PyPhonPlan: Simulating phonetic planning with dynamic neural fields and task dynamics

Sam Kirkham

cs.CLMar 17, 2026

PyPhonPlan: Simulating phonetic planning with dynamic neural fields and task dynamics

Authors: Sam Kirkham

Organizations: Phonetics Laboratory, Lancaster University, United Kingdom

Abstract

We introduce PyPhonPlan, a Python toolkit for implementing dynamical models of phonetic planning using coupled dynamic neural fields and task dynamic simulations. The toolkit provides modular components for defining planning, perception and memory fields, as well as between-field coupling, gestural inputs, and using field activation profiles to solve tract variable trajectories. We illustrate the toolkit's capabilities through an example application: simulating production/perception loops with a coupled memory field, which demonstrates the framework's ability to model interactive speech dynamics using representations that are temporally-principled, neurally-grounded, and phonetically-rich. PyPhonPlan is released as open-source software and contains executable examples to promote reproducibility, extensibility, and cumulative computational development for speech communication research.

Figures & tables

Figure 1: Heatmap of the planning-layer of a DNF with two competing inputs at $x=-5$ and $x=+5$ . The red line indicates the start of above-threshold activation; the dotted white line tracks the location of above-threshold peak activation.

Figure 2: Memory field coupled to the planning field in Figure 1 . The memory trace reflects a history of above-threshold planning activation, with a broader spatial profile due to convolution with

w(x-x^{\prime})

Figure 3: Shadowing simulation results. Left: planning peak position across baseline with no perceptual input (BL), 10 shadowing trials with strong perceptual input (S1–S10), and washout with no perceptual input (WO). Right: tract variable trajectories for baseline versus washout, showing subtle articulatory consequences of memory-driven convergence.

Computational Psycholinguistics

Computational psycholinguistics models how people learn, produce, and understand language using formal theories, algorithms, and behavioral data.

See all 89 papers

VoxReason: Auditing Source-Grounded Speech Plans Before Synthesis
Sep 2, 2026Mengzhe GengSpeech Generation Evaluation Controllable Speech Generation
National Research Council Canada
Physics-Informed Neural Operator for Speech Production Analysis
Jun 21, 2026Kazuya Yokota, Xinmeng Luan, Debasish Ray Mohapatra +2Speech Processing PDE Operator Learning
Department of Mechanical Engineering, Nagaoka University of Technology, Japan · Schulich School of Music, McGill University, Canada · Department of Electrical and Computer Engineering, University of British Columbia, Canada
Vocal-Tract Acoustics, Aerodynamics, and Overlap-and-Add Synthesis in DYNARTmo
Oct 17, 2026Bernd J. KrögerControllable Speech Generation TTS Synthesis
Medical School, RWTH Aachen University, Aachen, Germany · Kröger Lab, Belgium, www.speechtrainer.eu

Sep 2, 2026J/K move · Enter open · S save

VoxReason: Auditing Source-Grounded Speech Plans Before Synthesis

Mengzhe Geng

National Research Council Canada

Speech Generation Evaluation Controllable Speech Generation

Open paper AskPDF arXiv

Plan accuracy alone cannot show whether a speech-delivery decision follows its source: a fixed prior may match the original label yet fail to respond appropriately when a cue changes. VoxReason provides a 100-case verifier benchmark that holds each utterance fixed, edits one designated source-label cue, and scores cited evidence, eight plan fields, and the permitted response. On a source-key-disjoint test of 24 cases, a source-emotion prior reaches plan-slot accuracy 0.958, but none of the 24 edited neutral targets appears in its training labels; its required-change accuracy is 0.000. This diagnoses the support boundary of this prior, not its performance on supported edits. In a complementary 32-case emotion-disjoint test, the prior has seen all edited neutral targets but neither original test emotion; its plan-slot accuracy is 0.219 and required-change accuracy is 1.000. The partitions reuse and overlap the same 100 cases, so these deterministic diagnostics are not independent cohorts or learned-planner results. The benchmark evaluates derived labels and structured plans, not audio input, generated speech, or listener judgments.

Abstract

Computational Psycholinguistics

Explore similar work

VoxReason: Auditing Source-Grounded Speech Plans Before Synthesis

Physics-Informed Neural Operator for Speech Production Analysis

Vocal-Tract Acoustics, Aerodynamics, and Overlap-and-Add Synthesis in DYNARTmo