AVI to DVD MPEG (CCE or TMPGenc encoders)

[DJRumpy]

lol...I ended up pasting my entire guide as the reply

{*sigh*} Fat fingered the paste key I guess.

Let me try this again...

[DJRumpy]

Phoenix2443:
The difference that the additional bitrate from a mono audio stream will make in a DVD project is negligable. You can safely follow the guide instructions as they are and expect good results. If you still decide you want to compensate for Mono audio, then multiply your audio's current bitrate times 2 (96x2=192), select that bitrate in the CCE audio settings, and then select "Monaural". If you don't know the bitrate (or your audio is wave), then just pick 224.

You should definately resample your audio to 48Khz, either with CCE, or with the ResampleAudio(48000) command in AVISynth. Either method will work just fine.

Jon J:
You caught quite a few week spots in the guide, which I've corrected. There are also a few spots that it appears you just missed. I'll try to answer all of your issues. Just let me know if I missed any.

The guide is intended for current versions of CCE (2.6x). CCE 2.50 is a couple of years out of date, which is why you noticed the differences. I put a note in the guide suggesting a minimum version of CCE.

Regarding your ECL question, you don't need to do anything specific with them in this guide. Just ensure your CCE Multipass VBR settings are set to 'Create New File', and not 'Use Existing File', as is shown in the guide. Getting this setting wrong will generate the VAF error your seeing.

You also forgot to include either the ConvertToYUY2 or ConvertToRGB commands at the end of your script. These would eliminate the YV12 and YUY2 errors you were getting. You'll see that step just before the "Encoding your MPEG" step. I am curious that you received a YUY2 error though. CCE supports the YUY2 colorspace, although it will generate a YV12 error if you don't include the ConvertToYUY2() command at the end of your script.

The MPV file is just an M2V (MPEG Video Stream) file with a different extension. You can change that in the CCE options to produce an M2V file instead of an MPV file. Thats just a personal preference of the output filename.

The alternate method mentioned in the guide was actually placed in my MPEG to DIVX guide, and not this one. Senior moment on my part I guess. I've also placed it here. Just check the section describing how to extract your audio from the AVI. You'll see a new alternate option (#4) for handling VBR MP3.

To see what kind of skew you have in you audio, play the MPA/MP2 file and look at the duration. Then do the same for the MPV/M2V file and look at hte duration. Let me know what each was. Just to verify, your AVI was only in one part right? Try the alternate VBR MP3 method to see if that helps to alleviate your problem, and let me know how it goes!

[Jon_J]

DJRumpy, Thank you for your quick reply.
Video is only one episode. (There is only one AVI file)

I played each file seperatly in media player classic. They both displayed the durations.
movie.mpa: 42:40
movie.m2v: 42:40

I tried pausing and starting playing at 6 minutes into the video. (To check speech sync). That didn't seem to work, so I just started the video from the beginning and waited till there were close-up conversations.
The audio is still out of sync, (though the sync seems closer than it was on my first encode). The conversations are lagging behind by about 1 to 3 seconds.

This is off topic:
Earlier I used a small free proggy called Divx To DVD. It seemed to keep the picture/audio in sync,.
But the picture quality is a whole lot better with the CCE encoded video. I figured this would be the case, and is why I'd rather use CCE.

Thanks again,
Jon

[DJRumpy]

Try playing just the AVSynth script in Media Player and see if the audio is in sync.

[Jon_J]

I played the AVSynth script in media player and the audio is just slightly out of sync. Not near as much as it is when playing the m2v file.
I also played the original AVI file to check it. It seems OK. The persons lips match the audio. I thought it may have been a rescource problem with my PC, because I had eMule running in the background. (this proggy sometimes freezes the taskbar). I rebooted and played the AVISynth script again, and it showed the same sync mismatch. Also when playing the AVS file, there are some places where the picture pixalates badly. (this only occurs when playing the AVS script)
My computer should be capable. It's an Intel P4 3Ghz with 1GB ram. 120GB SATA HDD. (Just defragged earlier). Win XP Home edition with sp2
Could the original video, (AVI), be bad and causing this? It plays just great on the computer. These are episodes of a TV program, supposedly taken in HDTV format. The local station has it's HDTV logo on it at the beginning of each episode.

These episodes are 346MB to 349MB each. If all else fails I can archive all 22 of them onto 2 DVDRs Too bad someone doesn't make a standalone DVD player that would play AVI files on a CDR, or DVDR. I can fit 11 - 12 of these on one disc in that format.
Thank you,
Jon

[DJRumpy]

Neither the speed of your computer, or the drive fragmentation, has any effect on audio sync. If the AVISynth script plays out of sync, then the audio stream you stripped from the AVI has corruption in it (in other words, the embedded MP3 in the AVI has bad frames). You can try to fix it by offsetting the audio, but it can be tricky. You use the DelayAudio command to bump the audio back or forward. You can specify a negative or a positive number. You can also use decimal values like 1.2, -.7, .3, etc.

DelayAudio(-1.3)

Just put that somewhere after your audiodub command, and play with the number value.

Before you do all of that though, try the method #4 I mentioned earlier. Just note that when you use the EnsureMP3VBRSync() command, that seeking is very slow, but it should play without issue in media player.

If the audio slowly strays more and more into the video (the end is more out of sync then the beginning), the you'll have to splice the video into multiple parts using the TRIM command, and then delaying the audio for each TRIM'ed section. Look at my MPEG to DivX guide for a better description and examples on how to use the TRIM command. It also has examples of the DelayAudio command.

The pixelated video is not due to AVISynth. It's in your source. AVISynth is just playing back the AVI as is, it is not recompressing it. It's no different then playing the AVI itself in Media Player (albeit without the resizing). If you see pixelation, then it's in your source AVI. It may just be more noticable because your enlarging your video. You can try using Half D1 (352x480) instead of full D1 (720x480). It won't magnify defects as much with the smaller resolution.

[midnightsun]

I would just like to point out that the command ConvertToRGB in Avisynth will convert to RGB32 which apparently is of no use for tmpgenc (every time I've tried it I've always gotten a black picture instead of the actual movie). In order to have RGB24 input (which is the colorspace tmpgenc works in) for tmpgenc the command ConvertToRGB24 is needed. Checking out the Avisynth documentation, it reads that ConvertToRGB will convert to RGB24 only if the argument of the function is RGB24 already, i.e. you will have to use ConvertToRGB24 (if the source is RGB24 evidently there's no need to convert it to RGB24!).

ConvertToRGB however works fine with virtualdub(mod).

[Phoenix2443]

when I resample the audio to 48000 with the encoder the audio ends up as chipmunk vocals if you know what I mean. The time length isn't right either, it's reduced to something like 50 minutes.

[DJRumpy]

Changing the frequency should have no effect on the playback speed or duration of the audio. Just use the ResampleAudio(48000) command in the AVISynth script, per the guide. If your audio is being sped up, it's due to something else. I would need to know the specifics of your conversion process.

[Jon_J]

Thank you for all your help DJRumpy.
I did place the command: EnsureVBRMP3Sync() into my AVS file and am still having out of sync audio.
I didn't realize just how many errors that these AVI files have until I used Divx to DVD program.
Divx to DVD saves a log with each conversion. I have done four conversions so far, (with Divx to DVD), and all four logs have numerous errors such as:
"video packet dropped due to error on frame xxxxx"
The first episode, (the one I'm having audio sync probs with), has 106 of these above errors.
The second episode has 88 of the above errors.
The third and fourth episodes also have as many errors as above.
The fourth episode also has eleven of these errors:
"MSG Level #0 - header damaged"

I was told these AVIs were excelent and played very nicely. (The AVI files look pretty good on a PC in media player).
I used to rip DVDs and convert to SVCD, several years ago. But my source was known and of commercial quality.
I don't think I'm going to try to convert all 22 episodes with CCE, since the source quality is unknown.
The ones I converted with Divx to DVD are acceptable and are as good, if not better quality than regular broadcast TV using a rooftop antenna. And these have little, or, no noticable audio sync problems.

Again, I thank you for all your help, and when I get some "Clean" video files to convert, I'll use this guide to the letter. I have saved the webpage, (with the guide on it), so I can use it when I'm offline.
Jon

[DJRumpy]

No worries. Just keep in mind that you can fix the audio, but it takes a bit of work. To do so, you split your input video up into virtual sections, and then adjust the audio for each section like so:

vid=avisource("input.avi")
aud=wavsource("audio.wav")
audiodub(vid,aud)
clip1=Trim(0,4999).DelayAudio(1.3)
clip2=Trim(5000,9999).DelayAudio(.7)
clip3=Trim(10000,14999).DelayAudio(.4)
clip4=Trim(15000,19999).DelayAudio(.1)
clip4=Trim(20000,0).DelayAudio(.1)
clip1++clip2++clip3++clip4

You get the idea. you use the Trim command to create virtual clips of your input video, and tack on the .DelayAudio at the end of that line to adjust the audio for each section. To test each section, simply remark out the last line ( clip1++clip2++clip3++clip4 ) by placing a pound sign (#) in front of it, and then put the variable assigned to the clip your currently testing. For example:

vid=avisource("input.avi")
aud=wavsource("audio.wav")
audiodub(vid,aud)
clip1=Trim(0,4999).DelayAudio(1.3)
clip2=Trim(5000,9999).DelayAudio(.7)
clip3=Trim(10000,14999).DelayAudio(.4)
clip4=Trim(15000,19999).DelayAudio(.1)
clip4=Trim(20000,0).DelayAudio(.1)
#clip1++clip2++clip3++clip4
clip2

This script would return only the video assigned to the clip2 variable (frames 10000 through frame 14999). You would tweak the delayaudio command on that line until you get it sounding good at both the beginning of the clip and the end, and then move on to the next section. When your done, just remove your ClipX variable from the end of the script, and unremark your line that joins all of your clips together. If you find that the audio wanders too much even in a small clip like the ones above, then you'll need to break each clip into even smaller numbers of frames (say 1000 frames per clipX).

Good luck!

[Jon_J]

Thank you. That does make a lot of sense, and after I read your reply twice I got it fully. You do a very good job putting these tasks into words that neophytes like me can understand. (Were you an instructer at one time?)

Anyway, I've been saving all your instructions and when I get some more time, I'll try to repair some of these episodes using the above method to fix any "out of sync" problems these episodes have.

Thanks again,
Jon

[apalace]

First I thank you DJRumpy very much for this great guide..

I am newbie, and I followed your guide , I wasted really many hours, but the result is a .mpv file , and the video in it plays before the Audio .AC3, like you call it "out of sync" , when I test playing the video and the extracted audio (ac3) together, I don't notice this problem at the beginning, but minutes after minutes I notice that very clearly and the video plays much earlier than the audio does... I'll give all the infos about the avi file I converted and what I noticed that it is maybe the reason of this problem:

- The video was 2 avi files (xvid) , I joined them with virtualdubmod and removed (Cut) a short part from it, then I extracted the ac3 audio from it.

- I loaded the joined Avi file with GSpot :
===VIDEO===
Runtime = 02:13:05 (191,459 fr)
x:y = 608x256 (2.38:1) [=19:8]
bitrate = 1007 kb/s
FPS = 23.976
Qf = 0.270 bits/pixel
===AUDIO===
Name = ac3 (0x2000) Dolby Laboratories, Inc
Bitrate = 448 kb/s (5 ch) CBR
Fs = 48000 Hz
===========
This was what I got in GSpot..

- I fixed the AC3 Audio file wich I extracted from the final joined avi file, using AC3FIX.EXE and yes there was error.

- I created my AviSynth script, and the code is the following:

AVISource("video.avi",false)
LanczosResize(720,360) 
AddBorders(0,60,0,60)
ConvertToYUY2()

I have a feel that there is something in the code that doesn't match my avi's bitrate, but I am not sure... I used LanczosResize(720,360)
AddBorders(0,60,0,60) and you said in your guide that it is for AVI which use a 2.35 aspect ratio, but my AVI has 2.38 aspect ratio as I posted it above. This is the first thing I thought that it is behind the issue.

- I loaded the AviSynth script in VirtualDub and saw the AVI Information, it was right like what you show in your guide, and I played the video to be sure that it works ok.

- I used the Bitrate Calculator to check for the calculated Bitrate :
Video length = 2 Hours 13 Mins 5 Secs
DVD 1 x 4.37(DVD5) GB
Audio Bitrate 448 kbit/s
Type DVD
Calculated Bitrate 4124 kbit/s
DVD Max Bitrate 9341 kbit/s

- I opened the CCE SP programm, and loaded the AviSynth file in it..
Here in this step I figured two things which I think maybe they are behind the problem.. First thing is that after I droped the avisynth file in CCE, it showed the Duration time different than the original time I figured out from GSpot, in GSpot says = 02:13:05 as I posted above, but in CCE it says = 02:12:57:11
The other thing is that when you click on VIDEO button in CCE, there is a "Frame rate" which we can change, you let it in your Guide : 24 not 23.976 like the source... And this is why I changed it too in CCE from 23.976 to 24.. And everything else in CCE I did it exactly like what you said and showed in screenshots, the "Bit Rate : 24" too.

Now after all this explanaition, what do you think that is behind the issue ? I mean the Audio is out of sync ... The Video plays earlier than Audio and it goes earlier and earlier each time ??

I want let you know that I tested the AviSynth file and loaded the AC3 Audio file with it too using BSplayer and looks ok without this problem... But when I open the DVD file , I mean the mpv file after encoding the AVI file, and load the AC3 file with it I get this Out of sync problem .

I am really tired about that and hoped that after many hours the result comes perfect, but no it didn't.

Any idea please to correct the file ? What is behind this problem ?

Thanks alot for any help!

[DJRumpy]

Sorr for the delay in getting back to you. I tend to be a bit busy on the weekends

I can't believe this bug in the guide has slipped past me and everyone else all this time.
The video rate indeed should be set to match your video. It should be set to 23.976, not 24 fps. Change that and your auido sync issue should go away. I've updated the guide with the correct information. Just give me a shout if you have any more problems, and I'll be happy to help.

[DJRumpy]

midnightsun, I didn't see your post up there. You are correct, and I've updated the guide as a result, however, VirtualDub will read YUY2 input, and VirtualDubMod will read YV12 output, so it does not require the colorspace to be changed. CCE requires YUY2 or RGB, however, I would convert it to YUY2 which is faster.

[apalace]

Thanks alot DJRumpy for your answer.

I encoded the video again but this time I changed the bitrate to 23.976 but the problem about the sound out of sync still there

Is that maybe a bug of CCE ?

Thanks for any help!

[DJRumpy]

No, it's most likely just bad frames in your AC3 files. You can try just opening your AC3 audio in WinDVD or another player that will play ac3 audio, and check the duration. Then check the duration of your joined AVI and see if they match.

[apalace]

OK here are new news about all of this ...

When I wrote my last reply above, the TMPGen programm was still encoding, and it finished hour ago ...

The file is working perfect , there is no such a "out of sync" problem, which means that it is really CEE problem... But the worst thing about TMPGen is tht it takes long time to finish, it took 5 hours and 30 mins to encode this avi file to DVD , but CEE took only 2 hours... I would like to use CEE to save times but if it gives this problem I'll rather stuck with TMPGen untill I find a programm that works friendly with AviSynth and can encode avi's to DVDs like CEE and TMPGen.

And yes, the problem really not from my Audio AC3 file, I wrote in my first reply that I tested the AVS script file with the AC3 file together and there was no such a problem. If you can figure out why the problem happened by using CEE I'll be thankful because I find it better and faster than TMPGen but only the "out of sync" problem makes it bad.

Last question about AviSynth , I want to use it many times to remove parts from the Avi files, by creating the script and load it in VirtualDub, but is there a code allows me to save the edited file in it's original file ? I mean not to change the quality and let the codec exactly like the original avi ?
As example : I tried to cut something from the Avi file , I wrote the code for it and saved it as AVS file and loaded it in VirtualDub, but when I wanted to save the file as AVI in VirtualDub, the file has ben saved as Uncompressed and was in huge size, over 7 GB, and I don't want to use any codec on it from the VirtualDub codec list, even if I check "Direct stream Copy" the saved file will be uncompressed if I use AviSynth script.. And my question is if there is a code available and I can add in the script which let the saved avi file be the same like the original one except what I removed from it ?

Thanks again for your help and great Guide!

[DJRumpy]

How did you play the AC3 file with AVISynth? I didn't show you how to use DirectShowSource in this guide. Your also forgetting that CCE doesn't do anything with an AC3 audio source. That part is done outside of the encoding process and muxed back in later. Did you convert your audio to WAV?

Regarding your other problem, VDub saves the file as uncompressed because the avi is decoded when it's frameserved to VDub. VDub no longer has the original codec information, so it can only save it as uncompressed. The only way to do what your trying might be to use the cut/splice functions in VDub. Really not a topic for this guide though. Try posting your question in the main forum for that one.

[apalace]

I loaded the AVS script and the AC3 file with a nice programm called BSplayer, this programm is for testing the Video with an external audio files ( like AC3 ) and subtitles too. It can play AVS scripts too. This is what I did. And no I didn't convert it to wav, it is still AC3.

About why I didn't use the Cut function in VDub itselves, because the part I wanted to cut was at the beginning of the Video file and it is a short part of a scene, if I was trying to cut this short part and save the file, I notice that the part hasn't been cut and the video starts from the beginning of the scene and ignores what I removed in VDub. This is why I am searching for a good code that can be used in AviSynth beside using the "Trim" code to save my file without changing anything of the original file.

Thanks for your Reply!

[DJRumpy]

If you use AVISynth, you'll have to re-encode. Only by physically slicing the AVI can you avoid re-encoding.

Does the output M2V duration match the input AVS duration? Also, what version of CCE are you using? Also, can you drop your AVS script onto VirtualDub, and tell me what Framerate it reports?

[apalace]

>Does the output M2V duration match the input AVS duration?

Do you mean the output of CEE or TMPGen ? If you mean TMPGen I got a MPG output file not M2V, and yes as what I saw they match, I opened the MPG (DVD) output file using the BSplayer and the duration time says : 2:13:5 , then I opened the AVS file using the same player and it showed the same duration time 2:13:5
But if I drop the AVS file in the main menu of CEE I notice that it calculates the duration time little different, it shows 02:12:57:09 I don't know if this is normal or not but I wanted to let you know, maybe it helps to figure out what the wrong is.

>Also, what version of CCE are you using?

The version is SP 2.67.00.27

>Also, can you drop your AVS script onto VirtualDub, and tell me what Framerate it reports?

Sure, here are the infos :
- Frame size, fps = 720x480, 23.976 fps (41708 us)
- Length = 191457 frames (2:13:05.36)
- FourCC code = FourCC : [YUY2]
YUV 4:2:2 as for UYVY but with different
- Decompressor = YUV 4:2:2 (YUY2)
- Number of key frames = 191457
- Min/avg/max/total key frame size = 691200/691200/691200 (129233475K)
- Min/avg/max/total delta frame size = (no delta frames)
- Data rate = 132578 kbps (0.00% overhead)

This are all infos , I am sure that you don't need them all, but I wrote them all just to be sure that I didn't mess something you need to know.

Thanks for your help!

[DJRumpy]

No ideas. If VDub, and TMPGenc are both reporting a different duration than CCE, either CCE doesn't like your source, or there is a bug in your build of CCE. I would report the issue to the CCE support forum and see if anyone has any ideas. I haven't experienced the problem on 2.27.00.29, and I've been using it for many months.

[apalace]

If you use different version, I would really like to test it and see if it gives the same duration or the correct one.
Is there any possibility to download the version like what you own ?

Thanks for your help!

[midnightsun]

midnightsun, I didn't see your post up there. You are correct, and I've updated the guide as a result, however, VirtualDub will read YUY2 input, and VirtualDubMod will read YV12 output, so it does not require the colorspace to be changed. CCE requires YUY2 or RGB, however, I would convert it to YUY2 which is faster.

No problem about the late reply. I see you're being very busy with this thread

As for the RGB32 thing, I just meant that RGB32 can be read by virtualdub(mod) in case anyone had a source in that colorspace (unlikely but you never know...) and didn't know what to do with it . Of course YUY2 is preferred

[DJRumpy]

midnightsun: No worries. I modified the guide regardless just to make that point a bit clearer

apalace: You should be able to get updates from the vendors website. Good luck!

[apalace]

DJRumpy, mine is newer than your version, that mean I have to search for this old version and uninstall mine to install the old one, the one you have.

[Jonas_Vedder]

I must say first of all, that this is a great guide.
But yesterday I had a problem when encoding a video .avi to .mpv.
These were the specs of the video seeing in GSpot:

video:
aspect ratio: 544x304
frames/sec: 29.971
audio:
0x0055(MP3, ISO) MPEG-1 Layer 3
48000Hz 133 kb/s total (2 chnls) LAME3.89

So I followed this guide, and my avisynth script went like this:

clip=AVISource("C:\movie.avi", False)
aud=WAVSource("C:\audio.wav")
AudioDub(clip,aud)
ResampleAudio(48000)
LanczosResize(720,576)
AssumeFPS(25, True)
ConvertToYUY2()

Like you can see I converted from NTSC to PAL with CCE, exactly the way this guide tells to, and when I had the movie encoded, the audio (audio.mpa) was in synchro with the movie, but when I compared the audio from the original .avi, i noticed that the audio from the encoded file was delayed..
Why did this happen and what do I have to do to solve this problem? I think this is because the original .avi was 29.971fps and I converted it to 25fps. I've followed this guide the same way when encoding 23.976fps movies and I didn't have any problem..this is the first time I encoded a 29.971fps movie..
Can you please help?

[DJRumpy]

apalace: Sorry about that. I had a typo in my reply. My version is 2.67.00.29

Not 2.27.00.29

Jonas_Vedder: When you convert from PAL to NTSC or vice versa, your changing the duration of both your audio and video. In your case, your slowing it down. I should warn you though, that slowing down your video from 29.97 fps to 25 fps is very noticable. The audio will sound very slow and low. You typically do not want to do this type of conversion. The only exception is when your source is telecined, meaning it has duplicated frames in it to convert it from 23.976 to 29.97 fps. You can remove these duplicate frames to return it to it's original 23.976 frames per second. You should not slow it down using the AssumeFPS method.

Do you know if your source file is telecined?

[Jonas_Vedder]

Thanks for the fast reply DJRumpy!
First of all, I have to say that I understood now..when I'm making that AssumeFPS method from a 29.97 fps to 25 fps I'm slowing down the movie..but for example when converting a 23.976fps to 25fps It's not noticable right? I've done this and I didnt' notice any difference.

Ok..but I don't know if my movie is telecined...I think it's not telecined..is there a way to know that?

But assuming that the movie is telecined..where and how can I remove those duplicate frames to return it to it's original 23.976fps?

Thanks again!