ffmpeg-skill 3.1.4 → 3.1.6

package/scripts/sync.py

@@ -34,6 +34,7 @@ from typing import Dict, List
 
 from _common import video_args, add_common, apply_common, emit, aac_args, audio_codec_for, default_output, die, ffmpeg_base, info, probe, require_tool, run, run_analysis, x264_args, decode_pcm_mono, rms_envelope
 from _common.emit import record_skip, warn
+from _common.decision import stream_starts
 
 SR = 8000  # decode sample rate
 
@@ -355,6 +356,11 @@ def main() -> int:
         if abs(drift_ratio - 1.0) > 1e-7:
             # the second file runs long by drift_ratio -> play it faster by that ratio (pitch shift is ~ppm, inaudible)
             drift_af = [f"asetrate={sec_sr * drift_ratio:.6f}", f"aresample={sec_sr}"]
+        # the offset is measured between the files' clocks (decode_pcm_mono pads a late sound), so
+        # the second's sound is placed on its own clock first: one that starts 0.3 s after its
+        # file's start keeps that 0.3 s instead of being delayed from its first sample (#409).
+        # Only then, so the command for an ordinary file is the one it always was.
+        on_clock = ["aresample=async=1:first_pts=0"] if stream_starts(second_meta)[1] > 0.0005 else []
 
         if args.replace_audio:
             cmd = ffmpeg_base() + ["-i", args.reference]
@@ -367,7 +373,7 @@ def main() -> int:
             # its length instead of -shortest, which cut the reference's tail whenever the second
             # file started earlier (head_trim > 0) or was simply shorter
             ref_dur = probe(args.reference).get("duration")
-            af_parts = drift_af + ([f"adelay={delay_ms}:all=1"] if delay_ms > 0 else []) + ["apad"]
+            af_parts = on_clock + drift_af + ([f"adelay={delay_ms}:all=1"] if delay_ms > 0 else []) + ["apad"]
             cmd += ["-af", ",".join(af_parts)]
             cmd += aac_args() + (["-t", f"{ref_dur:.3f}"] if ref_dur else ["-shortest"]) + [output]
             proc = run(cmd, check=False)
@@ -391,7 +397,7 @@ def main() -> int:
                 if head_trim > 0:
                     cmd += ["-ss", f"{head_trim:.4f}"]
                 cmd += ["-i", args.second]
-                af_parts = list(drift_af)
+                af_parts = on_clock + drift_af
                 if delay_ms > 0:
                     af_parts.append(f"adelay={delay_ms}:all=1")
                 if has_video:

package/scripts/verify.py

@@ -26,7 +26,7 @@ import time
 from pathlib import Path
 from typing import Any, Dict, List, Optional, Tuple
 
-from _common import STATE, add_common, apply_common, die, emit, info, probe
+from _common import STATE, add_common, apply_common, die, emit, guard_sidecar, info, probe
 
 HERE = Path(__file__).resolve().parent
 MEDIA_EXT = {".mp4", ".mov", ".m4v", ".mkv", ".webm", ".avi", ".mts", ".m2ts", ".mxf", ".wav", ".m4a", ".mp3", ".flac", ".aac"}
@@ -160,9 +160,14 @@ def main() -> int:
     files = collect(args.paths)
     tmp = None
     if args.out:
+        # made first, so a --report inside a new --out folder passes the guard below (#413 review)
         outdir = Path(args.out)
         outdir.mkdir(parents=True, exist_ok=True)
-    else:
+    if args.report:
+        # a file this run writes itself (#408): not one of the media it verifies (a folder's
+        # files included), nor an existing file without --overwrite
+        guard_sidecar(args.report, *args.paths, *map(str, files))
+    if not args.out:
         tmp = tempfile.TemporaryDirectory(prefix="ffskill_verify_")
         outdir = Path(tmp.name)
 

package/scripts/waveform.py

@@ -35,11 +35,12 @@ import os
 import sys
 
 from _platforms import PLATFORMS, PLATFORM_CHOICES, resolve as resolve_platform
-from _common.probe import SILENT_PEAK_DBFS
+from _common.probe import SILENT_PEAK_DBFS, square_pixels, square_pixels_note
+from _common import still_input_args  # noqa: E402
 from _common import (add_common, apply_common, aac_args, default_output, die, draws_chroma, emit, ffmpeg_base,
                      info, load_brand, pad_filters, probe, rgb_chain_to_bt709, run, STATE, validate_color, video_args, X264_PRESETS,
-                     fmt_secs, child_args, run_tool, measured_level_dbfs)
-from _common.emit import absorb_stage_hw, warn
+                     fmt_secs, child_args, run_tool, measured_level_dbfs, sibling_temp)
+from _common.emit import absorb_stage_hw, note, warn
 
 WAVEFORM_MODES = ["point", "line", "p2p", "cline"]
 
@@ -183,7 +184,12 @@ def main() -> int:
             plate = pad_filters(args.width, args.height,
                                 "blur" if args.image_fit == "blur" else "color",
                                 args.background, 20)
-        graph = f"[1:v]{plate},format=rgba[plate];[0:a:{args.audio_stream}]{vf},format=rgba"
+        # a non-square-pixel still is squared at its display aspect before it is fitted (#411)
+        plate_v = probe(args.image, role="input").get("video") if not STATE.dry_run else None
+        plate_sq = square_pixels(plate_v)
+        if plate_sq:
+            note(square_pixels_note(plate_v, "--image"))
+        graph = f"[1:v]{plate_sq + ',' if plate_sq else ''}{plate},format=rgba[plate];[0:a:{args.audio_stream}]{vf},format=rgba"
         if args.opacity < 1.0:
             graph += f",colorchannelmixer=aa={args.opacity:g}"
         # the composite is RGBA (the plate is): converted to the BT.709 the output is tagged with,
@@ -207,14 +213,15 @@ def main() -> int:
     if args.title or args.srt or args.text:
         # the visualisation is rendered first and the label/captions are drawn on it by the tools
         # that own those code paths, so neither is re-implemented here
-        stem, ext = os.path.splitext(output)
-        render = stem + "_vis" + ext
+        # hidden, pid-named siblings (#408): a fixed `<out>_vis` name was removed afterwards
+        # whoever had written it, the input or another file of the caller's included
+        render = sibling_temp(output, "vis")
     cmd = ffmpeg_base() + ["-i", args.input]
     if args.image:
         # -framerate before -loop: a looped still defaults to 25 fps, and overlay takes its rate
         # from the FIRST input -- so without this the file came out 25 fps however loud --fps or
         # --platform said otherwise (the colour-plate path never had the bug: `color=` carries r=).
-        cmd += ["-framerate", f"{args.fps:g}", "-loop", "1", "-i", args.image]
+        cmd += ["-framerate", f"{args.fps:g}", "-loop", "1", *still_input_args(args.image), "-i", args.image]
     cmd += ["-filter_complex", vf]
     if args.image:
         cmd += ["-map", "[v]"]
@@ -227,15 +234,20 @@ def main() -> int:
     if meta.get("duration"):
         cmd += ["-t", f"{float(meta['duration']):.3f}"]
     cmd += ["-shortest", render]
-    stem, ext = os.path.splitext(output)
-    intermediates = [p for p in (render, stem + "_titled" + ext) if p != output]
+    titled_tmp = sibling_temp(output, "titled")
+    intermediates = [p for p in (render, titled_tmp) if p != output]
     try:
         run(cmd)
 
         stages = ["waveform"]
         current = render
         if args.title:
-            titled = stem + "_titled" + ext if (args.srt or args.text) else output
+            titled = titled_tmp if (args.srt or args.text) else output
+            if titled == titled_tmp:
+                # graphics.py's libass route writes <its output stem>_gfx.ass beside it (#413 review)
+                titled_ass = os.path.splitext(titled)[0] + "_gfx.ass"
+                if not os.path.exists(titled_ass):
+                    intermediates.append(titled_ass)
             # a dry run rendered no _vis file for graphics.py to probe, so its default --end (the
             # clip's end) came out 0 and the plan failed: give it the source's duration, which
             # is what -t caps the render at. A real run is unchanged.