Running time video file problem

[Harm]

I'm only experimenting with authoring for 1 month now and I seem to be totally stuck!

Let me tell you a story:
I read all kinds of tutorials on how to do it (DIF4U, Reauthorist, DOCCE4U, Scenarist) but kept on getting problems in the encoding phase.
At first CCE hung at the end of each pass (albeit one or multipass). I changed the mpeg2dec3.dll for that to get CCE stop (or sometimes hung) at the end of the first pass of the multipass phase (so the initial VAF pass would be OK). I tried BatchCCEWS, another version of CCE (Trial) to no avail.
Then I tried other encoders (like Mainconcept) to find it would encode fine but not to the full running time of the videofile.
I already had installed Elecard MPG player so I played the videofiles in question to find out that the running time would be to short..... but all the material is there! The player runs on after total time is reached! This also happens in Windows Media Player and JPlayer. If I play the file by AVS and Media Player the time is right.
At first I thought maybe the DVD Decrypter phase would be at fault but the same thing happens with files from another source.
So I tried to run VStrip myself (this is all very educational!).

The end result of that being that if I demux my videofile to a m2v the running time shows up as too short; when I save to a vob running time is OK. But...... if I encode this VOB file to a MPG the same error happens!

So does anyone know why the running time is off after VStrip?
Is this some codec problem?
To give you some idea of the running time differences:
24:02 official --> 14:24 M2V after VStrip or MPG after encoding VOB
11:46 official --> 06:36 same
00:34 official --> 00:12 same

Help! Ofcourse I got the most recent versions of the software involved (At least I think so).

[Harm]

I did another experiment with one of the wrongly timed M2V files. I converted it into AVI (DVD2AVI) and after that encoded it into MPG (Mainconcept). Besides the awful videoresult the running time of both resulting files is OK!
I am guessing AVI creates it's own header?
Anyway maybe someone knows of an alternative for VStrip? Or maybe I should replace the codecs for M2V and MPG?

[manono]

Hi and welcome to the forum-

I don't understand. What do you care if the players show the wrong length, as long as it's all there? Why are you messing with VStrip at all? Get just the vobs from DVDDecrypter (no Stream Processing). Then set up a project file in DVD2AVI and get the audio. Then you'll have both the vob files and the AC3 audio. And don't enable the audio in the encoder.

But like I said, I don't understand.

[Harm]

Thanks for the welcome!

As I want to make a backup of an original DVD using the so called big 3 (DIF4U, CCE, Scenarist) and I can't get succesfully past the CCE part I started investigating why this happens. As the CCE step is needed to encode (and compress) the videofile I started looking to other encoders to find that they encode succesfully but use the shorter running time. Apparently they don't take into account that there is more info! I am also guessing this is the reason why CCE fails to complete succesfully.
As I also want to include an extra subtitle and try and edit the menu I really have to demux, encode and author.
As you can read in my story above I looked at all the steps DIF4U is doing for you. The result being me running VStrip myself to find that after this step the running times are wrong for all videofiles.
(I don't encode audio in these steps)

I just want to understand what is going wrong. Is there some info in the M2V header I can look at? Or is just a codec problem? And why is the MPG file encoded from the VOB also wrong? And why is the AVI length correct? There has to be a reason for this!
Is there a way I can fool the encoders that the specified running time should really be something else?
Questions, questions, questions ;)

[manono]

Hi-

So, you don't want to start off slowly by using DVDShrink or something similar? You just want to jump right into the deep end, eh? What I might suggest is to try DVD2SVCD which allows the use of CCE. Doom9 has a guide:

http://www.doom9.org/mpg/dvd2svcd.htm

It allows for encoding to DVD specs now. When you install it you can tell it that you'll be using it for DVD creation. Maybe do one or two with it, and study the .avs it creates, and the processes it employs, to try and learn where you're going wrong.

And you've seen Doom9's CCE guide?

http://www.doom9.org/mpg/cce_guide.htm

He's got a bunch of CCE guides, as well as other guides (some of which you've seen, I guess):

http://www.doom9.org/doc-overview.htm

But for some of your questions, a DVD expert will have to step in. I've never made one, although I've used CCE probably a hundred times without problems.

[Harm]

Hi to you to!

Directly of the deep end? Well yes and no.
It only took me several days to get the hang of the shrinkers (I personally used InstantCopy several times). But I wasn't totally satisfied with the picture quality of especially the music DVDs.
I also wondered why the IFO contains all these parms so when I stumbled on the doom9 sites (I'm also registered at the Dutch site) and the guides they provide my printer had to make overtime.
Also I wanted to be able to change the look and feel of the subtitles (and even typo's!) of the Dutch subtitles and maybe even add my own. You can guess that my choice to embark on the authoring path was a logical one because I thought I could take care of all these inconveniences in one go!
The DVD in question I have problems with is a multi PGC one with a complicated menu structure within the main movie. All the shrinkers I got (most recommended ones) can't seem to copy this one or do a very bad job. This seems logical after reading one of the guides concerning the multi PGC in one VTS. Just rip the largest one and use the special parm in IFOUpdate to take care of PGC and cells. Works great!

I also stumbled on the DVD2SVCD guide you point out. Maybe this is my ignorance but I don't want to make a SVCD but a DVD-R. Or are they partly the same?

As I said DIF4U is a fantastic tool and Scenarist gives you many possibilities (some are hard to understand I might add) but CCE is still a major problem for me. If it would work I wouldn't contemplate of using other encoders or VStrip or whatever.
Let me elaborate on my CCE problems:
CCE V2.5 seems to encode small files without any problems even using a 5 multipass. Bigger files halted after the first pass as soon as in the ECL the creation of the MPV was ordered. Win XP told me an error had occured in NTDLL.DLL and CCE had to be closed. This also happened with CCE SP 2.67.00.23 (and some of its predecessors).
As these videofiles where PAL interlaced (according to bitrateviewer) I guessed changing the MPEG2DEC3.DLL could make a difference. I downloaded the recommended one from warpenterprises to find now that CCE would just run pass 1 from the VBR multipass and then stop without giving me any information whatsoever. The only giveaway was that it didn't enter an endresultentry in its logfile.
So I was stuck! Can you understand why I felt forced to go in even deeper???? When I stumbled on the running time difference the thought crossed my mind that this could also be the reason of my CCE problems. Or couldn't it?

For your information my current AVS file:
import("F:\Tools\DoItFast4U\new.avs\addaudio.avs")
LoadPlugin("F:\Tools\GordianKnot\mpeg2dec3.dll")
Mpeg2Source("VTS__02_P01.I-BFF.4~3_1.d2v",idct=0)
LoadPlugin("F:\Tools\DoItFast4U\new.avs\decomb.dll")
FieldDeinterlace(blend=false)
AddAudio()
ConvertToYUY2(interlaced=true)

... I keep on typing many lines on each answer..... :D

[manono]

Hi-

I haven't found out much about the ntdll.dll error on this site. Just this:

http://forum.doom9.org/showthread.php?s=&threadid=55519&highlight=error+has+occurred+in+NTDLL.DLL

A Google search turned up this:

http://forum.digital-digest.com/archive/topic/26347-1.html

It seems to have to do with a DirectShow Priority. And make sure you haven't installed any Filter Packs, such as the Nimo Filter Pack.

Maybe this is my ignorance but I don't want to make a SVCD but a DVD-R. Or are they partly the same?

Although the Doom9 guide doesn't mention it, DVD2SVCD is also for DVD creation. Since you have Scenarist, it can go all the way and do the authoring (don't know about your subtitles, though). But as I said earlier, you can tell it that you'll be encoding for DVD, and those will be the priorities and the settings you'll see. It's easy-give it a try. And it will create an .avs that you can steal from later on.

Here's his DVD creation guide using DVD2SVCD:

http://www.doom9.org/mpg/dvd2svcd-dvdr.htm

Looking at your .avs, most all PAL DVDs are not interlaced. BitRate Viewer is wrong. Don't deinterlace unless you're absolutely sure it's really interlaced. Comment out the FieldDeinterlace line and open up the .avs in VDubMod and have a look at the frames. Make sure you have the right version of AviSynth to use the Mpeg2Dec3.dll. It should be 2.5 or better. To find out, make an .avs entitled Version.avs. Inside have just:

Version().

Open that in VDubMod and it'll tell you.

I keep on typing many lines on each answer

Me too. :) But like I said, give DVD2SVCD a try. Can't hurt. You're not getting anywhere now. Then try and learn from it. Good Luck.

[Harm]

I had a look at the 2 links you pasted and I think my NTDLL.DLL problems were different. The error occured at the end of the first pass in a multipass and only when the vid_out parm was turned on.
With my changing of the mpeg2dec3.dll that error disappeared (only in BatchCCEWS I might add) but no additional passes occured.
As I still have to find a complete manual for the syntax of all available ECL parms I'm still struggling with that.
Maybe you could shed some light on the quality of a CCE encode of just 1 pass or maybe I should try a CBR pass instead?
Do I really need this ffdshow? What's it for? And I can't change any priority in CCE for that.

I think I have found something myself for the running time problem. As soon as I create a DVD2AVI project for the file in question and use it in an eligible encoder (TMPGEnc) the running time of the resulting file is fine!
Sadly for example Mainconcepts encoder can't read those D2V files.

I also got a reply of the author of VStrip stating that it changes nothing to the M2V as compared to the VOB and that the running time normally is detected from the IFO. So maybe you always need a DVD2AVI project?

Obviously I wanted to test this on my failing files. Then I found a difference between the files made by DIF4U (which are used via the AVS) and the ones TMPGEnc wants to read. There's a parm in there called Luminance which should read Luminance_Filter.
When I tried CCE again with this change it turned out that this change resulted in an error in the resolution detected by DOCCE4U/BatchCCEWS so no go there (it changed into 1032x52!).
Currently I am running a test with 1 of the failing files with TMPGEnc. Better start looking for an encoding guide for that.
Is there much difference in encoding quality between CCE and TMPGEnc?

I am definitely going to test DVD2SVCD (I already found out after downloading it that it's also DVD2DVD) later! Currently I don't want to leave my track of demuxing with DUF4U, encoding somewhere and authoring with Scenarist too much.

[manono]

Hi-

I can't answer everything, as I don't make DVD-Rs, but I know some of the things about which you asked.

Maybe you could shed some light on the quality of a CCE encode of just 1 pass or maybe I should try a CBR pass instead?

I always run 4 passes (3 plus the vaf creation pass). But some of the real MPEG2 pros run a 1-pass quality based encode:

http://forum.doom9.org/showthread.php?s=&threadid=64886

But a regular 1-pass encode isn't nearly as good as either of those 2 methods.

Do I really need this ffdshow?

Don't really know. It's for decoding .avi's. It's good and very configurable.

There's a parm in there called Luminance which should read Luminance_Filter.

There are many different versions of DVD2AVI floating around. And not all applications can use all of them. This could be the cause of some of your problems. The best is DVD2AVIdg (number 6):

http://forum.doom9.org/showthread.php?s=&threadid=59272

If you use it, do not get your audio from DVDDecrypter, but from DVD2AVIdg. Otherwise you're liable to get audio asynch.

(it changed into 1032x52!).

I've read about that problem before (but I don't think it's 1032, as a search doesn't turn up anything). When CCE crashes or freezes, it doesn't really tell you what's wrong. I think the problem is related to the .avs. Open it in VDubMod. It will often give you an error message that will help pinpoint the problem.

Is there much difference in encoding quality between CCE and TMPGEnc?

Besides being a heck of a lot faster, CCE also produces better quality for MPEG2, I believe.

For more on CCE problems, please see Q 60 of the FAQ:

http://forum.doom9.org/showthread.php?s=&threadid=21859

[Harm]

The threads you point me to are rather complicated......
I didn't know there was such a difference between the DVD2AVIs. This doesn't make it any more easier...... :(

I want to follow the 4 pass CCE rule to. I can run succesfull the first pass (vaf creation) and the first of the other 3. Should the VAF_out creation parm also be on then (as it is now)? If so do you suppose I can rerun it by hand to get the second pass and then again for the third? In other words does the encode-info get stored in the VAF during those passes or does it work in another way?

I ran TMPGEnc in 2 pass VBR with the bitrates DIF4U calculated for me but I'm not that satified by the result. It's more blocky than the one pass of CCE. As you seem to know your way around this forum: is there also a TMPGEnc guide?

And does there also exist a manual for ECL (CCE Encoder Control List)?

I had a brief look at DVD2SVCD (or DVD2DVD in my case); I downloaded the D2SRoba plugin and wanted to do a testrun. It looks like only the main movie is ripped and converted (only 1 IFO is selected). Is this true? If so is it also possible to copy the extras to the same DVD?

And again..... many thanks for your help!

[manono]

Hi-

If so do you suppose I can rerun it by hand to get the second pass and then again for the third? In other words does the encode-info get stored in the VAF during those passes or does it work in another way?

Once you have a vaf file, then you don't have to get it again. You can set it up to do the whole encode, but only the "real" passes are done. But if you're asking if after a first pass your computer or CCE crashes, will it remember the data for that first pass, so you can go to the second pass, then I think the answer is no, it will do all the "real" passes over again.

is there also a TMPGEnc guide?

Yes, of course:

http://www.doom9.org/mpg/tmpg-dvdencoding.htm

And DVDRHelp.com is loaded with guides. Try here (http://www.dvdrhelp.com/tmpgencdvd.htm), here (http://dvd-hq.info/Compression.html), and here (http://www.games-realm.com/encoding/tmpgenc_quick.html).

And does there also exist a manual for ECL (CCE Encoder Control List)?

There's probably something about it in the CCE help. But other than that, I don't know. But when you open it up, it's pretty easy to understand it.

is it also possible to copy the extras to the same DVD?

I'm sure it is, but as I said before that I don't make DVDs. You might have to encode them all separately, and then author them yourself.

[oddyseus]

Just read this interesting thread and I would like to add something if I may.

First let me describe how CCE works for me.
I have 2.64 version running on winXP pro on a P4 1.8
Since May 2002, when I set up this system, CCE never crashes and always manages to perform all the passes I tell her to perform. Usually I do the initial pass to get the vaf file and one more, unless it is a very long film, more than 2.5 hours or a grainy one, where I do 3 passes. The difference in quality is not noticable, at least at my tv set.

My system is not overclocked, although it performs at .7 real time most of the time.

I am not messing around with divx or similar encodes, so my system is clear of their codecs.

I use dvd2avidg for demuxing the audio and producing the d2v project.
For frameserver I am using vfapi instead of avisynth. I know it is outdated but it frameserves perfectly slow to CCE.

I am encoding in a clean system that has nothing but the absolutely necessary for encoding.

This system was ghosted when installed and I am replacing it quite often, having always a clean system to work with.

This is located in a second hard disk, the encoding hd. My applications and stuff r in the main hd of the computer.

Now to your problems.
If u have tried to remove the CCE logo, pls reconsider. Check if the evaluation version, as downloaded, freezes your machine.

If it is, use a second disk, or a partition to install a fresh copy of windows and test there. Install only the necessary tools to test, nothing else.

If so do you suppose I can rerun it by hand to get the second pass and then again for the third? In other words does the encode-info get stored in the VAF during those passes or does it work in another way?
CCE updates the vaf file after each pass. So running the passes one by one in actually the same as running them all together. I have checked various demanding parts of a movie after each pass and saw a few differences.

is it also possible to copy the extras to the same DVD?
Yes it is. Although if u r following the path u r now, u must re-encode all of them and re-author.

[manono]

Thanks oddyseus-

Finally someone that knows what he's talking about when it comes to DVD creation.

CCE updates the vaf file after each pass.

Gee, I was wrong and didn't know that. Learn something new every day.

Although if u r following the path u r now, u must re-encode all of them and re-author.

At least I got that part right. Thanks for stepping in. Maybe you can keep an eye on this thread until Harm gets things sorted out. I'm sweating bullets here. :)

[Harm]

Well I'm glad with the input of you both!

Just an update on what I'm doing/trying:
- I am using mpeg2dec3.dll of Gordian Knot (latest version) now to encode the M2V to MPV
- I've added parm ResampleAudio(44100) to all AVS files
- In one of the links Manono gave me someone talked about raising the clustersize from 4K to 8K; so I did (Partition Magic crashed in the process and it took me several hours to repair the partition, move most of the data to another partition and running PM again followed by another move of data......pffeeew)
- I took Reauthorist out of the sequence and let DIF4U start up DOCCE4U right away.
- I am not concentrating on the running time anymore because DIF4U uses DVD2AVI to create a project which in turn is used as input for the encoding process; so the running time shouldn't be the problem
- I didn't bother with interlace or not thinking that this couldn't be the reason for my missing multipass

Than I ran the whole process from scratch. And to no avail. I still got CCE stopping (not crashing) after pass 1 in the VBR multipass. I don't know if I already told you so but this only happens with the larger files. I got 5 files to encode.
This is the CCEDATA.TXT file (I've replaced the real filename):
"AVSname1",0,7880,1536,0,4
"AVSname2",0,8072,1536,0,4
"AVSname3",0,9800,1850,0,5
"AVSname4",0,9800,1850,0,5
"AVSname5",0,9800,1850,0,5
Sizes of the inputfiles:
M2V1 - 3,58 Gb / M2V2 - 823 Mb / M2V3 - 378 Mb / M2V4 - 288 Mb / M2V5 - 475 Mb
As I already said the results of M2V3-5 have always been the same: succesfull encode with 5 VBR multipass.
M2V1 and 2 run the initial pass and just 1 of the supposed multipass.
I don't know if it is relevant but the first 2 files have chapters (although I think these are irrelevant in this phase)
Here is the ECL which is generated for M2V2:

; Cinema Craft Encoder SP -- Encoder Control List

[item]
title=VTS__02_P01.I-BFF.4~3_1
vid_out=1
vaf_out=1
aud_out=0
vid_file0=MPVname2
vid_file1=MPVname2
vaf_file=VAFname2
aud_file=MPAname2
file_focused=0
encode_mode=0
packet_size=2048
timecode=0x0
width=720
height=576
vmode=2
frame_rate_idx=3
cbr_brate=1536
opv_q_factor=40
vbr_bias=30
opv_brate_min=0
opv_brate_max=8072
vbr_brate_avg=1536
vbr_brate_min=0
vbr_brate_max=8072
vbr_pass=4
quality_prec=16
use_filter=0
filter_val=6
seq_endcode=1
dvd=0
half_width=0
half_height=0
fast_mode=0
progressive=1
alternate_scan=0
non_linear=1
top_first=0
lum_level=1
intra_dc_prec=2
aspect_ratio=2
gop_m=3
gop_nm=4
gop_hdr=12
seq_hdr=1
all_closed_gop=1
fix_gop_length=0
mpeg1=0
mpeg1_cps=0
samples_per_sec=44100
stereo=2
brate_idx=7
crc=1

[file]
name=AVSname2
type=0
frame_first=36064
frame_last=36064
encode_first=0
encode_last=36064
chapter=192:
chapter=7100:
chapter=7278:
chapter=14235:
chapter=14415:
chapter=21335:
chapter=21527:
chapter=28460:
chapter=28640:

I've colored the parms of which I'm wondering red.
1. Should that CBR_RATE parm be there while I'm running a VBR multipass?
2. As you can see there is 4 multipass requested
3. I've noticed frame_first isn't 0; I remember rerunning one of my earlier test with 0 with the same result; the same goes for missing of chapter=0 or leaving all the chapters out

But I think the ECL parms are not that self-explanatory!
Maybe you could have a look if you see something strange?
And maybe you can enlighten me about exactly what software is used by CCE to encode? And also which codecs? So I can investigate them and maybe downgrade them with older versions.

As I refuse to give in what I'm gonna do is:
1. Reinstall CCE this time with the logo and rerun
2. Run the multipasses myself by hand and see if the picture gets better (not that I don't believe you Oddyseus but I'd like to see it with my own eyes)
3. Read up on the links about TMPGEnc and try it that way

I think reinstalling XP on a clean disc is a bit to far for me (just yet).

Keep your eyes on this space! If others can do this, so can I!!!!!

[oddyseus]

I am indeed very glad when I read u writing
As I refuse to give in what I'm gonna do is: ...

Don't worry, we will get to the bottom of this. This is a doom9 forum.

Some questions that I thought during the day.
I trust that u r in a NTFS enviroment. I see nothing to substantiate that u r working with a NTFS disk. Nevermind u probably r.

Here is an ecl file of one of my encodes.
The ecl file was produced by CCE 2.64

title=adder-2_d2v_vfapi.avi
vid_out=1
vaf_out=1
aud_out=0
aud_mode=0
vid_file0=C:\Name1.mpv
vid_file1=C:\Name1.mpv
vaf_file=C:\Name1.vaf
aud_file=C:\Name1.mpa
file_focused=0
video_type=2
packet_size=2048
timecode=0x0
width=720
height=576
frame_rate_idx=3
cbr_brate=6000
opv_q_factor=75
opv_brate_min=0
opv_brate_max=9800
vbr_brate_avg=6000
vbr_brate_min=2000
vbr_brate_max=9000
vbr_bias=30
vbr_pass=3
quality_prec=5
use_filter=0
filter_val=6
seq_endcode=1
dvd=1
half_width=0
half_height=0
drop_frame=0
progressive=0
alternate_scan=1
non_linear=0
top_first=0
lum_level=1
intra_dc_prec=2
aspect_ratio=2
gop_m=3
gop_nm=4
gop_hdr=12
seq_hdr=1
all_closed_gop=0
fix_gop_length=0
qmat_idx=1
samples_per_sec=0
stereo=2
brate_idx=7
crc=1

[file]
name=D:\1d.Movies\Adder-2\d2v\adder-2_d2v_vfapi.avi
type=0
frame_first=0
frame_last=263526
encode_first=0
encode_last=263526
chapter=0:
chapter=40200:
chapter=85570:
chapter=130706:
chapter=175673:
chapter=219535:

as u can clearly see there r distinct differences that worth investigating.

I want to point out that I never saw an ecl file that had frame_first other than 0. Having the first frame to encode the same with the end frame, strikes me as odd.

Studying the differernces between the two ecl files we can see that
1. There is a command aud_mode=0 since I have disabled the creation of the audio file in CCE.
2. The command video_type=2 is missing from your file. Instead u have a similar one (video_mode=2) and in different place. I realy haven't searched what actually it does, but it is present in every ecl I have ever made and they r a lot of them.
3. The audio command samples_per_sec=0 is also present in your ecl, but with an argument of 44100. Have u disabled audio or not?
4. The main difference ofcourse is the abnormality(?) of frame_first.

Trying to pinpoint what can go wrong in a complicated situation like this is a painful process.

As a first measure for testing purposes is to install the demo version of CCE and try to encode a multi pass vbr no chapters at first. See what happens.

[Harm]

Well you gave me some more problems Oddyseus! ;)
I can imagine your ECL being very different from mine.
You got CCE 2.64 (which is rather old but hey it functions...) and I got CCE 2.5 (for use with DOCCE4U) and CCE 2.67 several subversions (for use with BatchCCEWS). I created an ECL exactly as yours (except for the filenames ofcourse) but both my versions wouldn't read your ECL parms.
So I thought uninstalling 2.67 and installing 2.64.01.02 (is this the same version as yours?) would do the trick.
Your ECL parm is read but execution brings nothing. So I investigated further and as it turns out 2.64 won't take AVS/D2V/M2V combination files as input! You have to have an AVI in there somewhere. This is not the case with 2.67! Also BatchCCEWS won't work together with 2.64
So can you tell me how to bypass this? I think this could be the reason of the different value in video_type

Yes I use NTFS. And I have audio disabled.
I am going to leave 2.64 for now and gonna try 2.5 again. As it turns out my base version of that was already logo less so I'm trying to resolve that to. ECL files of 2.5 also differ from 2.67 so I'd better try to get my software totally in sync with yours otherwise any comparison will go wrong.

Which version of ECLCCE and Avisynth are you using? I got 1.7b and 2.5.4.0

[Harm]

Well I ran the multipass again on file 2 and now with CCE 2.5 real demo version...... and it still won't go after the first one of the multipass.
I am now going to fiddle with the created ECL and run it again directly from CCE. I'll reinstall the demo version of 2.67 to so I can test that one also.

Another question: when you encode you leave your virusscanner and firewall active?

[Harm]

And another update........

I forgot to add this on my last respons: several days ago I tried encoding the first 2500 frames of file2 and then the multipass went fine. But I can't remember exactly in what conditions I ran that one.
Can there be a difference because of that? Is it just the size?

You say that if you run the multi pass as singles pass one after another it will get me the same effect as a multipass. Are you sure? Because when the pass starts the mpv file is gone and is build again! So where is this accumulating info stored? In the VAF? Or should I always have different filenames on vid_file0 and vid_file1?

The ECL file I put in here is a CCE 2.5 version so maybe that is another explanation for the differences?
And before you ask: I applied the 2.5 CRC patch.

So I ran the encode again with these changes:
- frame_first=0 + chapter=0: added --> same result
- removed all chapter lines --> same result

to be continued.....

[Harm]

In the guide "Getting the best out of CCE" I found that CCE 2.50, CCE 2.64.01.10 and CCE 2.66 (and presumably also higher) accept AVI projects (D2V) as encodable input. As I got 2.64.01.02 I downloaded 2.64.01.10 and this one is waiting for further testing. What version do you got Oddyseus?

Anyway more testing results:
- samples_per_sec=0 + intra_dc_prec=-1 (corresponds with Auto) --> same result
- changes to AVS: added ResampleAudio(44100) and changed MPEG2DEC3.DLL from DIF4U (new.avs) version to Gordian Knot version --> same result
- noticed that "Getting the best out of CCE" uses MPEG2DEC.DLL so used that one of GK --> full run but file is much smaller and only contains red blurr at the bottom of the screen; played AVS with Media Player to find out Avisynth gives error: Not an AVS plugin
- downloaded recommended MPEG2DEC.DLL from warpenterprises and changed AVS script accordingly; try to play file -- get error about IDCT parm; removed it (idct=0) and played again; play is OK
- assuming that the VAF file needs to recreated after the last avisynth errorrun so rebuilding with still changed AVS script (via DOCCE4U so other corrections are undone)
- waiting for multipass to start............

[oddyseus]

This is very intersting. Quite info u supply in these posts.

Unfortunately I am still at work, its 4pm in Greece right now.

I will carefully read your posts and get back to u later this evening.

[Harm]

I can't believe my eyes...... full multipass run!
Resulting file is slightly smaller than the first test. Although I think the video quality is somewhat less.
So what's exactly the difference between those dll's? It seems like the MPEG2DEC.DLL version is older looking at the filedate.

I also read something about VFAPI converter or is this only needed when AVS can't be read by the encoder?

Gonna install CCE 2.67 again and test it there!

[Harm]

More info from what seems to be a giant test site..... :D

I succesfully encoded file2 with CCE2.5 although I think sometimes CCE doesn't always run all its multipasses (e.g. I tell it to do 4 and sometimes I suspect it by looking at the running times it has only done 2 or 3 -- I only wait for it to pass into the second one) but......
encoding file1 (the big one) hasn't been succesfull yet!
The first multipass run just plainly stopped at 25% with a CPU% of 100% for hours; a rerun also stopped at about 7% (again CPU 100% for a long time); several reruns (multipass of 1) with some different parms (from "Getting the best out of CCE" and "RoBa CCE encoding") gave me normal ending but with different to small filesizes; when I played these resulting files there was always the same type of Avisynth message to be seen: "caught an access violation ...... write error... to 0x0ffb00fe); when I tried to play the same part from the input file I wouldn't get an avisynth error; then I tried running it with "real" trialversion 2.67.0.23 which is rather difficult because this version won't let you load/save ECL files; on the VAF only creation phase I got some kind of weird error shown by CCE after over 60000 frames: "CCE encoding failed: OPV VBV ovf frame#..........." which, according to the net, could be because of some kind of buffer overflow.
So I changed the mpeg dll in AVS script again to the mpeg2dec3.dll belonging to avisynth and now I got a good VAF. As we speak I'm embarking on the multipass of 1 using that VAF as input.....

So more questions: could these hangs and/or write errors be resulting from me using the 8 Kb clustersize partition?
Or should I be encoding on a partition (or maybe a separate HD) where CCE and the other involved tools aren't located?
Is there a way to check the integrity of the M2V input files or the D2V files for that matter (like Avidefreezer)?

I try to leave the encoding process alone (I got another PC.... ;) )
This is becoming a very long stretch......

[Harm]

Next results for file1 (the big one):
- ran a multipass of 1 with version 2.67 by adding the file to CCE --> OK
- ran a multipass of 1 with version 2.67b by starting ECLCCE from DOS prompt (you can save the ECL changes then) --> OK
- ran a multipass of 3 with version 2.67b (ECLCCE) --> just 1 pass and then nothing
- did this again with the same result
- going to run multipass of 3 with version 2.67.........

Still wondering why this difference between MPEG2DEC3.DLL and MPEG2DEC.DLL?
And are you sure if you do several runs after another then eventually I get the same results as a multipass of 4?

I see someone as editted my post........ :D

[Harm]

OK these are the last testresults (until someone points me to something I could test):
- multipass of 3 with version 2.67 and ECLCCE --> just 1 pass

My conclusion for now is that either the problem is in the MPEGxxx.DLL or there is something wrong in my system itself.

So maybe anyone who reads this can tell me with what MPEG2DEC3.DLL they run there CCE encodes? My last ones where with the recommended one from www.avisynth.org/warpenterprises

[Harm]

Are you still investigating this problem Oddyseus?
It's been a while since you replied in here.

Or should I better open up a new thread purely concerning the problem why the multipass won't work?

[oddyseus]

I am still with you Harm. I am monitoring this thread closely. I can't be of more help since I don;t use avisynth and any version of mpeg2decX.dll at all. I am using vfapi and it suits me at all levels.

A nice summation and a new thread about this problem is a good idea indeed.

[manono]

Hi-

If you're creating your DVD2AVI project file with DVD2AVIdg, then you should also be using the version of the MPEG2DEC3.dll that's included with DVD2AVIdg. I don't think that's your problem, but it'll prevent other problems from occuring.

All of the MPEG2DEC3.dlls are for use with AViSynth 2.50 and later. The regular MPEG2DEC.dll is for the earlier versions of AviSynth.

[Harm]

Originally posted by oddyseus
A nice summation and a new thread about this problem is a good idea indeed.
Alright I'll do that Oddyseus........

I guess there is a guide about using VFAPI to?

[oddyseus]

Yes there is.

You can check on its faq (http://www.doom9.org/mpg/vfp-faq.htm), Basic Vfapi Guide (http://www.doom9.org/sbc-basicsetup.htm), Vfapi Installation notes (http://www.doom9.org/install.htm) and the vfapi guide (http://www.doom9.org/mpg/vfapi_reader_codec.htm).

All the above u can find at the Guides section of the site (http://www.doom9.org/doc-overview.htm)