Duck soundtrack when using clip audio

movmaker.py

@@ -50,6 +50,11 @@ DEFAULT_AUDIO_FADE_IN = 2.0
 DEFAULT_AUDIO_FADE_OUT = 10.0
 DEFAULT_VIDEO_FADE = 1.0
 INTRO_VIDEO_FADE_ENABLED = False
+CLIP_AUDIO_FADE = 0.75
+CLIP_AUDIO_SIDECHAIN_THRESHOLD = 0.04
+CLIP_AUDIO_SIDECHAIN_RATIO = 8.0
+CLIP_AUDIO_SIDECHAIN_ATTACK = 5.0
+CLIP_AUDIO_SIDECHAIN_RELEASE = 250.0
 JAMENDO_ENV = Path.home() / ".config" / "movmaker" / "jamendo.env"
 LOG_PATH: Path | None = None
 
@@ -1109,7 +1114,7 @@ def build_video(
         f"fontsize={stamp_size}:fontcolor=white:borderw=1:bordercolor=black[{final_label}]"
     )
 
-    audio_mix_labels: list[str] = []
+    bg_audio_label: str | None = None
     if audio:
         audio_index = len(clips)
         audio_fade_out_start = max(0.0, total_duration - audio_fade)
@@ -1117,7 +1122,8 @@ def build_video(
             f"[{audio_index}:a:0]atrim=0:{total_duration:.3f},asetpts=PTS-STARTPTS,"
             f"afade=t=in:st=0:d={audio_fade_in:.3f},afade=t=out:st={audio_fade_out_start:.3f}:d={audio_fade:.3f}[bg_audio]"
         )
-        audio_mix_labels.append("[bg_audio]")
+        bg_audio_label = "bg_audio"
+    clip_audio_labels: list[str] = []
     clip_audio_input = clip_audio_input_start
     for idx, clip_audio in enumerate(clip_audio_paths):
         if clip_audio is None:
@@ -1138,7 +1144,7 @@ def build_video(
         if segment <= 0:
             clip_audio_input += 1
             continue
-        fade_len = min(fade, segment / 2)
+        fade_len = min(CLIP_AUDIO_FADE, segment / 2)
         delay_ms = max(0, int(start * 1000))
         label = f"clip_audio_{idx}"
         filters.append(
@@ -1146,10 +1152,38 @@ def build_video(
             f"afade=t=in:st=0:d={fade_len:.3f},afade=t=out:st={max(0.0, segment - fade_len):.3f}:d={fade_len:.3f},"
             f"adelay={delay_ms}:all=1[{label}]"
         )
-        audio_mix_labels.append(f"[{label}]")
+        clip_audio_labels.append(label)
         clip_audio_input += 1
-    if audio_mix_labels:
-        filters.append("".join(audio_mix_labels) + f"amix=inputs={len(audio_mix_labels)}:duration=longest:normalize=0[aout]")
+    clip_mix_label: str | None = None
+    if clip_audio_labels:
+        if len(clip_audio_labels) == 1:
+            clip_mix_label = clip_audio_labels[0]
+        else:
+            inputs = "".join(f"[{name}]" for name in clip_audio_labels)
+            filters.append(inputs + f"amix=inputs={len(clip_audio_labels)}:duration=longest:normalize=0[clip_mix]")
+            clip_mix_label = "clip_mix"
+
+    final_bg_label = bg_audio_label
+    final_clip_label = clip_mix_label
+
+    if final_bg_label and final_clip_label:
+        sc_label = f"{final_clip_label}_sc"
+        mix_label = f"{final_clip_label}_main"
+        filters.append(f"[{final_clip_label}]asplit=2[{sc_label}][{mix_label}]")
+        filters.append(
+            f"[{final_bg_label}][{sc_label}]sidechaincompress=threshold={CLIP_AUDIO_SIDECHAIN_THRESHOLD}:"
+            f"ratio={CLIP_AUDIO_SIDECHAIN_RATIO}:attack={CLIP_AUDIO_SIDECHAIN_ATTACK}:"
+            f"release={CLIP_AUDIO_SIDECHAIN_RELEASE}[bg_ducked]"
+        )
+        final_bg_label = "bg_ducked"
+        final_clip_label = mix_label
+
+    if final_bg_label and final_clip_label:
+        filters.append(f"[{final_bg_label}][{final_clip_label}]amix=inputs=2:duration=longest:normalize=0[aout]")
+    elif final_bg_label:
+        filters.append(f"[{final_bg_label}]anull[aout]")
+    elif final_clip_label:
+        filters.append(f"[{final_clip_label}]anull[aout]")
 
     cmd += ["-filter_complex", ";".join(filters), "-map", f"[{final_label}]"]
     metadata = {
@@ -1168,7 +1202,7 @@ def build_video(
         if value:
             cmd += ["-metadata", f"{key}={value}"]
 
-    if audio_mix_labels:
+    if final_bg_label or final_clip_label:
         cmd += ["-map", "[aout]", "-c:a", "aac", "-b:a", "192k", "-shortest"]
     else:
         cmd += ["-an"]