3 # smooth-dl - download videos served using Smooth Streaming technology
5 # Copyright (C) 2010 Antonio Ospite <ospite@studenti.unina.it>
7 # This program is free software: you can redistribute it and/or modify
8 # it under the terms of the GNU General Public License as published by
9 # the Free Software Foundation, either version 3 of the License, or
10 # (at your option) any later version.
12 # This program is distributed in the hope that it will be useful,
13 # but WITHOUT ANY WARRANTY; without even the implied warranty of
14 # MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
15 # GNU General Public License for more details.
17 # You should have received a copy of the GNU General Public License
18 # along with this program. If not, see <http://www.gnu.org/licenses/>.
22 # - Handle HTTP errors:
23 # "Connection reset by peer"
24 # "Resource not available"
26 # - Support more Manifest formats:
27 # WaveFormatEx attribute instead of PrivateCodecdata
28 # 'd' and other attributes in chunk element ('i', 's', 'q')
30 # basically, write a proper implementation of manifest parsing and chunk
34 __description = "Download videos served using Smooth Streaming technology"
36 __author_info = "Written by Antonio Ospite http://ao2.it"
40 import xml.etree.ElementTree as etree
44 from optparse import OptionParser
47 def get_chunk_data(data):
49 moof_size = struct.unpack(">L", data[0:4])[0]
50 mdat_size = struct.unpack(">L", data[moof_size:moof_size + 4])[0]
52 data_start = moof_size + 4 + len('mdat')
53 data_size = mdat_size - 4 - len('mdat')
55 #print len(data[data_start:]), \
56 # len(data[data_start:data_start + data_size]), data_size
58 assert(len(data[data_start:]) == data_size)
60 return data[data_start:data_start + data_size]
63 def hexstring_to_bytes(hex_string):
65 for i in range(0, len(hex_string), 2):
66 res += chr(int(hex_string[i:i + 2], 16))
71 def write_wav_header(out_file, fmt, codec_private_data, data_len):
73 extradata = hexstring_to_bytes(codec_private_data)
75 fmt['cbSize'] = len(extradata)
76 fmt_len = 18 + fmt['cbSize']
77 wave_len = len("WAVEfmt ") + 4 + fmt_len + len('data') + 4
79 out_file.write("RIFF")
80 out_file.write(struct.pack('<L', wave_len))
81 out_file.write("WAVEfmt ")
82 out_file.write(struct.pack('<L', fmt_len))
83 out_file.write(struct.pack('<H', fmt['wFormatTag']))
84 out_file.write(struct.pack('<H', fmt['nChannels']))
85 out_file.write(struct.pack('<L', fmt['nSamplesPerSec']))
86 out_file.write(struct.pack('<L', fmt['nAvgBytesPerSec']))
87 out_file.write(struct.pack('<H', fmt['nBlockAlign']))
88 out_file.write(struct.pack('<H', fmt['wBitsPerSample']))
89 out_file.write(struct.pack('<H', fmt['cbSize']))
90 out_file.write(extradata)
91 out_file.write("data")
92 out_file.write(struct.pack('<L', data_len))
95 def get_manifest(base_url, dest_dir=tempfile.gettempdir(),
96 manifest_file='Manifest'):
97 """Returns the manifest and the new URL if this is changed"""
99 if os.path.exists(dest_dir) == False:
100 os.mkdir(dest_dir, 0755)
102 if base_url.startswith('http://'):
104 manifest_url = base_url
105 if not manifest_url.lower().endswith(('/manifest', '.ismc', '.csm')):
106 manifest_url += '/Manifest'
108 response = urllib2.urlopen(manifest_url)
109 data = response.read()
111 manifest_path = os.path.join(dest_dir, manifest_file)
112 f = open(manifest_path, "w")
116 manifest_path = base_url
118 manifest = etree.parse(manifest_path)
120 version = manifest.getroot().attrib['MajorVersion']
122 raise Exception('Only Smooth Streaming version 2 supported')
125 # if some intermediate client Manifest is used, like in Rai Replay
126 clip = manifest.find("Clip")
127 actual_manifest_url = clip.attrib["Url"]
128 base_url = actual_manifest_url.lower().replace("/manifest", "")
132 return (manifest, base_url)
135 def print_manifest_info(manifest):
137 streams = manifest.findall('.//StreamIndex')
139 for i, s in enumerate(streams):
140 stream_type = s.attrib["Type"]
141 url = s.attrib["Url"]
143 print "Stream: %s Type: %s" % (i, stream_type)
145 print "\tQuality Levels:"
146 qualities = s.findall("QualityLevel")
147 for i, q in enumerate(qualities):
148 bitrate = q.attrib["Bitrate"]
149 fourcc = q.attrib["FourCC"]
151 if stream_type == "video":
152 size = "%sx%s" % (q.attrib["MaxWidth"], q.attrib["MaxHeight"])
153 print "\t%2s: %4s %10s @ %7s bps" % (i, fourcc, size, bitrate)
154 if stream_type == "audio":
155 channels = q.attrib["Channels"]
156 sampling_rate = q.attrib["SamplingRate"]
157 bits_per_sample = q.attrib["BitsPerSample"]
158 print "\t%2s: %4s %sHz %sbits %sch @ %7s bps" % (i, fourcc,
159 sampling_rate, bits_per_sample, channels, bitrate)
164 def get_chunk_quality_string(stream, quality_level):
165 quality = stream.findall("QualityLevel")[quality_level]
166 bitrate = quality.attrib["Bitrate"]
168 quality_attributes = quality.findall("CustomAttributes/Attribute")
169 custom_attributes = ""
170 for i in quality_attributes:
171 custom_attributes += "%s=%s," % (i.attrib["Name"], i.attrib["Value"])
172 custom_attributes = custom_attributes.rstrip(',')
174 # Assume URLs are in this form:
175 # Url="QualityLevels({bitrate})/Fragments(video={start time})"
177 # Url="QualityLevels({bitrate},{CustomAttributes})/Fragments(video={start time})"
178 url = stream.attrib["Url"]
180 chunks_quality = url.split('/')[0].replace("{bitrate}", bitrate)
181 chunks_quality = chunks_quality.replace("{CustomAttributes}", custom_attributes)
183 return chunks_quality
186 def get_chunk_name_string(stream, chunk):
187 t = chunk.attrib["t"]
188 url = stream.attrib["Url"]
189 chunk_name = url.split('/')[1].replace("{start time}", t)
194 def download_chunks(base_url, manifest, stream_index, quality_level, dest_dir):
196 if os.path.exists(dest_dir) == False:
197 os.mkdir(dest_dir, 0755)
199 stream = manifest.findall('.//StreamIndex')[stream_index]
201 chunks_quality = get_chunk_quality_string(stream, quality_level)
203 chunks_dest_dir = os.path.join(dest_dir, chunks_quality)
204 if os.path.exists(chunks_dest_dir) == False:
205 os.mkdir(chunks_dest_dir, 0755)
207 chunks = stream.findall("c")
209 print "\nDownloading Stream %d" % stream_index
210 print "\tChunks %10d/%-10d" % (0, len(chunks)), "\r",
212 for i, c in enumerate(chunks):
214 chunk_name = get_chunk_name_string(stream, c)
215 chunk_file = os.path.join(dest_dir, chunks_quality, chunk_name)
217 if os.path.exists(chunk_file) == False:
218 chunk_url = base_url + '/' + chunks_quality + '/' + chunk_name
219 response = urllib2.urlopen(chunk_url)
220 data = response.read()
222 f = open(chunk_file, "wb")
226 f = open(chunk_file, "rb")
230 data_size += len(data)
231 print "\tChunks %10d/%-10d" % (i + 1, len(chunks)), "\r",
233 print "\tDownloaded size:", data_size
236 def rebuild_stream(manifest, stream_index, quality_level, src_dir,
237 dest_file_name, final_dest_file=None):
239 if final_dest_file == None:
240 final_dest_file = dest_file_name
242 stream = manifest.findall('.//StreamIndex')[stream_index]
244 chunks_quality = get_chunk_quality_string(stream, quality_level)
246 chunks_src_dir = os.path.join(src_dir, chunks_quality)
248 dest_file = open(dest_file_name, "wb")
250 chunks = stream.findall("c")
252 print "\nRebuilding Stream %d" % stream_index
253 print "\tChunks %10d/%-10d" % (0, len(chunks)), "\r",
255 for i, c in enumerate(chunks):
257 chunk_name = get_chunk_name_string(stream, c)
258 chunk_file = os.path.join(chunks_src_dir, chunk_name)
260 f = open(chunk_file, "rb")
261 data = get_chunk_data(f.read())
263 dest_file.write(data)
264 data_size += len(data)
265 print "\tChunks %10d/%-10d" % (i + 1, len(chunks)), "\r",
268 # Add a nice WAV header
269 if stream.attrib['Type'] == "audio":
270 quality = stream.findall("QualityLevel")[quality_level]
271 codec_private_data = quality.attrib['CodecPrivateData']
274 fmt['wFormatTag'] = int(quality.attrib['AudioTag'])
275 fmt['nChannels'] = int(quality.attrib['Channels'])
276 fmt['nSamplesPerSec'] = int(quality.attrib['SamplingRate'])
277 fmt['nAvgBytesPerSec'] = int(quality.attrib['Bitrate']) / 8
278 fmt['wBitsPerSample'] = int(quality.attrib['BitsPerSample'])
279 fmt['nBlockAlign'] = int(quality.attrib['PacketSize'])
282 f = open(final_dest_file, "wb")
283 write_wav_header(f, fmt, codec_private_data, data_size)
285 dest_file = open(dest_file_name, "rb")
286 f.write(dest_file.read())
291 print "Stream %d, actual data size: %d\n" % (stream_index, data_size)
294 def calc_tracks_delay(manifest, stream1_index, stream2_index):
295 streams = manifest.findall('.//StreamIndex')
297 s1 = streams[stream1_index]
298 s2 = streams[stream2_index]
300 s1_start_chunk = s1.find("c")
301 s2_start_chunk = s2.find("c")
303 s1_start_time = int(s1_start_chunk.attrib['t'])
304 s2_start_time = int(s2_start_chunk.attrib['t'])
306 s1_timescale = float(s1.attrib['TimeScale'])
307 s2_timescale = float(s2.attrib['TimeScale'])
309 # calc difference in seconds
310 delay = s2_start_time / s2_timescale - \
311 s1_start_time / s1_timescale
316 def get_clip_duration(manifest):
317 # TODO: use <Clip ClipBegin="" ClipEnd=""> if Duration is not available
318 duration = manifest.getroot().attrib['Duration']
320 return float(duration) / 10000000 # here is the default timescale
323 def smooth_download(url, manifest, dest_dir=tempfile.gettempdir(),
324 video_stream_index=0, audio_stream_index=1,
325 video_quality_level=0, audio_quality_level=0,
326 chunks_dir=None, download=True,
327 out_video_file='_video.vc1', out_audio_file='_audio.raw'):
329 if chunks_dir == None:
330 chunks_dir = dest_dir
333 download_chunks(url, manifest, video_stream_index,
334 video_quality_level, chunks_dir)
335 download_chunks(url, manifest, audio_stream_index,
336 audio_quality_level, chunks_dir)
338 dest_video = os.path.join(dest_dir, out_video_file)
339 dest_audio = os.path.join(dest_dir, out_audio_file)
341 rebuild_stream(manifest, video_stream_index, video_quality_level,
342 chunks_dir, dest_video)
343 rebuild_stream(manifest, audio_stream_index, audio_quality_level,
344 chunks_dir, dest_audio, dest_audio + '.wav')
346 #duration = get_clip_duration(manifest)
348 delay = calc_tracks_delay(manifest, video_stream_index,
351 # optionally encode audio to vorbis:
352 # ffmpeg -i _audio.raw.wav -acodec libvorbis -aq 60 audio.ogg
353 mux_command = ("ffmpeg -i %s \\\n" +
354 " -itsoffset %f -async 1 -i %s \\\n" +
355 " -vcodec copy -acodec copy ffout.mkv") % \
356 (dest_video, delay, dest_audio + '.wav')
361 def options_parser():
362 version = "%%prog %s" % __version
363 usage = "usage: %prog [options] <manifest URL or file>"
364 parser = OptionParser(usage=usage, version=version,
365 description=__description, epilog=__author_info)
366 parser.add_option("-i", "--info",
367 action="store_true", dest="info_only",
368 default=False, help="print Manifest info and exit")
369 parser.add_option("-m", "--manifest-only",
370 action="store_true", dest="manifest_only",
371 default=False, help="download Manifest file and exit")
372 parser.add_option("-n", "--no-download",
373 action="store_false", dest="download",
374 default=True, help="disable downloading chunks")
375 parser.add_option("-s", "--sync-delay",
376 action="store_true", dest="sync_delay",
377 default=False, help="show the sync delay between the given streams and exit")
378 parser.add_option("-d", "--dest-dir", metavar="<dir>",
379 dest="dest_dir", default=tempfile.gettempdir(),
380 help="destination directory")
381 parser.add_option("-c", "--chunks-dir", metavar="<dir>",
382 dest="chunks_dir", default=None,
383 help="directory containing chunks, if different from destination dir")
384 parser.add_option("-v", "--video-stream", metavar="<n>",
385 type="int", dest="video_stream_index", default=0,
386 help="index of the video stream")
387 parser.add_option("-a", "--audio-stream", metavar="<n>",
388 type="int", dest="audio_stream_index", default=1,
389 help="index of the audio stream")
390 parser.add_option("-q", "--video-quality", metavar="<n>",
391 type="int", dest="video_quality_level", default=0,
392 help="index of the video quality level")
393 parser.add_option("-Q", "--audio-quality", metavar="<n>",
394 type="int", dest="audio_quality_level", default=0,
395 help="index of the audio quality level")
400 if __name__ == "__main__":
402 parser = options_parser()
403 (options, args) = parser.parse_args()
410 manifest, url = get_manifest(url, options.dest_dir)
412 if options.manifest_only:
415 if options.sync_delay:
416 print calc_tracks_delay(manifest,
417 options.video_stream_index,
418 options.audio_stream_index)
421 if options.info_only:
422 print_manifest_info(manifest)
425 print_manifest_info(manifest)
427 smooth_download(url, manifest, options.dest_dir,
428 options.video_stream_index, options.audio_stream_index,
429 options.video_quality_level, options.audio_quality_level,
430 options.chunks_dir, options.download)