[extractor/common, downloader/ism] Extract SSTR subtitle tracks

_parse_ism_formats was extended into _parse_ism_formats_and_subtitles;
all direct users were updated, though _extract_ism_formats was left
as a compatibility wrapper.

The SSTR downloader was also modified in order to prepare for muxing
subtitle streams, although no support for any subtitle codecs was
added in this commit.

yt_dlp/downloader/ism.py

@@ -48,7 +48,7 @@ def write_piff_header(stream, params):
     language = params.get('language', 'und')
     height = params.get('height', 0)
     width = params.get('width', 0)
-    is_audio = width == 0 and height == 0
+    stream_type = params['stream_type']
     creation_time = modification_time = int(time.time())
 
     ftyp_payload = b'isml'  # major brand
@@ -77,7 +77,7 @@ def write_piff_header(stream, params):
     tkhd_payload += u32.pack(0) * 2  # reserved
     tkhd_payload += s16.pack(0)  # layer
     tkhd_payload += s16.pack(0)  # alternate group
-    tkhd_payload += s88.pack(1 if is_audio else 0)  # volume
+    tkhd_payload += s88.pack(1 if stream_type == 'audio' else 0)  # volume
     tkhd_payload += u16.pack(0)  # reserved
     tkhd_payload += unity_matrix
     tkhd_payload += u1616.pack(width)
@@ -93,19 +93,35 @@ def write_piff_header(stream, params):
     mdia_payload = full_box(b'mdhd', 1, 0, mdhd_payload)  # Media Header Box
 
     hdlr_payload = u32.pack(0)  # pre defined
-    hdlr_payload += b'soun' if is_audio else b'vide'  # handler type
-    hdlr_payload += u32.pack(0) * 3  # reserved
-    hdlr_payload += (b'Sound' if is_audio else b'Video') + b'Handler\0'  # name
+    if stream_type == 'audio':  # handler type
+        hdlr_payload += b'soun'
+        hdlr_payload += u32.pack(0) * 3  # reserved
+        hdlr_payload += b'SoundHandler\0'  # name
+    elif stream_type == 'video':
+        hdlr_payload += b'vide'
+        hdlr_payload += u32.pack(0) * 3  # reserved
+        hdlr_payload += b'VideoHandler\0'  # name
+    elif stream_type == 'text':
+        hdlr_payload += b'subt'
+        hdlr_payload += u32.pack(0) * 3  # reserved
+        hdlr_payload += b'SubtitleHandler\0'  # name
+    else:
+        assert False
     mdia_payload += full_box(b'hdlr', 0, 0, hdlr_payload)  # Handler Reference Box
 
-    if is_audio:
+    if stream_type == 'audio':
         smhd_payload = s88.pack(0)  # balance
         smhd_payload += u16.pack(0)  # reserved
         media_header_box = full_box(b'smhd', 0, 0, smhd_payload)  # Sound Media Header
-    else:
+    elif stream_type == 'video':
         vmhd_payload = u16.pack(0)  # graphics mode
         vmhd_payload += u16.pack(0) * 3  # opcolor
         media_header_box = full_box(b'vmhd', 0, 1, vmhd_payload)  # Video Media Header
+    elif stream_type == 'text':
+        sthd_payload = u16.pack(0) * 2
+        media_header_box = full_box(b'sthd', 0, 1, sthd_payload)  # Subtitle Media Header
+    else:
+        assert False
     minf_payload = media_header_box
 
     dref_payload = u32.pack(1)  # entry count
@@ -117,7 +133,7 @@ def write_piff_header(stream, params):
 
     sample_entry_payload = u8.pack(0) * 6  # reserved
     sample_entry_payload += u16.pack(1)  # data reference index
-    if is_audio:
+    if stream_type == 'audio':
         sample_entry_payload += u32.pack(0) * 2  # reserved
         sample_entry_payload += u16.pack(params.get('channels', 2))
         sample_entry_payload += u16.pack(params.get('bits_per_sample', 16))
@@ -127,7 +143,7 @@ def write_piff_header(stream, params):
 
         if fourcc == 'AACL':
             sample_entry_box = box(b'mp4a', sample_entry_payload)
-    else:
+    elif stream_type == 'video':
         sample_entry_payload += u16.pack(0)  # pre defined
         sample_entry_payload += u16.pack(0)  # reserved
         sample_entry_payload += u32.pack(0) * 3  # pre defined
@@ -155,6 +171,10 @@ def write_piff_header(stream, params):
             avcc_payload += pps
             sample_entry_payload += box(b'avcC', avcc_payload)  # AVC Decoder Configuration Record
             sample_entry_box = box(b'avc1', sample_entry_payload)  # AVC Simple Entry
+        else:
+            assert False
+    else:
+        assert False
     stsd_payload += sample_entry_box
 
     stbl_payload = full_box(b'stsd', 0, 0, stsd_payload)  # Sample Description Box