feat: completed with all features

video-generator/scripts/make_voiceover_macos.py

@@ -0,0 +1,107 @@
+import argparse
+import os
+import re
+import subprocess
+from pathlib import Path
+
+SRT_BLOCK_RE = re.compile(
+    r"\s*(\d+)\s+([\d:,]+)\s+-->\s+([\d:,]+)\s+(.+?)(?=\n\s*\d+\s+\d\d:|\Z)",
+    re.S,
+)
+
+
+def to_seconds(timestamp: str) -> float:
+    h, m, rest = timestamp.split(":")
+    s, ms = rest.split(",")
+    return int(h) * 3600 + int(m) * 60 + int(s) + int(ms) / 1000
+
+
+def read_srt(path: Path) -> list[tuple[float, float, str]]:
+    text = path.read_text(encoding="utf-8-sig")
+    captions = []
+    for _, start, end, body in SRT_BLOCK_RE.findall(text):
+        clean_text = " ".join(line.strip() for line in body.strip().splitlines())
+        captions.append((to_seconds(start), to_seconds(end), clean_text))
+    if not captions:
+        raise ValueError(f"No captions found in {path}")
+    return captions
+
+
+def resolve_path(value: str) -> Path:
+    return Path(value).expanduser().resolve()
+
+
+def media_duration(path: Path) -> float:
+    result = subprocess.run(
+        ["ffprobe", "-v", "error", "-show_entries", "format=duration", "-of", "default=noprint_wrappers=1:nokey=1", str(path)],
+        check=True,
+        capture_output=True,
+        text=True,
+    )
+    return float(result.stdout.strip())
+
+
+def run(command: list[str]) -> None:
+    print("Running:", " ".join(command))
+    subprocess.run(command, check=True)
+
+
+def main() -> None:
+    parser = argparse.ArgumentParser(description="Create a timed voiceover WAV/MP3 from an SRT file on macOS.")
+    parser.add_argument("--srt", required=True, help="Input SRT file")
+    parser.add_argument("--voice", default="Daniel", help="macOS voice name, e.g. Daniel or Reed (English (US))")
+    parser.add_argument("--rate", type=int, default=185, help="Speech rate. Higher is faster. Example: 180-200")
+    parser.add_argument("--wav", required=True, help="Output WAV file")
+    parser.add_argument("--mp3", default=None, help="Optional output MP3 file")
+    args = parser.parse_args()
+
+    srt_path = resolve_path(args.srt)
+    wav_output = resolve_path(args.wav)
+    mp3_output = resolve_path(args.mp3) if args.mp3 else None
+    work_dir = Path(__file__).resolve().parent / "voice_parts"
+
+    if not srt_path.exists():
+        raise FileNotFoundError(f"SRT not found: {srt_path}")
+
+    work_dir.mkdir(parents=True, exist_ok=True)
+    wav_output.parent.mkdir(parents=True, exist_ok=True)
+    if mp3_output:
+        mp3_output.parent.mkdir(parents=True, exist_ok=True)
+
+    captions = read_srt(srt_path)
+    segment_files = []
+
+    for i, (start, end, text) in enumerate(captions, 1):
+        duration = end - start
+        raw_aiff = work_dir / f"raw_{i:03d}.aiff"
+        segment_wav = work_dir / f"seg_{i:03d}.wav"
+
+        run(["say", "-v", args.voice, "-r", str(args.rate), "-o", str(raw_aiff), text])
+
+        spoken_duration = media_duration(raw_aiff)
+        if spoken_duration > duration:
+            print(f"Warning: caption {i} speech is {spoken_duration:.2f}s but slot is {duration:.2f}s; it will be trimmed.")
+
+        run([
+            "ffmpeg", "-y",
+            "-i", str(raw_aiff),
+            "-af", f"apad,atrim=0:{duration}",
+            "-ar", "44100",
+            "-ac", "2",
+            str(segment_wav),
+        ])
+        segment_files.append(segment_wav)
+
+    concat_list = work_dir / "list.txt"
+    concat_list.write_text("".join(f"file '{os.path.abspath(file)}'\n" for file in segment_files), encoding="utf-8")
+
+    run(["ffmpeg", "-y", "-f", "concat", "-safe", "0", "-i", str(concat_list), "-c:a", "pcm_s16le", str(wav_output)])
+    print(f"Done: {wav_output}")
+
+    if mp3_output:
+        run(["ffmpeg", "-y", "-i", str(wav_output), "-codec:a", "libmp3lame", "-b:a", "192k", str(mp3_output)])
+        print(f"Done: {mp3_output}")
+
+
+if __name__ == "__main__":
+    main()