When a Warning Is Not the Root Cause: The I2S Mode1 Conflict
An I2S Mode1 conflict warning looked suspicious enough to become a candidate explanation for crackle.
evidence
Why PCAP Evidence Cannot Prove Acoustic Latency
It was tempting to use server packet timing as proof of the complete mouth-to-ear delay.
The Measurement Stack That Made Subjective Audio Debuggable
Words like robotic, delayed and crackly were useful user reports but poor root-cause evidence.
The Audio Bug Was Not One Bug
Robotic speech, cutouts, lag and echo initially collapsed into one vague complaint called bad audio.
How I Separate Confirmed Facts, Inferences and Failed Experiments
Fast-moving firmware work made it easy for an attractive theory to become remembered as fact.
Similar Patterns Do Not Prove the Same Process
When similar patterns appear at very different scales, we naturally look for a connection. That instinct can begin an investigation, but visual similarity does not prove that the same mechanism produced both structures.
Documentation Should Preserve the Path From Claim to Evidence
Writing that a system works is different from documenting how that conclusion was reached. Without evidence, the next person cannot reproduce the claim. Conditions, tests, and limits make the result traceable.
A Green Badge Needs an Expiration Time
Operational evidence should age out automatically rather than remaining green until someone notices it is old.
Evidence Freshness Belongs in the Operator UI
A production acceptance result should carry age and policy, not just a green badge.
Reconstructing V115 as a Known-Good Behavioral Control
Later experiments had changed queues, PLC, timing and logs until nobody could safely say which behavior belonged to the last clear build.