Apple TV: Audio tracks are merged

temp1

Downloads from Apple TV with multiple audio tracks will play with the audio tracks merged into one (in different players, e.g. QuickTime Player, although the audio tracks can be selected in VLC). Does anyone else experience this?

0xFeedBeef

What do you mean by “merged”? It’s physically impossible… what does mediainfo say about the tracks?

temp1

Both audio tracks are played simultaneously (at least in QuickTime Player and Jellyfin), and they cannot be selected – the menu / context menu do not list the audio tracks. The tracks can be selected and played separately in VLC. The problem does not occur with downloads from Netflix.

0xFeedBeef

temp1 possible that the audio tracks are each in their own alternative groups instead of all being in one alternative group; without mediainfo output, or any other MP4 parsing tool to tell us what you’ve got, I’m just guessing…

temp1

0xFeedBeef mediainfo’s output looks good to me, but Track 2 seems to be missing:

Duration:` 2048.756 seconds (2048756/1000)
Track count: 4
Track 1: Video 'vide'
	Enabled: Yes
	Format Description 1:
		Format: Dolby Vision HEVC 'dvh1'
		Dimensions: 1920 x 1080
		Encoded Pixels: 1920 x 1080
		Presentation Dimensions: 1920 x 1080
		Warning: Color Primaries were not specified in the format description
		Warning: Transfer function was not specified in the format description
		Warning: YCbCr matrix was not specified in the format description
	System support for decoding this track: Yes
	Data size: 2132265524 bytes
	Media time scale: 24000
	Duration: 2048.756 seconds
	Estimated data rate: 8326.093 kbit/s
	Nominal frame rate: 23.976 fps
	Language code: eng
	Minimum sample duration: 1001/24000 seconds
	Frame reordering required
	1 segment present
	Index   Media Start  Media Duration   Track Start  Track Duration 
	    1  00:00:00.042    00:34:08.756  00:00:00.000    00:34:08.756
Track 3: Sound 'soun'
	Enabled: Yes
	Format Description 1:
		Format: MPEG-4 HE AAC V2 'aacp'
		Channel Layout: Stereo (L R)
		Sample rate: 44100.0
		Bytes per packet: 0
		Frames per packet: 2048
		Bytes per frame: 0
		Channels per frame: 2
		Bits per channel: 0
	System support for decoding this track: Yes
	Data size: 8193128 bytes
	Media time scale: 44100
	Duration: 2048.803 seconds
	Estimated data rate: 31.992 kbit/s
	Language code: deu
	1 segment present
	Index   Media Start  Media Duration   Track Start  Track Duration 
	    1  00:00:00.000    00:34:08.803  00:00:00.000    00:34:08.803
	Member of alternate group 0: (3)
Track 4: Sound 'soun'
	Enabled: Yes
	Format Description 1:
		Format: MPEG-4 HE AAC V2 'aacp'
		Channel Layout: Stereo (L R)
		Sample rate: 44100.0
		Bytes per packet: 0
		Frames per packet: 2048
		Bytes per frame: 0
		Channels per frame: 2
		Bits per channel: 0
	System support for decoding this track: Yes
	Data size: 8188744 bytes
	Media time scale: 44100
	Duration: 2048.803 seconds
	Estimated data rate: 31.975 kbit/s
	Language code: eng
	1 segment present
	Index   Media Start  Media Duration   Track Start  Track Duration 
	    1  00:00:00.000    00:34:08.803  00:00:00.000    00:34:08.803
	Member of alternate group 1: (4)
Track 5: Subtitle 'sbtl'
	Enabled: Yes
	Format Description 1:
		Format: Subtitle 'tx3g'
		Dimensions: 0 x 0
		DisplayFlags: N/A 
		Justification: Horizontal 1, Vertical -1 
		Default Text Box: N/A
		Default Style: Local Font ID: 1 Size: 16 Color (RGBA): 1 1 1 1
		Font Name: Arial
	System support for decoding this track: Yes
	Data size: 6637 bytes
	Media time scale: 1000000
	Duration: 1918.043 seconds
	Estimated data rate: 0.028 kbit/s
	Nominal frame rate: 0.247 fps
	Language code: eng
	1 segment present
	Index   Media Start  Media Duration   Track Start  Track Duration 
	    1  00:00:00.000    00:31:58.043  00:00:00.000    00:31:58.043
	Member of alternate group 2: (5)

0xFeedBeef

temp1 the problem is that track 3 is in alt group 0, whereas track 4 is in alt group 1, meaning they are independent of each other, so an intelligent player might reasonably decide that they are to be played concurrently

temp1

temp1 This is the output for a Netflix file (which does not exhibit the audio track issue):

Duration: 7749.440 seconds (7749440/1000)
Track count: 4
Track 1: Video 'vide'
	Enabled: Yes
	Format Description 1:
		Format: H.264 'avc1'
		Dimensions: 1920 x 1080
		Encoded Pixels: 1920 x 1080
		Presentation Dimensions: 1920 x 1080
		Warning: Color Primaries were not specified in the format description
		Warning: Transfer function was not specified in the format description
		Warning: YCbCr matrix was not specified in the format description
	System support for decoding this track: Yes
	Data size: 1640869931 bytes
	Media time scale: 24000
	Duration: 7749.209 seconds
	Estimated data rate: 1693.974 kbit/s
	Nominal frame rate: 23.998 fps
	Minimum sample duration: 1000/24000 seconds
	Frame reordering required
	1 segment present
	Index   Media Start  Media Duration   Track Start  Track Duration 
	    1  00:00:00.083    02:09:09.209  00:00:00.000    02:09:09.209
Track 3: Sound 'soun'
	Enabled: No
	Format Description 1:
		Format: Dolby Digital Plus 'ec-3'
		Channel Layout: 5.1 (L C R Ls Rs LFE)
		Sample rate: 48000.0
		Bytes per packet: 0
		Frames per packet: 1536
		Bytes per frame: 0
		Channels per frame: 6
		Bits per channel: 0
	System support for decoding this track: Yes
	Data size: 619955200 bytes
	Media time scale: 48000
	Duration: 7749.440 seconds
	Estimated data rate: 640.000 kbit/s
	Language code: deu
	1 segment present
	Index   Media Start  Media Duration   Track Start  Track Duration 
	    1  00:00:00.000    02:09:09.440  00:00:00.000    02:09:09.440
	Member of alternate group 0: (3, 4)
Track 4: Sound 'soun'
	Enabled: Yes
	Format Description 1:
		Format: Dolby Digital Plus 'ec-3'
		Channel Layout: 5.1 (L C R Ls Rs LFE)
		Sample rate: 48000.0
		Bytes per packet: 0
		Frames per packet: 1536
		Bytes per frame: 0
		Channels per frame: 6
		Bits per channel: 0
	System support for decoding this track: Yes
	Data size: 619955200 bytes
	Media time scale: 48000
	Duration: 7749.440 seconds
	Estimated data rate: 640.000 kbit/s
	Language code: eng
	1 segment present
	Index   Media Start  Media Duration   Track Start  Track Duration 
	    1  00:00:00.000    02:09:09.440  00:00:00.000    02:09:09.440
	Member of alternate group 0: (3, 4)
Track 5: Subtitle 'sbtl'
	Enabled: Yes
	Format Description 1:
		Format: Subtitle 'tx3g'
		Dimensions: 0 x 0
		DisplayFlags: N/A 
		Justification: Horizontal 1, Vertical -1 
		Default Text Box: N/A
		Default Style: Local Font ID: 1 Size: 16 Color (RGBA): 1 1 1 1
		Font Name: Arial
	System support for decoding this track: Yes
	Data size: 68194 bytes
	Media time scale: 1000000
	Duration: 7593.958 seconds
	Estimated data rate: 0.072 kbit/s
	Nominal frame rate: 0.478 fps
	Language code: eng
	1 segment present
	Index   Media Start  Media Duration   Track Start  Track Duration 
	    1  00:00:00.000    02:06:33.958  00:00:00.000    02:06:33.958
	Member of alternate group 1: (5)

I see the difference now – the audio tracks are not members of the same alternate group for the Apple TV file (as opposed to the Netflix file). Is this something that StreamFab could/should handle differently when processing Apple TV files?

temp1

0xFeedBeef Thanks, I verified this with a Netflix file (added in a separate post). Is this something that the streaming service decides, or could StreamFab process the content differently?

0xFeedBeef

temp1 as you can see from the output, the netflix download is mixed so tracks 3 and 4 are members of the same alt group 0, meaning they are mutually exclusive.

0xFeedBeef

temp1 I’m guessing it’s the way SF multiplexes the final file once the tracks are decrypted. I’ll add it to the list of open problem reports, as I have confirmed the same thing is happening on my end, and maybe @Wilson.Wang could get the dev to look into that at some point.

0xFeedBeef

temp1 can you try the beta from this post and see if you get different results: https://forum.dvdfab.cn/d/472011-7022-not-picking-up-dd51-at-all/23

Wilson.Wang

Apple TV
Track 3: Sound ‘soun’
Enabled: Yes
Track 4: Sound ‘soun’
Enabled: Yes

Netflix
Track 3: Sound ‘soun’
Enabled: No
Track 4: Sound ‘soun’
Enabled: Yes

The dev is checking for the problem, the issue is caused by that the both audio tracks are default play back audio in Apple TV.

Wilson

Wilson.Wang

Test version in the post: https://forum.dvdfab.cn/d/472120-analysis-hard-coded-limit-of-100-episodes-/3