Repository navigation
Self-contained mic vs system-audio skew probe - #6
Merged
Merged
Conversation
…s both channels Co-Authored-By: Claude Sonnet 5.5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01Pop6MMPof79Dd7oPc93QUj
…tputs Co-Authored-By: Claude Sonnet 5.5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01Pop6MMPof79Dd7oPc93QUj
Its base, the old work branch, was merged into main (PR #3). Merged cleanly; the probe tool builds with 0 warnings and all 304 Core tests pass. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
CakeRepository
changed the base branch from
claude/project-thread-gmksub
to
main
September 30, 2026 05:00
CakeRepository
marked this pull request as ready for review
September 30, 2026 05:00
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
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
Before: measuring the skew needed a recording where the microphone clearly heard the speakers; two Windows recordings had correlation 0.05 and 0.23, so the numbers meant nothing.
After:
ChannelSkew --probeopens the app's combined lease, plays a known noise-burst train through the default output, and finds it in each channel with a matched filter, so no app UI or loud recording is needed. It prints each side's correlation and says when a result is unusable.How:
ReferenceLocator(FFT cross-correlation, unit tested with faint and noisy signals),Probe.cs, and aTonePlayer(WASAPI on Windows, paplay elsewhere). Linux null-sink check: +14 ms. Not run on Windows.🤖 Generated with Claude Code
https://claude.ai/code/session_01Pop6MMPof79Dd7oPc93QUj
Generated by Claude Code