]> Raphaƫl G. Git Repositories - youtubedl/blobdiff - youtube_dl/postprocessor/ffmpeg.py
Update upstream source from tag 'upstream/2020.06.16'
[youtubedl] / youtube_dl / postprocessor / ffmpeg.py
index 5ed723bc6cae2a464e34fad6c48b5ba71e62dab9..5f7298345b3f550c9a8d8b7ba2b602d2f084ed36 100644 (file)
@@ -4,13 +4,11 @@ import io
 import os
 import subprocess
 import time
+import re
 
 
 from .common import AudioConversionError, PostProcessor
 
-from ..compat import (
-    compat_subprocess_get_DEVNULL,
-)
 from ..utils import (
     encodeArgument,
     encodeFilename,
@@ -22,9 +20,33 @@ from ..utils import (
     subtitles_filename,
     dfxp2srt,
     ISO639Utils,
+    replace_extension,
 )
 
 
+EXT_TO_OUT_FORMATS = {
+    'aac': 'adts',
+    'flac': 'flac',
+    'm4a': 'ipod',
+    'mka': 'matroska',
+    'mkv': 'matroska',
+    'mpg': 'mpeg',
+    'ogv': 'ogg',
+    'ts': 'mpegts',
+    'wma': 'asf',
+    'wmv': 'asf',
+}
+ACODECS = {
+    'mp3': 'libmp3lame',
+    'aac': 'aac',
+    'flac': 'flac',
+    'm4a': 'aac',
+    'opus': 'libopus',
+    'vorbis': 'libvorbis',
+    'wav': None,
+}
+
+
 class FFmpegPostProcessorError(PostProcessingError):
     pass
 
@@ -52,7 +74,21 @@ class FFmpegPostProcessor(PostProcessor):
 
     def _determine_executables(self):
         programs = ['avprobe', 'avconv', 'ffmpeg', 'ffprobe']
-        prefer_ffmpeg = self._downloader.params.get('prefer_ffmpeg', False)
+        prefer_ffmpeg = True
+
+        def get_ffmpeg_version(path):
+            ver = get_exe_version(path, args=['-version'])
+            if ver:
+                regexs = [
+                    r'(?:\d+:)?([0-9.]+)-[0-9]+ubuntu[0-9.]+$',  # Ubuntu, see [1]
+                    r'n([0-9.]+)$',  # Arch Linux
+                    # 1. http://www.ducea.com/2006/06/17/ubuntu-package-version-naming-explanation/
+                ]
+                for regex in regexs:
+                    mobj = re.match(regex, ver)
+                    if mobj:
+                        ver = mobj.group(1)
+            return ver
 
         self.basename = None
         self.probe_basename = None
@@ -60,6 +96,7 @@ class FFmpegPostProcessor(PostProcessor):
         self._paths = None
         self._versions = None
         if self._downloader:
+            prefer_ffmpeg = self._downloader.params.get('prefer_ffmpeg', True)
             location = self._downloader.params.get('ffmpeg_location')
             if location is not None:
                 if not os.path.exists(location):
@@ -84,26 +121,25 @@ class FFmpegPostProcessor(PostProcessor):
                 self._paths = dict(
                     (p, os.path.join(location, p)) for p in programs)
                 self._versions = dict(
-                    (p, get_exe_version(self._paths[p], args=['-version']))
-                    for p in programs)
+                    (p, get_ffmpeg_version(self._paths[p])) for p in programs)
         if self._versions is None:
             self._versions = dict(
-                (p, get_exe_version(p, args=['-version'])) for p in programs)
+                (p, get_ffmpeg_version(p)) for p in programs)
             self._paths = dict((p, p) for p in programs)
 
-        if prefer_ffmpeg:
-            prefs = ('ffmpeg', 'avconv')
-        else:
+        if prefer_ffmpeg is False:
             prefs = ('avconv', 'ffmpeg')
+        else:
+            prefs = ('ffmpeg', 'avconv')
         for p in prefs:
             if self._versions[p]:
                 self.basename = p
                 break
 
-        if prefer_ffmpeg:
-            prefs = ('ffprobe', 'avprobe')
-        else:
+        if prefer_ffmpeg is False:
             prefs = ('avprobe', 'ffprobe')
+        else:
+            prefs = ('ffprobe', 'avprobe')
         for p in prefs:
             if self._versions[p]:
                 self.probe_basename = p
@@ -125,6 +161,48 @@ class FFmpegPostProcessor(PostProcessor):
     def probe_executable(self):
         return self._paths[self.probe_basename]
 
+    def get_audio_codec(self, path):
+        if not self.probe_available and not self.available:
+            raise PostProcessingError('ffprobe/avprobe and ffmpeg/avconv not found. Please install one.')
+        try:
+            if self.probe_available:
+                cmd = [
+                    encodeFilename(self.probe_executable, True),
+                    encodeArgument('-show_streams')]
+            else:
+                cmd = [
+                    encodeFilename(self.executable, True),
+                    encodeArgument('-i')]
+            cmd.append(encodeFilename(self._ffmpeg_filename_argument(path), True))
+            if self._downloader.params.get('verbose', False):
+                self._downloader.to_screen(
+                    '[debug] %s command line: %s' % (self.basename, shell_quote(cmd)))
+            handle = subprocess.Popen(
+                cmd, stderr=subprocess.PIPE,
+                stdout=subprocess.PIPE, stdin=subprocess.PIPE)
+            stdout_data, stderr_data = handle.communicate()
+            expected_ret = 0 if self.probe_available else 1
+            if handle.wait() != expected_ret:
+                return None
+        except (IOError, OSError):
+            return None
+        output = (stdout_data if self.probe_available else stderr_data).decode('ascii', 'ignore')
+        if self.probe_available:
+            audio_codec = None
+            for line in output.split('\n'):
+                if line.startswith('codec_name='):
+                    audio_codec = line.split('=')[1].strip()
+                elif line.strip() == 'codec_type=audio' and audio_codec is not None:
+                    return audio_codec
+        else:
+            # Stream #FILE_INDEX:STREAM_INDEX[STREAM_ID](LANGUAGE): CODEC_TYPE: CODEC_NAME
+            mobj = re.search(
+                r'Stream\s*#\d+:\d+(?:\[0x[0-9a-f]+\])?(?:\([a-z]{3}\))?:\s*Audio:\s*([0-9a-z]+)',
+                output)
+            if mobj:
+                return mobj.group(1)
+        return None
+
     def run_ffmpeg_multiple_files(self, input_paths, out_path, opts):
         self.check_version()
 
@@ -139,10 +217,13 @@ class FFmpegPostProcessor(PostProcessor):
                 encodeArgument('-i'),
                 encodeFilename(self._ffmpeg_filename_argument(path), True)
             ])
-        cmd = ([encodeFilename(self.executable, True), encodeArgument('-y')] +
-               files_cmd +
-               [encodeArgument(o) for o in opts] +
-               [encodeFilename(self._ffmpeg_filename_argument(out_path), True)])
+        cmd = [encodeFilename(self.executable, True), encodeArgument('-y')]
+        # avconv does not have repeat option
+        if self.basename == 'ffmpeg':
+            cmd += [encodeArgument('-loglevel'), encodeArgument('repeat+info')]
+        cmd += (files_cmd
+                + [encodeArgument(o) for o in opts]
+                + [encodeFilename(self._ffmpeg_filename_argument(out_path), True)])
 
         if self._downloader.params.get('verbose', False):
             self._downloader.to_screen('[debug] ffmpeg command line: %s' % shell_quote(cmd))
@@ -161,7 +242,8 @@ class FFmpegPostProcessor(PostProcessor):
         # Always use 'file:' because the filename may contain ':' (ffmpeg
         # interprets that as a protocol) or can start with '-' (-- is broken in
         # ffmpeg, see https://ffmpeg.org/trac/ffmpeg/ticket/2127 for details)
-        return 'file:' + fn
+        # Also leave '-' intact in order not to break streaming to stdout.
+        return 'file:' + fn if fn != '-' else fn
 
 
 class FFmpegExtractAudioPP(FFmpegPostProcessor):
@@ -173,31 +255,6 @@ class FFmpegExtractAudioPP(FFmpegPostProcessor):
         self._preferredquality = preferredquality
         self._nopostoverwrites = nopostoverwrites
 
-    def get_audio_codec(self, path):
-
-        if not self.probe_available:
-            raise PostProcessingError('ffprobe or avprobe not found. Please install one.')
-        try:
-            cmd = [
-                encodeFilename(self.probe_executable, True),
-                encodeArgument('-show_streams'),
-                encodeFilename(self._ffmpeg_filename_argument(path), True)]
-            if self._downloader.params.get('verbose', False):
-                self._downloader.to_screen('[debug] %s command line: %s' % (self.basename, shell_quote(cmd)))
-            handle = subprocess.Popen(cmd, stderr=compat_subprocess_get_DEVNULL(), stdout=subprocess.PIPE, stdin=subprocess.PIPE)
-            output = handle.communicate()[0]
-            if handle.wait() != 0:
-                return None
-        except (IOError, OSError):
-            return None
-        audio_codec = None
-        for line in output.decode('ascii', 'ignore').split('\n'):
-            if line.startswith('codec_name='):
-                audio_codec = line.split('=')[1].strip()
-            elif line.strip() == 'codec_type=audio' and audio_codec is not None:
-                return audio_codec
-        return None
-
     def run_ffmpeg(self, path, out_path, codec, more_opts):
         if codec is None:
             acodec_opts = []
@@ -223,7 +280,7 @@ class FFmpegExtractAudioPP(FFmpegPostProcessor):
                 acodec = 'copy'
                 extension = 'm4a'
                 more_opts = ['-bsf:a', 'aac_adtstoasc']
-            elif filecodec in ['aac', 'mp3', 'vorbis', 'opus']:
+            elif filecodec in ['aac', 'flac', 'mp3', 'vorbis', 'opus']:
                 # Lossless if possible
                 acodec = 'copy'
                 extension = filecodec
@@ -242,8 +299,8 @@ class FFmpegExtractAudioPP(FFmpegPostProcessor):
                     else:
                         more_opts += ['-b:a', self._preferredquality + 'k']
         else:
-            # We convert the audio (lossy)
-            acodec = {'mp3': 'libmp3lame', 'aac': 'aac', 'm4a': 'aac', 'opus': 'opus', 'vorbis': 'libvorbis', 'wav': None}[self._preferredcodec]
+            # We convert the audio (lossy if codec is lossy)
+            acodec = ACODECS[self._preferredcodec]
             extension = self._preferredcodec
             more_opts = []
             if self._preferredquality is not None:
@@ -265,9 +322,12 @@ class FFmpegExtractAudioPP(FFmpegPostProcessor):
         prefix, sep, ext = path.rpartition('.')  # not os.path.splitext, since the latter does not work on unicode in all setups
         new_path = prefix + sep + extension
 
+        information['filepath'] = new_path
+        information['ext'] = extension
+
         # If we download foo.mp3 and convert it to... foo.mp3, then don't delete foo.mp3, silly.
-        if (new_path == path or
-                (self._nopostoverwrites and os.path.exists(encodeFilename(new_path)))):
+        if (new_path == path
+                or (self._nopostoverwrites and os.path.exists(encodeFilename(new_path)))):
             self._downloader.to_screen('[ffmpeg] Post-process file %s exists, skipping' % new_path)
             return [], information
 
@@ -286,9 +346,6 @@ class FFmpegExtractAudioPP(FFmpegPostProcessor):
                 new_path, time.time(), information['filetime'],
                 errnote='Cannot update utime of audio file')
 
-        information['filepath'] = new_path
-        information['ext'] = extension
-
         return [path], information
 
 
@@ -317,17 +374,34 @@ class FFmpegVideoConvertorPP(FFmpegPostProcessor):
 
 class FFmpegEmbedSubtitlePP(FFmpegPostProcessor):
     def run(self, information):
-        if information['ext'] not in ['mp4', 'mkv']:
-            self._downloader.to_screen('[ffmpeg] Subtitles can only be embedded in mp4 or mkv files')
+        if information['ext'] not in ('mp4', 'webm', 'mkv'):
+            self._downloader.to_screen('[ffmpeg] Subtitles can only be embedded in mp4, webm or mkv files')
             return [], information
         subtitles = information.get('requested_subtitles')
         if not subtitles:
             self._downloader.to_screen('[ffmpeg] There aren\'t any subtitles to embed')
             return [], information
 
-        sub_langs = list(subtitles.keys())
         filename = information['filepath']
-        sub_filenames = [subtitles_filename(filename, lang, sub_info['ext']) for lang, sub_info in subtitles.items()]
+
+        ext = information['ext']
+        sub_langs = []
+        sub_filenames = []
+        webm_vtt_warn = False
+
+        for lang, sub_info in subtitles.items():
+            sub_ext = sub_info['ext']
+            if ext != 'webm' or ext == 'webm' and sub_ext == 'vtt':
+                sub_langs.append(lang)
+                sub_filenames.append(subtitles_filename(filename, lang, sub_ext, ext))
+            else:
+                if not webm_vtt_warn and ext == 'webm' and sub_ext != 'vtt':
+                    webm_vtt_warn = True
+                    self._downloader.to_screen('[ffmpeg] Only WebVTT subtitles can be embedded in webm files')
+
+        if not sub_langs:
+            return [], information
+
         input_files = [filename] + sub_filenames
 
         opts = [
@@ -336,14 +410,16 @@ class FFmpegEmbedSubtitlePP(FFmpegPostProcessor):
             # Don't copy the existing subtitles, we may be running the
             # postprocessor a second time
             '-map', '-0:s',
+            # Don't copy Apple TV chapters track, bin_data (see #19042, #19024,
+            # https://trac.ffmpeg.org/ticket/6016)
+            '-map', '-0:d',
         ]
         if information['ext'] == 'mp4':
             opts += ['-c:s', 'mov_text']
         for (i, lang) in enumerate(sub_langs):
             opts.extend(['-map', '%d:0' % (i + 1)])
-            lang_code = ISO639Utils.short2long(lang)
-            if lang_code is not None:
-                opts.extend(['-metadata:s:s:%d' % i, 'language=%s' % lang_code])
+            lang_code = ISO639Utils.short2long(lang) or lang
+            opts.extend(['-metadata:s:s:%d' % i, 'language=%s' % lang_code])
 
         temp_filename = prepend_extension(filename, 'temp')
         self._downloader.to_screen('[ffmpeg] Embedding subtitles in \'%s\'' % filename)
@@ -357,23 +433,41 @@ class FFmpegEmbedSubtitlePP(FFmpegPostProcessor):
 class FFmpegMetadataPP(FFmpegPostProcessor):
     def run(self, info):
         metadata = {}
-        if info.get('title') is not None:
-            metadata['title'] = info['title']
-        if info.get('upload_date') is not None:
-            metadata['date'] = info['upload_date']
-        if info.get('artist') is not None:
-            metadata['artist'] = info['artist']
-        elif info.get('uploader') is not None:
-            metadata['artist'] = info['uploader']
-        elif info.get('uploader_id') is not None:
-            metadata['artist'] = info['uploader_id']
-        if info.get('description') is not None:
-            metadata['description'] = info['description']
-            metadata['comment'] = info['description']
-        if info.get('webpage_url') is not None:
-            metadata['purl'] = info['webpage_url']
-        if info.get('album') is not None:
-            metadata['album'] = info['album']
+
+        def add(meta_list, info_list=None):
+            if not info_list:
+                info_list = meta_list
+            if not isinstance(meta_list, (list, tuple)):
+                meta_list = (meta_list,)
+            if not isinstance(info_list, (list, tuple)):
+                info_list = (info_list,)
+            for info_f in info_list:
+                if info.get(info_f) is not None:
+                    for meta_f in meta_list:
+                        metadata[meta_f] = info[info_f]
+                    break
+
+        # See [1-4] for some info on media metadata/metadata supported
+        # by ffmpeg.
+        # 1. https://kdenlive.org/en/project/adding-meta-data-to-mp4-video/
+        # 2. https://wiki.multimedia.cx/index.php/FFmpeg_Metadata
+        # 3. https://kodi.wiki/view/Video_file_tagging
+        # 4. http://atomicparsley.sourceforge.net/mpeg-4files.html
+
+        add('title', ('track', 'title'))
+        add('date', 'upload_date')
+        add(('description', 'comment'), 'description')
+        add('purl', 'webpage_url')
+        add('track', 'track_number')
+        add('artist', ('artist', 'creator', 'uploader', 'uploader_id'))
+        add('genre')
+        add('album')
+        add('album_artist')
+        add('disc', 'disc_number')
+        add('show', 'series')
+        add('season_number')
+        add('episode_id', ('episode', 'episode_id'))
+        add('episode_sort', 'episode_number')
 
         if not metadata:
             self._downloader.to_screen('[ffmpeg] There isn\'t any metadata to add')
@@ -381,17 +475,40 @@ class FFmpegMetadataPP(FFmpegPostProcessor):
 
         filename = info['filepath']
         temp_filename = prepend_extension(filename, 'temp')
+        in_filenames = [filename]
+        options = []
 
         if info['ext'] == 'm4a':
-            options = ['-vn', '-acodec', 'copy']
+            options.extend(['-vn', '-acodec', 'copy'])
         else:
-            options = ['-c', 'copy']
+            options.extend(['-c', 'copy'])
 
         for (name, value) in metadata.items():
             options.extend(['-metadata', '%s=%s' % (name, value)])
 
+        chapters = info.get('chapters', [])
+        if chapters:
+            metadata_filename = replace_extension(filename, 'meta')
+            with io.open(metadata_filename, 'wt', encoding='utf-8') as f:
+                def ffmpeg_escape(text):
+                    return re.sub(r'(=|;|#|\\|\n)', r'\\\1', text)
+
+                metadata_file_content = ';FFMETADATA1\n'
+                for chapter in chapters:
+                    metadata_file_content += '[CHAPTER]\nTIMEBASE=1/1000\n'
+                    metadata_file_content += 'START=%d\n' % (chapter['start_time'] * 1000)
+                    metadata_file_content += 'END=%d\n' % (chapter['end_time'] * 1000)
+                    chapter_title = chapter.get('title')
+                    if chapter_title:
+                        metadata_file_content += 'title=%s\n' % ffmpeg_escape(chapter_title)
+                f.write(metadata_file_content)
+                in_filenames.append(metadata_filename)
+                options.extend(['-map_metadata', '1'])
+
         self._downloader.to_screen('[ffmpeg] Adding metadata to \'%s\'' % filename)
-        self.run_ffmpeg(filename, temp_filename, options)
+        self.run_ffmpeg_multiple_files(in_filenames, temp_filename, options)
+        if chapters:
+            os.remove(metadata_filename)
         os.remove(encodeFilename(filename))
         os.rename(encodeFilename(temp_filename), encodeFilename(filename))
         return [], info
@@ -462,6 +579,21 @@ class FFmpegFixupM4aPP(FFmpegPostProcessor):
         return [], info
 
 
+class FFmpegFixupM3u8PP(FFmpegPostProcessor):
+    def run(self, info):
+        filename = info['filepath']
+        if self.get_audio_codec(filename) == 'aac':
+            temp_filename = prepend_extension(filename, 'temp')
+
+            options = ['-c', 'copy', '-f', 'mp4', '-bsf:a', 'aac_adtstoasc']
+            self._downloader.to_screen('[ffmpeg] Fixing malformed AAC bitstream in "%s"' % filename)
+            self.run_ffmpeg(filename, temp_filename, options)
+
+            os.remove(encodeFilename(filename))
+            os.rename(encodeFilename(temp_filename), encodeFilename(filename))
+        return [], info
+
+
 class FFmpegSubtitlesConvertorPP(FFmpegPostProcessor):
     def __init__(self, downloader=None, format=None):
         super(FFmpegSubtitlesConvertorPP, self).__init__(downloader)
@@ -478,30 +610,32 @@ class FFmpegSubtitlesConvertorPP(FFmpegPostProcessor):
             self._downloader.to_screen('[ffmpeg] There aren\'t any subtitles to convert')
             return [], info
         self._downloader.to_screen('[ffmpeg] Converting subtitles')
+        sub_filenames = []
         for lang, sub in subs.items():
             ext = sub['ext']
             if ext == new_ext:
                 self._downloader.to_screen(
-                    '[ffmpeg] Subtitle file for %s is already in the requested'
-                    'format' % new_ext)
+                    '[ffmpeg] Subtitle file for %s is already in the requested format' % new_ext)
                 continue
-            new_file = subtitles_filename(filename, lang, new_ext)
+            old_file = subtitles_filename(filename, lang, ext, info.get('ext'))
+            sub_filenames.append(old_file)
+            new_file = subtitles_filename(filename, lang, new_ext, info.get('ext'))
 
-            if ext == 'dfxp' or ext == 'ttml':
+            if ext in ('dfxp', 'ttml', 'tt'):
                 self._downloader.report_warning(
                     'You have requested to convert dfxp (TTML) subtitles into another format, '
                     'which results in style information loss')
 
-                dfxp_file = subtitles_filename(filename, lang, ext)
-                srt_file = subtitles_filename(filename, lang, 'srt')
+                dfxp_file = old_file
+                srt_file = subtitles_filename(filename, lang, 'srt', info.get('ext'))
 
-                with io.open(dfxp_file, 'rt', encoding='utf-8') as f:
+                with open(dfxp_file, 'rb') as f:
                     srt_data = dfxp2srt(f.read())
 
                 with io.open(srt_file, 'wt', encoding='utf-8') as f:
                     f.write(srt_data)
+                old_file = srt_file
 
-                ext = 'srt'
                 subs[lang] = {
                     'ext': 'srt',
                     'data': srt_data
@@ -509,15 +643,15 @@ class FFmpegSubtitlesConvertorPP(FFmpegPostProcessor):
 
                 if new_ext == 'srt':
                     continue
+                else:
+                    sub_filenames.append(srt_file)
 
-            self.run_ffmpeg(
-                subtitles_filename(filename, lang, ext),
-                new_file, ['-f', new_format])
+            self.run_ffmpeg(old_file, new_file, ['-f', new_format])
 
             with io.open(new_file, 'rt', encoding='utf-8') as f:
                 subs[lang] = {
-                    'ext': ext,
+                    'ext': new_ext,
                     'data': f.read(),
                 }
 
-        return [], info
+        return sub_filenames, info