Merge pull request #3 from unit193/master

[youtubedl] / youtube_dl / postprocessor / ffmpeg.py

diff --git a/youtube_dl/postprocessor/ffmpeg.py b/youtube_dl/postprocessor/ffmpeg.py
index 083c79592b54f069cc25512017260f10882fb76b..fd3f921a8a11da2e8c31573889ea4d7f5a9fea25 100644 (file)
--- a/youtube_dl/postprocessor/ffmpeg.py
+++ b/youtube_dl/postprocessor/ffmpeg.py
@@ -1,39 +1,50 @@
+from __future__ import unicode_literals
+
+import io

 import os
 import os

-import re

 import subprocess
 import subprocess

-import sys

 import time
 import time

+import re

 
 
 from .common import AudioConversionError, PostProcessor
 
 from ..utils import (
 
 
 from .common import AudioConversionError, PostProcessor
 
 from ..utils import (

-    compat_subprocess_get_DEVNULL,

     encodeArgument,
     encodeFilename,
     encodeArgument,
     encodeFilename,

+    get_exe_version,

     is_outdated_version,
     PostProcessingError,
     prepend_extension,
     shell_quote,
     subtitles_filename,
     is_outdated_version,
     PostProcessingError,
     prepend_extension,
     shell_quote,
     subtitles_filename,

+    dfxp2srt,
+    ISO639Utils,
+    replace_extension,

 )
 
 
 )
 
 

-def get_version(executable):
-    """ Returns the version of the specified executable,
-    or False if the executable is not present """
-    try:
-        out, err = subprocess.Popen(
-            [executable, '-version'],
-            stdout=subprocess.PIPE, stderr=subprocess.STDOUT).communicate()
-    except OSError:
-        return False
-    firstline = out.partition(b'\n')[0].decode('ascii', 'ignore')
-    m = re.search(r'version\s+([0-9._-a-zA-Z]+)', firstline)
-    if not m:
-        return u'present'
-    else:
-        return m.group(1)
+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):
 
 
 class FFmpegPostProcessorError(PostProcessingError):


@@ -41,83 +52,198 @@ class FFmpegPostProcessorError(PostProcessingError):
 
 
 class FFmpegPostProcessor(PostProcessor):
 
 
 class FFmpegPostProcessor(PostProcessor):

-    def __init__(self, downloader=None, deletetempfiles=False):
+    def __init__(self, downloader=None):

         PostProcessor.__init__(self, downloader)
         PostProcessor.__init__(self, downloader)

-        self._versions = self.get_versions()
-        self._deletetempfiles = deletetempfiles
+        self._determine_executables()

 
     def check_version(self):
 
     def check_version(self):

-        if not self._executable:
-            raise FFmpegPostProcessorError(u'ffmpeg or avconv not found. Please install one.')
+        if not self.available:
+            raise FFmpegPostProcessorError('ffmpeg or avconv not found. Please install one.')

 
 

-        REQUIRED_VERSION = '1.0'
+        required_version = '10-0' if self.basename == 'avconv' else '1.0'

         if is_outdated_version(
         if is_outdated_version(

-                self._versions[self._executable], REQUIRED_VERSION):
-            warning = u'Your copy of %s is outdated, update %s to version %s or newer if you encounter any errors.' % (
-                self._executable, self._executable, REQUIRED_VERSION)
+                self._versions[self.basename], required_version):
+            warning = 'Your copy of %s is outdated, update %s to version %s or newer if you encounter any errors.' % (
+                self.basename, self.basename, required_version)

             if self._downloader:
                 self._downloader.report_warning(warning)
 
     @staticmethod
             if self._downloader:
                 self._downloader.report_warning(warning)
 
     @staticmethod

-    def get_versions():
-        programs = ['avprobe', 'avconv', 'ffmpeg', 'ffprobe']
-        return dict((program, get_version(program)) for program in programs)
+    def get_versions(downloader=None):
+        return FFmpegPostProcessor(downloader)._versions

 
 

-    @property
-    def _executable(self):
-        if self._downloader.params.get('prefer_ffmpeg', False):
-            prefs = ('ffmpeg', 'avconv')
-        else:
+    def _determine_executables(self):
+        programs = ['avprobe', 'avconv', 'ffmpeg', 'ffprobe']
+        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
+
+        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):
+                    self._downloader.report_warning(
+                        'ffmpeg-location %s does not exist! '
+                        'Continuing without avconv/ffmpeg.' % (location))
+                    self._versions = {}
+                    return
+                elif not os.path.isdir(location):
+                    basename = os.path.splitext(os.path.basename(location))[0]
+                    if basename not in programs:
+                        self._downloader.report_warning(
+                            'Cannot identify executable %s, its basename should be one of %s. '
+                            'Continuing without avconv/ffmpeg.' %
+                            (location, ', '.join(programs)))
+                        self._versions = {}
+                        return None
+                    location = os.path.dirname(os.path.abspath(location))
+                    if basename in ('ffmpeg', 'ffprobe'):
+                        prefer_ffmpeg = True
+
+                self._paths = dict(
+                    (p, os.path.join(location, p)) for p in programs)
+                self._versions = dict(
+                    (p, get_ffmpeg_version(self._paths[p])) for p in programs)
+        if self._versions is None:
+            self._versions = dict(
+                (p, get_ffmpeg_version(p)) for p in programs)
+            self._paths = dict((p, p) for p in programs)
+
+        if prefer_ffmpeg is False:

             prefs = ('avconv', 'ffmpeg')
             prefs = ('avconv', 'ffmpeg')

+        else:
+            prefs = ('ffmpeg', 'avconv')

         for p in prefs:
             if self._versions[p]:
         for p in prefs:
             if self._versions[p]:

-                return p
-        return None
+                self.basename = p
+                break

 
 

-    @property
-    def _probe_executable(self):
-        if self._downloader.params.get('prefer_ffmpeg', False):
-            prefs = ('ffprobe', 'avprobe')
-        else:
+        if prefer_ffmpeg is False:

             prefs = ('avprobe', 'ffprobe')
             prefs = ('avprobe', 'ffprobe')

+        else:
+            prefs = ('ffprobe', 'avprobe')

         for p in prefs:
             if self._versions[p]:
         for p in prefs:
             if self._versions[p]:

-                return p
-        return None
+                self.probe_basename = p
+                break
+
+    @property
+    def available(self):
+        return self.basename is not None
+
+    @property
+    def executable(self):
+        return self._paths[self.basename]

 
 

-    def _uses_avconv(self):
-        return self._executable == 'avconv'
+    @property
+    def probe_available(self):
+        return self.probe_basename is not None
+
+    @property
+    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()
 
 
     def run_ffmpeg_multiple_files(self, input_paths, out_path, opts):
         self.check_version()
 

+        oldest_mtime = min(
+            os.stat(encodeFilename(path)).st_mtime for path in input_paths)
+
+        opts += self._configuration_args()
+

         files_cmd = []
         for path in input_paths:
         files_cmd = []
         for path in input_paths:

-            files_cmd.extend(['-i', encodeFilename(path, True)])
-        cmd = ([self._executable, '-y'] + files_cmd
-               + [encodeArgument(o) for o in opts] +
-               [encodeFilename(self._ffmpeg_filename_argument(out_path), True)])
+            files_cmd.extend([
+                encodeArgument('-i'),
+                encodeFilename(self._ffmpeg_filename_argument(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):
 
         if self._downloader.params.get('verbose', False):

-            self._downloader.to_screen(u'[debug] ffmpeg command line: %s' % shell_quote(cmd))
-        p = subprocess.Popen(cmd, stdout=subprocess.PIPE, stderr=subprocess.PIPE)
+            self._downloader.to_screen('[debug] ffmpeg command line: %s' % shell_quote(cmd))
+        p = subprocess.Popen(cmd, stdout=subprocess.PIPE, stderr=subprocess.PIPE, stdin=subprocess.PIPE)

         stdout, stderr = p.communicate()
         if p.returncode != 0:
             stderr = stderr.decode('utf-8', 'replace')
             msg = stderr.strip().split('\n')[-1]
             raise FFmpegPostProcessorError(msg)
         stdout, stderr = p.communicate()
         if p.returncode != 0:
             stderr = stderr.decode('utf-8', 'replace')
             msg = stderr.strip().split('\n')[-1]
             raise FFmpegPostProcessorError(msg)

-        if self._deletetempfiles:
-            for ipath in input_paths:
-                os.remove(ipath)
+        self.try_utime(out_path, oldest_mtime, oldest_mtime)

 
     def run_ffmpeg(self, path, out_path, opts):
         self.run_ffmpeg_multiple_files([path], out_path, opts)
 
     def _ffmpeg_filename_argument(self, fn):
 
     def run_ffmpeg(self, path, out_path, opts):
         self.run_ffmpeg_multiple_files([path], out_path, opts)
 
     def _ffmpeg_filename_argument(self, fn):

-        # ffmpeg broke --, see https://ffmpeg.org/trac/ffmpeg/ticket/2127 for details
-        if fn.startswith(u'-'):
-            return u'./' + fn
-        return fn
+        # 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)
+        # Also leave '-' intact in order not to break streaming to stdout.
+        return 'file:' + fn if fn != '-' else fn

 
 
 class FFmpegExtractAudioPP(FFmpegPostProcessor):
 
 
 class FFmpegExtractAudioPP(FFmpegPostProcessor):


@@ -129,29 +255,6 @@ class FFmpegExtractAudioPP(FFmpegPostProcessor):
         self._preferredquality = preferredquality
         self._nopostoverwrites = nopostoverwrites
 
         self._preferredquality = preferredquality
         self._nopostoverwrites = nopostoverwrites
 

-    def get_audio_codec(self, path):
-
-        if not self._probe_executable:
-            raise PostProcessingError(u'ffprobe or avprobe not found. Please install one.')
-        try:
-            cmd = [
-                self._probe_executable,
-                '-show_streams',
-                encodeFilename(self._ffmpeg_filename_argument(path), True)]
-            handle = subprocess.Popen(cmd, stderr=compat_subprocess_get_DEVNULL(), stdout=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 = []
     def run_ffmpeg(self, path, out_path, codec, more_opts):
         if codec is None:
             acodec_opts = []


@@ -168,17 +271,16 @@ class FFmpegExtractAudioPP(FFmpegPostProcessor):
 
         filecodec = self.get_audio_codec(path)
         if filecodec is None:
 
         filecodec = self.get_audio_codec(path)
         if filecodec is None:

-            raise PostProcessingError(u'WARNING: unable to obtain file audio codec with ffprobe')
+            raise PostProcessingError('WARNING: unable to obtain file audio codec with ffprobe')

 
 

-        uses_avconv = self._uses_avconv()

         more_opts = []
         if self._preferredcodec == 'best' or self._preferredcodec == filecodec or (self._preferredcodec == 'm4a' and filecodec == 'aac'):
             if filecodec == 'aac' and self._preferredcodec in ['m4a', 'best']:
                 # Lossless, but in another container
                 acodec = 'copy'
                 extension = 'm4a'
         more_opts = []
         if self._preferredcodec == 'best' or self._preferredcodec == filecodec or (self._preferredcodec == 'm4a' and filecodec == 'aac'):
             if filecodec == 'aac' and self._preferredcodec in ['m4a', 'best']:
                 # Lossless, but in another container
                 acodec = 'copy'
                 extension = 'm4a'

-                more_opts = ['-bsf:a' if uses_avconv else '-absf', 'aac_adtstoasc']
-            elif filecodec in ['aac', 'mp3', 'vorbis', 'opus']:
+                more_opts = ['-bsf:a', 'aac_adtstoasc']
+            elif filecodec in ['aac', 'flac', 'mp3', 'vorbis', 'opus']:

                 # Lossless if possible
                 acodec = 'copy'
                 extension = filecodec
                 # Lossless if possible
                 acodec = 'copy'
                 extension = filecodec


@@ -193,362 +295,352 @@ class FFmpegExtractAudioPP(FFmpegPostProcessor):
                 more_opts = []
                 if self._preferredquality is not None:
                     if int(self._preferredquality) < 10:
                 more_opts = []
                 if self._preferredquality is not None:
                     if int(self._preferredquality) < 10:

-                        more_opts += ['-q:a' if uses_avconv else '-aq', self._preferredquality]
+                        more_opts += ['-q:a', self._preferredquality]

                     else:
                     else:

-                        more_opts += ['-b:a' if uses_avconv else '-ab', self._preferredquality + 'k']
+                        more_opts += ['-b:a', self._preferredquality + 'k']

         else:
         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:
                 # The opus codec doesn't support the -aq option
                 if int(self._preferredquality) < 10 and extension != 'opus':
             extension = self._preferredcodec
             more_opts = []
             if self._preferredquality is not None:
                 # The opus codec doesn't support the -aq option
                 if int(self._preferredquality) < 10 and extension != 'opus':

-                    more_opts += ['-q:a' if uses_avconv else '-aq', self._preferredquality]
+                    more_opts += ['-q:a', self._preferredquality]

                 else:
                 else:

-                    more_opts += ['-b:a' if uses_avconv else '-ab', self._preferredquality + 'k']
+                    more_opts += ['-b:a', self._preferredquality + 'k']

             if self._preferredcodec == 'aac':
                 more_opts += ['-f', 'adts']
             if self._preferredcodec == 'm4a':
             if self._preferredcodec == 'aac':
                 more_opts += ['-f', 'adts']
             if self._preferredcodec == 'm4a':

-                more_opts += ['-bsf:a' if uses_avconv else '-absf', 'aac_adtstoasc']
+                more_opts += ['-bsf:a', 'aac_adtstoasc']

             if self._preferredcodec == 'vorbis':
                 extension = 'ogg'
             if self._preferredcodec == 'wav':
                 extension = 'wav'
                 more_opts += ['-f', 'wav']
 
             if self._preferredcodec == 'vorbis':
                 extension = 'ogg'
             if self._preferredcodec == 'wav':
                 extension = 'wav'
                 more_opts += ['-f', 'wav']
 

-        prefix, sep, ext = path.rpartition(u'.') # not os.path.splitext, since the latter does not work on unicode in all setups
+        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
 
         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 we download foo.mp3 and convert it to... foo.mp3, then don't delete foo.mp3, silly.

-        if new_path == path:
-            self._nopostoverwrites = True
+        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

 
         try:
 
         try:

-            if self._nopostoverwrites and os.path.exists(encodeFilename(new_path)):
-                self._downloader.to_screen(u'[youtube] Post-process file %s exists, skipping' % new_path)
-            else:
-                self._downloader.to_screen(u'[' + self._executable + '] Destination: ' + new_path)
-                self.run_ffmpeg(path, new_path, acodec, more_opts)
-        except:
-            etype,e,tb = sys.exc_info()
-            if isinstance(e, AudioConversionError):
-                msg = u'audio conversion failed: ' + e.msg
-            else:
-                msg = u'error running ' + self._executable
-            raise PostProcessingError(msg)
+            self._downloader.to_screen('[ffmpeg] Destination: ' + new_path)
+            self.run_ffmpeg(path, new_path, acodec, more_opts)
+        except AudioConversionError as e:
+            raise PostProcessingError(
+                'audio conversion failed: ' + e.msg)
+        except Exception:
+            raise PostProcessingError('error running ' + self.basename)

 
         # Try to update the date time for extracted audio file.
         if information.get('filetime') is not None:
 
         # Try to update the date time for extracted audio file.
         if information.get('filetime') is not None:

-            try:
-                os.utime(encodeFilename(new_path), (time.time(), information['filetime']))
-            except:
-                self._downloader.report_warning(u'Cannot update utime of audio file')
+            self.try_utime(
+                new_path, time.time(), information['filetime'],
+                errnote='Cannot update utime of audio file')

 
 

-        information['filepath'] = new_path
-        return self._nopostoverwrites,information
+        return [path], information

 
 
 
 

-class FFmpegVideoConvertor(FFmpegPostProcessor):
-    def __init__(self, downloader=None,preferedformat=None):
-        super(FFmpegVideoConvertor, self).__init__(downloader)
-        self._preferedformat=preferedformat
+class FFmpegVideoConvertorPP(FFmpegPostProcessor):
+    def __init__(self, downloader=None, preferedformat=None):
+        super(FFmpegVideoConvertorPP, self).__init__(downloader)
+        self._preferedformat = preferedformat

 
     def run(self, information):
         path = information['filepath']
 
     def run(self, information):
         path = information['filepath']

-        prefix, sep, ext = path.rpartition(u'.')
-        outpath = prefix + sep + self._preferedformat

         if information['ext'] == self._preferedformat:
         if information['ext'] == self._preferedformat:

-            self._downloader.to_screen(u'[ffmpeg] Not converting video file %s - already is in target format %s' % (path, self._preferedformat))
-            return True,information
-        self._downloader.to_screen(u'['+'ffmpeg'+'] Converting video from %s to %s, Destination: ' % (information['ext'], self._preferedformat) +outpath)
-        self.run_ffmpeg(path, outpath, [])
+            self._downloader.to_screen('[ffmpeg] Not converting video file %s - already is in target format %s' % (path, self._preferedformat))
+            return [], information
+        options = []
+        if self._preferedformat == 'avi':
+            options.extend(['-c:v', 'libxvid', '-vtag', 'XVID'])
+        prefix, sep, ext = path.rpartition('.')
+        outpath = prefix + sep + self._preferedformat
+        self._downloader.to_screen('[' + 'ffmpeg' + '] Converting video from %s to %s, Destination: ' % (information['ext'], self._preferedformat) + outpath)
+        self.run_ffmpeg(path, outpath, options)

         information['filepath'] = outpath
         information['format'] = self._preferedformat
         information['ext'] = self._preferedformat
         information['filepath'] = outpath
         information['format'] = self._preferedformat
         information['ext'] = self._preferedformat

-        return False,information
+        return [path], information

 
 
 class FFmpegEmbedSubtitlePP(FFmpegPostProcessor):
 
 
 class FFmpegEmbedSubtitlePP(FFmpegPostProcessor):

-    # See http://www.loc.gov/standards/iso639-2/ISO-639-2_utf-8.txt
-    _lang_map = {
-        'aa': 'aar',
-        'ab': 'abk',
-        'ae': 'ave',
-        'af': 'afr',
-        'ak': 'aka',
-        'am': 'amh',
-        'an': 'arg',
-        'ar': 'ara',
-        'as': 'asm',
-        'av': 'ava',
-        'ay': 'aym',
-        'az': 'aze',
-        'ba': 'bak',
-        'be': 'bel',
-        'bg': 'bul',
-        'bh': 'bih',
-        'bi': 'bis',
-        'bm': 'bam',
-        'bn': 'ben',
-        'bo': 'bod',
-        'br': 'bre',
-        'bs': 'bos',
-        'ca': 'cat',
-        'ce': 'che',
-        'ch': 'cha',
-        'co': 'cos',
-        'cr': 'cre',
-        'cs': 'ces',
-        'cu': 'chu',
-        'cv': 'chv',
-        'cy': 'cym',
-        'da': 'dan',
-        'de': 'deu',
-        'dv': 'div',
-        'dz': 'dzo',
-        'ee': 'ewe',
-        'el': 'ell',
-        'en': 'eng',
-        'eo': 'epo',
-        'es': 'spa',
-        'et': 'est',
-        'eu': 'eus',
-        'fa': 'fas',
-        'ff': 'ful',
-        'fi': 'fin',
-        'fj': 'fij',
-        'fo': 'fao',
-        'fr': 'fra',
-        'fy': 'fry',
-        'ga': 'gle',
-        'gd': 'gla',
-        'gl': 'glg',
-        'gn': 'grn',
-        'gu': 'guj',
-        'gv': 'glv',
-        'ha': 'hau',
-        'he': 'heb',
-        'hi': 'hin',
-        'ho': 'hmo',
-        'hr': 'hrv',
-        'ht': 'hat',
-        'hu': 'hun',
-        'hy': 'hye',
-        'hz': 'her',
-        'ia': 'ina',
-        'id': 'ind',
-        'ie': 'ile',
-        'ig': 'ibo',
-        'ii': 'iii',
-        'ik': 'ipk',
-        'io': 'ido',
-        'is': 'isl',
-        'it': 'ita',
-        'iu': 'iku',
-        'ja': 'jpn',
-        'jv': 'jav',
-        'ka': 'kat',
-        'kg': 'kon',
-        'ki': 'kik',
-        'kj': 'kua',
-        'kk': 'kaz',
-        'kl': 'kal',
-        'km': 'khm',
-        'kn': 'kan',
-        'ko': 'kor',
-        'kr': 'kau',
-        'ks': 'kas',
-        'ku': 'kur',
-        'kv': 'kom',
-        'kw': 'cor',
-        'ky': 'kir',
-        'la': 'lat',
-        'lb': 'ltz',
-        'lg': 'lug',
-        'li': 'lim',
-        'ln': 'lin',
-        'lo': 'lao',
-        'lt': 'lit',
-        'lu': 'lub',
-        'lv': 'lav',
-        'mg': 'mlg',
-        'mh': 'mah',
-        'mi': 'mri',
-        'mk': 'mkd',
-        'ml': 'mal',
-        'mn': 'mon',
-        'mr': 'mar',
-        'ms': 'msa',
-        'mt': 'mlt',
-        'my': 'mya',
-        'na': 'nau',
-        'nb': 'nob',
-        'nd': 'nde',
-        'ne': 'nep',
-        'ng': 'ndo',
-        'nl': 'nld',
-        'nn': 'nno',
-        'no': 'nor',
-        'nr': 'nbl',
-        'nv': 'nav',
-        'ny': 'nya',
-        'oc': 'oci',
-        'oj': 'oji',
-        'om': 'orm',
-        'or': 'ori',
-        'os': 'oss',
-        'pa': 'pan',
-        'pi': 'pli',
-        'pl': 'pol',
-        'ps': 'pus',
-        'pt': 'por',
-        'qu': 'que',
-        'rm': 'roh',
-        'rn': 'run',
-        'ro': 'ron',
-        'ru': 'rus',
-        'rw': 'kin',
-        'sa': 'san',
-        'sc': 'srd',
-        'sd': 'snd',
-        'se': 'sme',
-        'sg': 'sag',
-        'si': 'sin',
-        'sk': 'slk',
-        'sl': 'slv',
-        'sm': 'smo',
-        'sn': 'sna',
-        'so': 'som',
-        'sq': 'sqi',
-        'sr': 'srp',
-        'ss': 'ssw',
-        'st': 'sot',
-        'su': 'sun',
-        'sv': 'swe',
-        'sw': 'swa',
-        'ta': 'tam',
-        'te': 'tel',
-        'tg': 'tgk',
-        'th': 'tha',
-        'ti': 'tir',
-        'tk': 'tuk',
-        'tl': 'tgl',
-        'tn': 'tsn',
-        'to': 'ton',
-        'tr': 'tur',
-        'ts': 'tso',
-        'tt': 'tat',
-        'tw': 'twi',
-        'ty': 'tah',
-        'ug': 'uig',
-        'uk': 'ukr',
-        'ur': 'urd',
-        'uz': 'uzb',
-        've': 'ven',
-        'vi': 'vie',
-        'vo': 'vol',
-        'wa': 'wln',
-        'wo': 'wol',
-        'xh': 'xho',
-        'yi': 'yid',
-        'yo': 'yor',
-        'za': 'zha',
-        'zh': 'zho',
-        'zu': 'zul',
-    }
-
-    def __init__(self, downloader=None, subtitlesformat='srt'):
-        super(FFmpegEmbedSubtitlePP, self).__init__(downloader)
-        self._subformat = subtitlesformat
-
-    @classmethod
-    def _conver_lang_code(cls, code):
-        """Convert language code from ISO 639-1 to ISO 639-2/T"""
-        return cls._lang_map.get(code[:2])
-

     def run(self, information):
     def run(self, information):

-        if information['ext'] != u'mp4':
-            self._downloader.to_screen(u'[ffmpeg] Subtitles can only be embedded in mp4 files')
-            return True, information
-        if not information.get('subtitles'):
-            self._downloader.to_screen(u'[ffmpeg] There aren\'t any subtitles to embed') 
-            return True, information
-
-        sub_langs = [key for key in information['subtitles']]
+        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
+

         filename = information['filepath']
         filename = information['filepath']

-        input_files = [filename] + [subtitles_filename(filename, lang, self._subformat) for lang in sub_langs]

 
 

-        opts = ['-map', '0:0', '-map', '0:1', '-c:v', 'copy', '-c:a', 'copy']
+        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 = [
+            '-map', '0',
+            '-c', 'copy',
+            # 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):
         for (i, lang) in enumerate(sub_langs):

-            opts.extend(['-map', '%d:0' % (i+1), '-c:s:%d' % i, 'mov_text'])
-            lang_code = self._conver_lang_code(lang)
-            if lang_code is not None:
-                opts.extend(['-metadata:s:s:%d' % i, 'language=%s' % lang_code])
-        opts.extend(['-f', 'mp4'])
-
-        temp_filename = filename + u'.temp'
-        self._downloader.to_screen(u'[ffmpeg] Embedding subtitles in \'%s\'' % filename)
+            opts.extend(['-map', '%d:0' % (i + 1)])
+            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)

         self.run_ffmpeg_multiple_files(input_files, temp_filename, opts)
         os.remove(encodeFilename(filename))
         os.rename(encodeFilename(temp_filename), encodeFilename(filename))
 
         self.run_ffmpeg_multiple_files(input_files, temp_filename, opts)
         os.remove(encodeFilename(filename))
         os.rename(encodeFilename(temp_filename), encodeFilename(filename))
 

-        return True, information
+        return sub_filenames, information

 
 
 class FFmpegMetadataPP(FFmpegPostProcessor):
     def run(self, info):
         metadata = {}
 
 
 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('uploader') is not None:
-            metadata['artist'] = info['uploader']
-        elif info.get('uploader_id') is not None:
-            metadata['artist'] = info['uploader_id']
+
+        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
+
+        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')

 
         if not metadata:
 
         if not metadata:

-            self._downloader.to_screen(u'[ffmpeg] There isn\'t any metadata to add')
-            return True, info
+            self._downloader.to_screen('[ffmpeg] There isn\'t any metadata to add')
+            return [], info

 
         filename = info['filepath']
         temp_filename = prepend_extension(filename, 'temp')
 
         filename = info['filepath']
         temp_filename = prepend_extension(filename, 'temp')

+        in_filenames = [filename]
+        options = []

 
 

-        if info['ext'] == u'm4a':
-            options = ['-vn', '-acodec', 'copy']
+        if info['ext'] == 'm4a':
+            options.extend(['-vn', '-acodec', 'copy'])

         else:
         else:

-            options = ['-c', 'copy']
+            options.extend(['-c', 'copy'])

 
         for (name, value) in metadata.items():
             options.extend(['-metadata', '%s=%s' % (name, value)])
 
 
         for (name, value) in metadata.items():
             options.extend(['-metadata', '%s=%s' % (name, value)])
 

-        self._downloader.to_screen(u'[ffmpeg] Adding metadata to \'%s\'' % filename)
-        self.run_ffmpeg(filename, temp_filename, options)
+        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_multiple_files(in_filenames, temp_filename, options)
+        if chapters:
+            os.remove(metadata_filename)

         os.remove(encodeFilename(filename))
         os.rename(encodeFilename(temp_filename), encodeFilename(filename))
         os.remove(encodeFilename(filename))
         os.rename(encodeFilename(temp_filename), encodeFilename(filename))

-        return True, info
+        return [], info

 
 
 class FFmpegMergerPP(FFmpegPostProcessor):
     def run(self, info):
         filename = info['filepath']
 
 
 class FFmpegMergerPP(FFmpegPostProcessor):
     def run(self, info):
         filename = info['filepath']

-        args = ['-c', 'copy', '-map', '0:v:0', '-map', '1:a:0', '-shortest']
-        self._downloader.to_screen(u'[ffmpeg] Merging formats into "%s"' % filename)
-        self.run_ffmpeg_multiple_files(info['__files_to_merge'], filename, args)
-        return True, info
+        temp_filename = prepend_extension(filename, 'temp')
+        args = ['-c', 'copy', '-map', '0:v:0', '-map', '1:a:0']
+        self._downloader.to_screen('[ffmpeg] Merging formats into "%s"' % filename)
+        self.run_ffmpeg_multiple_files(info['__files_to_merge'], temp_filename, args)
+        os.rename(encodeFilename(temp_filename), encodeFilename(filename))
+        return info['__files_to_merge'], info
+
+    def can_merge(self):
+        # TODO: figure out merge-capable ffmpeg version
+        if self.basename != 'avconv':
+            return True
+
+        required_version = '10-0'
+        if is_outdated_version(
+                self._versions[self.basename], required_version):
+            warning = ('Your copy of %s is outdated and unable to properly mux separate video and audio files, '
+                       'youtube-dl will download single file media. '
+                       'Update %s to version %s or newer to fix this.') % (
+                           self.basename, self.basename, required_version)
+            if self._downloader:
+                self._downloader.report_warning(warning)
+            return False
+        return True
+
+
+class FFmpegFixupStretchedPP(FFmpegPostProcessor):
+    def run(self, info):
+        stretched_ratio = info.get('stretched_ratio')
+        if stretched_ratio is None or stretched_ratio == 1:
+            return [], info
+
+        filename = info['filepath']
+        temp_filename = prepend_extension(filename, 'temp')
+
+        options = ['-c', 'copy', '-aspect', '%f' % stretched_ratio]
+        self._downloader.to_screen('[ffmpeg] Fixing aspect ratio in "%s"' % filename)
+        self.run_ffmpeg(filename, temp_filename, options)
+
+        os.remove(encodeFilename(filename))
+        os.rename(encodeFilename(temp_filename), encodeFilename(filename))

 
 

+        return [], info

 
 

-class FFmpegAudioFixPP(FFmpegPostProcessor):
+
+class FFmpegFixupM4aPP(FFmpegPostProcessor):

     def run(self, info):
     def run(self, info):

+        if info.get('container') != 'm4a_dash':
+            return [], info
+

         filename = info['filepath']
         temp_filename = prepend_extension(filename, 'temp')
 
         filename = info['filepath']
         temp_filename = prepend_extension(filename, 'temp')
 

-        options = ['-vn', '-acodec', 'copy']
-        self._downloader.to_screen(u'[ffmpeg] Fixing audio file "%s"' % filename)
+        options = ['-c', 'copy', '-f', 'mp4']
+        self._downloader.to_screen('[ffmpeg] Correcting container in "%s"' % filename)

         self.run_ffmpeg(filename, temp_filename, options)
 
         os.remove(encodeFilename(filename))
         os.rename(encodeFilename(temp_filename), encodeFilename(filename))
 
         self.run_ffmpeg(filename, temp_filename, options)
 
         os.remove(encodeFilename(filename))
         os.rename(encodeFilename(temp_filename), encodeFilename(filename))
 

-        return True, info
+        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)
+        self.format = format
+
+    def run(self, info):
+        subs = info.get('requested_subtitles')
+        filename = info['filepath']
+        new_ext = self.format
+        new_format = new_ext
+        if new_format == 'vtt':
+            new_format = 'webvtt'
+        if subs is None:
+            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)
+                continue
+            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 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 = old_file
+                srt_file = subtitles_filename(filename, lang, 'srt', info.get('ext'))
+
+                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
+
+                subs[lang] = {
+                    'ext': 'srt',
+                    'data': srt_data
+                }
+
+                if new_ext == 'srt':
+                    continue
+                else:
+                    sub_filenames.append(srt_file)
+
+            self.run_ffmpeg(old_file, new_file, ['-f', new_format])
+
+            with io.open(new_file, 'rt', encoding='utf-8') as f:
+                subs[lang] = {
+                    'ext': new_ext,
+                    'data': f.read(),
+                }
+
+        return sub_filenames, info