Question about editing out commercials for TV capture

[Acceler8n]

Hi,
I tried to look under the sections and couldn't decide where this question would go under, so I placed it here, which is what I use to make the videos.

My question is that I use a TV card to capture shows on TV in mpeg2 format. I then use GordianKot to convert them to DivX. Then I use Nandub to try to cut out the commercials...but when I do this, it does not cut where I want it to and I end up losing a couple of seconds of video (from 5 to sometimes up to 30 seconds), or get parts of the commercial still. Right now Im just cutting a bit before where I intended to... to avoid losing and part of the show, but then I get a couple seconds of the commercial still. Does anyone know how I can cut out the commercials with getting any commercials or losing any part of the show? Your replies are greatly appreciated.

[JohnMK]

Correct me if I'm wrong, but I think you can't cut in-between key-frames, so there's no way, at your stage in the process, to get things as clean as you want them to be.

[manono]

Hi and welcome to the forums-

What you're asking isn't that easy to accomplish. If you're willing to use DivX3, then you can set keyframes where ever you want in the Stats File Editor Tab after the first pass. Then you can cut out the video and audio you don't want from the completed .avi later. To do it before encoding (or if using DivX5), you'd better know your way around AviSynth. But for some help, see hakko504's comments in this thread:

http://forum.doom9.org/showthread.php?s=&threadid=36218

[Swan]

There is also the possibility to edit the Mpeg-2 files first, then treating the remaining, commercial-free, segments as you would vob-files. Just name then 1, 2, 3, etc and load then into DVD2Avi, then create a d2v-project, load that into GordianKnot and you're on your way. Convert the mpa (Mpeg audio) that Dvd2Avi created to wav, then load the wav and d2v project into Gordian and you're on your way. :-)
Doing it this way instead of editing the final DivX file will also give you greater quality on your final DivX file, as you won't waste any valuable bits on the commercials.

I do this every day and it works great. It enables me to use GordianKnot fully, run a compressibility test, crop and resize with the correct AR and all the other great things Gordian offers.

There are limitations in how precise your cuts can be, though. And you may get audio video sync problems if you drop frames while capturing, or use a poor Mpeg-2 capturing program, especially if you capture very long shows.

Mpeg-2 is not a format designed with editing in mind. It's a distribution format. But there are software titles available that will let you cut exactly where you want, if you just take a few important steps before you capture.
My best tip, if you want to try this, is set your Mpeg-2 capturing software to capture using I and P frames only. I-frames are Mpeg-keyframes, encoded without any reference to the previous or following frame. P-frames are predicted frames and the reason you should use at least 1 is that in my tests, I have found that just using I-frames produces a lesser quality than 1 I and 1 P.
Whenever you use B-frames, you may not be able to get precision cuts. It can differ on a few frames, but certainly not as much as you describe using the method you're using now.
Disabling B-frames will increase the file size, but if you're going to capture, encode to DivX and throw the Mpeg away, then this is perhaps not be a problem for you?

You can use either M2 Edit Pro, Womble Mpeg-VCR or TMPGEnc to edit out the commercials (without re-encoding, the editing is lossless).

There was a thread about his in the "Capturing Video" forum not too long ago that you may want to check out.
http://forum.doom9.org/showthread.php?s=&threadid=34728

/Swan

[Acceler8n]

Thank you for your replies...I am using an AverMedia AverTV Studio card and I didn't see an option to change any I or P frames. Ill try to edit the mpg2 first then covert them to DivX. Thank you once again.

[Swan]

I used to have an Aver TV Studio card too.:)
Are you using "AverTV" to capture with?
Are you sure there is no such setting if you select a customized setting for capturing?
I am almost 99% sure you can change it. It's under the Customized capture settings. Look for the option to set the Gops there (Group Of Pictures):)

Even if you can't set this, you can make a test capture and check if you think you get the precision cuts you aim at. Usually, you will have to sacrifice a few frames or let a few unwanted frames be included in your edits if you're using B-frames (and I'm sure Avermedia's software uses them by default). But even this is better than the method you use now, isn't it?
Also, you could switch to another software. I highly recommend Ulead VideoStudio 6.

/Swan

[kempodragon]

I answered a similar question awhile back. Here's the short version:

1. Capture Mpeg2 in I-frames only. This allows precision cuts, no having to leave unwanted frames.
2. Use DVD2AVI and creat project file and audio file.
3. Use Winamp to change audio file to wav
4. Create AVS with Gordian Knot and add wav file with AudioDub function of Avisynth.
5. Open avs with VirtualDub and delete frames with commercials. Save VirtualDub settings.
6. Save edited sound into new wav file and close VirtualDub
7. Use Lame to compress edited wav file
8. Now go back to your avs, remove the audio, reopen VirtualDub, load your avs and your settings, and compress the video normally. Use Nandub to mux the edited sound, and voila a capture with no commercials.

[Acceler8n]

Hi Swan,
I found the settings...its under GOP. Now what do I set it to? There is alos a setting for sub GOP. Also, does ULead work with my existing TV card?

[bb]

You should know a tool called PVACut. It can cut at keyframes only, but it's free and good. If your format is not PVA, then use mpg2pva.

See also this site (in German): http://www.pctvsat.com/html/mpeg_ii_schneiden.html

bb

[Swan]

@Acceler8n
I found the settings...its under GOP. Now what do I set it to? There is alos a setting for sub GOP. Also, does ULead work with my existing TV card?
I don't remember exactly how they should be set. Look in the manual, it's there. I believe it was a matter of setting GOP and Subgop to 0 that produced I-frames only, but I won't swear on it.

But I suggest you use at least 1 P frame. I did some extensive testing a short time ago (capturing a test image) with various compressions (HuffYuv, Mjpeg, Mpeg-2) to determine what quality I could achive and I noticed one vital thing:
Using I-frames only produces more artifacts than using 1 P frame and 1 I-frame in the Gop settings. I tested with PowerVCR (Deluxe version) and Ulead VideoStudio 6.01.
Both softwares produced better looking Mpeg-2 video when 1 P frame was added. This setting is what I use now. For me, it can sometimes mean a loss of max 3 frames, or keeping three frames I don't want. You have to decide for yourself what you prefer; better looking compression or on-the-frame precision cuts.

As there often are a few frames of pure black between a show and the commercial, I don't get a lot of junk with my setting. You'll just have to try and see what you think. If you run into audio/video going out of sync, first test editing the Mpeg-2 with a different software. If that doesn't help, then test capturing with VideoStudio. If you don't have any problems using AverTV, there's no reason to switch software.

Yes, Ulead VideoStudio will work with the drivers for your Avermedia card. You should try to find the 6.0 version and then upgrade it to 6.01. Otherwise it will not work. 6.0 does not work with AverTV Studio's drivers. You need an ftp-client to download the 6.01 patch. Ulead has taken it off their web site, now they only offer a patch with version number 6.02. This one screwed up VideoStudio for me and for some other users in the "Capturing Video" forum too. But if you go to ftp.ulead.com you will find the 6.01 patch there.

kempodragon's suggestions sounds great to me. The only negative thing as I can see it is that you won't be able to perform a compressibility test in GordianKnot, as the commercials are included in the d2v project. Also, the bitrate calculation may be tricky. How do you solve that, kempodragon?

Acceler8n, when using VirtualDub as kempodragon does, you won't be using 2-pass encoding (Nandub) which is what GordianKnot uses when compressing to DivX 3. It will be 1-pass encoding and that does not give the same quality, IMHO.
But perhaps you use DivX 5?

@BB
You should know a tool called PVACut.
I don't speak German, but the tool looks interesting.
Thanks for the tip. Have you used it?

Also, another idea.. What if one opened the Mpeg-2 in DVD2Avi and used the start and end markers there? And just create a DVd2Avi project for each commercial-free segment?
I know the audio offset will differ between each segment, so it's not a good idea to slap the commercial-free segments together again after encoding to DivX. It's probably a bad idea, but I just thought I'd throw it out here and see if someone else has tested it.

/Swan

[kempodragon]

I don't exactly do a compressability test with captures. I've done enough vidcaps that I can generally guess how compressable an episode is. For stuff that I haven't done, like the Iron Chef France special, which was filmed live and couldn't be IVTC'ed, I ran a first pass only with XVID, and loaded it into Gordian Knot. I'd already made the mp3 file, so coming up with filesize was straightforward. Basically with captures, if you wish to test the compressability, you can either do a first pass only with XVID or Divx3(NanDub) and load the stat file. You'll have to adjust the # of frames, and also the framerate if you IVTC'ed. The other way, which is what I did, is save the edited file to Huffyuv, and then ran my first pass on that. The reason for this is because it's alot easier to load the Huffyuv file into VirtualDub and get the number of frames. Sorry, but I don't use Divx5 so I can't help with compressablility in that way

[Swan]

@kempodragon
you can either do a first pass only with XVID or Divx3(NanDub) and load the stat file.
Ok, so you perform the first pass, either with Xvid (in VirtualDub I presume?) or DivX 3 (Nandub).
The entire first pass? Not an estimation of the compressibility, by using SelectRangeEvery(280,14)in the .avs?

Can an Xvid stats file be loaded into GordianKnot?
Does GordianKnot "understand" it?

/Swan

[manono]

Hey Swan-

GKnot reads the XviD stats file fine-either the compress test ones or the entire first pass ones. Just push "Load" when done. I do it all the time.

[bb]

Originally posted by Swan
@BB

I don't speak German, but the tool looks interesting.
Thanks for the tip. Have you used it?

Also, another idea.. What if one opened the Mpeg-2 in DVD2Avi and used the start and end markers there? And just create a DVd2Avi project for each commercial-free segment?
I know the audio offset will differ between each segment, so it's not a good idea to slap the commercial-free segments together again after encoding to DivX. It's probably a bad idea, but I just thought I'd throw it out here and see if someone else has tested it.

/Swan [/B]
For a translation use http://babelfish.altavista.com .
And yes, I have used it, but for DVB sources. Was easy to use, and it worked reliably.

Your proposed DVD2AVI method sounds interesting. Maybe you could use AviSynth somehow to add the pieces together, and compensate for the audio delays at the same time.

bb

[Swan]

@bb And yes, I have used it, but for DVB sources. Was easy to use, and it worked reliably.

That's great. My files are mpeg-2 program streams. It would be great if the mpg2pva cutter worked on those. I've never seen any freeware Mpeg-2 editors before, so thanks BB!

Your proposed DVD2AVI method sounds interesting. Maybe you could use AviSynth somehow to add the pieces together, and compensate for the audio delays at the same time.
If I had advanced over the pure basics of Avisynth.. Or was more used to programming. :) I can only handle simple stuff, but over at the Avisynth forums I see posters who would probably be able to pull it off with ease. :-)
If I only had more time.....

For everyone scared to install VideoStudio 6 Update Patch (Build Number: 6.02, Released Date: 2002/10/07)

This morning, I visited Ulead's site to see if there were any new updates: http://www.ulead.com/tech/vs/vs_ftp.htm).
I noticed that they had changed the *release date* on the 6.02 patch. I downloaded and installed the 6.02 patch in late September, definitely not in October. It was on Friday, September 27, 2002, 11:19:23 to be exact. :)
It screwed up VideoStudio to the point where I could no longer select to capture interlaced. On-the-fly-deinterlacing was enforced.

I installed the "new" version of the patch and captured an hour. It worked like a charm. Ulead must have done some changes in the patch but tries to hide it by using the same build number. Just thought I'd share this with everyone who uses VideoStudio and has been scared off from installing the 6.02 patch by my and other's posts about it.
I have not yet rebooted after installing the patch, but the installer didn't ask me to either. If I run into problems after rebooting, I will edit this post.

BB, I hope it won't be considered cross-posting if I post about it in the Capturing Video forum too?

/Swan

[Acceler8n]

One other thing, I noticed that when converting these mpg2 to DivX, there is an audio sync. When using GordianKnot and setting the audio encoding parameters from the original mp3 parameter to expert, the audio sync problem is reduced, but its still there. I read something about adding a command to the lame line, but I dont understand exactly wqhat to put. I tried a couple. Is there a way to fix this? I capture the shows at 224 kb/s for the audio. How would you fix this? Thanks ahead.

[Swan]

@Acceler8n
One other thing, I noticed that when converting these mpg2 to DivX, there is an audio sync. When using GordianKnot and setting the audio encoding parameters from the original mp3 parameter to expert, the audio sync problem is reduced, but its still there.
Does your Mpeg-2 file have perfect synch between audio and video, but your finished DivX file with mp3 does not?

Did this start to happen now, after you edited the Mpeg-2 files, or did you have this problem when you encoded like you used to (trying to edit the finished DivX in Nandub)?

Audio and video going out of sync on analog captures is a common problem, at least it seems from what I've seen on the Capturing Video forum, if one does not capture in a lossless format like HuffYuv, but uses Mpeg-2.

A couple of more questions:
Are you using the DVD2avi that came with GordianKnot or a newer version?

Do you double check that Gordian fills in the correct audio offset value? It reads this number from the file name of the wav.

Are you down or upsampling the audio (conversions between 44.000 Hz and 48.000 Hz can cause audio video sync problems).

Have you tried encoding the wav to Mp3 with a Lame with, for example the Razorlame gui?

If the problem started after you began editing the Mpeg-2 files, see what I wrote in an earlier reply in this thread. :-)

/Swan

[Acceler8n]

Actually its been out of sync ever since i started coverting mp2 shows to DivX. The mpg2 output file is really out of sync (like hald a second) but when i covert them to DivX, the sync problem is reduced, but its stil there, and gets a bit worse as the movie progresses. Im using GordianKnot 26.1 and did not upgrade any programs, just what it came with. Is there i site that shows the command lines you can enter in for lame converting?

[Swan]

Actually its been out of sync ever since i started coverting mp2 shows to DivX. The mpg2 output file is really out of sync (like hald a second) but when i covert them to DivX, the sync problem is reduced, but its stil there
This is seriously bad :-O
You will never be able to get perfect audio and video synch on your DivX video if not even your original Mpeg-2 file is in sync.
You need to look at gettig perfect audio/video sync on your Mpeg-2 files before attempting to convert them to anything else.

What you say goes against anything I've ever experienced; that your mpeg-2 file is out of sync, yet the problem is reduced when you convert to DivX. It sounds weird. It's not a probable scenario. It's always the other way around; the Mpeg-2 is in sync, the DivX is not.

Perhaps this is a playback problem? What software player are you using to play your Mpeg-2's with? Have you tested any other players than the one you're using?

/Swan

[Acceler8n]

Hey Swan,

Yea it sounds weird, but thats what happens. I think AverMedia's software has problems, sometimes the sync is perfect but most of the time it is off when capturing MPEG2. I use Windows Media Player to play all my files.

Can you recommend me any capturing software that is a freeware that I can use to capture videos from my TV card? Thanks Swan.

[Suzahara]

Isn't everyone forgetting about good old Trim()? Just make a d2v, put the d2v in and avs with MPEG2Source, open up that avs in vdub, get the frame numbers of stuff you want to keep and put those in Trim(start,end), and a perfectly cut, easy way of trimming and converting. And yeah, just demux and mp2 or whatever format your audio is, convert it to pcm, and use AudioDub and it should trim your audio as well.

E.g.
Video = Mpeg2Source("C:\source.d2v")
Audio = WavSource("C:\source.wav")
AV = AudioDub(video,audio)
A = Trim(AV,start,end)
B = Trim(AV,start2,end2)
C = Trim(AV,start3,end3)
Return A+B+C

[Swan]

@ Acceler8n

Yea it sounds weird, but thats what happens. I think AverMedia's software has problems, sometimes the sync is perfect but most of the time it is off when capturing MPEG2. I use Windows Media Player to play all my files.
OK. But what you are saying doesn't make sense. If the audio is off in the Mpeg-2 file, then converting that to Mp3 using any experts parameters will not fix the problem.

So, I can only see one explanation: The playback software (the filters used to decode your Mpeg-2 file's video and audio) doesn't sync audio and video correctly. I have seen this on my own captures on many occasions. I captured a video with PowerVCR for testing purposes. Played it with MediaPlyer 6.4 and Elecard's filters and the sync was off. Not as much as half a second, but it was off. Disabled Elecard and played the file through Ligos' filters and the sync was perfect.

If you have not got any DVD software player installed, but just AverTV, it installs Intervideo's Mpeg-2 playback filters. If I were you, I'd install one or more software DVD player such as PowerDVD, Sonic Cineplayer or WinDVD and open your captured videos in either of them (not in Mediaplayer) and check the sync.
If the sync is still off, then the sync is probably off in the file.

There are no freeware Mpeg-2 capturing softwares that I know of, since the Mpeg-2 technology requires licensing.

@Suzahara
This sounds to easy to work. :)
I will check it out.

/Swan

[Swan]

@suzahara
I tested your script and it worked as I wanted, partially.:)
Because I want to perform a compressibility test in GordianKnot before encoding, I need to know where and how to put SelectRangeEvery.
Here is the script I made:

LoadPlugin("c:\program files\gordianknot\mpeg2dec.dll")
LoadPlugin("C:\Program Files\GordianKnot\decomb.dll")
Video = Mpeg2Source("movie.d2v")
Audio = WavSource("movie.wav")
AV = AudioDub(video,audio)
AV = DelayAudio(AV, -0.182)
AV = FieldDeinterlace(AV)
AV = crop(AV, 11,13,687,552)
AV = BicubicResize (AV, 544,400,0,0.5)
A = Trim (AV, 9050,38236)
B = Trim (AV, 39198,66742)
C = Trim (AV, 67704,85916)
Return A+B+C
AV = SelectRangeEvery(AV, 280,14)

This does not produce the desired effect. When opening the avs, the video plays as normal, not jumping from scene to scene, which is what Gordian does to check compressibility in various scenes of the movie.

If I put AV = SelectRangeEvery(AV, 280,14)
on the line below AV = BicubicResize (AV, 544,400,0,0.5) (before the trim lines), I get an odd looking video with just three frames where parts that I wish to trim away are included (3 frames from the end of the video which I do not want in my final file).

How and where do I add AV = SelectRangeEvery(AV, 280,14) so the script will enable GordianKnot to jump and check compressibility in scenes in the *trimmed video*, not going through the parts I have elected to cut away?

Another question. When I open the avs in Nandub, I can save the trimmed audio as a wav and encode this to Mp3.
How should this be muxed into the final DivX?
With what audio skew correction?
As you can see my Mpeg-2 file had an audio delay of -182 ms. Is this also a valid number for the trimmed video, when I mux the mp3 into the final DivX?

/Swan

[Suzahara]

Hmm, if you do a D = A+B+C, and then do a SelectEvery(D,280,14), that should probably work since it's doing it off one clip.

The -182 ms is just for audio. You can just put this number in the delay box in besweet, or gordianknot will also make the audio for you if you open a wav or ac3 and it will mux it.
Otherwise: If you're planning to do a cbr mp3, then that can be muxed with vdub, otherwise you'll need to use nandub to mux a vbr mp3.

[manono]

Hi Swan-

By using Return in the command, nothing happens if you put anything after it. I think I'd do it something like this:

A+B+C
SelectRangeEvery(280,14)

But I think Suzahara's D = A+B+C.SelectEvery(D,280,14) should work as well.

[Swan]

Thank you both! SelectRangeEvery works now!
I'll experiment with this some more.

The -182 ms is just for audio. You can just put this number in the delay box in besweet, or gordianknot will also make the audio for you if you open a wav or ac3 and it will mux it.
I always create the Mp3 first, using Lame (and the Razorlame GUI) and mux it in via Nandub (or let Gordian do it). I know about using Interleave --> Audio Skew in Nandub. :) I was just getting confused as to whether I should honor the audio delay of -182 ms or not, when I muxed the *trimmed audio* with the finished DivX video. Is this number still valid, after trimming the d2v in the avs? I guess that's what I'm asking. :)

/Swan

[Suzahara]

Well, that value probably should actually be applied before it's used in the avs, since it's a delay from the original source. You can just use besweet to transcode the ac3 to wav and put the delay in. Then use that for the wavsource.

[Swan]

@Suzahara
I am using your script on Mpeg-2 files from my own TV-captures.
The audio format is mpa, not AC3 on those.
I will apply the delay and see what happens. Since I mux in the audio in the final DivX video, I will hopefully see if the delay is necessary or not. I too would assume it is.

/Swan

[Suzahara]

Originally posted by Swan
@Suzahara
I am using your script on Mpeg-2 files from my own TV-captures.
The audio format is mpa, not AC3 on those.
I will apply the delay and see what happens. Since I mux in the audio in the final DivX video, I will hopefully see if the delay is necessary or not. I too would assume it is.

/Swan

Ah yes, forgot what we were dealing with. Principle is the same though if it's necessary to apply a delay. Only time you never need to is if the delay is 0 ms (of course :p ), or if you cap to a format that doesn't need the delay (or the software automatically applies it) like huffyuv or mjpeg with pcm audio capped in vdub, I never need to apply delays. MPEG2 however needs a delay, not sure exactly why (anyone care to comment on why?).

[Swan]

When I experiemented with setting start and end markers at various points in the video using DVD2Avi I noticed that I get a different audio delay value, depending on the start marker's position.
This is also why I am a bit worried as to what happens when I start selecting start and stop points (trim points) in VirtualDub, using your avs, Suzahara. What does Vdub know about the specific audio delay on the markers I set? How can Avisynth correctly adjust this or know this?
Does the audio delay vary in an Mpeg-2 file, since this is what DVD2Avi seems to tell me?

I have noticed that if an audio delay is present at the very beginning of a Mpeg-2 file (unedited, straight from the capture software) depends on the software. With some, I get a delay, some have 0 delay. The same goes for Mpeg-2 editing software. When trimming clips in some softwares, a delay is introduced, in others, the trimmed file always has 0 delay, even if the orginal video the clip was trimmed from had a delay. This is mysterious stuff indeed.

I do not wish to capture to HuffYuv or Mjpeg without any audio compression, for various reasons. But I would assume the audio/video sync problems are fewer in these formats, provided one has the sufficient hardware. Mpeg-2 is superior for my purposes.

/Swan

[Suzahara]

Yes, I believe that delays do vary in the mpeg2 depending on where you select. However, the number in the filename is all you need to apply to that section that you saved. Avisynth will trim the audio that you have dubbed into your file accordingly. And as long as the lengths are the same, you'll get a perfectly synced file.

[Acceler8n]

hey Swan,

I tried PowerDVD and the mpeg2 audio sync is good. Thanks a bunch Swan! I didn't know Media Player was causing the mpeg2 sync problem, I kept thinking it was the AverMedia video capture software.

Now how should I encode the audio so that it syncs with the video for DivX5? I noticed that using the expert mp3 settings makes the sync a bit better than the standard one. I capture the audio at 224 kb/s but once it is coverted to divx, the audio bitrate is 96 to keep the file size down. Could the reduction in the bitrate be the problem? Thank again for your responses Swan.

[Swan]

@ Acceler8n I tried PowerDVD and the mpeg2 audio sync is good. Thanks a bunch Swan! I didn't know Media Player was causing the mpeg2 sync problem, I kept thinking it was the AverMedia video capture software
You're welcome. Thanks for starting the thread, manono and Suzahara has helped me with stuff, because of it! ;)

MediaPlayer itself is not causing the problems, Acceler8n.
This is how it works:
When MediaPlayer gets a video clip, in this case your Mpeg-2, it searches for a filter, a codec, which is suitable for decoding the video: i.e. playing it.

When you installed AverTV, it installed filters from Intervideo for Mpeg-2 encoding and decoding (recording Mpeg-2 and playing Mpeg-2).
These should work fine.
Perhaps you had some other Mpeg-2 filter in your system that was interfering, so the Intervideo filters were not used, as Aver had imagined.

Anwyay, when you installed PowerDVD, it installed filters too (but by Cyberlink). When you open your Mpeg-2 file in PowerDVD, it uses Cyberlink's filters to play the file. And hey presto! You get sync!

Now, in MediaPlayer, probably some other filter but Intervideo's was used and it was not suitable or poorer than Cyberlink's. Always, always test your Mpeg-2 files with different software (other Mpeg-2 playback filters) to determine if audio/video not in sync is a playback or a capturing problem. It took me a while to figure this out and it was a relief when I saw that it was not the capturing that had gone wrong..:)

You don't need any special settings for the audio to sync on your DivX. Just pay attention to the audio delay that DVD2Avi gives you.

But if you care about the audio quality and use AverTV, set it to use 384 kbps for audio bitrate. I remember being frustrated at the poor quality of the 224 kbps audio it always selects and which is used in all the presets. It has a nasty, ringing, metallic sound to it.
I doubt your choice of 96 kbps for Mp3 audio has any effect on the audio and video sync issue. But you can always test encoding the audio yourself with Lame and Razorlame, instead of letting Gknot do it, just for fun. I personally never go below 128-192 VBR Mp3 for stereo audio and 64-96 for mono.


/Swan

[Acceler8n]

Hey Swan,

I have another problem. After coverting the shows to DivX, the volume of the audio is too low...how can I increase the volume of the audio?

[Suzahara]

Originally posted by Acceler8n
Hey Swan,

I have another problem. After coverting the shows to DivX, the volume of the audio is too low...how can I increase the volume of the audio?

This can easily be done in besweet. Just select one of the premade profiles and it will put the gain on. Or there's a volume control under vdub and ndub's audio menu, just crank that up.

[Acceler8n]

To Swan:

I use the mpa audio file that DVD2AVI creates then use GKnot to convert it to mp3. Why do I have to convert it to wav first? Is that necessary and what are the benefits of that?


I also tried using just I and P frames you suggested, and the mpg2 videos can be cut with more precision. Thanks for the tip.

What bitrate and settings do you use to capture your mpg2 videos using PowerVCR II that gives you good DivX coverted videos and good audio sync? I think my settings may be wrong. Right now (using PowerVCR II) here are my settings:

MPEG2
320 x 240
video: 3000 kbps
audio: 256 kbps
and use default settings for the GOP

Thanks Swan.

[Swan]

I use the mpa audio file that DVD2AVI creates then use GKnot to convert it to mp3. Why do I have to convert it to wav first? Is that necessary and what are the benefits of that?
You don't have to at all. :) I do it because I want to edit the audio sometimes, fade it in and out, stuff like that.

What bitrate and settings do you use to capture your mpg2 videos using PowerVCR II that gives you good DivX coverted videos and good audio sync?
I capture at 720 x 576, bitrate 10.000, highest quality setting, audio at 384 and for the GOPS 1 I and 1 P frame. But this is irrevelant. Your settings are fine. Do try another software, preferably one that will let you see what happens during capturing. Perhaps you are getting frame drops during capturing?
This can cause audio and video to go out of sync.
You won't know until you use a software that will tell you about this as it happens. Try another software, like Ulead VideoStudio.

Is this the first Mpeg you've had this problem with?
Or does it happen all the time, every video you want to convert to DivX has this problem?
Are we still talking about an unedited file, straight from PowerVCR?
If you have edited the file and audio and video is out of sync, then test another editing software (see the other post where you asked about that).

/Swan

[Acceler8n]

When I used AverMedia's capturing software, although the settings are limited, the audio sync was much better. Recently I just got PowerVCR and tried to convert mpg2 captured sing PowerVCR to DivX, and the audio sync was really bad. And yes all these files are unedited, straight from the capuring software.

What do you mean by dropped frames? I tried to use VirtualDub VCR but my TV tuer card was compatible with the software. The funny thing is that the video that PowerVCR captures is in perfect sync in mpg2 format, but the converted DivX file is out of sync.

[Swan]

When I used AverMedia's capturing software, although the settings are limited, the audio sync was much better. Recently I just got PowerVCR and tried to convert mpg2 captured sing PowerVCR to DivX, and the audio sync was really bad. And yes all these files are unedited, straight from the capuring software.
Ok, then try using a different software.
I have found that the files I capture with Ulead VideoStudio are much more robust, i.e. audio and video does not go out of sync when editing them and producing DivX files from the Mpeg-2 files.
But I can make DivX with good audio/video sync out of stuff captured with PowerVCR, using Mpeg-2 editing tools.

Don't be afraid to try other programs!:)

The funny thing is that the video that PowerVCR captures is in perfect sync in mpg2 format, but the converted DivX file is out of sync.
It's like what I wrote before "I have not completely solved the mystery as to why audio and video may be in perfect sync on the Mpeg-2, but goes out of sync when converted to another format. But I am leaning towards that the software we use produce poor Mpeg-2 files, files that are not fully compliant to the standard. Or that DVD2avi and Mpeg2dec.dll are not as kind and accepts flaws in the Mpeg-2 file as a software DVD player obviously does. There may be some error correction in the DVD software players that is not included in DVD2Avi or Mpeg2dec." See http://forum.doom9.org/showthread.php?s=&threadid=37986.

Did you make sure that the correct audio delay was set in GordianKnot?
Did you mux in the audio manually in Nandub to see if this was the problem?

If yes, and the problem persists, you have two options:
A: Start using a different capturing software, see if that helps.
B: Edit the files you already have encoded with PowerVCR or AverTV, using various softwares. Some work better than others, you really have to test for yourself, Acceler8n.

How long are the shows you're having trouble converting? In minutes?

What do you mean by dropped frames?
This means that during the capturing, the software omits frames.

They are "skips in your video or audio, like stutters or missing chunks of time. They can also appear to make the audio and video be out of synch" (http://artserver2.art.columbia.edu/DMC/docs/FCP-QS5-DropFrames.html)
You can also have some use of this (although you capture in Mpeg-2):
http://forum.doom9.org/showthread.php?s=&threadid=32575

/Swan

[Acceler8n]

The shows I capture are 60 minutes long. For example the show "24". Since I am not home when the show is on, I set the recorder to capture the shows but the mpeg2 files are so big thats why I want them in DivX. I dont usually watch them till later so I archive them until I finish watching them, thats why I prefer the smaller file sizes.

I just took your advice and captured using AverMedia with only I and P frames. The resulting DivX file can be cut using NanDub with more precision, which helped edit out the commercials a lot. Thanks for the help! The only problem is that AverMedia's VCR scheduler doesn't allow for scheduled captures using custom settings, it only has the pull-down tab for its pre-set settings, (ie. best, good, long, etc.) So I have to manually start the capture using custom settings.

I will try to get ULead to give that a try.

[Swan]

@Acceler8n
Believe me, I know how tricky this stuff is.
The lack of a good selection of software to capture Mpeg-2 using a scheduler is the problem I'm having too.
PowerVCR worked better for me than AverTV, and I still use PowerVCR, now with my Pinnacle card, but it's still not a perfect software. The files it generates are somehow badly encoded, since on and off, I get the problem of audio and video not being in sync on a DivX file encoded from the Mpeg-2 material. I have battled it every time, though, by editing the files into suitable portions. Normally, I also keep the Mpeg-2 files, and the audio and video is always in sync on those.

I do recall how irritated I was by the lack of customized presets in AverTV, it was not configurable at all. Sadly, Ulead VideoStudio does not have a scheduler.
There is one other Mpeg-2 capturing software with scheduler you can try: Intervideo WinDVR.

The shows I capture are 60 minutes long.
Yeah, the longer the video is, the more likely it is that you'll get bad sync between audio and video when encoding it to another format.

I have another idea for you. Since you capture at such a small resolution and low bitrate anyway, why not test capturing in Mpeg-1 format using the VCD preset in AverTV or PowerVCR?
I believe Nandub and VirtualDub can open these files directly, as opposed to Mpeg-2.

just took your advice and captured using AverMedia with only I and P frames. The resulting DivX file can be cut using NanDub with more precision, which helped edit out the commercials a lot
????
Capturing in Mpeg-2 format with only I and P frames is good when you want to edit the *Mpeg-2* file. It does not have ANY effect on where the keyframes are placed in the final DivX-file you encoded from the Mpeg-2 material.

By the way, have you made sure that PowerVCR does not split your recordings into several 600 MB files?
When you capture an hour with it, how many files do you get?


/Swan

[Acceler8n]

When I capture an hour long show, I believe I get 3 mpg2 files. how do I set it so that it does not split the files?

Using I and P frames does not have an effect on the keyframes of the DivX file? I think I may be mistaken, but I I noticed that the keyframes where in smaller gaps when I capture with just I and P frames, Ill go check it over.

Can you still convert to DivX using mpeg1? I don't capture using mpeg1 because I cant get DVD2AVI to open an mpeg1 file.

[Swan]

When I capture an hour long show, I believe I get 3 mpg2 files. how do I set it so that it does not split the files?
That may be the cause of your problems.
Imagine this: PowerVCR creates 3 separate mpeg-2 files for your 60 minute capture. Each Mpeg-2 has a different delay (audio and video does not start at the exact same time).
You open Mpeg #1 in DVD2Avi, it loads Mpeg-file #2 and #3 automatically.
Correct me if I'm wrong, but DVD2Avi only reads the delay value from the first file (mpeg file #1) an that's the delay it writes in the file name for the mpa file.
So it sets, for example, -100 on the mpa audio file it creates, even though file #2 and file #3 may have a delay of a different value, for example -160 or (+)250.
Do you see the problem that may arise from that?

What you need to do is make some changes in PowerVCR's registry entries. I can send you a registry file that you just need to click on, if you feel uncomfortable messing with the registry. You can screw up Windows, unless you know what you're doing. Send me a PM with your e-mail address.
You can also contact Cyberlink's support and they'll send you the registry file.
What operating system are you using and are you using NTFS or Fat32?

Can you still convert to DivX using mpeg1? I don't capture using mpeg1 because I cant get DVD2AVI to open an mpeg1 file.Yes you can, but you're right, DVD2Avi doesn't seem to like Mpeg-1.
And since you use GordianKnot, it's perhaps best to stick to Mpeg-2.

/Swan

[stax76]

I suggest to use DVX instead of Gordian Knot because GKnot is a straight DVD tool. In DVX you can cut frame accurate with VirtualDub before you start the encoding process.

[Swan]

@Dolemite
I'll check DVX out. Thanks for the tip.
I see that you're the developer. ;)
In what way is DVX better than GordianKnot?
Does it have the same features?
You mention VirtualDub. Can it also be used with Nandub and DivX 3.11?

The most important ones in my view are cropping and resizing to be sure of correct AR (getting help with this without the use of a regular calculator) and the compressibility test. Are these features available in DVX?

How come it is not listed on Doom9's download page?

/Swan

[stax76]

In what way is DVX better than GordianKnot?

there are a few advantages, a feature list can be found on my homepage

Does it have the same features?

the features are similar, I've learned a lot from GK

Can it also be used with Nandub and DivX 3.11?

it supports DivX5 and all XviD releases, different codec settings for first and second pass are possible

The most important ones in my view are cropping and resizing to be sure of correct AR (getting help with this without the use of a regular calculator) and the compressibility test. Are these features available in DVX?

I released yesterday a version that takes care of ITU-R BT.601 and aspect error calculation same as in GK, Comp Test is currently only for DivX5 available, not for XviD

How come it is not listed on Doom9's download page?

DVX is not very popular, maybe I'm doing something wrong, suggestions would be highly appreciated. The biggest problems are probably .NET and people don't know how to use the program. Scripts, commandlines and so called macros used in the scripts and commandlines might confuse some people, but the scripts and commandlines don't have to be edited by hand, I never do so, there are menus and profile systems for everything

Regards,
Dolemite

[Acceler8n]

To Swan,

Even when I capture a 30 minute clip and get just one mpeg2 file...the audio is still out of sync. I think I am still doing something wrong because you get good audio sync with PowerVCR but I can't. I will try using your registry to see if it cures the problem. I also made sure the audio delay is set correctly when putting in the audio parameters.

To Dolemite,

I will also try your program out.

[Swan]

Even when I capture a 30 minute clip and get just one mpeg2 file...the audio is still out of sync.
Well then, the Mpeg-2 files PowerVCR outputs must be badly encoded.
They are, I know they are. Running them through Mpeg-2 analysis tools reveals that the time codes are off. I get by, because I edit the Mpeg-2 files. And, when that doesn't work, I encode the trimmed Mpeg-2 files separately (making a DivX file out of each one).

Increasing the maximum recording time (to prevent PowerVCR from chunking files) won't help you.
I can only suggest one thing: try Intervideo's WinDVR. It's the only Mpeg-2 capturing software I know of, besides PowerVCR, that has a scheduler.

/Swan