diff options
Diffstat (limited to 'CLAUDE.md')
| -rw-r--r-- | CLAUDE.md | 29 |
1 files changed, 26 insertions, 3 deletions
@@ -127,9 +127,27 @@ downloaded at once, Facebook videos are supported, the library ranks shows by th and opt-in skipping of breaks the publisher marked `toc: false`. Nothing is guessed from chapter titles. -**Tier 3 skip detection — audio fingerprinting of theme music — is still not built.** Peter -deferred it; the design plan puts it in Phase 3 and wants a desktop prototype against real -episodes before any Android code. +**Tier 3 skip detection — audio fingerprinting — is built, in two parts.** Both were desktop- +prototyped against real episodes before any Android code, per the design plan's own instruction. + +- **Intro/outro** (`IntroOutroDetector`, `AudioCorrelation`, `MediaCodecClipSource`, + `FingerprintUpdater`): one fixed candidate per show per end, needing 3+ independently-fetched + episodes to agree before it is trusted -- the confirmation threshold exists because two episodes + fetched close together can agree on a segment that is a dynamically-inserted ad, not the show's + real intro (found prototyping against Darknet Diaries). Triggered per download via + `FingerprintWorker`. +- **Midroll ad breaks** (`MidrollScanner`, `MidrollDetector`, `MidrollUpdater`): a show can have + several distinct midroll breaks, at a different timestamp in every episode, so this scans a + whole episode body rather than a fixed head/tail window -- `MidrollScanner.findSharedSegments` + finds every segment two episodes share, at whatever position each happens to have it, and + `MidrollUpdater` composes a coarse (decimated) pass with a native-resolution verify so that scan + is fast enough to run on a phone. Prototyping this against Darknet Diaries found a real recurring + segment (not the fixed intro) independently present across episodes nine years apart. Runs + nightly while charging (`MidrollWorker`), not per-download: a full-body scan is too expensive for + that, and unlike a download it never touches the network. No per-show toggle, same as + intro/outro -- the confirmation threshold is the safety mechanism. + +Neither has been verified on a device yet -- see the known-gaps list below. **ID3 `CHAP` frames are dropped, not deferred** — decided 2026-09-26 by measuring Peter's actual 61 subscriptions rather than guessing. Every feed's newest episode was fetched far enough to read @@ -161,6 +179,11 @@ by default and correct; it is simply waiting for a publisher who marks breaks. - **Android Auto is unverified in a real car.** The session protocol is proven by `MediaBrowseTest`; whether Auto's own compatibility checks pass on a head unit is untested, and a sideloaded media app may need Auto's developer mode enabled. +- **Neither half of Tier 3 fingerprinting has run on a device.** The decode/decimate/correlate + pipeline is proven against synthetic fixtures and, separately, against real downloaded episodes + in a throwaway desktop prototype -- never against `MediaCodecClipSource`'s actual + MediaExtractor/MediaCodec decode loop on real hardware, and the `MigrationTest` 7→8 case has + only run on the JVM-backed `MigrationTestHelper`, not an emulator or device. ## Domain notes worth keeping |