[extractor] Detect `sttp` as subtitles in MPD

Closes #656
Solution by: fstirlitz

yt_dlp/extractor/common.py

@@ -2596,7 +2596,16 @@ class InfoExtractor(object):
                     mime_type = representation_attrib['mimeType']
                     content_type = representation_attrib.get('contentType', mime_type.split('/')[0])
 
-                    if content_type in ('video', 'audio', 'text') or mime_type == 'image/jpeg':
+                    codecs = representation_attrib.get('codecs', '')
+                    if content_type not in ('video', 'audio', 'text'):
+                        if mime_type == 'image/jpeg':
+                            content_type = 'image/jpeg'
+                        if codecs.split('.')[0] == 'stpp':
+                            content_type = 'text'
+                        else:
+                            self.report_warning('Unknown MIME type %s in DASH manifest' % mime_type)
+                            continue
+
                     base_url = ''
                     for element in (representation, adaptation_set, period, mpd_doc):
                         base_url_e = element.find(_add_ns('BaseURL'))
@@ -2634,14 +2643,14 @@ class InfoExtractor(object):
                             'filesize': filesize,
                             'container': mimetype2ext(mime_type) + '_dash',
                         }
-                            f.update(parse_codecs(representation_attrib.get('codecs')))
+                        f.update(parse_codecs(codecs))
                     elif content_type == 'text':
                         f = {
                             'ext': mimetype2ext(mime_type),
                             'manifest_url': mpd_url,
                             'filesize': filesize,
                         }
-                        elif mime_type == 'image/jpeg':
+                    elif content_type == 'image/jpeg':
                         # See test case in VikiIE
                         # https://www.viki.com/videos/1175236v-choosing-spouse-by-lottery-episode-1
                         f = {
@@ -2803,8 +2812,7 @@ class InfoExtractor(object):
                         formats.append(f)
                     elif content_type == 'text':
                         subtitles.setdefault(lang or 'und', []).append(f)
-                    else:
-                        self.report_warning('Unknown MIME type %s in DASH manifest' % mime_type)
+
         return formats, subtitles
 
     def _extract_ism_formats(self, *args, **kwargs):