, whether or not it is ever done. Engine claims are from reading the tree at 6a3fede9 on main. Instrument claims are from that document's research, read from the manuals by research agents and not re-checked; the sources are listed there. Nothing here is built, measured or listened to.

Summary

The waveguide layer's lanes are pressure, tension, pitch, level, noise amount, return and loss (WaveguideLayer::Destination in dsp/waveguide_layer.h), over reed, jet, lip and burst drives. The VL70-m adds, ahead of the driver, a throat formant ("the simulated player's lungs, trachea, and oral cavity ... only applies to some reed-type voices") and, after the resonator, a harmonic enhancer, both with key scaling; its damping and absorption are the loss and return lanes and its growl an LFO on pressure, so those fit. Scream, "chaotic oscillation", is #146 and the bow #143.

What the tree has

The unvoiced engine's three formant bands over an input (UnvoicedLayer with ExcitationSource::kInput), which a second layer can put on the waveguide's tap after the fact, not in the breath path.

The work, as a sketch

A formant filter on the pressure path before the drive, one band with frequency, bandwidth and level lanes; and an enhancer stage, a high-passed drive summed back, after the loop.

Open

What the VL70-m's enhancer is; its manual says "actually only used on a few voices".

bcox asked on 2026-10-06 for a ticket for each gap `docs/feature-survey.md` names without one, whether or not it is ever done. Engine claims are from reading the tree at `6a3fede9` on `main`. Instrument claims are from that document's research, read from the manuals by research agents and not re-checked; the sources are listed there. **Nothing here is built, measured or listened to.** ## Summary The waveguide layer's lanes are pressure, tension, pitch, level, noise amount, return and loss (`WaveguideLayer::Destination` in `dsp/waveguide_layer.h`), over reed, jet, lip and burst drives. The VL70-m adds, ahead of the driver, a throat formant ("the simulated player's lungs, trachea, and oral cavity ... only applies to some reed-type voices") and, after the resonator, a harmonic enhancer, both with key scaling; its damping and absorption are the loss and return lanes and its growl an LFO on pressure, so those fit. Scream, "chaotic oscillation", is #146 and the bow #143. ## What the tree has - The unvoiced engine's three formant bands over an input (`UnvoicedLayer` with `ExcitationSource::kInput`), which a second layer can put on the waveguide's tap after the fact, not in the breath path. ## The work, as a sketch - A formant filter on the pressure path before the drive, one band with frequency, bandwidth and level lanes; and an enhancer stage, a high-passed drive summed back, after the loop. ## Open What the VL70-m's enhancer is; its manual says "actually only used on a few voices".