Ading cleanup and small improvements

gnommo/cli.py

@@ -2185,22 +2185,54 @@ def _prune_manifest(
         if (key.lower() if lowercase_keys else key) not in keep_ids
     ]
 
-    if not to_remove:
+    # Case-collision duplicates: when keys are case-insensitive (videos), several
+    # keys can share the same lowercase form (e.g. "KnightRotating" and
+    # "knightrotating"). Markers resolve by lowercased id, so only the exact
+    # lowercase key is ever reachable — the other case-variants are dead weight.
+    # Collapse each surviving collision group to its lowercase canonical.
+    dup_remove: list[tuple[str, str]] = []  # (removed_key, kept_canonical_key)
+    if lowercase_keys:
+        remove_set = set(to_remove)
+        groups: dict[str, list[str]] = {}
+        for key in data:
+            if key in remove_set:
+                continue
+            groups.setdefault(key.lower(), []).append(key)
+        for low, keys in groups.items():
+            if len(keys) < 2:
+                continue
+            # Keep the already-lowercase key (what the renderer resolves); else
+            # the first-seen, so at least one entry survives.
+            canonical = next((k for k in keys if k == low), keys[0])
+            for key in keys:
+                if key != canonical:
+                    dup_remove.append((key, canonical))
+
+    if not to_remove and not dup_remove:
         if verbose:
             print(f"  {path.name}: all {len(data)} {label} entries in use.")
         return 0
 
-    print(f"  {path.name}: removing {len(to_remove)} unused {label} entr"
-          f"{'y' if len(to_remove) == 1 else 'ies'}:")
-    for key in to_remove:
-        print(f"    - {key}")
+    if to_remove:
+        print(f"  {path.name}: removing {len(to_remove)} unused {label} entr"
+              f"{'y' if len(to_remove) == 1 else 'ies'}:")
+        for key in to_remove:
+            print(f"    - {key}")
+
+    if dup_remove:
+        print(f"  {path.name}: removing {len(dup_remove)} case-duplicate {label} entr"
+              f"{'y' if len(dup_remove) == 1 else 'ies'}:")
+        for key, canonical in dup_remove:
+            print(f"    - {key}  (duplicate of '{canonical}')")
 
     if not dry_run:
         for key in to_remove:
             del data[key]
+        for key, _ in dup_remove:
+            del data[key]
         _write_json_preserve(path, data)
 
-    return len(to_remove)
+    return len(to_remove) + len(dup_remove)
 
 
 def _prune_narration(project_path: Path, verbose: bool, dry_run: bool) -> int:

gnommo/renderer.py

@@ -1019,7 +1019,7 @@ def build_filter_complex(
         zoomed_width = int(cut_width * zoom)
         zoomed_height = int(cut_height * zoom)
 
-        if not plan.narration_pauses or narration_concat:
+        if not plan.narration_pauses:
             video_label = f"av{i}"
             filters.append(
                 f"{narr_src}fps={plan.config.fps},setpts=PTS-STARTPTS,"
@@ -1035,9 +1035,19 @@ def build_filter_complex(
             )
             current_label = next_label
         else:
+            # Freeze-splice: cut the narration video at each pause and re-time the
+            # remainder so it resumes exactly where it stopped (the cutscene owns
+            # the frame during the gap). narr_src may be the concat filter output
+            # [narrsrc_v], which is single-consumer, so split it into one branch
+            # per segment before trimming.
             segments = _build_narration_segments(
                 plan.narration_pauses, plan.total_duration
             )
+            split_labels = [f"nvsplit{i}_{k}" for k in range(len(segments))]
+            filters.append(
+                f"{narr_src}split={len(segments)}"
+                + "".join(f"[{lbl}]" for lbl in split_labels)
+            )
 
             for seg_idx, (src_start, src_end, out_start, out_end) in enumerate(
                 segments
@@ -1045,7 +1055,7 @@ def build_filter_complex(
                 seg_label = f"av{i}_seg{seg_idx}"
                 pts_offset = out_start
                 filters.append(
-                    f"{narr_src}trim={src_start:.3f}:{src_end:.3f},"
+                    f"[{split_labels[seg_idx]}]trim={src_start:.3f}:{src_end:.3f},"
                     f"setpts=PTS-STARTPTS+{pts_offset:.3f}/TB,"
                     f"format=yuva444p10le,"
                     f"scale={zoomed_width}:{zoomed_height}:force_original_aspect_ratio=increase,"
@@ -1275,7 +1285,7 @@ def build_filter_complex(
             plan.narration_end_time if plan.outro_events else plan.total_duration
         )
 
-        if not plan.narration_pauses or narration_concat:
+        if not plan.narration_pauses:
             # Simple case: trim main audio to end before outro (with optional channel and volume filters)
             filter_parts = []
             if channel_filter:
@@ -1299,8 +1309,17 @@ def build_filter_complex(
             else:
                 audio_labels_to_mix.append(f"{_main_aud}")
         else:
-            # Complex case: segment the narration audio for pauses
+            # Freeze-splice: cut the narration audio at each pause and delay the
+            # remainder so it resumes on the exact sample it stopped on (the
+            # cutscene's own audio fills the gap). _main_aud may be the concat
+            # output [narrsrc_a], which is single-consumer, so asplit it into one
+            # branch per segment first.
             segments = _build_narration_segments(plan.narration_pauses, audio_end_time)
+            asplit_labels = [f"narr_aud_split{k}" for k in range(len(segments))]
+            filters.append(
+                f"{_main_aud}asplit={len(segments)}"
+                + "".join(f"[{lbl}]" for lbl in asplit_labels)
+            )
             for seg_idx, (src_start, src_end, out_start, out_end) in enumerate(
                 segments
             ):
@@ -1317,7 +1336,7 @@ def build_filter_complex(
                 if volume_filter:
                     filter_parts.append(volume_filter)
                 filters.append(
-                    f"{_main_aud}{','.join(filter_parts)}[{seg_label}]"
+                    f"[{asplit_labels[seg_idx]}]{','.join(filter_parts)}[{seg_label}]"
                 )
                 audio_labels_to_mix.append(f"[{seg_label}]")