Narrated audio track (nar) aka 'Audio Described'

Help/How To
1

Hello. Apologies if I’ve missed this in the FAQs or elsewhere on the forum but I’m editing and converting some .ts files to .mp4. I’ve got the editing side of it sorted but sometimes the .ts contains both the ‘primary’ audio track and a second, narrated, track. And sometimes Shotcut seems to choose the primary track and sometimes the narrated track - irrespective of which is Track 1 and which Track 2. (Though it is always the same choice for any particular video.) So sometimes I end up with the narration when I don’t want it.

I can deal with this easily enough. Before starting editing, with MediaInfo I can see if the .ts has more than one audio track. If so then, usually, one will have the language ‘English’ and the other ‘nar’. Then in tsMuxer I can remove the nar track.

But I’d like to know how Shotcut decides which audio track to use. I’d guess it would be something to do with a choice of language (‘nar’ taking the role of a ‘language’). But I cannot see where that might be set (other than under Settings/Language - where nar does not appear).

Is there some logic to this that I haven’t perceived?

Bob

1 Like
2

Here is an animated GIF to see where you can select the audio tracks of a clip.

1 Like
3

Now the opposite:

For future reference (after I get my H8+, LOL), can one Export from Shotcut with two audio versions on one mp4, Stereo and 5.1? If so, how?

4

Sorry @kagsundaram, I will look at this during the day today. :sweat_smile:

1 Like
5

Thanks ejmillan. I’ve marked your reply as ‘solution’ because I think it is near enough for me.

Usually the ‘nar’ track is indeed the original audio plus, for those hard of hearing, a dubbed overlay reading screen captions or describing what is going on when there is no dialogue. E.g. “Early morning; man walks by the edge of the sea”. And when the recording is released, it will have two audio tracks, the original and the dubbed nar track. Typically they are numbered in that order and, frequently, the nar track will be of lower quality than the original. Here, for instance, courtesy of MediaInfo, are the characteristics (is this what you meant?) of the two tracks on an example that was puzzling me:

Audio #1
ID : 2354 (0x932)
Menu ID : 9220 (0x2404)
Format : MPEG Audio
Format version : Version 1
Format profile : Layer 2
Codec ID : 4
Duration : 2 h 26 min
Bit rate mode : Constant
Bit rate : 192 kb/s
Maximum bit rate : 31.0 Mb/s
Channel(s) : 2 channels
Sampling rate : 48.0 kHz
Frame rate : 41.667 FPS (1152 SPF)
Compression mode : Lossy
Delay relative to video : -160 ms
Stream size : 202 MiB (6%)
Language : English

Audio #2
ID : 2355 (0x933)
Menu ID : 9220 (0x2404)
Format : MPEG Audio
Format version : Version 1
Format profile : Layer 2
Format settings : Joint stereo
Codec ID : 4
Duration : 2 h 26 min
Bit rate mode : Constant
Bit rate : 128 kb/s
Maximum bit rate : 31.0 Mb/s
Channel(s) : 2 channels
Sampling rate : 48.0 kHz
Frame rate : 41.667 FPS (1152 SPF)
Compression mode : Lossy
Delay relative to video : -160 ms
Stream size : 134 MiB (4%)
Language : nar

I think media players will usually select the first audio track unless directed to do otherwise. And I think Shotcut does so too - though that was kind of my original question.

But in one or two cases it seems to have muxed in the nar track rather than the original. I mustn’t rule out the possibility that I messed up Shotcut’s default setting and by mistake actually forced the choice of the nar track.

Your little gif has been a great help in helping me to investigate what has been going on. Though in some way it raises yet another puzzle! I’ll explain …

First I reopened the saved Shotcut project (.mlt) file with the original .ts source. Guided by your gif I then looked at which audio track(s) had been selected. And it was indeed Track 1. BUT, despite the nar track being Track 2, as above, according to both MediaInfo and tsMuxer, it is very clear both from the waveforms and by listening that Shotcut thinks the nar track is Track 1.

So, with your help, I now understand WHAT happened with this particular video. And I think I have now answered my own question: I think Shotcut defaults to Track 1 but sometimes seems to confuse Track 1 and Track 2. At least I now know how to check these things.

1 Like
6

Wow. It’s great to have detailed feedback so there is more information on this particular for everyone.
I didn’t know what the nar track was, thanks for the explanation.

I linked my animated GIF that I recorded for another query to avoid recording again. In this GIF you can see how to select one of the audio tracks included with the video. In that case there was a stereo audio track and another audio track with 5.1 channels.
I make my video recordings playing bass) with 3 audio tracks (webcam mic, bass, backing track and bass). For some reason I haven’t figured out, the W10 player (movies and TV) plays most of the time track 2 by default (bass track only), when track 1 is the default setting (backing track and bass).
However VLC always plays correctly by selecting track 1 (as configured).
So it’s a mystery to me, and I don’t worry too much either because the continuous changes in W10 apps bring these things.
Regarding Shotcut, maybe something in the video (timestamps, or the TS format itself) modifies the criteria Shotcut has to choose the default track.
However you already have a solution to choose the right track in all scenarios.

7

Yes, that has been my emerging hunch as to what is going on. Shotcut does, like most media players (usually, it seems from your W10 example!) default to what it thinks is Track 1. But it looks like there is something in the source file, or something about its structure, that gets it confused. I’ve crawled through the detailed output from Media Info and cannot see what it might be. But as you say, as long as we know of the risk we can manage it.

Postscript. I’ve now looked yet again through the MediaInfo info for a number of .ts files, some of which were fine and some of which confused nar and Eng tracks. And right at the bottom of the detailed info is the ‘Menu’ data. Here is an example:

Menu
ID: 259 (0x103)
Menu ID: 6316 (0x18AC)
Duration: 1 h 50 min
List: 5200 (0x1450) (MPEG Video) / 5202 (0x1452) (MPEG Audio, nar) / 5201 (0x1451) (MPEG Audio, English)
/ 5203 (0x1453)-888 (Teletext Subtitle, en) / 5204 (0x1454) (DVB Subtitle, English)
/ 3851 (0xF0B) () / 3844 (0xF04) () / 3843 (0xF03) () / 3842 (0xF02) () / 3841 (0xF01) () / 3840 (0xF00) ()
/ 2346 (0x92A) () / 2323 (0x913) () / 2321 (0x911) () / 2306 (0x902) () / 3849 (0xF09) () / 3847 (0xF07) ()
/ 3846 (0xF06) ()
Language: / nar / English / English / English

(Sorry if the layout has gone wrong)

The ‘List’ line shows the two audio tracks (5202 and 5201) in the ‘wrong’ order.

The ‘Language’ line matches the List line but again it has the audio tracks the opposite way round to 'Track 1; Track 2. (The other two are the subtitle tracks)

Now I know that correlation is not causation, but each example that I have looked at so far where Shotcut picked up the ‘nar’ track as Track 1, had this curious ordering. It’s only a small sample - a handful - but maybe, just maybe, Shotcut is making use of this information and assuming that the ordering in the ‘List’ is the track order. Maybe.

Enough!

8

I couldn’t find a way to do that in Shotcut. Maybe there is some parameter that can be included in the “other” tab of the custom export settings.
Searching a bit I found this, but I don’t know if it can be added in Shotcut or if it is only via commands.

ffmpeg - video conversion with multiple audio tracks

1 Like
9

Sometimes, in a video, I need to mark another audio as default, add another audio source, or even remove an existing one.
I don’t know anything about commands, so my usual way for this is through avidemux.
Also Handbrake or MKVtoolnix allow to do this via GUI.

10

This topic was automatically closed after 90 days. New replies are no longer allowed.