Removed extra scene-detection step

MkvOpusEnc.py

@@ -85,7 +85,9 @@ def convert_audio_track(stream_index, channels, temp_dir, source_file, should_do
     # Step 3: Encode to Opus with the correct bitrate
     bitrate = "192k"  # Fallback
     
-    if final_channels == 2:
+    if final_channels == 1:
+        bitrate = "64k"
+    elif final_channels == 2:
         bitrate = "128k"
     elif final_channels == 6:
         bitrate = "256k"

anime_audio_encoder.py

@@ -19,6 +19,21 @@ DIR_CONV_LOGS = Path("conv_logs") # Directory for conversion logs
 
 REMUX_CODECS = {"aac", "opus"}  # Using a set for efficient lookups
 
+SVT_AV1_PARAMS = {
+    "speed": "slower",  # "slower", "slow", "medium", "fast", "faster"
+    "quality": "medium",  # "higher", "high", "medium", "low", "lower"
+    "film-grain": 6,
+    "color-primaries": 1,
+    "transfer-characteristics": 1,
+    "matrix-coefficients": 1,
+    "scd": 0,  # Scene change detection OFF for Av1an use
+    "keyint": 0,  # Keyframe interval, 0 disables automatic keyframes placement at a constant interval
+    "lp": 2, # Level of parallelism
+    "auto-tiling": 1,  # Auto tiling ON
+    "tune": 1,  # 0 = VQ, 1 = PSNR, 2 = SSIM
+    "progress": 2,  # Detailed progress output
+}
+
 def check_tools():
     for tool in REQUIRED_TOOLS:
         if shutil.which(tool) is None:
@@ -40,7 +55,7 @@ def convert_audio_track(index, ch, lang, audio_temp_dir, source_file, should_dow
 
     print(f"    - Extracting Audio Track #{index} to FLAC...")
     ffmpeg_args = [
-        "ffmpeg", "-v", "quiet", "-stats", "-y", "-i", str(source_file), "-map", f"0:{index}"
+        "ffmpeg", "-v", "quiet", "-stats", "-y", "-i", str(source_file), "-map", f"0:{index}", "-map_metadata", "-1"
     ]
     if should_downmix and ch >= 6:
         if ch == 6:
@@ -68,14 +83,16 @@ def convert_audio_track(index, ch, lang, audio_temp_dir, source_file, should_dow
     else:
         # Not downmixing (or source is already stereo or less).
         # Base bitrate on the source channel count.
-        if ch == 2:      # Stereo
+        if ch == 1:      # Mono
+            bitrate = "64k"
+        elif ch == 2:    # Stereo
             bitrate = "128k"
         elif ch == 6:    # 5.1 Surround
             bitrate = "256k"
         elif ch == 8:    # 7.1 Surround
             bitrate = "384k"
-        else:            # Mono or other layouts
-            bitrate = "96k" # A sensible default for mono.
+        else:            # Other layouts
+            bitrate = "96k" # A sensible default for other/uncommon layouts.
 
     print(f"    - Encoding Audio Track #{index} to Opus at {bitrate}...")
     run_cmd([
@@ -86,7 +103,6 @@ def convert_audio_track(index, ch, lang, audio_temp_dir, source_file, should_dow
 def convert_video(source_file_base, source_file_full, is_vfr, target_cfr_fps_for_handbrake, autocrop_filter=None):
     print("  --- Starting Video Processing ---")
     # source_file_base is file_path.stem (e.g., "my.anime.episode.01")
-    scene_file = Path(f"{source_file_base}.txt")
     vpy_file = Path(f"{source_file_base}.vpy")
     ut_video_file = Path(f"{source_file_base}.ut.mkv")
     encoded_video_file = Path(f"temp-{source_file_base}.mkv")
@@ -159,39 +175,19 @@ clip.set_output()
     with vpy_file.open("w", encoding="utf-8") as f:
         f.write(vpy_script_content)
 
-    if not scene_file.exists():
-        print("    - Performing scene detection with av1an...")
-        av1an_sc_args = [
-            "av1an", "-i", str(vpy_file), "-s", str(scene_file), "--sc-only", "--verbose"
-        ]
-        run_cmd(av1an_sc_args)
-    else:
-        print("    - Found existing scene file, skipping detection.")
-
     print("    - Starting AV1 encode with av1an (this will take a long time)...")
     total_cores = os.cpu_count() or 4 # Fallback if cpu_count is None
-    workers = max(total_cores - 2, 1) # Ensure at least 1 worker
-    print(f"    - Using {workers} workers for av1an (Total Cores: {total_cores}).")
+    workers = max(1, (total_cores // 2) - 1) # Half the cores minus one, with a minimum of 1 worker.
+    print(f"    - Using {workers} workers for av1an (Total Cores: {total_cores}, Logic: (Cores/2)-1).")
 
-    svt_av1_params = {
-        "preset": 2,
-        "crf": 27,
-        "film-grain": 6,
-        "lp": 1,
-        "tune": 1,
-        "keyint": -1,
-        "color-primaries": 1,
-        "transfer-characteristics": 1,
-        "matrix-coefficients": 1,
-    }
     # Create the parameter string for av1an's -v option, which expects a single string.
-    av1an_video_params_str = " ".join([f"--{key} {value}" for key, value in svt_av1_params.items()])
+    av1an_video_params_str = " ".join([f"--{key} {value}" for key, value in SVT_AV1_PARAMS.items()])
     print(f"    - Using SVT-AV1 parameters: {av1an_video_params_str}")
 
     av1an_enc_args = [
-        "av1an", "-i", str(vpy_file), "-o", str(encoded_video_file), "-s", str(scene_file), "-n",
+        "av1an", "-i", str(vpy_file), "-o", str(encoded_video_file), "-n",
         "-e", "svt-av1", "--resume", "--sc-pix-format", "yuv420p", "-c", "mkvmerge",
-        "--set-thread-affinity", "1", "--pix-format", "yuv420p10le", "--force",
+        "--set-thread-affinity", "2", "--pix-format", "yuv420p10le", "--force",
         "-w", str(workers),
         "-v", av1an_video_params_str
     ]
@@ -270,17 +266,21 @@ def _snap_to_known_ar_cropdetect(w, h, x, y, video_w, video_h, tolerance=0.03):
         new_h = round(video_w / best_match['ratio'])
         if new_h % 8 != 0:
             new_h = new_h + (8 - (new_h % 8))
+        new_h = min(new_h, video_h)
         new_y = round((video_h - new_h) / 2)
         if new_y % 2 != 0:
             new_y -= 1
+        new_y = max(0, new_y)
         return f"crop={video_w}:{new_h}:0:{new_y}", best_match['name']
     if abs(h - video_h) < 16:
         new_w = round(video_h * best_match['ratio'])
         if new_w % 8 != 0:
             new_w = new_w + (8 - (new_w % 8))
+        new_w = min(new_w, video_w)
         new_x = round((video_w - new_w) / 2)
         if new_x % 2 != 0:
             new_x -= 1
+        new_x = max(0, new_x)
         return f"crop={new_w}:{video_h}:{new_x}:0", best_match['name']
     return f"crop={w}:{h}:{x}:{y}", None
 
@@ -433,8 +433,17 @@ def detect_autocrop_filter(input_file, significant_crop_threshold=5.0, min_crop=
         return None
     return _analyze_video_cropdetect(input_file, duration, width, height, max(1, os.cpu_count() // 2), significant_crop_threshold, min_crop, debug)
 
-def main(no_downmix=False, autocrop=False):
+def main(no_downmix=False, autocrop=False, speed=None, quality=None, grain=None):
     check_tools()
+
+    # Override default SVT-AV1 params if provided via command line
+    if speed:
+        SVT_AV1_PARAMS["speed"] = speed
+    if quality:
+        SVT_AV1_PARAMS["quality"] = quality
+    if grain is not None:
+        SVT_AV1_PARAMS["film-grain"] = grain
+
     current_dir = Path(".")
     files_to_process = sorted(
         f for f in current_dir.glob("*.mkv")
@@ -638,7 +647,6 @@ def main(no_downmix=False, autocrop=False):
 
                 print("Cleaning up persistent video temporary files (after successful processing)...")
                 video_temp_files_on_success = [
-                    current_dir / f"{file_path.stem}.txt",
                     current_dir / f"{file_path.stem}.vpy",
                     current_dir / f"{file_path.stem}.ut.mkv",
                     current_dir / f"temp-{file_path.stem}.mkv", # This is encoded_video_file
@@ -720,5 +728,8 @@ if __name__ == "__main__":
     parser = argparse.ArgumentParser(description="Batch-process MKV files with resumable video encoding, audio downmixing, per-file logging, and optional autocrop.")
     parser.add_argument("--no-downmix", action="store_true", help="Preserve original audio channel layout.")
     parser.add_argument("--autocrop", action="store_true", help="Automatically detect and crop black bars from video using cropdetect.")
+    parser.add_argument("--speed", type=str, help="Set the encoding speed. Possible values: slower, slow, medium, fast, faster.")
+    parser.add_argument("--quality", type=str, help="Set the encoding quality. Possible values: lowest, low, medium, high, higher.")
+    parser.add_argument("--grain", type=int, help="Set the film-grain value (number). Adjusts the film grain synthesis level.")
     args = parser.parse_args()
-    main(no_downmix=args.no_downmix, autocrop=args.autocrop)
+    main(no_downmix=args.no_downmix, autocrop=args.autocrop, speed=args.speed, quality=args.quality, grain=args.grain)

tv_audio_encoder.py

@@ -90,14 +90,16 @@ def convert_audio_track(index, ch, lang, audio_temp_dir, source_file, should_dow
     else:
         # Not downmixing (or source is already stereo or less).
         # Base bitrate on the source channel count.
-        if ch == 2:      # Stereo
+        if ch == 1:      # Mono
+            bitrate = "64k"
+        elif ch == 2:    # Stereo
             bitrate = "128k"
         elif ch == 6:    # 5.1 Surround
             bitrate = "256k"
         elif ch == 8:    # 7.1 Surround
             bitrate = "384k"
-        else:            # Mono or other layouts
-            bitrate = "96k" # A sensible default for mono.
+        else:            # Other layouts
+            bitrate = "96k" # A sensible default for other/uncommon layouts.
 
     print(f"    - Encoding Audio Track #{index} to Opus at {bitrate}...")
     run_cmd([