lyrics-transcriber 0.41.0__py3-none-any.whl → 0.42.0__py3-none-any.whl

lyrics_transcriber/lyrics/base_lyrics_provider.py

@@ -1,13 +1,14 @@
 from dataclasses import dataclass
 import logging
-from typing import Optional, Dict, Any
+from typing import Optional, Dict, Any, List
 import json
 import hashlib
 from pathlib import Path
 import os
 from abc import ABC, abstractmethod
-from lyrics_transcriber.types import LyricsData
+from lyrics_transcriber.types import LyricsData, LyricsSegment, Word
 from karaoke_lyrics_processor import KaraokeLyricsProcessor
+from lyrics_transcriber.utils.word_utils import WordUtils
 
 
 @dataclass
@@ -41,20 +42,31 @@ class BaseLyricsProvider(ABC):
 
         # Use artist and title for cache key instead of audio file hash
         cache_key = self._get_artist_title_hash(artist, title)
-        raw_cache_path = self._get_cache_path(cache_key, "raw")
 
-        # Try to load from cache first
+        # Check converted cache first
+        converted_cache_path = self._get_cache_path(cache_key, "converted")
+        converted_data = self._load_from_cache(converted_cache_path)
+        if converted_data:
+            self.logger.info(f"Using cached converted lyrics for {artist} - {title}")
+            return LyricsData.from_dict(converted_data)
+
+        # Check raw cache next
+        raw_cache_path = self._get_cache_path(cache_key, "raw")
         raw_data = self._load_from_cache(raw_cache_path)
-        if raw_data is not None:
-            self.logger.info(f"Using cached lyrics for {artist} - {title}")
-            return self._save_and_convert_result(cache_key, raw_data)
+        if raw_data:
+            self.logger.info(f"Using cached raw lyrics for {artist} - {title}")
+            converted_result = self._convert_result_format(raw_data)
+            self._save_to_cache(converted_cache_path, converted_result.to_dict())
+            return converted_result
 
         # If not in cache, fetch from source
         raw_result = self._fetch_data_from_source(artist, title)
         if raw_result:
             # Save raw API response
             self._save_to_cache(raw_cache_path, raw_result)
-            return self._save_and_convert_result(cache_key, raw_result)
+            converted_result = self._convert_result_format(raw_result)
+            self._save_to_cache(converted_cache_path, converted_result.to_dict())
+            return converted_result
 
         return None
 
@@ -100,18 +112,69 @@ class BaseLyricsProvider(ABC):
             self.logger.warning(f"Cache file {cache_path} is corrupted")
             return None
 
+    def _create_segments_with_words(self, text: str, is_synced: bool = False) -> List[LyricsSegment]:
+        """Create LyricsSegment objects with properly formatted words from text.
+
+        Args:
+            text: Raw lyrics text
+            is_synced: Whether timing information is available
+
+        Returns:
+            List of LyricsSegment objects with unique IDs and Word objects
+        """
+        segments = []
+        lines = text.strip().split("\n")
+
+        for line in lines:
+            if not line.strip():
+                continue
+
+            # Split line into words
+            word_texts = line.strip().split()
+            if not word_texts:
+                continue
+
+            words = []
+            for word_text in word_texts:
+                word = Word(
+                    id=WordUtils.generate_id(),
+                    text=word_text,
+                    start_time=0.0 if is_synced else None,
+                    end_time=0.0 if is_synced else None,
+                    confidence=1.0,  # Reference lyrics are considered ground truth
+                    created_during_correction=False,
+                )
+                words.append(word)
+
+            segment = LyricsSegment(
+                id=WordUtils.generate_id(),
+                text=line.strip(),
+                words=words,
+                start_time=words[0].start_time if is_synced else None,
+                end_time=words[-1].end_time if is_synced else None,
+            )
+            segments.append(segment)
+
+        return segments
+
     def _process_lyrics(self, lyrics_data: LyricsData) -> LyricsData:
-        """Process lyrics using KaraokeLyricsProcessor."""
+        """Process lyrics using KaraokeLyricsProcessor and create proper segments."""
+        # Concatenate all segment texts to get the full lyrics
+        full_lyrics = lyrics_data.get_full_text()
+
         processor = KaraokeLyricsProcessor(
             log_level=self.logger.getEffectiveLevel(),
             log_formatter=self.logger.handlers[0].formatter if self.logger.handlers else None,
-            input_lyrics_text=lyrics_data.lyrics,
+            input_lyrics_text=full_lyrics,
             max_line_length=self.max_line_length,
         )
         processed_text = processor.process()
 
-        # Create new LyricsData with processed text
-        return LyricsData(source=lyrics_data.source, lyrics=processed_text, segments=lyrics_data.segments, metadata=lyrics_data.metadata)
+        # Create segments with words from processed text
+        segments = self._create_segments_with_words(processed_text, is_synced=lyrics_data.metadata.is_synced)
+
+        # Create new LyricsData with processed text and segments
+        return LyricsData(source=lyrics_data.source, segments=segments, metadata=lyrics_data.metadata)
 
     def _save_and_convert_result(self, cache_key: str, raw_data: Dict[str, Any]) -> LyricsData:
         """Convert raw result to standardized format, process lyrics, save to cache, and return."""

lyrics_transcriber/lyrics/file_provider.py

@@ -67,7 +67,7 @@ class FileProvider(BaseLyricsProvider):
         self.logger.debug(f"Converting raw data to LyricsData format: {raw_data}")
 
         try:
-            # Create metadata object like Genius provider does
+            # Create metadata object
             metadata = LyricsMetadata(
                 source="file",
                 track_name=self.title,
@@ -78,10 +78,11 @@ class FileProvider(BaseLyricsProvider):
                 provider_metadata={"filepath": raw_data["filepath"]},
             )
 
-            lyrics_data = LyricsData(
-                source="file", lyrics=raw_data["text"], segments=[], metadata=metadata  # No timing information from file
-            )
-            self.logger.debug(f"Created LyricsData object: {lyrics_data}")
+            # Create segments with words from the processed text
+            segments = self._create_segments_with_words(raw_data["text"], is_synced=False)
+
+            lyrics_data = LyricsData(source="file", segments=segments, metadata=metadata)
+            self.logger.debug(f"Created LyricsData object with {len(segments)} segments")
             return lyrics_data
 
         except Exception as e:

lyrics_transcriber/lyrics/genius.py

@@ -77,8 +77,11 @@ class GeniusProvider(BaseLyricsProvider):
             },
         )
 
-        # Create result object
-        return LyricsData(source="genius", lyrics=lyrics, segments=[], metadata=metadata)
+        # Create segments with words from cleaned lyrics
+        segments = self._create_segments_with_words(lyrics, is_synced=False)
+
+        # Create result object with segments
+        return LyricsData(source="genius", segments=segments, metadata=metadata)
 
     def _clean_lyrics(self, lyrics: str) -> str:
         """Clean and process lyrics from Genius to remove unwanted content."""

lyrics_transcriber/lyrics/spotify.py

@@ -1,9 +1,11 @@
 import logging
 from typing import Optional, Dict, Any
 import syrics.api
+import time
 
-from lyrics_transcriber.types import LyricsData, LyricsMetadata, LyricsSegment
+from lyrics_transcriber.types import LyricsData, LyricsMetadata, LyricsSegment, Word
 from lyrics_transcriber.lyrics.base_lyrics_provider import BaseLyricsProvider, LyricsProviderConfig
+from lyrics_transcriber.utils.word_utils import WordUtils
 
 
 class SpotifyProvider(BaseLyricsProvider):
@@ -12,7 +14,22 @@ class SpotifyProvider(BaseLyricsProvider):
     def __init__(self, config: LyricsProviderConfig, logger: Optional[logging.Logger] = None):
         super().__init__(config, logger)
         self.cookie = config.spotify_cookie
-        self.client = syrics.api.Spotify(self.cookie) if self.cookie else None
+        self.client = None
+
+        if self.cookie:
+            max_retries = 5
+            retry_delay = 5  # seconds
+
+            for attempt in range(max_retries):
+                try:
+                    self.client = syrics.api.Spotify(self.cookie)
+                    break  # Successfully initialized
+                except Exception as e:
+                    if attempt == max_retries - 1:  # Last attempt
+                        self.logger.error(f"Failed to initialize Spotify client after {max_retries} attempts: {str(e)}")
+                        break
+                    self.logger.warning(f"Attempt {attempt + 1}/{max_retries} failed, retrying in {retry_delay} seconds...")
+                    time.sleep(retry_delay)
 
     def _fetch_data_from_source(self, artist: str, title: str) -> Optional[Dict[str, Any]]:
         """Fetch raw data from Spotify APIs using syrics library."""
@@ -45,24 +62,6 @@ class SpotifyProvider(BaseLyricsProvider):
         track_data = raw_data["track_data"]
         lyrics_data = raw_data["lyrics_data"]["lyrics"]
 
-        # Convert raw lines to LyricsSegment objects
-        segments = []
-        for line in lyrics_data.get("lines", []):
-            if not line.get("words"):
-                continue
-
-            # Skip lines that are just musical notes
-            if not self._clean_lyrics(line["words"]):
-                continue
-
-            segment = LyricsSegment(
-                text=line["words"],
-                words=[],  # TODO: Could potentially split words if needed
-                start_time=float(line["startTimeMs"]) / 1000 if line["startTimeMs"] != "0" else None,
-                end_time=float(line["endTimeMs"]) / 1000 if line["endTimeMs"] != "0" else None,
-            )
-            segments.append(segment)
-
         # Create metadata object
         metadata = LyricsMetadata(
             source="spotify",
@@ -83,7 +82,45 @@ class SpotifyProvider(BaseLyricsProvider):
             },
         )
 
-        return LyricsData(source="spotify", lyrics="\n".join(segment.text for segment in segments), segments=segments, metadata=metadata)
+        # Create segments with timing information
+        segments = []
+        for line in lyrics_data.get("lines", []):
+            if not line.get("words"):
+                continue
+
+            # Skip lines that are just musical notes
+            if not self._clean_lyrics(line["words"]):
+                continue
+
+            # Split line into words
+            word_texts = line["words"].strip().split()
+            if not word_texts:
+                continue
+
+            # Calculate approximate timing for each word
+            start_time = float(line["startTimeMs"]) / 1000 if line["startTimeMs"] != "0" else 0.0
+            end_time = float(line["endTimeMs"]) / 1000 if line["endTimeMs"] != "0" else 0.0
+            duration = end_time - start_time
+            word_duration = duration / len(word_texts)
+
+            words = []
+            for i, word_text in enumerate(word_texts):
+                word = Word(
+                    id=WordUtils.generate_id(),
+                    text=word_text,
+                    start_time=start_time + (i * word_duration),
+                    end_time=start_time + ((i + 1) * word_duration),
+                    confidence=1.0,
+                    created_during_correction=False,
+                )
+                words.append(word)
+
+            segment = LyricsSegment(
+                id=WordUtils.generate_id(), text=line["words"].strip(), words=words, start_time=start_time, end_time=end_time
+            )
+            segments.append(segment)
+
+        return LyricsData(source="spotify", segments=segments, metadata=metadata)
 
     def _clean_lyrics(self, lyrics: str) -> str:
         """Clean and process lyrics from Spotify to remove unwanted content."""

lyrics_transcriber/output/ass/config.py

@@ -4,7 +4,18 @@ from dataclasses import dataclass
 class ScreenConfig:
     """Configuration for screen timing and layout."""
 
-    def __init__(self, line_height: int = 50, max_visible_lines: int = 4, top_padding: int = None, video_width: int = 640, video_height: int = 360):
+    def __init__(
+        self,
+        line_height: int = 50,
+        max_visible_lines: int = 4,
+        top_padding: int = None,
+        video_width: int = 640,
+        video_height: int = 360,
+        screen_gap_threshold: float = 5.0,
+        post_roll_time: float = 1.0,
+        fade_in_ms: int = 200,
+        fade_out_ms: int = 300,
+    ):
         # Screen layout
         self.max_visible_lines = max_visible_lines
         self.line_height = line_height
@@ -12,10 +23,10 @@ class ScreenConfig:
         self.video_height = video_height
         self.video_width = video_width
         # Timing configuration
-        self.screen_gap_threshold = 5.0
-        self.post_roll_time = 1.0
-        self.fade_in_ms = 200
-        self.fade_out_ms = 300
+        self.screen_gap_threshold = screen_gap_threshold
+        self.post_roll_time = post_roll_time
+        self.fade_in_ms = fade_in_ms
+        self.fade_out_ms = fade_out_ms
 
 
 @dataclass

lyrics_transcriber/output/cdg.py

@@ -496,7 +496,7 @@ class CDGGenerator:
                 text = text[1:]
 
             current_line += text + " "
-            self.logger.debug(f"format_lyrics: Current line: '{current_line}'")
+            # self.logger.debug(f"format_lyrics: Current line: '{current_line}'")
 
             is_last_before_instrumental = any(
                 inst["sync"] > sync_times[i] and (i == len(sync_times) - 1 or sync_times[i + 1] > inst["sync"]) for inst in instrumentals

lyrics_transcriber/output/generator.py

@@ -97,28 +97,42 @@ class OutputGenerator:
     def generate_outputs(
         self,
         transcription_corrected: Optional[CorrectionResult],
-        lyrics_results: List[LyricsData],
+        lyrics_results: dict[str, LyricsData],
         output_prefix: str,
         audio_filepath: str,
         artist: Optional[str] = None,
         title: Optional[str] = None,
+        preview_mode: bool = False,
     ) -> OutputPaths:
         """Generate all requested output formats."""
         outputs = OutputPaths()
 
         try:
-            # Generate plain lyrics files for each provider
-            for lyrics_data in lyrics_results:
-                self.plain_text.write_lyrics(lyrics_data, output_prefix)
-
             # Only process transcription-related outputs if we have transcription data
             if transcription_corrected:
-                # Write original (uncorrected) transcription
-                outputs.original_txt = self.plain_text.write_original_transcription(transcription_corrected, output_prefix)
 
-                # Resize corrected segments to ensure none are longer than max_line_length
+                # Resize corrected segments
                 resized_segments = self.segment_resizer.resize_segments(transcription_corrected.corrected_segments)
                 transcription_corrected.resized_segments = resized_segments
+
+                # For preview, we only need to generate ASS and video
+                if preview_mode:
+                    # Generate ASS subtitles for preview
+                    outputs.ass = self.subtitle.generate_ass(transcription_corrected.resized_segments, output_prefix, audio_filepath)
+
+                    # Generate preview video
+                    outputs.video = self.video.generate_preview_video(outputs.ass, audio_filepath, output_prefix)
+
+                    return outputs
+
+                # Normal output generation (non-preview mode)
+                # Generate plain lyrics files for each provider
+                for name, lyrics_data in lyrics_results.items():
+                    self.plain_text.write_lyrics(lyrics_data, output_prefix)
+
+                # Write original (uncorrected) transcription
+                outputs.original_txt = self.plain_text.write_original_transcription(transcription_corrected, output_prefix)
+
                 outputs.corrections_json = self.write_corrections_data(transcription_corrected, output_prefix)
 
                 # Write corrected lyrics as plain text

lyrics_transcriber/output/plain_text.py

@@ -5,12 +5,13 @@ from typing import List, Optional
 from lyrics_transcriber.types import LyricsData, LyricsSegment
 from lyrics_transcriber.correction.corrector import CorrectionResult
 
+
 class PlainTextGenerator:
     """Handles generation of plain text output files for lyrics and transcriptions."""
 
     def __init__(self, output_dir: str, logger: Optional[logging.Logger] = None):
         """Initialize PlainTextGenerator.
-        
+
         Args:
             output_dir: Directory where output files will be written
             logger: Optional logger instance
@@ -24,11 +25,11 @@ class PlainTextGenerator:
 
     def write_lyrics(self, lyrics_data: LyricsData, output_prefix: str) -> str:
         """Write plain text lyrics file from provider data.
-        
+
         Args:
             lyrics_data: LyricsData from a lyrics provider
             output_prefix: Prefix for output filename
-            
+
         Returns:
             Path to generated file
         """
@@ -38,7 +39,9 @@ class PlainTextGenerator:
 
         try:
             with open(output_path, "w", encoding="utf-8") as f:
-                f.write(lyrics_data.lyrics)
+                # Join segment texts with newlines
+                lyrics_text = "\n".join(segment.text for segment in lyrics_data.segments)
+                f.write(lyrics_text)
             self.logger.info(f"Plain lyrics file generated: {output_path}")
             return output_path
         except Exception as e:
@@ -47,11 +50,11 @@ class PlainTextGenerator:
 
     def write_corrected_lyrics(self, segments: List[LyricsSegment], output_prefix: str) -> str:
         """Write corrected lyrics as plain text file.
-        
+
         Args:
             segments: List of corrected LyricsSegment objects
             output_prefix: Prefix for output filename
-            
+
         Returns:
             Path to generated file
         """
@@ -70,22 +73,24 @@ class PlainTextGenerator:
 
     def write_original_transcription(self, correction_result: CorrectionResult, output_prefix: str) -> str:
         """Write original (uncorrected) transcription as plain text.
-        
+
         Args:
             correction_result: CorrectionResult containing original transcription
             output_prefix: Prefix for output filename
-            
+
         Returns:
             Path to generated file
         """
         self.logger.info("Writing original transcription file")
         output_path = self._get_output_path(f"{output_prefix} (Lyrics Uncorrected)", "txt")
 
+        transcribed_text = " ".join(" ".join(w.text for w in segment.words) for segment in correction_result.original_segments)
+
         try:
             with open(output_path, "w", encoding="utf-8") as f:
-                f.write(correction_result.transcribed_text)
+                f.write(transcribed_text)
             self.logger.info(f"Original transcription file generated: {output_path}")
             return output_path
         except Exception as e:
             self.logger.error(f"Failed to write original transcription file: {str(e)}")
-            raise 
+            raise

lyrics_transcriber/output/segment_resizer.py

@@ -1,8 +1,9 @@
 import logging
 import re
-from typing import List, Optional, Tuple
+from typing import List, Optional
 
 from lyrics_transcriber.types import LyricsSegment, Word
+from lyrics_transcriber.utils.word_utils import WordUtils
 
 
 class SegmentResizer:
@@ -101,7 +102,13 @@ class SegmentResizer:
             Output: LyricsSegment(text="Hello World", words=[...])
         """
         cleaned_text = self._clean_text(segment.text)
-        return LyricsSegment(text=cleaned_text, words=segment.words, start_time=segment.start_time, end_time=segment.end_time)
+        return LyricsSegment(
+            id=segment.id,  # Preserve the original segment ID
+            text=cleaned_text,
+            words=segment.words,
+            start_time=segment.start_time,
+            end_time=segment.end_time,
+        )
 
     def _create_cleaned_word(self, word: Word) -> Word:
         """Create a new word with cleaned text."""
@@ -226,7 +233,13 @@ class SegmentResizer:
     def _create_segment_from_words(self, line: str, words: List[Word]) -> LyricsSegment:
         """Create a new segment from a list of words."""
         cleaned_text = self._clean_text(line)
-        return LyricsSegment(text=cleaned_text, words=words, start_time=words[0].start_time, end_time=words[-1].end_time)
+        return LyricsSegment(
+            id=WordUtils.generate_id(),  # Generate new ID for split segments
+            text=cleaned_text,
+            words=words,
+            start_time=words[0].start_time,
+            end_time=words[-1].end_time,
+        )
 
     def _process_segment_text(self, text: str) -> List[str]:
         """Process segment text to determine optimal split points."""

lyrics_transcriber/output/subtitles.py

@@ -44,7 +44,30 @@ class SubtitlesGenerator:
         self.font_size = font_size
         self.styles = styles
         self.subtitle_offset_ms = subtitle_offset_ms
-        self.config = ScreenConfig(line_height=line_height, video_width=video_resolution[0], video_height=video_resolution[1])
+        
+        # Create ScreenConfig with potential overrides from styles
+        karaoke_styles = styles.get("karaoke", {})
+        config_params = {
+            "line_height": line_height,
+            "video_width": video_resolution[0],
+            "video_height": video_resolution[1]
+        }
+        
+        # Add any overrides from styles
+        screen_config_props = [
+            "max_visible_lines",
+            "top_padding",
+            "screen_gap_threshold",
+            "post_roll_time",
+            "fade_in_ms",
+            "fade_out_ms"
+        ]
+        
+        for prop in screen_config_props:
+            if prop in karaoke_styles:
+                config_params[prop] = karaoke_styles[prop]
+        
+        self.config = ScreenConfig(**config_params)
         self.logger = logger or logging.getLogger(__name__)
 
     def _get_output_path(self, output_prefix: str, extension: str) -> str:
@@ -102,13 +125,16 @@ class SubtitlesGenerator:
             offset_seconds = self.subtitle_offset_ms / 1000.0
             segments = [
                 LyricsSegment(
+                    id=seg.id,  # Preserve original segment ID
                     text=seg.text,
                     words=[
                         Word(
+                            id=word.id,  # Preserve original word ID
                             text=word.text,
                             start_time=max(0, word.start_time + offset_seconds),
                             end_time=word.end_time + offset_seconds,
                             confidence=word.confidence,
+                            created_during_correction=getattr(word, "created_during_correction", False),  # Preserve correction flag
                         )
                         for word in seg.words
                     ],

lyrics_transcriber/output/video.py

@@ -88,6 +88,52 @@ class VideoGenerator:
                     pass
             raise
 
+    def generate_preview_video(self, ass_path: str, audio_path: str, output_prefix: str) -> str:
+        """Generate lower resolution MP4 preview video with lyrics overlay.
+
+        Args:
+            ass_path: Path to ASS subtitles file
+            audio_path: Path to audio file
+            output_prefix: Prefix for output filename
+
+        Returns:
+            Path to generated preview video file
+        """
+        self.logger.info("Generating preview video with lyrics overlay")
+        output_path = os.path.join(self.cache_dir, f"{output_prefix}_preview.mp4")
+
+        # Check input files exist before running FFmpeg
+        if not os.path.isfile(ass_path):
+            raise FileNotFoundError(f"Subtitles file not found: {ass_path}")
+        if not os.path.isfile(audio_path):
+            raise FileNotFoundError(f"Audio file not found: {audio_path}")
+
+        try:
+            # Create a temporary copy of the ASS file with a safe filename
+            temp_ass_path = os.path.join(self.cache_dir, "temp_preview_subtitles.ass")
+            import shutil
+
+            shutil.copy2(ass_path, temp_ass_path)
+            self.logger.debug(f"Created temporary ASS file: {temp_ass_path}")
+
+            cmd = self._build_preview_ffmpeg_command(temp_ass_path, audio_path, output_path)
+            self._run_ffmpeg_command(cmd)
+            self.logger.info(f"Preview video generated: {output_path}")
+
+            # Clean up temporary file
+            os.remove(temp_ass_path)
+            return output_path
+
+        except Exception as e:
+            self.logger.error(f"Failed to generate preview video: {str(e)}")
+            # Clean up temporary file in case of error
+            if "temp_ass_path" in locals():
+                try:
+                    os.remove(temp_ass_path)
+                except:
+                    pass
+            raise
+
     def _get_output_path(self, output_prefix: str, extension: str) -> str:
         """Generate full output path for a file."""
         return os.path.join(self.output_dir, f"{output_prefix}.{extension}")
@@ -181,7 +227,7 @@ class VideoGenerator:
             "-vf", f"ass={ass_path}",  # Add subtitles
             "-c:v", self._get_video_codec(),
             # Video quality settings
-            "-preset", "slow",  # Better compression efficiency
+            "-preset", "fast",  # Better compression efficiency
             "-b:v", "5000k",  # Base video bitrate
             "-minrate", "5000k",  # Minimum bitrate
             "-maxrate", "20000k",  # Maximum bitrate
@@ -196,6 +242,66 @@ class VideoGenerator:
 
         return cmd
 
+    def _build_preview_ffmpeg_command(self, ass_path: str, audio_path: str, output_path: str) -> List[str]:
+        """Build FFmpeg command for preview video generation with optimized settings."""
+        # Use 360p resolution for preview
+        width, height = 640, 360
+
+        # fmt: off
+        cmd = [
+            "ffmpeg",
+            "-hide_banner",
+            "-loglevel", "error",
+            "-r", "30",  # Set frame rate to 30 fps
+        ]
+
+        # Input source (background)
+        if self.background_image:
+            # Resize background image first
+            resized_bg = self._resize_background_image(self.background_image)
+            self.logger.debug(f"Using resized background image: {resized_bg}")
+            cmd.extend([
+                "-loop", "1",  # Loop the image
+                "-i", resized_bg,
+            ])
+        else:
+            self.logger.debug(
+                f"Using solid {self.background_color} background "
+                f"with resolution: {width}x{height}"
+            )
+            cmd.extend([
+                "-f", "lavfi",
+                "-i", f"color=c={self.background_color}:s={width}x{height}:r=30"
+            ])
+
+        # Add audio input and subtitle overlay
+        cmd.extend([
+            "-i", audio_path,
+            "-c:a", "aac",  # Use AAC for audio
+            "-b:a", "128k",  # Audio bitrate
+            "-vf", f"ass={ass_path}",  # Add subtitles
+            "-c:v", "libx264",  # Use H.264 codec
+            "-profile:v", "baseline",  # Most compatible H.264 profile
+            "-level", "3.0",  # Compatibility level
+            "-pix_fmt", "yuv420p",  # Required for browser compatibility
+            "-preset", "ultrafast",
+            "-b:v", "1000k",  # Slightly higher bitrate
+            "-maxrate", "1500k",
+            "-bufsize", "2000k",
+            "-movflags", "+faststart+frag_keyframe+empty_moov",  # Enhanced streaming flags
+            "-g", "30",  # Keyframe every 30 frames (1 second)
+            "-keyint_min", "30",  # Minimum keyframe interval
+            "-sc_threshold", "0",  # Disable scene change detection
+            "-shortest",
+            "-y"
+        ])
+        # fmt: on
+
+        # Add output path
+        cmd.append(output_path)
+
+        return cmd
+
     def _get_video_codec(self) -> str:
         """Determine the best available video codec."""
         # try: