diff options
| -rw-r--r-- | settings.py | 11 | ||||
| -rw-r--r-- | youtube/watch.py | 53 | ||||
| -rw-r--r-- | youtube/yt_data_extract/common.py | 7 | ||||
| -rw-r--r-- | youtube/yt_data_extract/watch_extraction.py | 8 |
4 files changed, 73 insertions, 6 deletions
diff --git a/settings.py b/settings.py index c97e3d8..fdaebc7 100644 --- a/settings.py +++ b/settings.py @@ -168,6 +168,17 @@ For security reasons, enabling this is not recommended.''', 'category': 'playback', }), + ('preferred_video_codec', { + 'type': int, + 'default': 0, + 'comment': '', + 'options': [ + (0, 'h.264'), + (1, 'AV1'), + ], + 'category': 'playback', + }), + ('prefer_uni_sources', { 'label': 'Prefer integrated sources', 'type': bool, diff --git a/youtube/watch.py b/youtube/watch.py index 7494b95..80885f9 100644 --- a/youtube/watch.py +++ b/youtube/watch.py @@ -56,7 +56,10 @@ def get_video_sources(info, target_resolution): continue # audio source - if fmt['acodec'] and not fmt['vcodec'] and fmt['audio_bitrate']: + if fmt['acodec'] and not fmt['vcodec'] and ( + fmt['audio_bitrate'] or fmt['bitrate']): + if fmt['bitrate']: # prefer this one, more accurate right now + fmt['audio_bitrate'] = int(fmt['bitrate']/1000) source = { 'type': 'audio/' + fmt['ext'], 'bitrate': fmt['audio_bitrate'], @@ -77,15 +80,51 @@ def get_video_sources(info, target_resolution): + source['vcodec'] + '"') video_only_sources.append(source) + # Remove alternative mp4 codecs from video sources + def codec_name(vcodec): + if vcodec.startswith('avc'): + return 'h.264' + elif vcodec.startswith('av01'): + return 'av1' + else: + return 'unknown' + quality_to_codecs = {} + for src in video_only_sources: + if src['quality'] in quality_to_codecs: + quality_to_codecs[src['quality']].add(codec_name(src['vcodec'])) + else: + quality_to_codecs[src['quality']] = {codec_name(src['vcodec'])} + i = 0 + while i < len(video_only_sources): + src = video_only_sources[i] + codecs_for_quality = quality_to_codecs[src['quality']] + have_both = ('h.264' in codecs_for_quality + and 'av1' in codecs_for_quality) + have_one = ('h.264' in codecs_for_quality + or 'av1' in codecs_for_quality) + name = codec_name(src['vcodec']) + if name == 'unknown' and have_one: + del video_only_sources[i] + continue + if not have_both: + i += 1 + continue + if name == 'av1' and settings.preferred_video_codec == 0: + del video_only_sources[i] + elif name == 'h.264' and settings.preferred_video_codec == 1: + del video_only_sources[i] + else: + i += 1 + audio_sources.sort(key=lambda source: source['audio_bitrate']) video_only_sources.sort(key=lambda src: src['quality']) uni_sources.sort(key=lambda src: src['quality']) for source in video_only_sources: # choose an audio source to go with it - # 0.15 is semiarbitrary empirical constant to spread audio sources + # 0.5 is semiarbitrary empirical constant to spread audio sources # between 144p and 1080p. Use something better eventually. - target_audio_bitrate = source['quality']*source.get('fps', 30)/30*0.15 + target_audio_bitrate = source['quality']*source.get('fps', 30)/30*0.5 compat_audios = [a for a in audio_sources if a['ext'] == source['ext']] if compat_audios: closest_audio_source = compat_audios[0] @@ -421,7 +460,13 @@ def video_quality_string(format): def short_video_quality_string(fmt): result = str(fmt['quality'] or '?') + 'p' if fmt['fps']: - result += ' ' + str(fmt['fps']) + 'fps' + result += str(fmt['fps']) + if fmt['vcodec'].startswith('av01'): + result += ' AV1' + elif fmt['vcodec'].startswith('avc'): + result += ' h264' + else: + result += ' ' + fmt['vcodec'] return result diff --git a/youtube/yt_data_extract/common.py b/youtube/yt_data_extract/common.py index ca999ba..f97597c 100644 --- a/youtube/yt_data_extract/common.py +++ b/youtube/yt_data_extract/common.py @@ -166,14 +166,17 @@ def extract_formatted_text(node): return [{'text': node['simpleText']}] return [] -def extract_int(string, default=None): +def extract_int(string, default=None, whole_word=True): if isinstance(string, int): return string if not isinstance(string, str): string = extract_str(string) if not string: return default - match = re.search(r'\b(\d+)\b', string.replace(',', '')) + if whole_word: + match = re.search(r'\b(\d+)\b', string.replace(',', '')) + else: + match = re.search(r'(\d+)', string.replace(',', '')) if match is None: return default try: diff --git a/youtube/yt_data_extract/watch_extraction.py b/youtube/yt_data_extract/watch_extraction.py index 43be909..e0af28e 100644 --- a/youtube/yt_data_extract/watch_extraction.py +++ b/youtube/yt_data_extract/watch_extraction.py @@ -445,6 +445,14 @@ def _extract_formats(info, player_response): for key, value in hardcoded_itag_info.items(): conservative_update(fmt, key, value) # prefer info from YouTube fmt['quality'] = hardcoded_itag_info.get('height') + conservative_update( + fmt, 'quality', + extract_int(yt_fmt.get('quality'), whole_word=False) + ) + conservative_update( + fmt, 'quality', + extract_int(yt_fmt.get('qualityLabel'), whole_word=False) + ) info['formats'].append(fmt) |