Audio delay reported in MKV and what to do with it when adding new audio

[pascal]

Hi there

When you get a reported audio delay (either by MediaInfo or Eac3to) for AAC stream in a MKV and then convert AAC to AC3 to add new AC3 stream to MKV using mkvmerge
Do you set audio delay in mkvmerge to value reported for AAC or set no delay and mkvmerge will deal with the new audio stream and sync it to video?

Thanks

[Baldrick]

You use popcorn mkv converter.

Or just try and then report back how it works.

[pascal]

Thanks Baldrick for taking the time.

I had the feeling my post was better placed in Editing. Anyway!
I'm not using popcorn mkv converter but foobar to convert AAC to AC3 (with Aften encoder).
Looks from the description like popcorn mkv converter is not the right tool for this type of conversion.

(PopCorn MKV AudioConverter is a Matroska MKV DTS to MKV AC3 converter. It can also remove DTS tracks and convert OGG audio to AC3,MP3,AAC,MP3)

But even if I was using another tool than foobar the point is what to do when remuxing AC3 stream in original MKV.
This is a question. Do I have to set the audio delay in mkvmerge to the value reported for original AAC stream.

Hope it's clear enough.

Thanks a lot.

[pascal]

My mistake!
After looking at the Tutorial I can see that PopCorn MKV AudioConverter supports AAC to AC3.
But still my question remains even if I use PopCorn MKV AudioConverter.

Cheers

[pascal]

Report for Baldrick

When using MKV with AVC/AAC as input and converting AAC to AC3

PopCorn MKV AudioConverter reports:

Progress log:

Extracting Information for file "T:\Temp\OpeningAndConcert.mkv"
Video Track:
Track Number: 1
Codec: V_MPEG4/ISO/AVC
Pixel Resolution: 1280x720
Aspect ratio: 1,78
Display Resolution: 1280x720
Frames per second: 29,970
Audio Track #1
Track Number: 2
Codec: A_AAC
Language: eng
SampleRate: 48000
Channels: 2

Skipping conversion - none of the audio tracks can be converted or the conversion has been disabled

The conversion has not been disabled. I've set up AAC to AC3.

Any idea?

Thanks

[Baldrick]

Nope.

[pascal]

So what about my first idea to convert AAC to AC3 with foobar? Works very well.
And mkvmerge is the perfect tool to add converted AC3 to MKV.
But still remains the audio delay
None or the same delay as reported for AAC?

Hope someone can help.

Cheers

[nic2k4]

Haven't tried to do what you're doing, but I can tell you that when you do a conversion from DVD to AVI with AutoGK the audio delay on the DVD is used in the muxing of the AVI. I believe what Baldrick is trying to say is try it to find out. Find a source with a reported delay greater than 60ms (that's the level when it becomes noticeable to the average human eye) and convert with no delay; you should see if it syncs or not.

[pascal]

Thanks nic2k4 for your input

So I need to find a source with a delay big enough to spot it.
Any help welcome!

Cheers

[nic2k4]

Use PgcDemux to check the delay from a DVD source (you'll probably have to rip the DVD first). You can get really large delay values if you edit a DVD; just use the remake a DVD option in DVDShrink (or NeroRecode) and trim a bit of the beginning of a clip.

[pascal]

Thanks again nic2k4
Will do as instructed

Cheers

[pascal]

First findings to report that raises a new question

I've used a DVD source as mentionned and PgcDemux reports a Audio/Video delay of -16ms when MediaInfo reports -96ms for the first VOB (ie start of movie)
Who would you trust more between PgcDemux & MediaInfo?

Thanks

[pascal]

And after removing first half second in DVD Shrink PgcDemux reports a Audio/Video delay of -176ms when MediaInfo reports -256ms. Any clue?

Cheers

[pascal]

Final report

With delaycut I've been able to create a delay of 2s which can easily be spotted.
I've then played with mkvmerge and I've got to the conclusion that you have to set up the audio delay in mkvmerge to take into account the delay in the audio. Sound like obvious? Well.

Thanks for your help and happy holidays to everyone

Cheers

[nic2k4]

Thanks for sharing your findings. I would trust PGCdemux more than Mediainfo for this; that's quite the discrepency. I'm not sure what would be the cause, but I've seen some oddities in IFO's before. I'm away from home and can't remember the details, but I think it was under PGCITI map in IFOedit. There were a number of cells with playback time and no lenght (LBA) or something like that. By adding up the playtime for each cells, I was able to find the exact amount of compensation to re-synch the audio. I'll have to check my notes when I get home.