Design: v1.1 video string-resolution for single-line#15
Merged
Conversation
Scopes the v1.1 milestone: use fretting-hand video to resolve the string that audio cannot, lifting single-line Tab F1 (~0.52 -> 0.94) and chord-instance accuracy (~0.48 -> 0.85). Key finding: the neck-anchor prior tiles its fret distribution across all strings (neck_prior.py:69), so it is string-agnostic; the string signal already exists in FrameFingering but is not wired into per-note resolution. Covers method, the eval-data gate (GuitarSet is audio-only), phased plan, acceptance test, and the SPEC 1.5 licensing check. Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
|
The latest updates on your projects. Learn more about Vercel for GitHub.
|
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Drafts the v1.1 milestone design: use fretting-hand video to resolve the string that audio cannot, lifting single-line Tab F1 (~0.52 -> 0.94) and chord-instance accuracy (~0.48 -> 0.85) -- the two numbers re-scoped from v1 to v1.1 in the 2026-06-03 acceptance.
Key finding grounding the design: the existing neck-anchor prior tiles its fret distribution across all strings (
neck_prior.py:69), so today''s video signal is string-agnostic -- it constrains the fret region, not the string. The string-discriminative signal already exists inFrameFingering(fingertip_to_fret) but is not wired into per-note resolution. v1.1 closes exactly that gap.Covers: method (per-note
FrameFingering-> candidate-restricted, confidence-gated string prior; the audio path never regresses), the eval-data gate (GuitarSet is audio-only, so v1.1 needs a video+string corpus -- synthetic-from-GuitarSet to validate the method, a license-clean public set as the gate), the phased plan, the acceptance test, and the SPEC 1.5 licensing check.Doc:
docs/plans/2026-06-03-v1.1-video-string-resolution-design.md. Design-only; no code.Generated with Claude Code