ffmpeg-skill 2.3.1 → 2.4.0

package/SKILL.md

@@ -5,13 +5,13 @@ description: 'Edit video and audio with local FFmpeg from natural-language reque
 
 # ffmpeg-skill
 
-Scripts live in `scripts/` next to this file; run them with `python3 <skill-dir>/scripts/<name>.py`, and delivery templates in `templates/`. This file is enough to do a job: the table below routes the request, and `--help` on the script about to run is the cheapest full flag list. A reference file costs as much as this one; open one only for a question you have: `references/scripts.md` (every flag of all 42 scripts), `references/devices.md` (iPhone HDR, GoPro, DJI, screen recordings, Zoom), `references/gotchas.md` (the long form of the one-line rules at the end). MCP: `tools/list` shows only the core 12 by default; the other 42 (per this table) stay callable by name via `tools/call`, described by `contract --json`; set `FFMPEG_SKILL_MCP_FULL=1` for all 42.
+Scripts live in `scripts/` next to this file; run them with `python3 <skill-dir>/scripts/<name>.py`, and delivery templates in `templates/`. This file is enough to do a job: the table below routes the request, and `--help` on the script about to run is the cheapest full flag list. A reference file costs as much as this one; open one only for a question you have: `references/scripts.md` (every flag of all 42 scripts), `references/devices.md` (iPhone HDR, GoPro, DJI, screen recordings, Zoom), `references/gotchas.md` (the long form of the one-line rules at the end). MCP: `tools/list` shows only the core 12 by default; the other 30 (per this table) stay callable by name via `tools/call`, described by `contract --json`; set `FFMPEG_SKILL_MCP_FULL=1` for all 42.
 
 Shared flags, on every script: `--dry-run`; `--json` (output path, a probe of the output, the commands run); `--json-brief` (status/output/verified plus a `summary`; prefer on writing steps); `--fast` (preview quality); `--progress`; `--timeout SECONDS` (`kind: timeout`, default 1800); `--overwrite` (step 7); `--plan FILE` (the dry run as a plan `render.py FILE` runs later; refuses if an input changed). Re-encoding tools also take `--codec h264|hevc|av1|prores` and `--quality N`: unset, SDR is x264, HDR is x265 Main10; `prores` needs `-o NAME.mov`, `h264` refuses HDR (`color.py --to-sdr` first).
 
 Writing tools run nothing under `--dry-run`; the measuring tools (`probe`, `check`, `sync`, `multicam`, `scenes`, `cropdetect`, `report`, `silence`, `loudness`, `stabilize`) may still run ffmpeg/ffprobe, skipping artifacts/side files (`--edl`, `--sheet`, a generated `.ass`); `verify` ignores the flag. Per tool: `contract --json` `dry_run`.
 
-## Workflow (always follow this order)
+## Workflow (in this order)
 
 0. **Environment, only on failure.** Never start a job with `doctor`: a broken machine fails with `kind: missing_tool` or an ffmpeg error naming the filter/encoder. Run `python3 <skill-dir>/scripts/_contract.py doctor` (or `npx ffmpeg-skill doctor`) after such a failure, or when asked what the machine can do: read `ok`/`usable`, report the missing capability. `contract --json`'s tool schema is for a *planning* agent, not this workflow.
 1. **Probe what you must plan from.** Run `probe.py` on each input you plan from — duration, fps, resolution, codecs, channels, `variable_frame_rate_suspected` — and whenever the user asks about a file. No separate probe before every edit: every writing tool's `--json` already carries its input and a probe of the output. Plan from real numbers, never assumptions.
@@ -96,7 +96,7 @@ Timestamp flags (`--start`, `--end`, `--at`, `--from`, `--duration`, `--offset`,
 | "add subtitles from this SRT", "burn in captions" | `caption.py input.mp4 --srt subs.srt` |
 | "caption it with these lines" (text with times) | `caption.py input.mp4 --text cues.txt` |
 | "keep the subtitles toggleable", "mux in an SRT" | `caption.py input.mp4 --srt subs.srt --mode mux`; repeat `--srt file:lang` for several languages, `.mkv` for more than two |
-| "the captions are tiny / three lines on a Short", "don't chop the sentence" | `caption.py` shrinks the size until the cue fits `--max-lines` before splitting it (`--fit-size off` = 1.16 behaviour, `--min-size` the floor). Keep the template's `--max-lines` (2 on a vertical) and let the size drop; raising it to dodge a shrink stacks two words per line. A word still wider than the column at the floor is sliced at the edge (hyphen preferred), never rewritten, automatic; `broken_inside_word` in the JSON counts it |
+| "the captions are tiny / three lines on a Short", "don't chop the sentence" | `caption.py` shrinks the size until the cue fits `--max-lines` before splitting it (`--fit-size off` keeps `--size` and splits instead, `--min-size` the floor). Keep the template's `--max-lines` (2 on a vertical) and let the size drop; raising it to dodge a shrink stacks two words per line. A word still wider than the column at the floor is sliced at the edge (hyphen preferred), never rewritten, automatic; `broken_inside_word` in the JSON counts it |
 | "transcribe it and caption it" | `caption.py input.mp4 --transcribe --animate pop --karaoke` (needs a local whisper; else `--text`) |
 | "TikTok-style captions with the words popping" | `caption.py input.mp4 --text cues.txt --animate pop --karaoke` |
 | "our logo top-right", "a watermark" | `overlay.py input.mp4 --image logo.png --position top-right --scale 200` |

package/docs/contract.md

@@ -21,7 +21,7 @@ The contract is derived from the code that runs, not maintained beside it:
 | Field | Meaning | Changes when |
 |---|---|---|
 | `contract_version` | shape of this document (`1.0`) | a key is renamed, removed or changes meaning |
-| `skill.version` | the npm / package.json version (`2.3.1`) | any release |
+| `skill.version` | the npm / package.json version (`2.4.0`) | any release |
 
 A release that adds a tool or a flag keeps `contract_version`; a breaking change to the
 ToolSpec shape bumps it. Consumers pin on `contract_version` and read `skill.version`
@@ -119,7 +119,7 @@ JSON would have been rewritten for that. The flat keys, typed per tool in `outpu
   "contract_version": "1.0",
   "deprecated": [],
   "removed": [{"what": "...", "since": "1.10.0", "replacement": "...", "removed_in": "2.0.0", "where": "cli | json | mcp | behaviour"}],
-  "skill": {"id": "ffmpeg-skill", "version": "2.3.1", "execution_mode": "local", "kind": "execution",
+  "skill": {"id": "ffmpeg-skill", "version": "2.4.0", "execution_mode": "local", "kind": "execution",
             "entrypoints": {"cli": "...", "mcp": "...", "contract": "...", "doctor": "..."},
             "not_provided": ["AI reasoning", "decisions", "production plans", "project IR", "approvals", "network access", "transcription engine"]},
   "requirements": {"python": ">=3.9 (standard library only)", "ffmpeg": ">=5.0", "ffprobe": ">=5.0"},
@@ -353,7 +353,10 @@ read (`ffmpeg -filters` in a layout the parser does not recognise, or ffmpeg exi
 non-zero), and it is never folded into `missing`, so an installed filter is not reported
 absent, nor into `available`, so a failed detection is not a pass. `detection` gives the
 status (`parsed`, `unparsed`, `failed`, `missing`), row count and detail of each listing;
-`errors` lists the unreadable ones. The filter parser recognises the FFmpeg 6/7 layout
+`errors` lists the unreadable ones. `filter:drawtext` is also rendered once, with the
+`fontfile=` the tools pass by default: a crash there makes it `missing`; a crash only on a
+bare fontconfig lookup leaves it `available` and adds a `notes` entry (`--font-file`, not a
+bare `--font <family>`, on that machine). The filter parser recognises the FFmpeg 6/7 layout
 (three flag characters, `..C acompressor A->A`) and the FFmpeg 8 layout (two, `T.
 acompressor A->A`) by the io-spec token, so the flag width does not matter; fixtures for
 both live in `tests/fixtures/`.
@@ -518,7 +521,7 @@ Per-tool keys added after 2.2.2, all additive:
 | key | tool | what it holds |
 |---|---|---|
 | `sdr_path`, `notes` | `color.py --to-sdr` | `sdr_path` is `"tonemap"` for PQ / HLG / Dolby Vision input (and `--force` on an untagged file) or `"gamut"` for BT.2020 primaries on an SDR transfer, which is converted to BT.709 without a tone map (`--tonemap`/`--peak`/`--desat` do not apply); `notes` says which path was taken and why |
-| `silent` | `join.py` (every run) | `[{index, path, peak_db}]`: inputs whose whole-file audio peak (volumedetect) is at or below `--silence-threshold` (default -50 dBFS) and that `--on-silent warn` (default) joined anyway, with a `notes` line; `[]` otherwise. `--on-silent fail` names them in the `kind: input` refusal's `problems` (reason `silent (peak -91.0 dBFS)`), `skip` lists them under `skipped`. An input with no audio stream is never silent; `verified` is unaffected |
+| `silent` | `join.py` (every run) | `[{index, path, peak_db, at, end}]`: inputs whose whole-file audio peak (volumedetect) is at or below `--silence-threshold` (default -50 dBFS) and that `--on-silent warn` (default) joined anyway, with a `notes` line; `at`/`end` are where the input sits in the output, in seconds (a transition overlaps its neighbours). An input named by `--allow-silent N[,N]` (1-based) is a planned pause: always joined, listed with `intended: true`, no warning or note, whatever `--on-silent` says. `[]` otherwise. `--on-silent fail` names them in the `kind: input` refusal's `problems` (reason `silent (peak -91.0 dBFS)`), `skip` lists them under `skipped`. An input with no audio stream is never silent; `verified` is unaffected |
 | `short_segments`, `duplicates` | `join.py` (every run) | warnings, never refusals, and the join command is unchanged: `short_segments: [{index, path, duration}]` names each measured input shorter than 2 frames at the join's fps (an audio-only join: shorter than 0.05 s); `duplicates: [{path, indices}]` names a path (compared resolved) listed more than once — repeating a clip can be intended. `[]` when none; each non-empty one adds a `notes` line |
 | `silent` | `waveform.py` (every run) | `true` when the input audio's whole-file peak is at or below `--silence-threshold` (default -50 dBFS), which draws a flat line; `--on-silent warn` (default) renders it anyway with a `notes` line, `fail` refuses (`kind: input`) before ffmpeg runs. `false` when audible, `null` under `--dry-run` (nothing measured). `verified` is unaffected |
 

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "ffmpeg-skill",
-  "version": "2.3.1",
+  "version": "2.4.0",
   "description": "Agent Skill that gives coding agents (Claude Code, Cursor, Codex) a local video editor: 42 FFmpeg tools with a machine-readable contract, contract-derived MCP server, FFmpeg capability detection, probe-first / verify-last workflow. Cut, join, silence removal, fit, captions and karaoke, overlays, motion graphics, HDR to SDR, LUTs, audio clean-up and typed dynamics, sync with drift correction, multicam, loudness, delivery checks, project rendering, batch. No API keys, no cloud, no dependencies.",
   "keywords": [
     "ffmpeg",

package/references/scripts.md

@@ -560,7 +560,7 @@ writes the resulting cut list exactly as it does today.
 ```
 join.py CLIP1 CLIP2 [...] [--transition fade|dissolve|wipeleft|slideleft|fadeblack|fadewhite|circleopen|none]
         [--duration 0.5] [--width W --height H] [--fps N] [--fit pad|crop] [-o OUT]
-join.py --list parts.txt [--on-missing fail|skip] [--on-silent warn|fail|skip] [--silence-threshold -50] [...]
+join.py --list parts.txt [--on-missing fail|skip] [--on-silent warn|fail|skip] [--allow-silent N[,N]] [--silence-threshold -50] [...]
 ```
 `--list FILE` (2.2) reads the clips from a file, one per line in order, paths relative to the
 file (blank lines and `#` comments ignored; ffmpeg's `file 'x.wav'` lines accepted), for the
@@ -579,9 +579,11 @@ frame, rate or xfade offsets that rest on a pending input's unmeasured placehold
 Every existing input with an audio stream is also measured (one volumedetect pass; not under `--dry-run`
 too): a peak at or below `--silence-threshold` (default -50 dBFS) is silent, the trace of a TTS
 call that wrote a valid but empty file. `--on-silent warn` (default) joins it and names it under
-`silent: [{index, path, peak_db}]` and in `notes`; `fail` refuses it in the same `kind: input`
-document (`silent (peak -91.0 dBFS)`); `skip` leaves it out under `skipped`. A clip without an
-audio stream is never silent.
+`silent: [{index, path, peak_db, at, end}]` (`at`/`end`: where it sits in the output, seconds)
+and in `notes`; `fail` refuses it in the same `kind: input` document (`silent (peak -91.0 dBFS)`);
+`skip` leaves it out under `skipped`. A pause that is meant to be silent: `--allow-silent 3,7`
+(1-based) joins those inputs whatever `--on-silent` says and lists them with `intended: true`,
+no warning. A clip without an audio stream is never silent.
 Two warnings never refuse or change the command: `short_segments: [{index, path, duration}]`
 names a clip shorter than 2 frames at the join's fps (audio-only: 0.05 s), and
 `duplicates: [{path, indices}]` a path listed more than once (a repeat can be intended); both

package/scripts/_ass_overlay.py

@@ -29,7 +29,7 @@ EMOJI_SENTINEL = "\ue000"
 
 
 def emoji_placeholder(box_px: float) -> str:
-    """The ASS override that reserves exactly `box_px` of advance and draws nothing.
+    r"""The ASS override that reserves exactly `box_px` of advance and draws nothing.
 
     Measured, not assumed (1.15 spec, open question 2). U+2588 FULL BLOCK is NOT 1.0 em: its
     advance measured 0.83 em in FreeSans, 0.79 in WenQuanYi Zen Hei and 0.66 in DejaVu Sans,

package/scripts/_common/drawtext.py

@@ -98,7 +98,7 @@ def escape_drawtext(text: str) -> str:
 
 
 def drawtext_text_opts(text: str, tmpdir: "Optional[str]" = None) -> str:
-    """`textfile=<path>:expansion=none` for drawtext -- the one route that is provably safe for
+    r"""`textfile=<path>:expansion=none` for drawtext -- the one route that is provably safe for
     every character on every build shape this repo uses.
 
     The filter-graph parser never sees the text at all: only the PATH is parsed, and

package/scripts/_contract.py

@@ -464,7 +464,7 @@ def output_schema(name: str, meta: Dict[str, Any]) -> Dict[str, Any]:
                  "sample_rate": {"type": "integer", "description": "audio mode only"}, "channels": {"type": "integer", "description": "audio mode only"},
                  "video": {"type": "boolean", "description": "false in audio mode: the output has no video stream"},
                  "skipped": {"type": "array", "description": "inputs --on-missing skip left out: [{index, path, reason}] ([] when none were)"},
-                 "silent": {"type": "array", "description": "inputs whose audio peak is at or below --silence-threshold (default -50 dBFS) and that --on-silent warn (default) joined anyway: [{index, path, peak_db}] ([] otherwise; fail refuses them under problems, skip lists them under skipped); an input with no audio stream is never silent"},
+                 "silent": {"type": "array", "description": "inputs whose audio peak is at or below --silence-threshold (default -50 dBFS) and that --on-silent warn (default) joined anyway, plus every --allow-silent input that was silent: [{index, path, peak_db, at, end, intended?}]; at/end are the input's start and end in output seconds (a transition overlaps neighbours), intended: true marks an --allow-silent input (joined whatever --on-silent says, no warning). [] otherwise; fail refuses unallowed ones under problems, skip lists them under skipped; an input with no audio stream is never silent"},
                  "short_segments": {"type": "array", "description": "warning, never a refusal: measured inputs shorter than 2 frames at the join's fps (audio-only join: shorter than 0.05 s), [{index, path, duration}] ([] when none), each also in notes; a pending --dry-run input is not measured"},
                  "duplicates": {"type": "array", "description": "warning, never a refusal (a repeat can be intended): a path (compared resolved) listed more than once and joined each time, [{path, indices}] with 0-based list positions ([] when none), also in notes"},
                  "pending": {"type": "array", "description": "--dry-run: inputs that do not exist yet, planned on as an earlier step's output and never also under skipped: [{index, path}] ([] otherwise)"},
@@ -668,22 +668,42 @@ def _drawtext_probe() -> Dict[str, Any]:
     present and doctor used to report the capability `available` anyway; every tool that actually
     used it (look, scenes --sheet, overlay --text, graphics) then crashed on first real use (#100).
 
-    This runs the cheapest real drawtext render there is: a one-frame synthetic clip, no font=
-    given at all (ffmpeg's own default resolution -- the same path that crashed). A clean exit
-    means drawtext genuinely works here. Anything that could not prove either way (no ffmpeg,
-    timeout, an ordinary nonzero exit with a real ffmpeg error) is `unknown`, same "unknown is not
-    missing" principle as every other capability here. A crash specifically -- killed by signal on
-    POSIX, or an unhandled access violation surfacing as a huge unsigned exit code on Windows -- is
-    the one case this function exists to catch, and folds into `missing`: the filter is present in
-    the build but cannot actually be used as ffmpeg's own default would use it.
+    This runs the cheapest real drawtext render there is: a one-frame synthetic clip, with the
+    fontfile= the tools themselves pass by default (_common.fonts.default_font_file), or no font
+    at all where none resolves. A clean exit means drawtext genuinely works for the tools here.
+    Anything that could not prove either way (no ffmpeg, timeout, an ordinary nonzero exit with a
+    real ffmpeg error) is `unknown`, same "unknown is not missing" principle as every other
+    capability here. A crash specifically -- killed by signal on POSIX, or an unhandled access
+    violation surfacing as a huge unsigned exit code on Windows -- folds into `missing`. When the
+    fontfile= render works but a bare one crashes, drawtext stays available and the result
+    carries a `note`: only a fontconfig lookup (`--font <family>` with no file) would crash.
     """
     exe = shutil.which("ffmpeg")
     if not exe:
         return {"status": "unknown", "detail": "ffmpeg not on PATH"}
+    # Probe the way the tools call drawtext: with the fontfile= they pick by default (#298). On a
+    # build whose fontconfig lookup crashes (#100) the tools still render, because they never ask
+    # fontconfig; probing only the bare `drawtext=text=x` reported such a machine as missing
+    # drawtext altogether. The bare render still runs, and its crash is reported as a note: it is
+    # the path `--font <family>` without a file would take.
+    from _common.fonts import default_font_file
+    from _common.decision import escape_filter_path
+    font_file = default_font_file(_default_font())
+    result = _drawtext_render(exe, f"drawtext=fontfile={escape_filter_path(font_file)}:text=x" if font_file else "drawtext=text=x")
+    if font_file and result["status"] == "available":
+        bare = _drawtext_render(exe, "drawtext=text=x")
+        if bare["status"] == "missing":
+            result["note"] = ("drawtext renders with fontfile= (what every drawtext tool passes by default) but crashes "
+                              "when it has to resolve a font through fontconfig: pass --font-file, not a bare --font "
+                              "<family>, on this machine (https://github.com/kajisho5/ffmpeg-skill/issues/100)")
+    return result
+
+
+def _drawtext_render(exe: str, vf: str) -> Dict[str, Any]:
     try:
         proc = subprocess.run(
             [exe, "-hide_banner", "-loglevel", "error", "-f", "lavfi", "-i", "color=c=black:s=64x64:d=1",
-             "-vf", "drawtext=text=x", "-frames:v", "1", "-f", "null", "-"],
+             "-vf", vf, "-frames:v", "1", "-f", "null", "-"],
             stdout=subprocess.PIPE, stderr=subprocess.PIPE, text=True, encoding="utf-8", errors="replace", timeout=_DETECT_TIMEOUT,
         )
     except subprocess.TimeoutExpired:
@@ -695,9 +715,8 @@ def _drawtext_probe() -> Dict[str, Any]:
     if proc.returncode < 0 or proc.returncode >= 0x80000000:
         return {"status": "missing",
                 "detail": f"drawtext render crashed (exit {proc.returncode}) instead of failing cleanly -- "
-                           "the filter is present in this build but cannot be used as-is, likely a fontconfig "
-                           "resolution crash (see https://github.com/kajisho5/ffmpeg-skill/issues/100); "
-                           "pass an explicit --font-file to every drawtext tool as a workaround"}
+                           "the filter is present in this build but cannot render even with the fontfile= the "
+                           "tools pass (see https://github.com/kajisho5/ffmpeg-skill/issues/100)"}
     tail = " ".join(proc.stderr.strip().splitlines()[-2:])
     return {"status": "unknown", "detail": f"drawtext probe exited {proc.returncode}: {tail}"}
 
@@ -809,6 +828,8 @@ def doctor() -> Dict[str, Any]:
                 # than downgrading it; see _drawtext_probe()'s own docstring for why a crash alone
                 # is the one case this exists to catch.
                 probe = _drawtext_probe()
+                if probe.get("note"):
+                    drawtext_probe = probe
                 if probe["status"] == "missing":
                     drawtext_probe = probe
                     state[cap] = "missing"
@@ -832,6 +853,7 @@ def doctor() -> Dict[str, Any]:
     errors = [f"{k}: {v['detail']}" for k, v in listings.items() if v["status"] in ("unparsed", "failed")]
     if drawtext_probe is not None and drawtext_probe["status"] != "available":
         errors.append(f"filter:drawtext: {drawtext_probe['detail']}")
+    notes = [f"filter:drawtext: {drawtext_probe['note']}"] if drawtext_probe and drawtext_probe.get("note") else []
     return {
         "version": skill_version(),
         "python": ".".join(str(x) for x in sys.version_info[:3]),
@@ -843,6 +865,7 @@ def doctor() -> Dict[str, Any]:
         "unknown": unknown,
         "detection": {k: {"status": v["status"], "count": len(v["names"]), "detail": v["detail"]} for k, v in listings.items()},
         "errors": errors,
+        **({"notes": notes} if notes else {}),
         "ok": not missing_required and not unknown_required,
         "tools": _tool_usability(state),
         "gpu_encoders": _gpu_encoders(listings["encoders"]),

package/scripts/caption.py

@@ -954,7 +954,7 @@ def main() -> int:
     sty.add_argument("--max-lines", type=int, default=2, help="most lines one cue may occupy; a longer cue is split into consecutive cues (default 2)")
     sty.add_argument("--fit-size", choices=["auto", "on", "off"], default="auto",
                      help="shrink the caption size until the cue fits --max-lines, BEFORE splitting it: "
-                          "'auto' (default) only when no --size was given, 'on' always, 'off' for 1.16 behaviour")
+                          "'auto' (default) only when no --size was given, 'on' always, 'off' keeps --size and splits the cue instead")
     sty.add_argument("--min-size", type=int, default=None,
                      help="smallest size --fit-size may use, in ASS points (default 13 = 4.5%% of the frame height, "
                           "the legibility floor)")

package/scripts/join.py

@@ -127,6 +127,7 @@ def join_audio(args: argparse.Namespace, metas: List[dict]) -> int:
     n = len(args.inputs)
     durs = [m.get("duration") or 0.0 for m in metas]
     d = args.duration if args.transition != "none" else 0.0
+    place_silent(durs, d)
     note_short_segments(args.inputs, metas, MIN_AUDIO_SEGMENT, f"{MIN_AUDIO_SEGMENT:g}s")
     for p, dur in zip(args.inputs, durs):
         if d and dur <= d * 2 and not STATE.dry_run:
@@ -258,6 +259,32 @@ def find_silent(paths: List[str], audible: List[int], threshold: float) -> List[
     return silent
 
 
+def place_silent(lens: List[float], d: float) -> None:
+    """Where each joined silent input sits in the output: `at`/`end` in output seconds. Input k
+    (in kept order) starts after the kept inputs before it, each overlapped by the transition d."""
+    for sil in STATE_SILENT:
+        if sil["index"] not in STATE_KEPT:
+            continue
+        k = STATE_KEPT.index(sil["index"])
+        if k >= len(lens):
+            continue
+        at = sum(lens[:k]) - k * d
+        sil["at"], sil["end"] = round(at, 3), round(at + lens[k], 3)
+
+
+def parse_allow_silent(value: Optional[str], n: int) -> List[int]:
+    """--allow-silent 3,7 -> 0-based indices; each must name an input (1..n)."""
+    if not value:
+        return []
+    out = []
+    for tok in value.split(","):
+        tok = tok.strip()
+        if not tok.isdigit() or not 1 <= int(tok) <= n:
+            die(f"--allow-silent takes input numbers 1..{n} separated by commas, got {tok!r}")
+        out.append(int(tok) - 1)
+    return out
+
+
 def main() -> int:
     ap = argparse.ArgumentParser(description=__doc__, formatter_class=argparse.RawDescriptionHelpFormatter)
     ap.add_argument("inputs", nargs="*", help="two or more clips in order (or --list FILE)")
@@ -270,6 +297,9 @@ def main() -> int:
                     help="an input whose audio peak is at or below --silence-threshold: warn (default; join it, "
                          "name it under `silent`), fail (refuse with the other input problems) or skip (leave it out, "
                          "under `skipped`); a clip with no audio stream is never silent")
+    ap.add_argument("--allow-silent", metavar="N[,N...]",
+                    help="input numbers (1-based) whose silence is intended, e.g. a planned pause: always joined, "
+                         "listed under `silent` with intended: true, no warning, whatever --on-silent says")
     ap.add_argument("--silence-threshold", type=float, default=-50.0, metavar="DB",
                     help="peak level in dBFS at or below which an input counts as silent (default -50)")
     ap.add_argument("-o", "--output", help="output file (default: <first>_joined.mp4)")
@@ -303,7 +333,10 @@ def main() -> int:
     orig_inputs = list(args.inputs)
     audible: List[int] = []
     problems, pending = preflight(args.inputs, audible)
-    silent = find_silent(args.inputs, audible, args.silence_threshold)
+    allowed = parse_allow_silent(args.allow_silent, len(args.inputs))
+    found = find_silent(args.inputs, audible, args.silence_threshold)
+    intended = [dict(s, intended=True) for s in found if s["index"] in allowed]
+    silent = [s for s in found if s["index"] not in allowed]
     silent_problems = [{"index": s["index"], "path": s["path"], "reason": f"silent (peak {s['peak_db']:.1f} dBFS)"}
                        for s in silent]
     refused = sorted((problems if args.on_missing == "fail" else [])
@@ -325,8 +358,8 @@ def main() -> int:
         args.inputs = [p for i, p in enumerate(args.inputs) if i not in bad]
         for p in dropped:
             info(f"skipping #{p['index'] + 1} {p['path']}: {p['reason']}")
+    STATE_SILENT[:] = sorted(intended + (silent if args.on_silent == "warn" else []), key=lambda s: s["index"])
     if silent and args.on_silent == "warn":
-        STATE_SILENT[:] = silent
         listing = ", ".join(f"#{s['index'] + 1} {s['path']} (peak {s['peak_db']:.1f} dBFS)" for s in silent)
         info(f"warning: {len(silent)} input(s) silent at or below {args.silence_threshold:g} dBFS, joined anyway: {listing}")
         STATE_NOTES.append(f"{len(silent)} input(s) have no audible sound (peak at or below {args.silence_threshold:g} dBFS) "
@@ -413,6 +446,7 @@ def main() -> int:
     # a crossfaded join gives each clip one length for both streams, so the offsets that place
     # every picture also place its sound; the plain cut (concat) keeps the streams paired itself
     lens = [clip_length(m, fps) for m in metas] if d else durs
+    place_silent(lens, d)
     for p, dur in zip(args.inputs, lens):
         if d and dur <= d * 2 and not STATE.dry_run:
             die(f"{p} is only {dur:.2f}s, too short for a {d:.2f}s transition; shorten --duration")