Summary
Todd Fisher of Philo Ventures presents a passion project that bridges electric guitar performance and AI-generated speech: a custom Logic Pro plugin that intercepts guitar notes and uses them as triggers to play back synthesized words and phrases, effectively making the guitar “speak.” The talk frames the project within a lineage of guitar innovation — from electrification to Peter Frampton’s talk box to modern DAW software — positioning AI as the next evolutionary step.
The engineering journey covers several iterations. Fisher started with Apple’s built-in TTS and the open-source Piper engine to generate audio clips, then tackled the harder problem of slicing continuous speech into per-word or per-syllable segments for real-time playback. He walks through two techniques he explored: energy-gap segmentation (cutting on near-silence between words) and sonority-peak syllabification (detecting vowel peaks to identify syllable boundaries), and explains where each breaks down with natural speech cadence.
Beyond the technical details, the talk serves as an encouragement for engineers to revisit their project backlogs now that AI tools lower the barrier to shipping unusual ideas quickly. Fisher demonstrates the plugin live in Logic Pro, triggering recognizable phrases note by note — a playful but technically substantive example of combining text-to-speech, digital signal processing, and DAW plugin architecture into a single creative instrument.
📺 Source: AI Engineer · Published August 18, 2026
🏷️ Format: Showcase
