AC3 delay as given by DVD Decrypter & DVD2AVI incorrect

[Wildfire]

Usually, DVD Decrypter (or DVD2AVI) correctly determine the delay of the AC3 audiotrack.

However, now I have an AC3 audiotrack of which the delay is -79ms according to both programs. However, when muxing I discovered it had to be +79ms. And on the second DVD, the delay was given as -74ms but when I used that value when muxing the audio would be approximately 1.5 second too early!

How can I find the actual delay? I'd prefer finding the correct delay because I don't like guessing, that way it'll never really be entirely synced...


-- Edit: well, I decided to play the guessing game anyway... I finally landed on a +800ms value for the delay instead of the -74ms as specified by DVD Decrypter.

I'd still like to know why DVD Decrypter and DVD2AVI made such a huge mistake - about 8 seconds off!

[Farfel]

Don't use DVDDecryptor to obtain the audio delay, or to demux the audio tracks for that reason. I've found the delay it gives you is never correct. Also, DVD2AVI seems to be inaccurate too, because of the dropped frame issue. For this reason, use DVD2AVIdg/DGIndex to obtain the delay, and then confirm the delay using VobEdit .6. Using this method, i've never had any synchronization issues, and thus has worked quite well. (VobEdit should have the real delay value, which DGIndex should match).

[Wildfire]

Originally posted by Farfel
Don't use DVDDecryptor to obtain the audio delay, or to demux the audio tracks for that reason. I've found the delay it gives you is never correct. Also, DVD2AVI seems to be inaccurate too, because of the dropped frame issue. For this reason, use DVD2AVIdg/DGIndex to obtain the delay, and then confirm the delay using VobEdit .6. Using this method, i've never had any synchronization issues, and thus has worked quite well. (VobEdit should have the real delay value, which DGIndex should match).

I should have been more specific -- I am using DGIndex (I wrote DVD2AVI as it's kindof the successor to DVD2AVI). And it gave an incorrect value.

Further more, delays given by DVD Decrypter are usually correct... at least, they are here.

[Nic]

Could you tell us more about what you are doing?

The delay is calculated by looking at the PTS time stamps of the video and audio streams. The delay should never be around 800ms, so something has gone wrong in the procedure you are using.

Some questions that will help find the solution...
How are you multiplexing your streams? What DVD are you ripping? What DVD mode (IFO, File, etc) are you using? What are you encoding to? And any other info that you think might be useful.

There is a chance that the start of the VOB the PTSs are not correct and then correct themselves, but this is unlikely....You could use VobEdit to view the first few PTSs of the Audio and Video stream and see how well they match....

-Nic

[Wildfire]

Originally posted by Nic
Could you tell us more about what you are doing?

The delay is calculated by looking at the PTS time stamps of the video and audio streams. The delay should never be around 800ms, so something has gone wrong in the procedure you are using.

Some questions that will help find the solution...
How are you multiplexing your streams? What DVD are you ripping? What DVD mode (IFO, File, etc) are you using? What are you encoding to? And any other info that you think might be useful.

There is a chance that the start of the VOB the PTSs are not correct and then correct themselves, but this is unlikely....You could use VobEdit to view the first few PTSs of the Audio and Video stream and see how well they match....

-Nic

The problematic DVDs are the four from Stephen King's Rose Red (Dutch PAL release). And the fourth was the worst, as stated before.

Decrypting to HD with DVD Decrypter, letting it handle the demuxing of the AC3 audiotrack. Muxing with AVIMux GUI 1.16.7 using default settings. Using IFO-mode to decrypt. Encoding of the AVI is done with GordianKnot 0.32 beta with no muxing done by GKnot (I do that myself with AVIMux GUI as I like to determine the 2CD splitpoint myself).

[Wildfire]

^^ kick ^^

It just happened again... decrypted a Star Trek Voyager DVD (European release) to HD with DVD Decrypter. It demuxed the AC3 audiotrack and specified a -80ms delay. However, when I let DVD Decrypter decrypt to HD while letting the AC3 audiotrack remain in the VOBs, VobEdit specified 0ms as audiodelay which was the correct value.

Is DVD Decrypter somehow messing up?