DVD-9 to KDVD DVD-R Using Ifoedit Remux?

Pied-Piper · #1 12-24-2003, 09:53 AM

Hi,

I always had very good results with the latest MA script when converting DVD to SVCD, but I wonder if that script is necessary while the project is a DVD-9 to DVD-R,
considering that
the loading of all those avisynth plug-ins doubles the encoding time.

let me explain the project;
To keep all the special features and original menus, IfoEdit does a great job,
At First I make a d2v of the main movie(Forced Film), MovieStacker predicts the final file size is just fine when KDVD Full-D1 NTSC Template is used.
here I realy get confused: is the default "Detect scene change" ok or should I force TMPGEnc to use the I-Frames template that IfoEdit makes for it?
Anyway, I load that template, and start the encoding and stop it after a few minutes to do a comparison;
Another time I load the d2v without the MA script and let it to proceed for the same amount of the movie length and then stop it.
I see no visible difference between the two m2v except in file size.
The first one is smaller!
The final m2v is to be remuxed by IfoEdit . What sholud I do? Please help.

Thanks.

Dialhot · #2 12-24-2003, 10:00 AM

I never see any file size prediction in your steps. Is that mean that you do not use CQ mode ?

If you use it and do not do prediction: that's normal to have a bigger file size as you probably encode to a CQ that is too hight for the job.

As you tell also "Moviestacker predict the final size" (sic !), I'm quite sure that you do not use CQ mode !

Pied-Piper · #3 12-24-2003, 10:44 AM

Quote:

Originally Posted by Dialhot

I never see any file size prediction in your steps. Is that mean that you do not use CQ mode ?

If you use it and do not do prediction: that's normal to have a bigger file size as you probably encode to a CQ that is too hight for the job.

As you tell also "Moviestacker predict the final size" (sic !), I'm quite sure that you do not use CQ mode !

I use Red-M's prediction formula

http://www.kvcd.net/guides/kguide/kvcdguide.htm),
The number of frames in the movie / 100 / 23.976 * 0.98 * the sample file size

the CQ number 80,
and the same KDVD Full-D1 template

with MA script: smaller size and double encoding time
without MA script: bigger size

But no visible difference between the two.

Is it worth using the MA script when doing such a DVD to DVD-R project that does not require so much compression?

Dialhot · #4 12-24-2003, 01:02 PM

Quote:

Originally Posted by Pied-Piper

The number of frames in the movie / 100 / 23.976 * 0.98 * the sample file size

Tahts is the formula to apply to find the final size after having encoding a sample with a given CQ, and to see how you must adjust tje CQ for the next sample. And so on until reach the perfect CQ : that what we call "file size prediction".

Quote:

the CQ number 80,

As you can see before, 80 isn't the correct CQ, but certainly the first CQ used for the first sample test. But the final CQ is surely different than this (and it's rarely a integer).

Quote:

with MA script: smaller size
without MA script: bigger size and double encoding time

Are you sure that the MA script you returns the same picture resolution than without ?
If you DOUBLE the encodign time without the script, that's surely because the picture isn't the same size.

Or did you make a mistake and inverted the two lines ? It seems more logical to have a longer time with a script than without.

Quote:

But no visible difference between the two.

Is it worth using the MA script when doing such a DVD to DVD-R project that does not require so much compression?

Below 10% of compression needed, use recoding methods (forget the one included in IfoEdit, tools like DVDShrink are more convenient and easy to use). Above 10%, I personnaly think reencoding in KVCD is better. And then, the MA script give a more better picture.

Pied-Piper · #5 12-24-2003, 03:08 PM

Thanks for your quick replies!

You're right, it was a mistake!

In fact it takes 20 hours to encode with the script and the encoding without using the script is done in less than 7-8 hours!

You see there is a huge difference in time! I wanna know is it worth it?

The CQ 80 and below it is perfect because the formula prediction says if I use that CQ the final file size is not bigger than 2.4 GB that is the size I need.

I extracted some frames from the 2 m2v and I see the picture with the script is blurred and not as sharp as the one without the script but it's not so important. When the filters are not applied the picture is more like the original DVD.
If I drop STMedianFilter and Temporal Smoother, the encoding time becomes much less! I am not very familiar with these filters, how do these two filters effect on the picture quality? does the script make the final file size smaller?

The main movie is 5.5GB and there is 1.83 GB Deleted Scenes, Extra Features and Menus.
I want to keep all of them on the DVD-R so I think the main movie compression should be over 50%.
I think there is nothing except IfoEdit that can keep the original DVD structure and it doesn't recode itself.
IfoEdit creates and stores all I-frames of the standard GOP structure in a text file, I meant should I use that TMPG I-frames template when encoding KDVD m2v that is to be remuxed by IfoEdit later.

again thanks for your replies.