2 ms delay

[shadell]

I've just noticed that Gordian Knot is setting a 2ms delay for the audio. Is this correct? If so, why?

The source is a wav file, converted to an mp3.

[apk]

i have this same problem - agk decides to delay by 30ms which causes huge sync issues for me.

[shadell]

I can now answer my own question.

It is not correct, it is a BUG.

The '2' from '2ch' in the file name is being interpreted by Gordian Knot as a delay. If I rename it to '4ch' it sets a delay of 4ms.

And the reason for this is that:
a) When the wav is extracted from the mpg, the text 'DELAY Xms' is not added to the end of the file name
and b) Gordian Knot simply takes the last number in the file and interprets it as a the delay, it does not check for the text 'ms' next to it or anything like that.

[CWR03]

Your explanation doesn't seem entirely likely. A 2ms delay in audio would be almost imperceptible, as it's only two 1000ths of a second, but if your solution works for you that's what really matters.

[shadell]

Your explanation doesn't seem entirely likely. A 2ms delay in audio would be almost imperceptible, as it's only two 1000ths of a second, but if your solution works for you that's what really matters.Why do you say my explanation doesn't seem likely?

Why are you talking about how perceptible the audio delay is? I didn't make any comment on how perceptible it is. I just said Gordian Knot was doing it.

To repeat and clarify: the figure from the '2ch' in my audio file name is being entered by Gordian Knot in the delay field (in the Encoder - Audio tab). This happens when there is no delay figure in the file name, which happens when I feed Gordian Knot mpg files with PCM audio. I would not say this was the case unless I had verified it. It is not a matter of how 'likely' it is - it is definitely happening.


Perhaps what you mean is it is not a likely explanation for apk's problem with Auto Gordian Knot. I can't comment on that, since I don't know how Auto Gordian Knot works internally. If apk doesn't have a number like 30 somewhere in his file names, it can certainly be ruled out as an explanation for his problem.

[Pulp Catalyst]

strange bug even i have to admit, i thought this would of been detected a long time ago, seems strange how this has just come to surface, will do some tests myself as soon as i get time,

even so though, 2ms wouldn't be detectable, so for you other people having sync issues, try not to get hang up on this event, very unlikely it's this causing the main brunt of A/V sync issues.

will post in the next few days if i get my holiday from work,

[BigDid]

Why do you say my explanation doesn't seem likely?

Why are you talking about how perceptible the audio delay is? I didn't make any comment on how perceptible it is. I just said Gordian Knot was doing it.
Hi,

:cool: Sir, CWR03 usually uses neutral formulation and I don't see any attack needing a counter-attack in what was written :rolleyes:
This is my opinion so you may compare.

Perhaps what you mean is it is not a likely explanation for apk's problem with Auto Gordian Knot. I can't comment on that, since I don't know how Auto Gordian Knot works internally. If apk doesn't have a number like 30 somewhere in his file names, it can certainly be ruled out as an explanation for his problem.
If you read this thread: http://forum.doom9.org/showthread.php?t=114034 , it may turn out there are other possible explanations regarding his problem (beside my misunderstanding).

Did

[jggimi]

Misreadings and minunderstandings can happen. Please, folks, try to keep Rule 4 (http://forum.doom9.org/forum-rules.htm) in mind. Thanks.

[shadell]

Sir, CWR03 usually uses neutral formulation and I don't see any attack needing a counter-attack in what was written Did
I think my 2nd posting made it clear that I had established without doubt what was happening. I said "If I rename it to '4ch' it sets a delay of 4ms" - what room for doubt does that leave?

To be told that what I had seen with my own eyes 'doesn't seem entirely likely', with no relevant reason given, is equivalent to saying I am seeing things. This surely merits a response questioning why he said what he said.

As to whether he or I was attacking, that's a subjective judgement. I just want to report a bug and make sure it's taken seriously.

I will give CWR03 the benefit of the doubt and assume he just didn't think about what he was writing on this occasion. Given what I had posted, his claim that it was not likely did not make any sense. He then went on to justify this by saying 'A 2ms delay in audio would be almost imperceptible', when I had never said that I was posting because of a noticeable sync error (although the encode does seem slightly off at some points - perhaps the result of accumulated errors in the production, initial mpeg editing, and now encoding).

If anyone is trying to reproduce this, please note that this only applies to source files with PCM audio as far as I know. It works fine for me with mp2 audio, and I imagine it would for ac3 too.

The reason my audio was PCM is that I had converted it to that from mp2 to ensure that the multiple edits I was making did not introduce sync errors (as can happen when you cut/join mpgs with mp2 audio). This was particularly important to be because I suspected the program I had originally used to convert the files to DVD format had already introduced a slight sync error, and I did not want to compound the problem. But because of the bug I ended up doing just that, and these files have now been distributed with no possibility of getting them back and fixing them :rolleyes:

[Pulp Catalyst]

look mate, this forum didn't make the software, it's a place where every one can express there opinion, and every one has there own way of writing, and like me, your quite new to the site, some people have a way of writing in a paticular way, and besides, he just said unlikely, that's not saying this didn't happen,

and reporting saying that this is a bug, i want it tacken seriously,,,man ease up, this is just a forum to discuss issues and problems with AGK, and CWR03 has been here longer than us, and has a lot of respect no doubt, no one was attacking you, and we don't have to take you serious, that is are free will, i don't see any rules saying that all posts has to be tacken serious, just let it lie, we all try and help the best we can, and we all have a way of speaking and phrasing, we all respect each other here, and there are mods that take care of any insults given,

personaly speaking, he may of been a bit hard, but some people are like that but with over 1,730 posts, he has the right in a way, to be straight, but i don't see him being disrespectful,

just take your time and explain the best you can, and at some point hopefully we can all help each other out,

so peace and respect to you all my friends,

p.s. wait and see what other people say first, as there are many varibles to consider, logic would suggest that when there are a handful of people (using mutiple setup's) conquer with what you say, then it's likely that there is a bug, but as of yet (1 person saying it is a bug) doesn't mean it is a bug,

but concerning the matter in question, i have no idea, but lets hope that all the experience people in the forum (that i for 1 respect your talents and knowledge) can show some light on the subject.

[jggimi]

This thread has wound down, with no further useful information, and I am closing it to avoid any further flame bait, flaming, or acrimony.

I think some level-setting is still in order:

Gordian Knot is an open source project, headed by len0x, and while it has not had a new release since 2003, bugs may be reported via its bug tracking database. Click on submit new in http://sourceforge.net/tracker/?atid=550084&group_id=77391&func=browse to do so.

Auto Gordian Knot is a closed source application, developed by len0x, and developement has officially ceased. Bugs may be reported in this forum, but there is no guarantee that len0x will address them.

Thread closed.

[len0x]

I'm reopening this to grab some more info from original posters (just for my information) and not for further flaming! :)

what is the exact file name of audio files that GK read faulty delay from? (I wanna check if AutoGK has the same problem).

*update* I had a look at the sources and indeed delay calculations are really weird (that part of code wasn't written by me). Autogk uses different method and is not affected.