diff --git a/anime_svtav1_encoder.py b/anime_audio_encoder.py similarity index 55% rename from anime_svtav1_encoder.py rename to anime_audio_encoder.py index 1d2a168..f09b70e 100644 --- a/anime_svtav1_encoder.py +++ b/anime_audio_encoder.py @@ -5,66 +5,47 @@ import subprocess import shutil import tempfile import json -import re +import re # Added for VFR frame rate parsing from datetime import datetime from pathlib import Path -# List of required external tools REQUIRED_TOOLS = [ "ffmpeg", "ffprobe", "mkvmerge", "mkvpropedit", - "sox", "opusenc", "mediainfo", "SvtAv1EncApp2", "HandBrakeCLI" + "sox", "opusenc", "mediainfo", "av1an", "HandBrakeCLI" # Added HandBrakeCLI ] +DIR_COMPLETED = Path("completed") +DIR_ORIGINAL = Path("original") +DIR_CONV_LOGS = Path("conv_logs") # Directory for conversion logs -# Set the SVT-AV1 encoder binary here for easy switching -SVTAV1_BIN = "SvtAv1EncApp2" +REMUX_CODECS = {"aac", "opus"} # Using a set for efficient lookups -# SVT-AV1-Essential (nekotrix) encoding parameters (edit here as needed) SVT_AV1_PARAMS = { - "speed": "slow", # "slower", "slow", "medium", "fast", "faster" - "quality": "high", # "higher", "high", "medium", "low", "lower" + "speed": "slow", # "slower", "slow", "medium", "fast", "faster" + "quality": "high", # "higher", "high", "medium", "low", "lower" "film-grain": 6, "color-primaries": 1, "transfer-characteristics": 1, "matrix-coefficients": 1, - "scd": 1, # Scene change detection ON (recommended) - "auto-tiling": 1, # Auto tiling ON (recommended) - "tune": 1, # 0 = VQ, 1 = PSNR, 2 = SSIM - "progress": 2, # Detailed progress output + "scd": 1, # Scene change detection ON + "auto-tiling": 1, # Auto tiling ON + "tune": 1, # 0 = VQ, 1 = PSNR, 2 = SSIM + "progress": 2, # Detailed progress output } -SVT_AV1_PARAMS_LIST = [] -for key, value in SVT_AV1_PARAMS.items(): - SVT_AV1_PARAMS_LIST.extend([f"--{key}", str(value)]) - -DIR_COMPLETED = Path("completed") -DIR_ORIGINAL = Path("original") -DIR_CONV_LOGS = Path("conv_logs") - -REMUX_CODECS = {"aac", "opus"} def check_tools(): - """Check if all required tools are available in PATH.""" for tool in REQUIRED_TOOLS: if shutil.which(tool) is None: print(f"Required tool '{tool}' not found in PATH.") sys.exit(1) def run_cmd(cmd, capture_output=False, check=True): - """Run a subprocess command, optionally capturing output.""" - try: - if capture_output: - result = subprocess.run(cmd, stdout=subprocess.PIPE, stderr=subprocess.PIPE, check=check, text=True) - return result.stdout - else: - subprocess.run(cmd, check=check) - except subprocess.CalledProcessError as e: - print(f"Command failed: {' '.join(map(str, cmd))}") - print(f"Return code: {e.returncode}") - if capture_output: - print(f"Output: {e.output}") - raise + if capture_output: + result = subprocess.run(cmd, stdout=subprocess.PIPE, stderr=subprocess.PIPE, check=check, text=True) + return result.stdout + else: + subprocess.run(cmd, check=check) def convert_audio_track(index, ch, lang, audio_temp_dir, source_file, should_downmix): - """Extract, normalize, and encode an audio track to Opus.""" audio_temp_path = Path(audio_temp_dir) temp_extracted = audio_temp_path / f"track_{index}_extracted.flac" temp_normalized = audio_temp_path / f"track_{index}_normalized.flac" @@ -74,13 +55,12 @@ def convert_audio_track(index, ch, lang, audio_temp_dir, source_file, should_dow ffmpeg_args = [ "ffmpeg", "-v", "quiet", "-stats", "-y", "-i", str(source_file), "-map", f"0:{index}" ] - # Downmix if needed if should_downmix and ch >= 6: if ch == 6: ffmpeg_args += ["-af", "pan=stereo|c0=c2+0.30*c0+0.30*c4|c1=c2+0.30*c1+0.30*c5"] elif ch == 8: ffmpeg_args += ["-af", "pan=stereo|c0=c2+0.30*c0+0.30*c4+0.30*c6|c1=c2+0.30*c1+0.30*c5+0.30*c7"] - else: + else: # Other multi-channel (e.g. 7ch, 10ch) ffmpeg_args += ["-ac", "2"] ffmpeg_args += ["-c:a", "flac", str(temp_extracted)] run_cmd(ffmpeg_args) @@ -90,19 +70,25 @@ def convert_audio_track(index, ch, lang, audio_temp_dir, source_file, should_dow "sox", str(temp_extracted), str(temp_normalized), "-S", "--temp", str(audio_temp_path), "--guard", "gain", "-n" ]) - # Set bitrate based on channel count and downmixing + # Set bitrate based on the final channel count of the Opus file. + # If we are downmixing, the result is stereo. + # If not, the result has the original channel count. is_being_downmixed = should_downmix and ch >= 6 + if is_being_downmixed: + # Downmixing from 5.1 or 7.1 results in a stereo track. bitrate = "128k" else: - if ch == 2: + # Not downmixing (or source is already stereo or less). + # Base bitrate on the source channel count. + if ch == 2: # Stereo bitrate = "128k" - elif ch == 6: + elif ch == 6: # 5.1 Surround bitrate = "256k" - elif ch == 8: + elif ch == 8: # 7.1 Surround bitrate = "384k" - else: - bitrate = "96k" + else: # Mono or other layouts + bitrate = "96k" # A sensible default for mono. print(f" - Encoding Audio Track #{index} to Opus at {bitrate}...") run_cmd([ @@ -111,30 +97,28 @@ def convert_audio_track(index, ch, lang, audio_temp_dir, source_file, should_dow return final_opus def convert_video(source_file_base, source_file_full, is_vfr, target_cfr_fps_for_handbrake, autocrop_filter=None): - """ - Convert video to AV1 using ffmpeg and SvtAv1EncApp. - If VFR, use HandBrakeCLI to convert to CFR first. - """ print(" --- Starting Video Processing ---") - encoded_video_file = Path(f"temp-{source_file_base}.ivf") + scene_file = Path(f"{source_file_base}.txt") # Not used anymore, but kept for compatibility + ut_video_file = Path(f"{source_file_base}.ut.mkv") + encoded_video_file = Path(f"temp-{source_file_base}.mkv") handbrake_cfr_intermediate_file = None - current_input = Path(source_file_full) - # VFR-to-CFR conversion using HandBrakeCLI if needed + current_input_for_utvideo = Path(source_file_full) + if is_vfr and target_cfr_fps_for_handbrake: print(f" - Source is VFR. Converting to CFR ({target_cfr_fps_for_handbrake}) with HandBrakeCLI...") handbrake_cfr_intermediate_file = Path(f"{source_file_base}.cfr_temp.mkv") handbrake_args = [ - "HandBrakeCLI", - "--input", str(source_file_full), + "HandBrakeCLI", + "--input", str(source_file_full), "--output", str(handbrake_cfr_intermediate_file), - "--cfr", + "--cfr", "--rate", str(target_cfr_fps_for_handbrake), "--encoder", "x264_10bit", "--quality", "0", "--encoder-preset", "superfast", "--encoder-tune", "fastdecode", - "--audio", "none", + "--audio", "none", "--subtitle", "none", "--crop-mode", "none" ] @@ -143,64 +127,65 @@ def convert_video(source_file_base, source_file_full, is_vfr, target_cfr_fps_for run_cmd(handbrake_args) if handbrake_cfr_intermediate_file.exists() and handbrake_cfr_intermediate_file.stat().st_size > 0: print(f" - HandBrake VFR to CFR conversion successful: {handbrake_cfr_intermediate_file}") - current_input = handbrake_cfr_intermediate_file + current_input_for_utvideo = handbrake_cfr_intermediate_file else: - print(f" - Warning: HandBrakeCLI VFR-to-CFR conversion failed or produced an empty file. Proceeding with original source.") + print(f" - Warning: HandBrakeCLI VFR-to-CFR conversion failed or produced an empty file. Proceeding with original source for UTVideo.") handbrake_cfr_intermediate_file = None except subprocess.CalledProcessError as e: print(f" - Error during HandBrakeCLI execution: {e}") - print(f" - Proceeding with original source.") + print(f" - Proceeding with original source for UTVideo.") handbrake_cfr_intermediate_file = None - print(" - Starting AV1 encode with ffmpeg and SvtAv1EncApp...") - - # Probe video info for width, height, and frame rate + print(" - Creating UTVideo intermediate file (overwriting if exists)...") ffprobe_cmd = [ "ffprobe", "-v", "error", "-select_streams", "v:0", - "-show_entries", "stream=width,height,r_frame_rate", "-of", "json", - str(current_input) + "-show_entries", "stream=codec_name", "-of", "default=noprint_wrappers=1:nokey=1", + str(current_input_for_utvideo) ] - video_info_json = run_cmd(ffprobe_cmd, capture_output=True, check=True) - video_info = json.loads(video_info_json)["streams"][0] - width = video_info["width"] - height = video_info["height"] - frame_rate = video_info["r_frame_rate"] + source_codec = run_cmd(ffprobe_cmd, capture_output=True, check=True).strip() + + video_codec_args = ["-c:v", "utvideo"] + if source_codec == "utvideo" and current_input_for_utvideo == Path(source_file_full): + print(" - Source is already UTVideo. Copying video stream...") + video_codec_args = ["-c:v", "copy"] - # Prepare ffmpeg command for yuv4mpegpipe ffmpeg_args = [ - "ffmpeg", "-hide_banner", "-v", "quiet", "-stats", "-y", "-i", str(current_input), - "-pix_fmt", "yuv420p10le", "-f", "yuv4mpegpipe", "-" + "ffmpeg", "-hide_banner", "-v", "quiet", "-stats", "-y", "-i", str(current_input_for_utvideo), + "-map", "0:v:0", "-map_metadata", "-1", "-map_chapters", "-1", "-an", "-sn", "-dn", ] if autocrop_filter: - match = re.match(r"crop=(\d+):(\d+):(\d+):(\d+)", autocrop_filter) - if match: - width = int(match.group(1)) - height = int(match.group(2)) - ffmpeg_args.insert(-2, "-vf") - ffmpeg_args.insert(-2, autocrop_filter) + ffmpeg_args += ["-vf", autocrop_filter] + ffmpeg_args += video_codec_args + [str(ut_video_file)] + run_cmd(ffmpeg_args) + ut_video_full_path = os.path.abspath(ut_video_file) - svt_enc_args = [ - SVTAV1_BIN, "-i", "stdin", - "--width", str(width), "--height", str(height), - "--fps-num", frame_rate.split('/')[0], "--fps-den", frame_rate.split('/')[1], - "-b", str(encoded_video_file) - ] + SVT_AV1_PARAMS_LIST + # --- SVT-AV1 ENCODING (ffmpeg pipe to SvtAv1EncApp) --- + print(" - Starting AV1 encode with ffmpeg -> SvtAv1EncApp pipe (this will take a long time)...") + svtav1_param_list = [] + for k, v in SVT_AV1_PARAMS.items(): + if isinstance(v, bool): + v = int(v) + svtav1_param_list.append(f"--{k} {v}") + svtav1_param_str = " ".join(svtav1_param_list) + print(f" - Using SvtAv1EncApp parameters: {svtav1_param_str}") - print(f" - FFmpeg command: {' '.join(ffmpeg_args)}") - print(f" - SvtAv1EncApp command: {' '.join(svt_enc_args)}") + # ffmpeg: always output yuv420p10le (10-bit), even if input is 8-bit + ffmpeg_pipe_cmd = [ + "ffmpeg", "-hide_banner", "-v", "quiet", "-stats", "-y", "-i", ut_video_full_path, + "-pix_fmt", "yuv420p10le", "-f", "yuv4mpegpipe", "-an", "-sn", "-dn", "-map", "0:v:0", "-" + ] + svtav1_cmd = [ + "SvtAv1EncApp2", "--input", "-", "--output", str(encoded_video_file) + ] + svtav1_param_list - # Pipe ffmpeg output to SvtAv1EncApp - ffmpeg_process = subprocess.Popen(ffmpeg_args, stdout=subprocess.PIPE) - svt_process = subprocess.Popen(svt_enc_args, stdin=ffmpeg_process.stdout) - - if ffmpeg_process.stdout: - ffmpeg_process.stdout.close() - - svt_process.wait() - ffmpeg_process.wait() - - if svt_process.returncode != 0: - raise subprocess.CalledProcessError(svt_process.returncode, svt_enc_args) + print(f" - Running: {' '.join(ffmpeg_pipe_cmd)} | {' '.join(svtav1_cmd)}") + # Use subprocess.Popen for piping + with subprocess.Popen(ffmpeg_pipe_cmd, stdout=subprocess.PIPE) as ffmpeg_proc: + with subprocess.Popen(svtav1_cmd, stdin=ffmpeg_proc.stdout) as svt_proc: + ffmpeg_proc.stdout.close() # Allow ffmpeg_proc to receive a SIGPIPE if svt_proc exits + svt_proc.communicate() + if svt_proc.returncode != 0: + raise RuntimeError(f"SvtAv1EncApp failed with exit code {svt_proc.returncode}") print(" --- Finished Video Processing ---") return encoded_video_file, handbrake_cfr_intermediate_file @@ -208,6 +193,7 @@ def convert_video(source_file_base, source_file_full, is_vfr, target_cfr_fps_for def is_ffmpeg_decodable(file_path): """Quickly check if ffmpeg can decode the input file.""" try: + # Try to decode a short segment of the first audio stream subprocess.run([ "ffmpeg", "-v", "error", "-i", str(file_path), "-map", "0:a:0", "-t", "1", "-f", "null", "-" ], check=True) @@ -215,10 +201,230 @@ def is_ffmpeg_decodable(file_path): except subprocess.CalledProcessError: return False -# --- Cropdetect logic omitted for brevity (unchanged) --- +# --- CROPDETECT LOGIC FROM cropdetect.py --- +import argparse as _argparse_cropdetect +import multiprocessing as _multiprocessing_cropdetect +from collections import Counter as _Counter_cropdetect + +COLOR_GREEN = "\033[92m" +COLOR_RED = "\033[91m" +COLOR_YELLOW = "\033[93m" +COLOR_RESET = "\033[0m" + +KNOWN_ASPECT_RATIOS = [ + {"name": "HDTV (16:9)", "ratio": 16/9}, + {"name": "Widescreen (Scope)", "ratio": 2.39}, + {"name": "Widescreen (Flat)", "ratio": 1.85}, + {"name": "IMAX Digital (1.90:1)", "ratio": 1.90}, + {"name": "Fullscreen (4:3)", "ratio": 4/3}, + {"name": "IMAX 70mm (1.43:1)", "ratio": 1.43}, +] + +def _check_prerequisites_cropdetect(): + for tool in ['ffmpeg', 'ffprobe']: + if not shutil.which(tool): + print(f"Error: '{tool}' command not found. Is it installed and in your PATH?") + return False + return True + +def _analyze_segment_cropdetect(task_args): + seek_time, input_file, width, height = task_args + ffmpeg_args = [ + 'ffmpeg', '-hide_banner', + '-ss', str(seek_time), + '-i', input_file, '-t', '1', '-vf', 'cropdetect', + '-f', 'null', '-' + ] + result = subprocess.run(ffmpeg_args, capture_output=True, text=True, encoding='utf-8') + if result.returncode != 0: + return [] + crop_detections = re.findall(r'crop=(\d+):(\d+):(\d+):(\d+)', result.stderr) + significant_crops = [] + for w_str, h_str, x_str, y_str in crop_detections: + w, h, x, y = map(int, [w_str, h_str, x_str, y_str]) + significant_crops.append((f"crop={w}:{h}:{x}:{y}", seek_time)) + return significant_crops + +def _snap_to_known_ar_cropdetect(w, h, x, y, video_w, video_h, tolerance=0.03): + if h == 0: return f"crop={w}:{h}:{x}:{y}", None + detected_ratio = w / h + best_match = None + smallest_diff = float('inf') + for ar in KNOWN_ASPECT_RATIOS: + diff = abs(detected_ratio - ar['ratio']) + if diff < smallest_diff: + smallest_diff = diff + best_match = ar + if not best_match or (smallest_diff / best_match['ratio']) >= tolerance: + return f"crop={w}:{h}:{x}:{y}", None + if abs(w - video_w) < 16: + new_h = round(video_w / best_match['ratio']) + if new_h % 8 != 0: + new_h = new_h + (8 - (new_h % 8)) + new_y = round((video_h - new_h) / 2) + if new_y % 2 != 0: + new_y -= 1 + 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_x = round((video_w - new_w) / 2) + if new_x % 2 != 0: + new_x -= 1 + return f"crop={new_w}:{video_h}:{new_x}:0", best_match['name'] + return f"crop={w}:{h}:{x}:{y}", None + +def _cluster_crop_values_cropdetect(crop_counts, tolerance=8): + clusters = [] + temp_counts = crop_counts.copy() + while temp_counts: + center_str, _ = temp_counts.most_common(1)[0] + try: + _, values = center_str.split('='); + cw, ch, cx, cy = map(int, values.split(':')) + except (ValueError, IndexError): + del temp_counts[center_str] + continue + cluster_total_count = 0 + crops_to_remove = [] + for crop_str, count in temp_counts.items(): + try: + _, values = crop_str.split('='); + w, h, x, y = map(int, values.split(':')) + if abs(x - cx) <= tolerance and abs(y - cy) <= tolerance: + cluster_total_count += count + crops_to_remove.append(crop_str) + except (ValueError, IndexError): + continue + if cluster_total_count > 0: + clusters.append({'center': center_str, 'count': cluster_total_count}) + for crop_str in crops_to_remove: + del temp_counts[crop_str] + clusters.sort(key=lambda c: c['count'], reverse=True) + return clusters + +def _parse_crop_string_cropdetect(crop_str): + try: + _, values = crop_str.split('='); + w, h, x, y = map(int, values.split(':')) + return {'w': w, 'h': h, 'x': x, 'y': y} + except (ValueError, IndexError): + return None + +def _calculate_bounding_box_cropdetect(crop_keys): + min_x = min_w = min_y = min_h = float('inf') + max_x = max_w = max_y = max_h = float('-inf') + for key in crop_keys: + parsed = _parse_crop_string_cropdetect(key) + if not parsed: + continue + w, h, x, y = parsed['w'], parsed['h'], parsed['x'], parsed['y'] + min_x = min(min_x, x) + min_y = min(min_y, y) + max_x = max(max_x, x + w) + max_y = max(max_y, y + h) + min_w = min(min_w, w) + min_h = min(min_h, h) + max_w = max(max_w, w) + max_h = max(max_h, h) + if (max_x - min_x) <= 2 and (max_y - min_y) <= 2: + return None + bounding_crop = f"crop={max_x - min_x}:{max_y - min_y}:{min_x}:{min_y}" + return bounding_crop + +def _analyze_video_cropdetect(input_file, duration, width, height, num_workers, significant_crop_threshold, min_crop, debug=False): + num_tasks = num_workers * 4 + segment_duration = max(1, duration // num_tasks) + tasks = [(i * segment_duration, input_file, width, height) for i in range(num_tasks)] + crop_results = [] + with _multiprocessing_cropdetect.Pool(processes=num_workers) as pool: + results_iterator = pool.imap_unordered(_analyze_segment_cropdetect, tasks) + for result in results_iterator: + crop_results.append(result) + all_crops_with_ts = [crop for sublist in crop_results for crop in sublist] + all_crop_strings = [item[0] for item in all_crops_with_ts] + if not all_crop_strings: + return None + crop_counts = _Counter_cropdetect(all_crop_strings) + clusters = _cluster_crop_values_cropdetect(crop_counts) + total_detections = sum(c['count'] for c in clusters) + significant_clusters = [] + for cluster in clusters: + percentage = (cluster['count'] / total_detections) * 100 + if percentage >= significant_crop_threshold: + significant_clusters.append(cluster) + for cluster in significant_clusters: + parsed_crop = _parse_crop_string_cropdetect(cluster['center']) + if parsed_crop: + _, ar_label = _snap_to_known_ar_cropdetect( + parsed_crop['w'], parsed_crop['h'], parsed_crop['x'], parsed_crop['y'], width, height + ) + cluster['ar_label'] = ar_label + else: + cluster['ar_label'] = None + if not significant_clusters: + return None + elif len(significant_clusters) == 1: + dominant_cluster = significant_clusters[0] + parsed_crop = _parse_crop_string_cropdetect(dominant_cluster['center']) + snapped_crop, ar_label = _snap_to_known_ar_cropdetect( + parsed_crop['w'], parsed_crop['h'], parsed_crop['x'], parsed_crop['y'], width, height + ) + parsed_snapped = _parse_crop_string_cropdetect(snapped_crop) + if parsed_snapped and parsed_snapped['w'] == width and parsed_snapped['h'] == height: + return None + else: + return snapped_crop + else: + crop_keys = [c['center'] for c in significant_clusters] + bounding_box_crop = _calculate_bounding_box_cropdetect(crop_keys) + if bounding_box_crop: + parsed_bb = _parse_crop_string_cropdetect(bounding_box_crop) + snapped_crop, ar_label = _snap_to_known_ar_cropdetect( + parsed_bb['w'], parsed_bb['h'], parsed_bb['x'], parsed_bb['y'], width, height + ) + parsed_snapped = _parse_crop_string_cropdetect(snapped_crop) + if parsed_snapped and parsed_snapped['w'] == width and parsed_snapped['h'] == height: + return None + else: + return snapped_crop + else: + return None + +def detect_autocrop_filter(input_file, significant_crop_threshold=5.0, min_crop=10, debug=False): + if not _check_prerequisites_cropdetect(): + return None + try: + probe_duration_args = [ + 'ffprobe', '-v', 'error', '-show_entries', 'format=duration', '-of', 'default=noprint_wrappers=1:nokey=1', + input_file + ] + duration_str = subprocess.check_output(probe_duration_args, stderr=subprocess.STDOUT, text=True) + duration = int(float(duration_str)) + probe_res_args = [ + 'ffprobe', '-v', 'error', + '-select_streams', 'v', + '-show_entries', 'stream=width,height,disposition', + '-of', 'json', + input_file + ] + probe_output = subprocess.check_output(probe_res_args, stderr=subprocess.STDOUT, text=True) + streams_data = json.loads(probe_output) + video_stream = None + for stream in streams_data.get('streams', []): + if stream.get('disposition', {}).get('attached_pic', 0) == 0: + video_stream = stream + break + if not video_stream or 'width' not in video_stream or 'height' not in video_stream: + return None + width = int(video_stream['width']) + height = int(video_stream['height']) + except Exception: + 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): - """Main batch-processing loop for MKV files.""" check_tools() current_dir = Path(".") files_to_process = sorted( @@ -285,7 +491,6 @@ def main(no_downmix=False, autocrop=False): is_vfr = False target_cfr_fps_for_handbrake = None video_track_info = None - # Find the video track in MediaInfo if media_info.get("media") and media_info["media"].get("track"): for track in media_info["media"]["track"]: if track.get("@type") == "Video": @@ -320,7 +525,7 @@ def main(no_downmix=False, autocrop=False): print(f" - Warning: Could not parse fractional FPS '{target_cfr_fps_for_handbrake}'. HandBrakeCLI might fail.") is_vfr = False else: - print(" - Warning: VFR detected, but could not determine target CFR from MediaInfo. Will attempt encode without HandBrake.") + print(" - Warning: VFR detected, but could not determine target CFR from MediaInfo. Will attempt standard UTVideo conversion without HandBrake.") is_vfr = False else: print(f" - Video appears to be CFR or FrameRate_Mode not specified as VFR/Variable by MediaInfo.") @@ -341,8 +546,10 @@ def main(no_downmix=False, autocrop=False): audio_tracks_to_remux = [] audio_streams = [s for s in ffprobe_info.get("streams", []) if s.get("codec_type") == "audio"] + # Build mkvmerge track mapping by track ID mkv_audio_tracks = {t["id"]: t for t in mkv_info.get("tracks", []) if t.get("type") == "audio"} + # Build mediainfo track mapping by StreamOrder media_tracks_data = media_info.get("media", {}).get("track", []) mediainfo_audio_tracks = {int(t.get("StreamOrder", -1)): t for t in media_tracks_data if t.get("@type") == "Audio"} @@ -352,23 +559,27 @@ def main(no_downmix=False, autocrop=False): channels = stream.get("channels", 2) language = stream.get("tags", {}).get("language", "und") + # Find mkvmerge track by matching ffprobe stream index to mkvmerge track's 'properties'->'stream_id' mkv_track = None for t in mkv_info.get("tracks", []): if t.get("type") == "audio" and t.get("properties", {}).get("stream_id") == stream_index: mkv_track = t break if not mkv_track: + # Fallback: try by position mkv_track = mkv_info.get("tracks", [])[stream_index] if stream_index < len(mkv_info.get("tracks", [])) else {} track_id = mkv_track.get("id", -1) track_title = mkv_track.get("properties", {}).get("track_name", "") + # Find mediainfo track by StreamOrder audio_track_info = mediainfo_audio_tracks.get(stream_index) track_delay = 0 delay_raw = audio_track_info.get("Video_Delay") if audio_track_info else None if delay_raw is not None: try: delay_val = float(delay_raw) + # If the value is a float < 1, it's seconds, so convert to ms. if delay_val < 1: track_delay = int(round(delay_val * 1000)) else: @@ -380,6 +591,7 @@ def main(no_downmix=False, autocrop=False): if codec in REMUX_CODECS: audio_tracks_to_remux.append(str(track_id)) else: + # Convert any codec that is not in REMUX_CODECS opus_file = convert_audio_track( stream_index, channels, language, audio_temp_dir, str(input_file_abs), not no_downmix ) @@ -392,6 +604,7 @@ def main(no_downmix=False, autocrop=False): print("--- Finished Audio Processing ---") + # Final mux print("Assembling final file with mkvmerge...") mkvmerge_args = ["mkvmerge", "-o", str(intermediate_output_file), str(encoded_video_file)] for file_info in processed_audio_files: @@ -409,15 +622,22 @@ def main(no_downmix=False, autocrop=False): mkvmerge_args += source_copy_args + [str(input_file_abs)] run_cmd(mkvmerge_args) + # Move files print("Moving files to final destinations...") shutil.move(str(file_path), DIR_ORIGINAL / file_path.name) shutil.move(str(intermediate_output_file), DIR_COMPLETED / file_path.name) print("Cleaning up persistent video temporary files (after successful processing)...") - video_temp_files_on_success = [encoded_video_file] + 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 + current_dir / f"{file_path.stem}.ut.mkv.lwi", + ] if handbrake_intermediate_for_cleanup and handbrake_intermediate_for_cleanup.exists(): video_temp_files_on_success.append(handbrake_intermediate_for_cleanup) - + for temp_vid_file in video_temp_files_on_success: if temp_vid_file.exists(): print(f" Deleting: {temp_vid_file}") @@ -426,40 +646,47 @@ def main(no_downmix=False, autocrop=False): print(f" Skipping (not found): {temp_vid_file}") except Exception as e: - print(f"ERROR: An error occurred while processing '{file_path.name}': {e}", file=sys.stderr) + print(f"ERROR: An error occurred while processing '{file_path.name}': {e}", file=sys.stderr) # Goes to log original_stderr_console.write(f"ERROR during processing of '{file_path.name}': {e}\nSee log '{log_file_path}' for details.\n") processing_error_occurred = True finally: + # This is the original 'finally' block. Its prints go to the log file. print("--- Starting Universal Cleanup (for this file) ---") print(" - Cleaning up disposable audio temporary directory...") if audio_temp_dir and Path(audio_temp_dir).exists(): shutil.rmtree(audio_temp_dir, ignore_errors=True) print(f" - Deleted audio temp dir: {audio_temp_dir}") - elif audio_temp_dir: + elif audio_temp_dir: # Was created but now not found print(f" - Audio temp dir not found or already cleaned: {audio_temp_dir}") - else: + else: # Was never created print(f" - Audio temp dir was not created.") - + print(" - Cleaning up intermediate output file (if it wasn't moved on success)...") - if intermediate_output_file.exists(): + if intermediate_output_file.exists(): # Check if it still exists (e.g. error before move) if processing_error_occurred: print(f" - WARNING: Processing error occurred. Intermediate output file '{intermediate_output_file}' is being preserved at its original path for inspection.") else: + # No processing error, so it should have been moved. + # If it's still here, it's unexpected but we'll clean it up. print(f" - INFO: Intermediate output file '{intermediate_output_file}' found at original path despite no errors (expected to be moved). Cleaning up.") - intermediate_output_file.unlink(missing_ok=True) + intermediate_output_file.unlink(missing_ok=True) # Only unlink if no error and it exists print(f" - Deleted intermediate output file from original path: {intermediate_output_file}") else: + # File does not exist at original path if not processing_error_occurred: print(f" - Intermediate output file successfully moved (not found at original path, as expected): {intermediate_output_file}") else: print(f" - Processing error occurred, and intermediate output file '{intermediate_output_file}' not found at original path (likely not created or cleaned by another step).") - - print(f"FINISHED LOG FOR: {file_path.name}") + # --- End of original per-file processing block --- - finally: + print(f"FINISHED LOG FOR: {file_path.name}") + # --- End of log-specific messages --- + + finally: # Outer finally for restoring stdout/stderr and closing log file runtime = datetime.now() - date_for_runtime_calc runtime_str = str(runtime).split('.')[0] + # This print goes to the log file, as stdout is not yet restored. print(f"\nTotal runtime for this file: {runtime_str}") if sys.stdout != original_stdout_console: @@ -469,6 +696,7 @@ def main(no_downmix=False, autocrop=False): if log_file_handle: log_file_handle.close() + # Announce to console (original stdout/stderr) that this file is done if processing_error_occurred: original_stderr_console.write(f"File: {file_path.name}\n") original_stderr_console.write(f"Log: {log_file_path}\n") @@ -480,7 +708,7 @@ def main(no_downmix=False, autocrop=False): if __name__ == "__main__": import argparse - parser = argparse.ArgumentParser(description="Batch-process MKV files with ffmpeg+SvtAv1EncApp, audio downmixing, per-file logging, and optional autocrop.") + 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.") args = parser.parse_args()