He makes his decisions about whether or not it's sourced from NTSC or PAL almost entirely by the film's length. Hence the mistake with the Suevia one. Never have I read of him mentioning anything about audio pitch. Unless you have perfect pitch (which I don't) or know the voice of the actor very well (only a couple for me), it can be very difficult to tell the difference unless you play the 2 one after the other.
As for the samples, the M2V has, as you said, a repeated frame after every four frames. If I were reencoding it for DVD I'd remove the dupe frames and encode as 23.976fps with pulldown applied. The AVI is the same.
That's one way, since you said you didn't want to reencode the 2 hour version. After being joined the differences will be so jarring when going from one to the other, for me it might not be worth the effort.Just add borders to the fullscreen and reencode the whole thing?
+ Reply to Thread
Results 91 to 105 of 105
-
-
VDub can detect (some) errors, but it's not good for fixing them.
The other utilities I mentioned are better for fixing, but you have to always check the result.
That's not an error, it's an option that helps Avidemux work better, so it can keep audio in sync. I don't think it actually makes a change to the file. -
manono, I tried deleting the duplicate frames but the video became jerky. What would be the appropriate script to do it?
In regards to combining the anamorphic widescreen with the fullscreen, you're probably right that it would look jarring... I was hoping the missing scenes would be long enough so the discontinuity would only happen a couple times. I'll do a test at some point.
AlanHk, thanks for explaining the index-fix message in Avedemux. Besides using Avisynth, what might be a good method for "demuxing" avi? I had previously been using Vdubmod to get just the audio (which is when I got the error correction message). -
Demuxing strictly means extracting the sound (or video, etc) as-is, without converting format.
Avisynth can be used to extract audio or video, but it always decompresses it in the process; i.e. to the equivalent of WAVE.
Vdub can do either -- i.e.e if the audio is AC3, you can get AC3.
Similarly Avidemux (as you might guess from the name), if used in "copy" mode can extract the original audio.
Also there are command line demuxers which are simple and efficient.
But if you need to process the audio in any way, you will have to decompress the audio (you can trim MP3 or AC3 in compressed format, but that's about all). -
I tried TDecimate, with and without assumefps(23.976). Previewing it in MPC, it played choppily at first, and it didn't help that the movie itself was abit choppy, so I watched the end credits for a more stable image; the silhouettes were definitely choppy. Previewing it again, it looks like the movie itself is improved (once Media player gets going). I guess the thing to do is exclude the end credits from the decimate script...but if I did that, it looks like Assumefps(23.976) would slow down the excluded end credits, so I should leave that line out entirely?
-
I don't know. I applied TDecimate() to your sample and it worked fine. If you have some other parts for which it doesn't work, post a sample. It's possible I suppose (although I don't really think it likely), that different parts of the movie were treated differently. I don't play them, but step through them in VDubMod. If a movie, and if it was deinterlaced so as to produce a dupe frame in every 5, the end credits should be the same.
-
There are the movie credits followed by the "show" credits. The silhouettes move thru both. The opening "show" credits are different and harder to tell if anything is amiss (so I didn't include them here). There might be times during the movie when the silhouettes move around to perform a comic skit (I haven't watched it thru yet), but if they don't, it would mainly be the end that was noticeable. If that's the case, and this brief jerkiness didn't indicate some other issue, it might be an acceptable compromise.
end credits.avi
In regards to the audio, can Avisynth output the audio "as is", without any format change? If so, what command would do that? -
That's odd. Here's a fast encode of the decimated "show" credits to illustrate what I'm saying about the jerky silhouettes:
tdecimated.m2v
AlanHK, sorry if I misunderstood. I wasn't sure if you were saying 'avisynth always decompresses to wave' when you tinker with it in a script, or just 'always', period. I'm surprised it won't output it "as is". Anyway, I'll use the recommended Avidemux. For the avi file I'm tinkering with, Avidemux produced a better synced audio file than Avisynth did. -
Oh, I didn't even get that far. I apologize for not checking the entire thing. That part is 29.97fps. If you had the source files you could probably do something if that part was interlaced. Not now, though. If you want to turn the vast majority of it that's native 23.976fps back to 23.976fps, then that end part will have to play jerky. Or just cut it out and pretend it was never there to begin with.
And I just noticed that in the earlier film credits, they had added little black figures crossing the screen at 29.97fps. So you also have mixed framerates there and there's no easy solution. If it were I, and given the crappy 320x240 source, I wouldn't worry about it if I absolutely had to have it. Of course, even with stuff I really wanted, if the source is hopelessly screwed up I often delete it.
Another alternative, assuming the bulk of it has those dupe frames, would be to encode that part at 23.976fps with pulldown while leaving the end credits at 29.97fps. Encode those parts separately and then join with Muxman at the authoring stage.
I apologize for missing those 29.97fps parts earlier. -
I looked at another dvd from the company, and the movie portion was the same. I dunno if this was their solution to combining the movie with silhouettes or what. I'm tempted to try what you suggest (process most of the movie, and marry the separate parts in Muxman) but it's watchable enough as is.
I've only thought of Muxman as a muxer rather than a "joiner" of clips. So variable frame rates would play okay in a dvd player?
Since you suggest Muxman, does that mean Avisynth (in this situation) could not output the movie with variable frame rates, as well as output synced audio thru Soundout? Just wondering. -
You encode the 2 parts separately, one at 23.976fps, the other at 29.97fps. When muxing the video using Muxman, you put in the first part (the main movie, I assume), and then use the 'Add' button to add in any additional videos (the end credits). The main movie will be 23.976fps with pulldown (making it 29.97), and the end credits will be native 29.97fps. For DVD purposes they're not really variable. By the time they're joined they're both 29.97fps. The length remains the same and the audio won't get unsynched. I do this quite a bit when the end credits are 29.97fps, or if I want to give the end credits a lower quality than the main movie.
If there's one fly in the ointment, it's sometimes at the join place. Usually movies have some black frames between the end of the movie and the time the credits begin. I choose a break with a few black frames at either side of the join. If yours has active video at both sides, then you're likely to get a couple of bad frames there. They'll look like colored squares. If using CCE, it might help to encode the last frame of part 1 as an I-Frame. Try it and see for yourself. Check the join place when done.
Since you're splitting the movie for encoding purposes, I suppose Soundout can also be used with the Trims. I always work with the audio separately myself, and don't know (or much care) what happens. But the Audio portion of Muxman also has an 'Add' button. -
Avisynth decompresses both audio and video; that's what the "source" functions do when they load a file. You can't use it to access the compressed data. (Well, not that I know of.)
As for keeping in sync, that therefore is down to what you get from the source function.
If using AVISource with an avi using MP3 audio, especially if you're doing any trims, add
EnsureVBRMP3sync() which does what it says, though it makes it slower to skip around the file.
Also you can try FFmpegSource2() instead of AVISource, that's very solid, though you get a delay when its building its indices the first time you run it.
Similar Threads
-
HCenc and PanScan
By Comp360 in forum Video ConversionReplies: 2Last Post: 14th Mar 2011, 17:35 -
Problem with HCenc 0.24.
By HolyFries in forum Video ConversionReplies: 2Last Post: 17th Jun 2010, 18:03 -
Problem with HCenc 0.24.
By HolyFries in forum Newbie / General discussionsReplies: 0Last Post: 17th Jun 2010, 13:18 -
Frameserve from VirtualDubMod to TMPGEnc
By pchan in forum DVD RippingReplies: 1Last Post: 25th Aug 2007, 08:13 -
HCEnc
By Belfour in forum DVD RippingReplies: 1Last Post: 28th Jun 2007, 09:57