From a95757d3b7ca2741aeca4d5aea8e1792f361d5b4 Mon Sep 17 00:00:00 2001 From: Peter Rowlands Date: Fri, 4 Oct 2024 18:04:52 +0900 Subject: [PATCH] [extractor] parse CENC + Clear Key information in DASH manifests --- test/test_InfoExtractor.py | 104 ++++++++++++++++++++++++++++ test/testdata/mpd/clearkey_cenc.mpd | 29 ++++++++ test/testdata/mpd/w3c_pssh.mpd | 13 ++++ yt_dlp/extractor/common.py | 98 ++++++++++++++++++++++++-- 4 files changed, 239 insertions(+), 5 deletions(-) create mode 100644 test/testdata/mpd/clearkey_cenc.mpd create mode 100644 test/testdata/mpd/w3c_pssh.mpd diff --git a/test/test_InfoExtractor.py b/test/test_InfoExtractor.py index 31e8f8244..3fee480f3 100644 --- a/test/test_InfoExtractor.py +++ b/test/test_InfoExtractor.py @@ -1369,6 +1369,110 @@ def test_parse_mpd_formats(self): }, ], }, + ), ( + # Clear Key with CENC default_KID + 'clearkey_cenc', + 'https://media.axprod.net/TestVectors/v7-MultiDRM-SingleKey/Manifest_1080p_ClearKey.mpd', # mpd_url + 'https://media.axprod.net/TestVectors/v7-MultiDRM-SingleKey/', # mpd_base_url + [{ + 'manifest_url': 'https://media.axprod.net/TestVectors/v7-MultiDRM-SingleKey/Manifest_1080p_ClearKey.mpd', + 'ext': 'mp4', + 'format_id': '1', + 'format_note': 'DASH video', + 'protocol': 'http_dash_segments', + 'acodec': 'none', + 'vcodec': 'avc1.64001f', + 'tbr': 389.802, + 'width': 512, + 'height': 288, + 'dash_cenc': { + 'laurl': 'https://drm-clearkey-testvectors.axtest.net/AcquireLicense', + 'key_ids': ['9eb4050de44b4802932e27d75083e266'], + }, + }, { + 'manifest_url': 'https://media.axprod.net/TestVectors/v7-MultiDRM-SingleKey/Manifest_1080p_ClearKey.mpd', + 'ext': 'mp4', + 'format_id': '2', + 'format_note': 'DASH video', + 'protocol': 'http_dash_segments', + 'acodec': 'none', + 'vcodec': 'avc1.64001f', + 'tbr': 764.935, + 'width': 640, + 'height': 360, + 'dash_cenc': { + 'laurl': 'https://drm-clearkey-testvectors.axtest.net/AcquireLicense', + 'key_ids': ['9eb4050de44b4802932e27d75083e266'], + }, + }, { + 'manifest_url': 'https://media.axprod.net/TestVectors/v7-MultiDRM-SingleKey/Manifest_1080p_ClearKey.mpd', + 'ext': 'mp4', + 'format_id': '3', + 'format_note': 'DASH video', + 'protocol': 'http_dash_segments', + 'acodec': 'none', + 'vcodec': 'avc1.640028', + 'tbr': 1120.439, + 'width': 852, + 'height': 480, + 'dash_cenc': { + 'laurl': 'https://drm-clearkey-testvectors.axtest.net/AcquireLicense', + 'key_ids': ['9eb4050de44b4802932e27d75083e266'], + }, + }, { + 'manifest_url': 'https://media.axprod.net/TestVectors/v7-MultiDRM-SingleKey/Manifest_1080p_ClearKey.mpd', + 'ext': 'mp4', + 'format_id': '4', + 'format_note': 'DASH video', + 'protocol': 'http_dash_segments', + 'acodec': 'none', + 'vcodec': 'avc1.640032', + 'tbr': 1945.258, + 'width': 1280, + 'height': 720, + 'dash_cenc': { + 'laurl': 'https://drm-clearkey-testvectors.axtest.net/AcquireLicense', + 'key_ids': ['9eb4050de44b4802932e27d75083e266'], + }, + }, { + 'manifest_url': 'https://media.axprod.net/TestVectors/v7-MultiDRM-SingleKey/Manifest_1080p_ClearKey.mpd', + 'ext': 'mp4', + 'format_id': '5', + 'format_note': 'DASH video', + 'protocol': 'http_dash_segments', + 'acodec': 'none', + 'vcodec': 'avc1.640033', + 'tbr': 2726.377, + 'width': 1920, + 'height': 1080, + 'dash_cenc': { + 'laurl': 'https://drm-clearkey-testvectors.axtest.net/AcquireLicense', + 'key_ids': ['9eb4050de44b4802932e27d75083e266'], + }, + }], + {}, + ), ( + # default CENC KID overridden via W3C PSSH box, no license server in manifest + 'w3c_pssh', + 'https://unknown/manifest.mpd', # mpd_url + 'https://unknown/', # mpd_base_url + [{ + 'manifest_url': 'https://unknown/manifest.mpd', + 'ext': 'mp4', + 'format_id': '1', + 'format_note': 'DASH video', + 'protocol': 'http_dash_segments', + 'acodec': 'none', + 'vcodec': 'avc1.64001f', + 'tbr': 389.802, + 'width': 512, + 'height': 288, + 'dash_cenc': { + 'key_ids': ['43215678123412341234123412341234'], + }, + 'has_drm': True, + }], + {}, ), ] diff --git a/test/testdata/mpd/clearkey_cenc.mpd b/test/testdata/mpd/clearkey_cenc.mpd new file mode 100644 index 000000000..40f212383 --- /dev/null +++ b/test/testdata/mpd/clearkey_cenc.mpd @@ -0,0 +1,29 @@ + + + + + + + + https://drm-clearkey-testvectors.axtest.net/AcquireLicense + + + + + + + + + + + diff --git a/test/testdata/mpd/w3c_pssh.mpd b/test/testdata/mpd/w3c_pssh.mpd new file mode 100644 index 000000000..d72cd866c --- /dev/null +++ b/test/testdata/mpd/w3c_pssh.mpd @@ -0,0 +1,13 @@ + + + + + + AAAANHBzc2gBAAAAEHfv7MCyTQKs4zweUuL7SwAAAAFDIVZ4EjQSNBI0EjQSNBI0AAAAAA== + + + + + + + diff --git a/yt_dlp/extractor/common.py b/yt_dlp/extractor/common.py index 3430036f4..1069686a9 100644 --- a/yt_dlp/extractor/common.py +++ b/yt_dlp/extractor/common.py @@ -14,12 +14,14 @@ import os import random import re +import struct import subprocess import sys import time import types import urllib.parse import urllib.request +import uuid import xml.etree.ElementTree from ..compat import ( @@ -258,6 +260,15 @@ class InfoExtractor: * ffmpeg_args_out Extra arguments for ffmpeg downloader (output) * is_dash_periods Whether the format is a result of merging multiple DASH periods. + * dash_cenc A dictionary of DASH CENC decryption information + used by the native DASH downloader when set. + * laurl The Clear Key license server URL from which + CENC keys will be downloaded. + * key_ids List of key IDs (as hex) to request from the ClearKey + license server. + * key The CENC key (as hex) used to decrypt fragments. + If `key` is given, any license server URL and + key IDs will be ignored. RTMP formats can also have the additional fields: page_url, app, play_path, tc_url, flash_version, rtmp_live, rtmp_conn, rtmp_protocol, rtmp_real_time @@ -2669,7 +2680,10 @@ def _merge_mpd_periods(self, periods): assert 'is_dash_periods' not in f, 'format already processed' f['is_dash_periods'] = True format_key = tuple(v for k, v in f.items() if k not in ( - ('format_id', 'fragments', 'manifest_stream_number'))) + ('format_id', 'fragments', 'manifest_stream_number', 'dash_cenc'))) + if 'dash_cenc' in f: + format_key = format_key + tuple( + tuple(v) if isinstance(v, list) else v for v in f['dash_cenc'].values()) if format_key not in formats: formats[format_key] = f elif 'fragments' in f: @@ -2703,8 +2717,18 @@ def _parse_mpd_periods(self, mpd_doc, mpd_id=None, mpd_base_url='', mpd_url=None def _add_ns(path): return self._xpath_ns(path, namespace) - def is_drm_protected(element): - return element.find(_add_ns('ContentProtection')) is not None + def extract_drm_info(element): + has_drm = False + cenc_info = {} + for cp_e in element.findall(_add_ns('ContentProtection')): + has_drm = True + self._extract_mpd_content_protection_info(cp_e, cenc_info) + info = {'dash_cenc': cenc_info} if cenc_info else {} + if has_drm and not ( + cenc_info.get('key') or cenc_info.get('laurl') and cenc_info.get('key_ids') + ): + info['has_drm'] = True + return info def extract_multisegment_info(element, ms_parent_info): ms_info = ms_parent_info.copy() @@ -2778,6 +2802,7 @@ def extract_Initialization(source): 'timescale': 1, }) for adaptation_set in period.findall(_add_ns('AdaptationSet')): + adaptation_set_drm_info = extract_drm_info(adaptation_set) adaption_set_ms_info = extract_multisegment_info(adaptation_set, period_ms_info) for representation in adaptation_set.findall(_add_ns('Representation')): representation_attrib = adaptation_set.attrib.copy() @@ -2864,8 +2889,8 @@ def extract_Initialization(source): 'acodec': 'none', 'vcodec': 'none', } - if is_drm_protected(adaptation_set) or is_drm_protected(representation): - f['has_drm'] = True + f.update(adaptation_set_drm_info) + f.update(extract_drm_info(representation)) representation_ms_info = extract_multisegment_info(representation, adaption_set_ms_info) def prepare_template(template_name, identifiers): @@ -3026,6 +3051,69 @@ def add_segment_url(): period_entry['subtitles'][lang or 'und'].append(f) yield period_entry + def _extract_mpd_content_protection_info(self, cp_e, cenc_info): + """ + Extract supported DASH-CENC parameters for an MPD ContentProtection element. + + Called multiple times per extracted format in an MPD (once per ContentProtection element + within AdaptationSet and Representation elements). Subclasses may override this method + when necessary (such as when the Clear Key license server URL is provided separately + from the manifest or when an extractor needs to process the optional data section in W3C + PSSH boxes). + + Note that the `has_drm` flag will be set for any format that does not meet one or more + of these conditions: + + * Both `laurl` and `key_ids` are set (indicating the native DASH downloader should + use the specified Clear Key server URL to retreive the CENC key for this format. + * `key_id` is set (indicating the native DASH downloader should use the specified + CENC key for this format). + + References: + 1. DASH-IF Content Protection Identifiers + https://dashif.org/identifiers/content_protection/ + 2. DASH-IF Content Protection Guidelines + https://dashif.org/docs/IOP-Guidelines/DASH-IF-IOP-Part6-v5.0.0.pdf + 3. W3C "cenc" Initialization Data Format + https://w3c.github.io/encrypted-media/format-registry/initdata/cenc.html + """ + scheme_id = cp_e.get('schemeIdUri') + if scheme_id == 'urn:mpeg:dash:mp4protection:2011': + if cp_e.get('value') == 'cenc': + # ISO/IEC 23009-1 MPEG Common Encryption (CENC) + if not cenc_info.get('key_ids'): + try: + default_kid = uuid.UUID(cp_e.get('{urn:mpeg:cenc:2013}default_KID')).hex + cenc_info['key_ids'] = [default_kid] + except (ValueError, TypeError): + pass + elif scheme_id == 'urn:uuid:e2719d58-a985-b3c9-781a-b030af78d30e': + # Clear Key DASH-IF + for tag, ns in itertools.product( + ('Laurl', 'laurl'), + ('https://dashif.org/CPS', 'http://dashif.org/guidelines/clearKey'), + ): + url_e = cp_e.find(self._xpath_ns(tag, ns)) + if url_e is not None: + cenc_info['laurl'] = url_e.text + break + elif scheme_id == 'urn:uuid:1077efec-c0b2-4d02-ace3-3c1e52e2fb4b': + # W3C Common System ID + pssh_e = cp_e.find(self._xpath_ns('pssh', 'urn:mpeg:cenc:2013')) + if pssh_e is not None: + # W3C PSSH box (may contain Clear Key KIDs but can also be used + # to store KIDs for other DRM systems) + try: + pssh_box = base64.b64decode(pssh_e.text) + kid_count, = struct.unpack('!L', pssh_box[28:32]) + kids = [] + for i in range(kid_count): + kid = pssh_box[32 + i * 16:32 + (i + 1) * 16] + kids.append(kid.hex()) + cenc_info['key_ids'] = kids + except (ValueError, TypeError, struct.error): + pass + def _extract_ism_formats(self, *args, **kwargs): fmts, subs = self._extract_ism_formats_and_subtitles(*args, **kwargs) if subs: