@torrent-tv/proxy 2.89.1 → 2.89.2

package/services/encode/OutputTimes.js

@@ -1,579 +1,579 @@
-/**
- * Where each segment of an output begins in the film, and how the cut table is
- * corrected from what the encoder actually produced.
- *
- * The published grid is what the playlist told the player; a run's landing and
- * the true start of each piece are what ffmpeg did. The two are compared here,
- * a disagreement is said, and a boundary the index got wrong is corrected once
- * for every output of the file.
- */
-
-import { segmentIndexForTime } from "./output/playlists.js";
-import { publishedGridFor as publishedGridOf, publishedStartTime } from "./run-command.js";
-import { liveRunsOf } from "./encode-run-state.js";
-// How many produced segments' true start times to remember, so a player's
-// report about one of them can be answered. Two hundred is about twenty
-// minutes of playback at these segment lengths — far more than the recent past
-// a stall report can be about, and small enough to be free.
-const TRUE_START_MEMORY = 200;
-// How far a segment's own recorded start may sit from the one the playlist
-// assigned it before the disagreement is worth a log line. The two are built
-// from the same keyframe index and normally match to the sample; a quarter of a
-// second is below any drift a viewer could notice, so anything above it is the
-// index being wrong about where a keyframe is rather than rounding.
-const SEGMENT_START_DISAGREEMENT_SEC = 0.25;
-/**
- * How far a fragment may land from where the playlist put it before the player
- * stops recognising it as buffered.
- *
- * hls.js's own `maxBufferHole`, whose default is 0.5 s: a gap smaller than this
- * is skipped, a gap larger is a hole, and a fragment appended across one is
- * judged not to have loaded — so the player asks for it again, and again.
- * Taken from the player's published default rather than chosen here.
- */
-export const PLAYER_BUFFER_HOLE_SEC = 0.5;
-/**
- * A live run of this session that begins exactly at this number, if there is
- * one.
- *
- * @param {object[] | Set<object>} runs
- * @param {number} index
- * @returns {object | null}
- */
-function runStartingAt(runs, index) {
-  return liveRunsOf(runs).find((run) => run.from === index) ?? null;
-}
-
-/**
- * How far the live boundary table has moved from the one the player holds, said
- * in words.
- *
- * The corrections are applied one boundary at a time and each is small enough
- * to look harmless; what nobody was watching is the total. It matters because a
- * run positioned on one table and cut on the other carries their difference into
- * every cut it makes — the fault of 2026-08-21, where the distance reached two
- * whole segments after one seek and four after the next. Printed beside each
- * correction so the total is visible while it is still small.
- *
- * @param {number[]} published - The table the playlist text was written from.
- * @param {number[]} live - The table corrected from produced segments.
- * @returns {string} A phrase, always readable, never throwing on odd input.
- */
-export function describeGridDrift(published, live) {
-  if (!Array.isArray(published) || !Array.isArray(live) || published.length === 0) {
-    return "not comparable";
-  }
-  if (published.length !== live.length) {
-    return `a different length (${published.length} against ${live.length})`;
-  }
-  let apart = 0;
-  let worst = 0;
-  let worstAt = -1;
-  for (let index = 0; index < published.length; index += 1) {
-    // Rounded to the millisecond BEFORE comparing, not only before printing.
-    // Two boundaries moved by the same amount differ in the last bits of a
-    // double, so an unrounded comparison picks between them by an accident
-    // invisible in the printed figure — and the line would name a boundary the
-    // reader cannot tell apart from the one before it. Rounded, ties keep the
-    // earliest, which is also the one worth looking at first.
-    const distance = Math.round(Math.abs(live[index] - published[index]) * 1000) / 1000;
-    if (distance <= 0.001) {
-      continue;
-    }
-    apart += 1;
-    if (distance > worst) {
-      worst = distance;
-      worstAt = index;
-    }
-  }
-  if (apart === 0) {
-    return "identical";
-  }
-  return `${apart} of ${published.length} boundaries apart, worst ${worst.toFixed(3)}s at #${worstAt}`;
-}
-
-export class OutputTimes {
-  #state = new WeakMap();
-  /** What this reads and asks of the rest of the proxy, and nothing else. @type {object} */
-  #host;
-
-  /**
-   * @param {object} host - `logger`, `runsOf`, `stopEncodeRun`, `planEncodersSoon`, `outputs`, `segmentDurationSec`
-   */
-  constructor(host) {
-    this.#host = host;
-  }
-
-  #stateFor(output) {
-    let state = this.#state.get(output);
-    if (!state) {
-      state = {
-        stampWarnedAt: new Map(),
-        landingReportedForRun: undefined,
-        trueStartByIndex: new Map(),
-        deviationWarnedAt: new Map()
-      };
-      this.#state.set(output, state);
-    }
-    return state;
-  }
-
-  trueStartAt(output, index) {
-    return this.#stateFor(output).trueStartByIndex.get(index);
-  }
-
-  /**
-   * Start time (seconds, 0-based) of segment `index`, from the session's
-   * boundary table. Clamped to valid range.
-   *
-   * @param {HlsSession} session
-   * @param {number} index
-   * @returns {number}
-   */
-  segmentStartTime(session, index) {
-    // The LIVE table, deliberately: this is where the file is cut now, which is
-    // a different question from what the player was told. The published table
-    // is passed as absent so that one arithmetic serves both.
-    return publishedStartTime({ boundaries: session.timeline.boundaries }, index, this.#host.segmentDurationSec);
-  }
-
-  /**
-   * The boundary table the player is working from: the one its playlist was
-   * written from, falling back to the live table when no playlist was built
-   * from a table at all (no duration, so no synthetic playlist — and then
-   * nothing the player holds contradicts it).
-   *
-   * @param {HlsSession} session
-   * @returns {number[]}
-   */
-  publishedGridFor(session) {
-    return publishedGridOf(session.timeline);
-  }
-
-  /**
-   * Where a run beginning at `index` must be positioned: the time the PLAYER
-   * was told that segment starts at.
-   *
-   * Public because it is the invariant this class has broken twice, and a
-   * private one cannot be pinned by a test. It must always be the table the cut
-   * list is taken from — see the comment where a run is started.
-   *
-   * @param {HlsSession} session
-   * @param {number} index
-   * @returns {number}
-   */
-  runStartTimeFor(session, index) {
-    return this.publishedStartTime(session, index);
-  }
-
-  publishedStartTime(session, index) {
-    return publishedStartTime(session.timeline, index, this.#host.segmentDurationSec);
-  }
-
-  /**
-   * Report a segment whose own timeline disagrees with the playlist by more
-   * than a player will bridge.
-   *
-   * Once per segment per five seconds, like every other repeating condition
-   * here: the same segment is requested again and again while it is refused,
-   * and a line each time buries the first one.
-   *
-   * @param {HlsSession} session
-   * @param {number} index
-   * @param {number} trueStart
-   * @param {number} publishedStart
-   * @returns {void}
-   */
-  notePlaylistDisagreement(session, index, trueStart, publishedStart) {
-    // Within what the player bridges it is no disagreement worth a line.
-    if (!(Math.abs(trueStart - publishedStart) > PLAYER_BUFFER_HOLE_SEC)) {
-      return;
-    }
-    const now = Date.now();
-    const state = this.#stateFor(session);
-    if (now - (state.stampWarnedAt.get(index) ?? 0) < 5_000) {
-      return;
-    }
-    state.stampWarnedAt.set(index, now);
-    this.#host.logger.warn(
-      `transcode ${session.id} segment #${index} carries ${trueStart.toFixed(3)}s while the playlist ` +
-      `the player holds says ${publishedStart.toFixed(3)}s — a gap of ` +
-      `${Math.abs(trueStart - publishedStart).toFixed(3)}s, beyond the ${PLAYER_BUFFER_HOLE_SEC}s a player ` +
-      "bridges; stamping it where the playlist says so the fragment lands where it was asked for"
-    );
-  }
-
-  /**
-   * Segment index whose span contains time `t` (0-based), via the boundary
-   * table.
-   *
-   * @param {HlsSession} session
-   * @param {number} t
-   * @returns {number}
-   */
-  segmentIndexForTime(session, t) {
-    // The player's grid, for the same reason the cut list uses it: the time
-    // being resolved came from the playlist the player holds, so the index it
-    // means is the index that playlist gives it.
-    return segmentIndexForTime(this.publishedGridFor(session), t, this.#host.segmentDurationSec);
-  }
-
-  /**
-   * Where the run REALLY began, against where it was asked to begin.
-   *
-   * The first piece a run produces is the only statement of this that exists,
-   * and until now nothing compared the two. They disagree whenever the seek
-   * lands somewhere other than the time asked for — which, before the landing
-   * offset, was every run on a Matroska source with B-frames, by exactly one
-   * keyframe interval. Every cut of the run then inherits it, because
-   * `-segment_times` is measured from the landing.
-   *
-   * Said once per run, and only when it matters: within what a player bridges
-   * there is nothing to report.
-   *
-   * @param {HlsSession} session
-   * @param {number} index
-   * @param {number} trueStart
-   * @returns {void}
-   */
-  noteRunLanding(session, index, trueStart) {
-    const state = this.#stateFor(session);
-    if (runStartingAt(this.#host.runsOf(session), index) === null || state.landingReportedForRun === index) {
-      return;
-    }
-    state.landingReportedForRun = index;
-    // What the run was ASKED for, taken from the run itself rather than looked
-    // up again in a table. The two used to be the same lookup; they stopped
-    // being so when a run began positioning on the published grid while this
-    // read the live one, which made a perfect landing report a drift equal to
-    // the distance between the tables — and cancelled a real landing error of
-    // the same size to zero. A run also has one legitimate position that is in
-    // no table at all: the realignment that starts the sound where the copied
-    // picture truly begins.
-    const asked = runStartingAt(this.#host.runsOf(session), index)?.progress?.startPositionSeconds ??
-      this.runStartTimeFor(session, index);
-    const drift = trueStart - asked;
-    if (!Number.isFinite(drift) || Math.abs(drift) <= PLAYER_BUFFER_HOLE_SEC) {
-      return;
-    }
-    this.#host.logger.warn(
-      `transcode ${session.id} run began at ${trueStart.toFixed(3)}s but was asked for ` +
-      `${asked.toFixed(3)}s — ${drift > 0 ? "+" : ""}${drift.toFixed(3)}s, and every cut of this ` +
-      "run is measured from where it began, so the whole run is that far from its playlist"
-    );
-  }
-
-  noteIndexAccuracy(session, index, trueStart, declaredStart) {
-    const deviation = Math.abs(trueStart - declaredStart);
-    // Where each produced segment truly began, kept so that a player reporting
-    // a stall can be ANSWERED rather than merely believed. Bounded: only the
-    // recent past can be the subject of such a report, and an unbounded map on
-    // a two-hour film is a leak.
-    const state = this.#stateFor(session);
-    state.trueStartByIndex.set(index, trueStart);
-    if (state.trueStartByIndex.size > TRUE_START_MEMORY) {
-      const oldest = state.trueStartByIndex.keys().next();
-      if (!oldest.done) {
-        state.trueStartByIndex.delete(oldest.value);
-      }
-    }
-    // ONE READING, AND WHOSE FACT IT IS DEPENDS ON HOW THIS OUTPUT IS MADE. The
-    // three branches below already say the distinction when they WARN; until
-    // 2026-09-15 what they COUNTED threw it away, into a tally held per (file,
-    // grid) that a picture and the soundtrack inside the same file share.
-    //
-    // Only a COPY can say anything about the file: it is cut at a keyframe of
-    // that file and nowhere else. So a copy's landing goes to the file's own
-    // table, where it outlives this grid and reaches every step and every later
-    // session of the file.
-    if (session.spec.carries !== "audio-only" && !session.spec.transcodesVideo) {
-      session.keyframes?.witness({
-        index,
-        trueStart,
-        deviationSec: deviation,
-        toleranceSec: SEGMENT_START_DISAGREEMENT_SEC
-      });
-    }
-    // And every output records its own landing against its own published grid,
-    // which is what says whether a step will splice and how far a soundtrack
-    // stands from the picture.
-    session.output?.noteLanding({ index, deviationSec: deviation, toleranceSec: SEGMENT_START_DISAGREEMENT_SEC });
-    if (deviation > SEGMENT_START_DISAGREEMENT_SEC) {
-      // Which boundary the true start DOES match, if any. This is what tells
-      // the two possible faults apart, and they need opposite fixes: matching
-      // boundary #N-1 means our numbering is shifted by one — a fault in this
-      // code, where the run begins — while matching nothing means the container
-      // index describes times the file does not have. Measured 2026-08-11,
-      // three samples all matched N-1, which is why the line now says so
-      // instead of leaving it to be inferred from the numbers.
-      const at = this.boundaryIndexAt(session, trueStart);
-      // Once per segment per five seconds, like `#notePlaylistDisagreement`
-      // beside it. The same segment is produced and served again and again
-      // while it is refused, and — since a run keeps cutting on the list it was
-      // launched with — a soundtrack whose grid has moved under it deviates on
-      // EVERY segment for the life of that run. A line each time buries the
-      // first one, which is the one somebody is reading the log for.
-      const state = this.#stateFor(session);
-      const lastWarnedAt = state.deviationWarnedAt.get(index) ?? 0;
-      if (Date.now() - lastWarnedAt >= 5_000) {
-        state.deviationWarnedAt.set(index, Date.now());
-        this.#host.logger.warn(
-        `transcode ${session.id} segment #${index} really starts at ` +
-        `${trueStart.toFixed(3)}s (boundary ${at === null ? "none" : `#${at}`}), ` +
-        `the grid says ${declaredStart.toFixed(3)}s — ` +
-        (session.spec.carries === "audio-only"
-          // A soundtrack is cut exactly where it was asked to be, so a
-          // disagreement here is not a reading about the file at all: it is the
-          // distance between this run's own cuts and a grid the picture has
-          // since corrected under it. Said plainly, because the same sentence
-          // used to claim a keyframe index was wrong when no keyframe was
-          // involved on this side of the stream.
-          ? "sound is cut where it is asked to be; this is the picture's grid having moved, not the index"
-          : session.spec.transcodesVideo
-            // A re-encode was TOLD to put a keyframe here and did not, so this
-            // rung's segments no longer stand where the stream it accompanies
-            // would have put them. That is a broken splice, not a wrong index.
-            ? "this rung did not cut where its grid says; a switch to it will not join cleanly"
-            // A copy. The two possible faults need opposite fixes and the
-            // numbers already tell them apart, so the sentence follows THEM
-            // rather than the branch it is printed from. Until 2026-08-21 it
-            // blamed the index either way — including through a session whose
-            // segments each held the boundary two, then four places before
-            // their own number, which is this code's own definition of a fault
-            // in this code. That sentence is what sent the reading of that
-            // session after the file instead of after the arithmetic.
-            : at === null
-              ? "the container's keyframe index disagrees with the file; using the file"
-              : `this began at another boundary of the same list, ${index - at} place(s) before ` +
-                "its own number — the numbering of this run is shifted, not the index")
-        );
-      }
-    }
-    // Said as the evidence accumulates, not only when the session is disposed.
-    // A proxy restart takes its sessions with it — every addon update does —
-    // and a summary that only ever appears at the end is a summary that is
-    // routinely never written. Twenty-five distinct boundaries is enough for
-    // the proportion to mean something and rare enough not to repeat itself.
-    const counted = session.output?.piecesLanded ?? 0;
-    if (counted > 0 && counted % 25 === 0) {
-      this.logIndexAccuracy(session);
-    }
-    this.correctBoundaryFromSegment(session, index, trueStart);
-  }
-
-  /**
-   * Replace a boundary the index got wrong with the time the file actually has.
-   *
-   * The grid of a copied stream comes from the container's keyframe index,
-   * because a copy can only be cut where a keyframe already is and nothing
-   * cheaper than the index can say where that is before a single byte is
-   * encoded. An index can be wrong — proven 2026-08-12 by reproducing both
-   * cases against the same file: with an honest index every produced segment
-   * started exactly where declared, and with one moved 1.8 s the segments
-   * started 1.8 s early, matching no boundary at all. The field showed the
-   * second shape.
-   *
-   * The truth arrives anyway, one segment at a time: a produced piece states
-   * where it really begins. Writing it back makes the grid describe the file
-   * instead of the index — and it is what lets a re-encoded rung be cut to
-   * match a copied one, because the rung is then forced onto times the copy
-   * really uses. The alternative considered and rejected was to stop offering
-   * quality on files with a bad index, which is not a fix but a withdrawal.
-   *
-   * The whole family shares one grid, so a correction reaches all of it: a rung
-   * created afterwards inherits a table that is true wherever anyone has looked.
-   *
-   * @param {HlsSession} session
-   * @param {number} index
-   * @param {number} trueStart
-   * @returns {void}
-   */
-  correctBoundaryFromSegment(session, index, trueStart) {
-    // Only the picture may move the grid, because the grid IS the picture's
-    // cut list: it is built from the container's keyframe index, and a copied
-    // stream can be cut nowhere else. A soundtrack has no keyframes — it is cut
-    // exactly where `-segment_times` asks, to within one audio frame — so its
-    // reading measures nothing about the grid and everything about itself.
-    //
-    // Writing it into the shared table is how one film ended up with two
-    // answers for one boundary, each side correctly describing its own stream
-    // and each overwriting the other: field 2026-08-20, segment #521 of
-    // "Minions.and.Monsters.1080p.mkv" corrected 2086.084s → 2084.082s by the
-    // picture at 11:14:16.939 and 2084.082s → 2086.033s by the sound 1.6s
-    // later — 1.951s apart, against the 0.25s that stops a correction and the
-    // 0.5s a player bridges. The next reading disagrees with the table again,
-    // so it never converges and never stops.
-    if (session.spec.carries === "audio-only") {
-      return;
-    }
-    const boundaries = session.timeline.boundaries;
-    if (!Array.isArray(boundaries) || index <= 0 || index >= boundaries.length - 1) {
-      // Index 0 is the start of the file and the last entry is its end; neither
-      // is a cut, and neither can be learned from a segment.
-      return;
-    }
-    if (Math.abs(boundaries[index] - trueStart) <= SEGMENT_START_DISAGREEMENT_SEC) {
-      return;
-    }
-    // A correction that would put this boundary at or past its neighbours is not
-    // a correction — it is a reading from a run that started somewhere else, and
-    // applying it would make the table describe nothing at all.
-    if (trueStart <= boundaries[index - 1] || trueStart >= boundaries[index + 1]) {
-      return;
-    }
-    const wasAt = boundaries[index];
-    // One write, because there is one table. Where a file is cut is a fact
-    // about the FILE, so every session of it holds the same array rather than a
-    // copy of it — which is what this loop used to keep in step, member by
-    // member, and only for members that happened to exist at the time. A
-    // session created afterwards used to inherit a copy taken at that moment;
-    // now it is handed the table itself.
-    boundaries[index] = trueStart;
-    this.#host.logger.info(
-      `transcode ${session.id} boundary #${index} corrected ${wasAt.toFixed(3)}s → ` +
-      `${trueStart.toFixed(3)}s from the file itself` +
-      `, and the live table is now ${describeGridDrift(this.publishedGridFor(session), boundaries)}` +
-      " from the one the player holds"
-    );
-    // And every OTHER member whose run begins at this very boundary is moved
-    // to the same instant.
-    //
-    // Why they were not there already: the two branches are asked for the same
-    // time and land in different places. The picture cannot begin anywhere but
-    // a real keyframe, and it may not begin before the time asked for — that
-    // content belongs to the previous segment — so it moves FORWARD to the next
-    // one, by up to the keyframe spacing (0.58-2.96 s measured 2026-08-17).
-    // A soundtrack has no keyframes: it begins exactly where asked, to within
-    // one audio frame. So after every restart the two runs of one film began up
-    // to three seconds apart, each correctly labelled with where it really was,
-    // and the viewer got sound with no new picture for the difference.
-    //
-    // The picture's true start is a MEASURED quantity — read from the piece it
-    // just produced, which is what the correction above is — so the soundtrack
-    // can be put exactly there instead of at the time the container's table
-    // claimed. It converges: once the boundary holds the true time, the next
-    // reading agrees with it and the guard above returns before doing anything.
-    for (const member of this.#host.outputs.familyOf(session)) {
-      if (member === session || runStartingAt(this.#host.runsOf(member), index) === null) {
-        continue;
-      }
-      this.#host.logger.info(
-        `transcode ${member.id} begins at #${index}, which really starts ` +
-        `${(trueStart - wasAt).toFixed(3)}s later than the table said — restarting it there ` +
-        `so picture and sound begin together`
-      );
-      // STOPPED, and its replacement is the plan's to place. What has been
-      // learned here is that this run is producing at the wrong instant, which
-      // nothing but the piece it produced could say — the plan reasons about
-      // numbers and cannot know it. So the fact is acted on where it is known,
-      // and only as far as it is known: the run that is wrong goes, the stretch
-      // it held returns to the map, and where the next one stands follows from
-      // where the viewers are.
-      //
-      // The corrected instant reaches that replacement through the live table,
-      // which this function has just written it into and which every session of
-      // the file shares. It used to be passed as an argument from here, so only
-      // a run started by this line ever had it.
-      this.#host.stopEncodeRun(member, `#${index} really begins ${(trueStart - wasAt).toFixed(3)}s later than the table said`);
-    }
-    this.#host.planEncodersSoon();
-  }
-
-  /**
-   * The boundary a time falls on, or null when it falls on none of them.
-   *
-   * Within the same tolerance a disagreement is judged by, so "matches boundary
-   * #N-1" and "matches nothing" mean what they say.
-   *
-   * @param {HlsSession} session
-   * @param {number} seconds
-   * @returns {number | null}
-   */
-  boundaryIndexAt(session, seconds, table) {
-    // The table is nameable because the two answer different questions. The
-    // LIVE one says "is this a cut this file actually has", which is what a
-    // reading taken off a produced segment is about. The PUBLISHED one says "is
-    // this a cut the player believes in", which is what a report from the player
-    // is about. Answering one with the other prints an index from one grid
-    // beside a time from the other.
-    const boundaries = Array.isArray(table) ? table : session.timeline.boundaries;
-    if (!Array.isArray(boundaries)) {
-      return null;
-    }
-    for (let index = 0; index < boundaries.length; index += 1) {
-      if (Math.abs(boundaries[index] - seconds) <= SEGMENT_START_DISAGREEMENT_SEC) {
-        return index;
-      }
-    }
-    return null;
-  }
-
-  /**
-   * Where this output's pieces have been landing, and — where this output is a
-   * copy — what that has shown about the FILE's own keyframe table.
-   *
-   * TWO LINES, because they are two facts with two owners and two lifetimes.
-   * The landing is this output's and dies with it. The table's accuracy belongs
-   * to the file, is learned only from a copy, and is the same finding for every
-   * step and every later session of that file — so it names the file and not
-   * the session, and it is written only by the output that can witness it.
-   *
-   * Written even when nothing disagreed, because that is the finding: with only
-   * the per-piece warning, silence could not be told from nobody having
-   * watched. Skipped where nothing was produced, which says neither.
-   *
-   * @param {HlsSession} session
-   * @returns {void}
-   */
-  logIndexAccuracy(session) {
-    const copiedPicture = session.spec.carries !== "audio-only" && !session.spec.transcodesVideo;
-    const landing = session.output?.landing ?? null;
-    if (landing) {
-      this.#host.logger.info(
-        // The session id, because without it this line cannot be attributed. A
-        // family produces one of these per member — the picture, each step,
-        // each soundtrack — and on 2026-08-17 the picture's was read as the
-        // sound's, from a neighbouring log line, and a roadmap item was written
-        // against the wrong half of the stream.
-        `landing ${session.id.slice(0, 8)} ` +
-        `${session.spec.carries === "audio-only" ? "sound" : copiedPicture ? "copied picture" : "re-encoded picture"} ` +
-        `"${session.file.name}": ${landing.disagreed} of ${landing.checked} produced pieces started ` +
-        `away from this output's own playlist, median ${landing.medianDeviationSec.toFixed(3)}s ` +
-        `worst ${landing.maxDeviationSec.toFixed(3)}s` +
-        (landing.firstDisagreementIndex >= 0 ? ` (first at #${landing.firstDisagreementIndex})` : "") +
-        ` [tolerance ${SEGMENT_START_DISAGREEMENT_SEC}s] — ` +
-        (session.spec.carries === "audio-only"
-          // A soundtrack is cut exactly where it is asked to be, so this is not
-          // a reading about any keyframe: it is how far this run's cuts stand
-          // from a grid the picture has corrected under it.
-          ? "sound is cut where it is asked to be, so this is the picture's grid having moved"
-          : copiedPicture
-            // A copy cannot be cut anywhere but a real keyframe, so this figure
-            // is also evidence about the file, reported on the line below.
-            ? "a copy is cut at the file's own keyframes, so this is also evidence about its table"
-            // A re-encode was TOLD to put a keyframe at each of these instants.
-            : "this rung was told where to cut; what it missed will not splice cleanly")
-      );
-    }
-    if (!copiedPicture) {
-      return;
-    }
-    const evidence = session.keyframes?.evidence ?? null;
-    if (!evidence) {
-      return;
-    }
-    this.#host.logger.info(
-      // Named by the FILE, with no session id in it, because that is what it is
-      // about: the same answer for every step and every viewer of these bytes.
-      `keyframe-index ${session.keyframes?.format ?? "unknown"} "${session.file.name}": ` +
-      `${evidence.disagreed} of ${evidence.checked} copied pieces began away from the table, ` +
-      `median ${evidence.medianDeviationSec.toFixed(3)}s worst ${evidence.maxDeviationSec.toFixed(3)}s` +
-      (evidence.firstDisagreementIndex >= 0 ? ` (first at #${evidence.firstDisagreementIndex})` : "") +
-      // The discriminator, stated in the same line as the count it explains: a
-      // piece that began at another time the SAME table names was not
-      // mis-described by the table — the grid was built over a gap in it.
-      `; ${evidence.landedOnAnotherKeyframe} of them began at another keyframe the table names` +
-      ` [tolerance ${SEGMENT_START_DISAGREEMENT_SEC}s, ${session.keyframes?.count ?? 0} keyframes read]`
-    );
-  }
-}
+/**
+ * Where each segment of an output begins in the film, and how the cut table is
+ * corrected from what the encoder actually produced.
+ *
+ * The published grid is what the playlist told the player; a run's landing and
+ * the true start of each piece are what ffmpeg did. The two are compared here,
+ * a disagreement is said, and a boundary the index got wrong is corrected once
+ * for every output of the file.
+ */
+
+import { segmentIndexForTime } from "./output/playlists.js";
+import { publishedGridFor as publishedGridOf, publishedStartTime } from "./run-command.js";
+import { liveRunsOf } from "./encode-run-state.js";
+// How many produced segments' true start times to remember, so a player's
+// report about one of them can be answered. Two hundred is about twenty
+// minutes of playback at these segment lengths — far more than the recent past
+// a stall report can be about, and small enough to be free.
+const TRUE_START_MEMORY = 200;
+// How far a segment's own recorded start may sit from the one the playlist
+// assigned it before the disagreement is worth a log line. The two are built
+// from the same keyframe index and normally match to the sample; a quarter of a
+// second is below any drift a viewer could notice, so anything above it is the
+// index being wrong about where a keyframe is rather than rounding.
+const SEGMENT_START_DISAGREEMENT_SEC = 0.25;
+/**
+ * How far a fragment may land from where the playlist put it before the player
+ * stops recognising it as buffered.
+ *
+ * hls.js's own `maxBufferHole`, whose default is 0.5 s: a gap smaller than this
+ * is skipped, a gap larger is a hole, and a fragment appended across one is
+ * judged not to have loaded — so the player asks for it again, and again.
+ * Taken from the player's published default rather than chosen here.
+ */
+export const PLAYER_BUFFER_HOLE_SEC = 0.5;
+/**
+ * A live run of this session that begins exactly at this number, if there is
+ * one.
+ *
+ * @param {object[] | Set<object>} runs
+ * @param {number} index
+ * @returns {object | null}
+ */
+function runStartingAt(runs, index) {
+  return liveRunsOf(runs).find((run) => run.from === index) ?? null;
+}
+
+/**
+ * How far the live boundary table has moved from the one the player holds, said
+ * in words.
+ *
+ * The corrections are applied one boundary at a time and each is small enough
+ * to look harmless; what nobody was watching is the total. It matters because a
+ * run positioned on one table and cut on the other carries their difference into
+ * every cut it makes — the fault of 2026-08-21, where the distance reached two
+ * whole segments after one seek and four after the next. Printed beside each
+ * correction so the total is visible while it is still small.
+ *
+ * @param {number[]} published - The table the playlist text was written from.
+ * @param {number[]} live - The table corrected from produced segments.
+ * @returns {string} A phrase, always readable, never throwing on odd input.
+ */
+export function describeGridDrift(published, live) {
+  if (!Array.isArray(published) || !Array.isArray(live) || published.length === 0) {
+    return "not comparable";
+  }
+  if (published.length !== live.length) {
+    return `a different length (${published.length} against ${live.length})`;
+  }
+  let apart = 0;
+  let worst = 0;
+  let worstAt = -1;
+  for (let index = 0; index < published.length; index += 1) {
+    // Rounded to the millisecond BEFORE comparing, not only before printing.
+    // Two boundaries moved by the same amount differ in the last bits of a
+    // double, so an unrounded comparison picks between them by an accident
+    // invisible in the printed figure — and the line would name a boundary the
+    // reader cannot tell apart from the one before it. Rounded, ties keep the
+    // earliest, which is also the one worth looking at first.
+    const distance = Math.round(Math.abs(live[index] - published[index]) * 1000) / 1000;
+    if (distance <= 0.001) {
+      continue;
+    }
+    apart += 1;
+    if (distance > worst) {
+      worst = distance;
+      worstAt = index;
+    }
+  }
+  if (apart === 0) {
+    return "identical";
+  }
+  return `${apart} of ${published.length} boundaries apart, worst ${worst.toFixed(3)}s at #${worstAt}`;
+}
+
+export class OutputTimes {
+  #state = new WeakMap();
+  /** What this reads and asks of the rest of the proxy, and nothing else. @type {object} */
+  #host;
+
+  /**
+   * @param {object} host - `logger`, `runsOf`, `stopEncodeRun`, `planEncodersSoon`, `outputs`, `segmentDurationSec`
+   */
+  constructor(host) {
+    this.#host = host;
+  }
+
+  #stateFor(output) {
+    let state = this.#state.get(output);
+    if (!state) {
+      state = {
+        stampWarnedAt: new Map(),
+        landingReportedForRun: undefined,
+        trueStartByIndex: new Map(),
+        deviationWarnedAt: new Map()
+      };
+      this.#state.set(output, state);
+    }
+    return state;
+  }
+
+  trueStartAt(output, index) {
+    return this.#stateFor(output).trueStartByIndex.get(index);
+  }
+
+  /**
+   * Start time (seconds, 0-based) of segment `index`, from the session's
+   * boundary table. Clamped to valid range.
+   *
+   * @param {HlsSession} session
+   * @param {number} index
+   * @returns {number}
+   */
+  segmentStartTime(session, index) {
+    // The LIVE table, deliberately: this is where the file is cut now, which is
+    // a different question from what the player was told. The published table
+    // is passed as absent so that one arithmetic serves both.
+    return publishedStartTime({ boundaries: session.timeline.boundaries }, index, this.#host.segmentDurationSec);
+  }
+
+  /**
+   * The boundary table the player is working from: the one its playlist was
+   * written from, falling back to the live table when no playlist was built
+   * from a table at all (no duration, so no synthetic playlist — and then
+   * nothing the player holds contradicts it).
+   *
+   * @param {HlsSession} session
+   * @returns {number[]}
+   */
+  publishedGridFor(session) {
+    return publishedGridOf(session.timeline);
+  }
+
+  /**
+   * Where a run beginning at `index` must be positioned: the time the PLAYER
+   * was told that segment starts at.
+   *
+   * Public because it is the invariant this class has broken twice, and a
+   * private one cannot be pinned by a test. It must always be the table the cut
+   * list is taken from — see the comment where a run is started.
+   *
+   * @param {HlsSession} session
+   * @param {number} index
+   * @returns {number}
+   */
+  runStartTimeFor(session, index) {
+    return this.publishedStartTime(session, index);
+  }
+
+  publishedStartTime(session, index) {
+    return publishedStartTime(session.timeline, index, this.#host.segmentDurationSec);
+  }
+
+  /**
+   * Report a segment whose own timeline disagrees with the playlist by more
+   * than a player will bridge.
+   *
+   * Once per segment per five seconds, like every other repeating condition
+   * here: the same segment is requested again and again while it is refused,
+   * and a line each time buries the first one.
+   *
+   * @param {HlsSession} session
+   * @param {number} index
+   * @param {number} trueStart
+   * @param {number} publishedStart
+   * @returns {void}
+   */
+  notePlaylistDisagreement(session, index, trueStart, publishedStart) {
+    // Within what the player bridges it is no disagreement worth a line.
+    if (!(Math.abs(trueStart - publishedStart) > PLAYER_BUFFER_HOLE_SEC)) {
+      return;
+    }
+    const now = Date.now();
+    const state = this.#stateFor(session);
+    if (now - (state.stampWarnedAt.get(index) ?? 0) < 5_000) {
+      return;
+    }
+    state.stampWarnedAt.set(index, now);
+    this.#host.logger.warn(
+      `transcode ${session.id} segment #${index} carries ${trueStart.toFixed(3)}s while the playlist ` +
+      `the player holds says ${publishedStart.toFixed(3)}s — a gap of ` +
+      `${Math.abs(trueStart - publishedStart).toFixed(3)}s, beyond the ${PLAYER_BUFFER_HOLE_SEC}s a player ` +
+      "bridges; stamping it where the playlist says so the fragment lands where it was asked for"
+    );
+  }
+
+  /**
+   * Segment index whose span contains time `t` (0-based), via the boundary
+   * table.
+   *
+   * @param {HlsSession} session
+   * @param {number} t
+   * @returns {number}
+   */
+  segmentIndexForTime(session, t) {
+    // The player's grid, for the same reason the cut list uses it: the time
+    // being resolved came from the playlist the player holds, so the index it
+    // means is the index that playlist gives it.
+    return segmentIndexForTime(this.publishedGridFor(session), t, this.#host.segmentDurationSec);
+  }
+
+  /**
+   * Where the run REALLY began, against where it was asked to begin.
+   *
+   * The first piece a run produces is the only statement of this that exists,
+   * and until now nothing compared the two. They disagree whenever the seek
+   * lands somewhere other than the time asked for — which, before the landing
+   * offset, was every run on a Matroska source with B-frames, by exactly one
+   * keyframe interval. Every cut of the run then inherits it, because
+   * `-segment_times` is measured from the landing.
+   *
+   * Said once per run, and only when it matters: within what a player bridges
+   * there is nothing to report.
+   *
+   * @param {HlsSession} session
+   * @param {number} index
+   * @param {number} trueStart
+   * @returns {void}
+   */
+  noteRunLanding(session, index, trueStart) {
+    const state = this.#stateFor(session);
+    if (runStartingAt(this.#host.runsOf(session), index) === null || state.landingReportedForRun === index) {
+      return;
+    }
+    state.landingReportedForRun = index;
+    // What the run was ASKED for, taken from the run itself rather than looked
+    // up again in a table. The two used to be the same lookup; they stopped
+    // being so when a run began positioning on the published grid while this
+    // read the live one, which made a perfect landing report a drift equal to
+    // the distance between the tables — and cancelled a real landing error of
+    // the same size to zero. A run also has one legitimate position that is in
+    // no table at all: the realignment that starts the sound where the copied
+    // picture truly begins.
+    const asked = runStartingAt(this.#host.runsOf(session), index)?.progress?.startPositionSeconds ??
+      this.runStartTimeFor(session, index);
+    const drift = trueStart - asked;
+    if (!Number.isFinite(drift) || Math.abs(drift) <= PLAYER_BUFFER_HOLE_SEC) {
+      return;
+    }
+    this.#host.logger.warn(
+      `transcode ${session.id} run began at ${trueStart.toFixed(3)}s but was asked for ` +
+      `${asked.toFixed(3)}s — ${drift > 0 ? "+" : ""}${drift.toFixed(3)}s, and every cut of this ` +
+      "run is measured from where it began, so the whole run is that far from its playlist"
+    );
+  }
+
+  noteIndexAccuracy(session, index, trueStart, declaredStart) {
+    const deviation = Math.abs(trueStart - declaredStart);
+    // Where each produced segment truly began, kept so that a player reporting
+    // a stall can be ANSWERED rather than merely believed. Bounded: only the
+    // recent past can be the subject of such a report, and an unbounded map on
+    // a two-hour film is a leak.
+    const state = this.#stateFor(session);
+    state.trueStartByIndex.set(index, trueStart);
+    if (state.trueStartByIndex.size > TRUE_START_MEMORY) {
+      const oldest = state.trueStartByIndex.keys().next();
+      if (!oldest.done) {
+        state.trueStartByIndex.delete(oldest.value);
+      }
+    }
+    // ONE READING, AND WHOSE FACT IT IS DEPENDS ON HOW THIS OUTPUT IS MADE. The
+    // three branches below already say the distinction when they WARN; until
+    // 2026-09-15 what they COUNTED threw it away, into a tally held per (file,
+    // grid) that a picture and the soundtrack inside the same file share.
+    //
+    // Only a COPY can say anything about the file: it is cut at a keyframe of
+    // that file and nowhere else. So a copy's landing goes to the file's own
+    // table, where it outlives this grid and reaches every step and every later
+    // session of the file.
+    if (session.spec.carries !== "audio-only" && !session.spec.transcodesVideo) {
+      session.keyframes?.witness({
+        index,
+        trueStart,
+        deviationSec: deviation,
+        toleranceSec: SEGMENT_START_DISAGREEMENT_SEC
+      });
+    }
+    // And every output records its own landing against its own published grid,
+    // which is what says whether a step will splice and how far a soundtrack
+    // stands from the picture.
+    session.output?.noteLanding({ index, deviationSec: deviation, toleranceSec: SEGMENT_START_DISAGREEMENT_SEC });
+    if (deviation > SEGMENT_START_DISAGREEMENT_SEC) {
+      // Which boundary the true start DOES match, if any. This is what tells
+      // the two possible faults apart, and they need opposite fixes: matching
+      // boundary #N-1 means our numbering is shifted by one — a fault in this
+      // code, where the run begins — while matching nothing means the container
+      // index describes times the file does not have. Measured 2026-08-11,
+      // three samples all matched N-1, which is why the line now says so
+      // instead of leaving it to be inferred from the numbers.
+      const at = this.boundaryIndexAt(session, trueStart);
+      // Once per segment per five seconds, like `#notePlaylistDisagreement`
+      // beside it. The same segment is produced and served again and again
+      // while it is refused, and — since a run keeps cutting on the list it was
+      // launched with — a soundtrack whose grid has moved under it deviates on
+      // EVERY segment for the life of that run. A line each time buries the
+      // first one, which is the one somebody is reading the log for.
+      const state = this.#stateFor(session);
+      const lastWarnedAt = state.deviationWarnedAt.get(index) ?? 0;
+      if (Date.now() - lastWarnedAt >= 5_000) {
+        state.deviationWarnedAt.set(index, Date.now());
+        this.#host.logger.warn(
+        `transcode ${session.id} segment #${index} really starts at ` +
+        `${trueStart.toFixed(3)}s (boundary ${at === null ? "none" : `#${at}`}), ` +
+        `the grid says ${declaredStart.toFixed(3)}s — ` +
+        (session.spec.carries === "audio-only"
+          // A soundtrack is cut exactly where it was asked to be, so a
+          // disagreement here is not a reading about the file at all: it is the
+          // distance between this run's own cuts and a grid the picture has
+          // since corrected under it. Said plainly, because the same sentence
+          // used to claim a keyframe index was wrong when no keyframe was
+          // involved on this side of the stream.
+          ? "sound is cut where it is asked to be; this is the picture's grid having moved, not the index"
+          : session.spec.transcodesVideo
+            // A re-encode was TOLD to put a keyframe here and did not, so this
+            // rung's segments no longer stand where the stream it accompanies
+            // would have put them. That is a broken splice, not a wrong index.
+            ? "this rung did not cut where its grid says; a switch to it will not join cleanly"
+            // A copy. The two possible faults need opposite fixes and the
+            // numbers already tell them apart, so the sentence follows THEM
+            // rather than the branch it is printed from. Until 2026-08-21 it
+            // blamed the index either way — including through a session whose
+            // segments each held the boundary two, then four places before
+            // their own number, which is this code's own definition of a fault
+            // in this code. That sentence is what sent the reading of that
+            // session after the file instead of after the arithmetic.
+            : at === null
+              ? "the container's keyframe index disagrees with the file; using the file"
+              : `this began at another boundary of the same list, ${index - at} place(s) before ` +
+                "its own number — the numbering of this run is shifted, not the index")
+        );
+      }
+    }
+    // Said as the evidence accumulates, not only when the session is disposed.
+    // A proxy restart takes its sessions with it — every addon update does —
+    // and a summary that only ever appears at the end is a summary that is
+    // routinely never written. Twenty-five distinct boundaries is enough for
+    // the proportion to mean something and rare enough not to repeat itself.
+    const counted = session.output?.piecesLanded ?? 0;
+    if (counted > 0 && counted % 25 === 0) {
+      this.logIndexAccuracy(session);
+    }
+    this.correctBoundaryFromSegment(session, index, trueStart);
+  }
+
+  /**
+   * Replace a boundary the index got wrong with the time the file actually has.
+   *
+   * The grid of a copied stream comes from the container's keyframe index,
+   * because a copy can only be cut where a keyframe already is and nothing
+   * cheaper than the index can say where that is before a single byte is
+   * encoded. An index can be wrong — proven 2026-08-12 by reproducing both
+   * cases against the same file: with an honest index every produced segment
+   * started exactly where declared, and with one moved 1.8 s the segments
+   * started 1.8 s early, matching no boundary at all. The field showed the
+   * second shape.
+   *
+   * The truth arrives anyway, one segment at a time: a produced piece states
+   * where it really begins. Writing it back makes the grid describe the file
+   * instead of the index — and it is what lets a re-encoded rung be cut to
+   * match a copied one, because the rung is then forced onto times the copy
+   * really uses. The alternative considered and rejected was to stop offering
+   * quality on files with a bad index, which is not a fix but a withdrawal.
+   *
+   * The whole family shares one grid, so a correction reaches all of it: a rung
+   * created afterwards inherits a table that is true wherever anyone has looked.
+   *
+   * @param {HlsSession} session
+   * @param {number} index
+   * @param {number} trueStart
+   * @returns {void}
+   */
+  correctBoundaryFromSegment(session, index, trueStart) {
+    // Only the picture may move the grid, because the grid IS the picture's
+    // cut list: it is built from the container's keyframe index, and a copied
+    // stream can be cut nowhere else. A soundtrack has no keyframes — it is cut
+    // exactly where `-segment_times` asks, to within one audio frame — so its
+    // reading measures nothing about the grid and everything about itself.
+    //
+    // Writing it into the shared table is how one film ended up with two
+    // answers for one boundary, each side correctly describing its own stream
+    // and each overwriting the other: field 2026-08-20, segment #521 of
+    // "Minions.and.Monsters.1080p.mkv" corrected 2086.084s → 2084.082s by the
+    // picture at 11:14:16.939 and 2084.082s → 2086.033s by the sound 1.6s
+    // later — 1.951s apart, against the 0.25s that stops a correction and the
+    // 0.5s a player bridges. The next reading disagrees with the table again,
+    // so it never converges and never stops.
+    if (session.spec.carries === "audio-only") {
+      return;
+    }
+    const boundaries = session.timeline.boundaries;
+    if (!Array.isArray(boundaries) || index <= 0 || index >= boundaries.length - 1) {
+      // Index 0 is the start of the file and the last entry is its end; neither
+      // is a cut, and neither can be learned from a segment.
+      return;
+    }
+    if (Math.abs(boundaries[index] - trueStart) <= SEGMENT_START_DISAGREEMENT_SEC) {
+      return;
+    }
+    // A correction that would put this boundary at or past its neighbours is not
+    // a correction — it is a reading from a run that started somewhere else, and
+    // applying it would make the table describe nothing at all.
+    if (trueStart <= boundaries[index - 1] || trueStart >= boundaries[index + 1]) {
+      return;
+    }
+    const wasAt = boundaries[index];
+    // One write, because there is one table. Where a file is cut is a fact
+    // about the FILE, so every session of it holds the same array rather than a
+    // copy of it — which is what this loop used to keep in step, member by
+    // member, and only for members that happened to exist at the time. A
+    // session created afterwards used to inherit a copy taken at that moment;
+    // now it is handed the table itself.
+    boundaries[index] = trueStart;
+    this.#host.logger.info(
+      `transcode ${session.id} boundary #${index} corrected ${wasAt.toFixed(3)}s → ` +
+      `${trueStart.toFixed(3)}s from the file itself` +
+      `, and the live table is now ${describeGridDrift(this.publishedGridFor(session), boundaries)}` +
+      " from the one the player holds"
+    );
+    // And every OTHER member whose run begins at this very boundary is moved
+    // to the same instant.
+    //
+    // Why they were not there already: the two branches are asked for the same
+    // time and land in different places. The picture cannot begin anywhere but
+    // a real keyframe, and it may not begin before the time asked for — that
+    // content belongs to the previous segment — so it moves FORWARD to the next
+    // one, by up to the keyframe spacing (0.58-2.96 s measured 2026-08-17).
+    // A soundtrack has no keyframes: it begins exactly where asked, to within
+    // one audio frame. So after every restart the two runs of one film began up
+    // to three seconds apart, each correctly labelled with where it really was,
+    // and the viewer got sound with no new picture for the difference.
+    //
+    // The picture's true start is a MEASURED quantity — read from the piece it
+    // just produced, which is what the correction above is — so the soundtrack
+    // can be put exactly there instead of at the time the container's table
+    // claimed. It converges: once the boundary holds the true time, the next
+    // reading agrees with it and the guard above returns before doing anything.
+    for (const member of this.#host.outputs.familyOf(session)) {
+      if (member === session || runStartingAt(this.#host.runsOf(member), index) === null) {
+        continue;
+      }
+      this.#host.logger.info(
+        `transcode ${member.id} begins at #${index}, which really starts ` +
+        `${(trueStart - wasAt).toFixed(3)}s later than the table said — restarting it there ` +
+        `so picture and sound begin together`
+      );
+      // STOPPED, and its replacement is the plan's to place. What has been
+      // learned here is that this run is producing at the wrong instant, which
+      // nothing but the piece it produced could say — the plan reasons about
+      // numbers and cannot know it. So the fact is acted on where it is known,
+      // and only as far as it is known: the run that is wrong goes, the stretch
+      // it held returns to the map, and where the next one stands follows from
+      // where the viewers are.
+      //
+      // The corrected instant reaches that replacement through the live table,
+      // which this function has just written it into and which every session of
+      // the file shares. It used to be passed as an argument from here, so only
+      // a run started by this line ever had it.
+      this.#host.stopEncodeRun(member, `#${index} really begins ${(trueStart - wasAt).toFixed(3)}s later than the table said`);
+    }
+    this.#host.planEncodersSoon();
+  }
+
+  /**
+   * The boundary a time falls on, or null when it falls on none of them.
+   *
+   * Within the same tolerance a disagreement is judged by, so "matches boundary
+   * #N-1" and "matches nothing" mean what they say.
+   *
+   * @param {HlsSession} session
+   * @param {number} seconds
+   * @returns {number | null}
+   */
+  boundaryIndexAt(session, seconds, table) {
+    // The table is nameable because the two answer different questions. The
+    // LIVE one says "is this a cut this file actually has", which is what a
+    // reading taken off a produced segment is about. The PUBLISHED one says "is
+    // this a cut the player believes in", which is what a report from the player
+    // is about. Answering one with the other prints an index from one grid
+    // beside a time from the other.
+    const boundaries = Array.isArray(table) ? table : session.timeline.boundaries;
+    if (!Array.isArray(boundaries)) {
+      return null;
+    }
+    for (let index = 0; index < boundaries.length; index += 1) {
+      if (Math.abs(boundaries[index] - seconds) <= SEGMENT_START_DISAGREEMENT_SEC) {
+        return index;
+      }
+    }
+    return null;
+  }
+
+  /**
+   * Where this output's pieces have been landing, and — where this output is a
+   * copy — what that has shown about the FILE's own keyframe table.
+   *
+   * TWO LINES, because they are two facts with two owners and two lifetimes.
+   * The landing is this output's and dies with it. The table's accuracy belongs
+   * to the file, is learned only from a copy, and is the same finding for every
+   * step and every later session of that file — so it names the file and not
+   * the session, and it is written only by the output that can witness it.
+   *
+   * Written even when nothing disagreed, because that is the finding: with only
+   * the per-piece warning, silence could not be told from nobody having
+   * watched. Skipped where nothing was produced, which says neither.
+   *
+   * @param {HlsSession} session
+   * @returns {void}
+   */
+  logIndexAccuracy(session) {
+    const copiedPicture = session.spec.carries !== "audio-only" && !session.spec.transcodesVideo;
+    const landing = session.output?.landing ?? null;
+    if (landing) {
+      this.#host.logger.info(
+        // The session id, because without it this line cannot be attributed. A
+        // family produces one of these per member — the picture, each step,
+        // each soundtrack — and on 2026-08-17 the picture's was read as the
+        // sound's, from a neighbouring log line, and a roadmap item was written
+        // against the wrong half of the stream.
+        `landing ${session.id.slice(0, 8)} ` +
+        `${session.spec.carries === "audio-only" ? "sound" : copiedPicture ? "copied picture" : "re-encoded picture"} ` +
+        `"${session.file.name}": ${landing.disagreed} of ${landing.checked} produced pieces started ` +
+        `away from this output's own playlist, median ${landing.medianDeviationSec.toFixed(3)}s ` +
+        `worst ${landing.maxDeviationSec.toFixed(3)}s` +
+        (landing.firstDisagreementIndex >= 0 ? ` (first at #${landing.firstDisagreementIndex})` : "") +
+        ` [tolerance ${SEGMENT_START_DISAGREEMENT_SEC}s] — ` +
+        (session.spec.carries === "audio-only"
+          // A soundtrack is cut exactly where it is asked to be, so this is not
+          // a reading about any keyframe: it is how far this run's cuts stand
+          // from a grid the picture has corrected under it.
+          ? "sound is cut where it is asked to be, so this is the picture's grid having moved"
+          : copiedPicture
+            // A copy cannot be cut anywhere but a real keyframe, so this figure
+            // is also evidence about the file, reported on the line below.
+            ? "a copy is cut at the file's own keyframes, so this is also evidence about its table"
+            // A re-encode was TOLD to put a keyframe at each of these instants.
+            : "this rung was told where to cut; what it missed will not splice cleanly")
+      );
+    }
+    if (!copiedPicture) {
+      return;
+    }
+    const evidence = session.keyframes?.evidence ?? null;
+    if (!evidence) {
+      return;
+    }
+    this.#host.logger.info(
+      // Named by the FILE, with no session id in it, because that is what it is
+      // about: the same answer for every step and every viewer of these bytes.
+      `keyframe-index ${session.keyframes?.format ?? "unknown"} "${session.file.name}": ` +
+      `${evidence.disagreed} of ${evidence.checked} copied pieces began away from the table, ` +
+      `median ${evidence.medianDeviationSec.toFixed(3)}s worst ${evidence.maxDeviationSec.toFixed(3)}s` +
+      (evidence.firstDisagreementIndex >= 0 ? ` (first at #${evidence.firstDisagreementIndex})` : "") +
+      // The discriminator, stated in the same line as the count it explains: a
+      // piece that began at another time the SAME table names was not
+      // mis-described by the table — the grid was built over a gap in it.
+      `; ${evidence.landedOnAnotherKeyframe} of them began at another keyframe the table names` +
+      ` [tolerance ${SEGMENT_START_DISAGREEMENT_SEC}s, ${session.keyframes?.count ?? 0} keyframes read]`
+    );
+  }
+}