Combing Effect Converting DV Video to DVD? - Page 2

Dialhot · #21 05-23-2004, 12:42 PM

Gripcrop is for resizing and you don't need it there. But FYI it is included ina filter called "gripfit". Do a research on the forum with "gripfit AND rar".

The same you don't need blindpp as it is for deblocking MPEG4 comrpessed source, that you don't have here. It is included in mpeg2de3.dll.

In fact you used my script for Divx but your source is far better than this and you should use the script optimal for DVD sources (see optimal script section of the forum.

Bleeder · #22 05-23-2004, 12:57 PM

Ok, do you think you could just post the script I should use here?
I saw the one that is optimal for clean DV video but just to be sure-could you paste the contents of the script I should use just so I don't screw everything up and lose anoter few hours and even more of your time with noob like problems?

Also, I read the guide and If I understand correctly:

1.I use TMPGenc and the template to get the KDVD video.
2.I use Virtualdub and the script (pointing to this KDVD video)?
3.Should I put audio into MP2 or AC3?

4.I will use DVDlab to author in the end.

The first three I stil don't quite get.

incredible · #23 05-23-2004, 01:03 PM

Quote:

SmoothDeinterlace(tff=false, doublerate=true, blend=false)

With only this you end up in a full fieldrate stream! Means FPS*2 !

Do apply behind that a "selectodd()" or a "selecteven()" depending on which results better.

BTW: DgBob() is faster and better, do take the needed values from its readme file which comes with the filter!

Bleeder · #24 05-23-2004, 01:19 PM

Well, since the script above is not the best one, can someone please please paste an appropriate script here. It does not need to include deinterlacing since combing was a strange error that disapeared now.

So basically I would need a good answer to my previous post.

Dialhot · #25 05-23-2004, 02:08 PM

Quote:

Originally Posted by Bleeder

So basically I would need a good answer to my previous post.

As you don't need any deinterlacing, use the optimal script for DVD source called "MA script" that you can find there :

http://www.kvcd.net/forum/viewtopic.php?t=3483

Bleeder · #26 05-23-2004, 05:30 PM

Is the following script OK?

Quote:

LoadPlugin("C:\Program Files\AviSynth 2.5\plugins\MPEG2Dec3.dll")
LoadPlugin("C:\Program Files\AviSynth 2.5\plugins\GripFit_YV12.dll")
LoadPlugin("C:\Program Files\AviSynth 2.5\plugins\STMedianFilter.dll")
LoadPlugin("C:\Program Files\AviSynth 2.5\plugins\asharp.dll")
LoadPlugin("C:\Program Files\AviSynth 2.5\plugins\unfilter.dll")
LoadPlugin("C:\Program Files\AviSynth 2.5\plugins\undot.dll")

Mpeg2Source("D:\New Folder\Untitled 44")

undot()
asharp(1, 4)
GripCrop(720, 576)
GripSize(resizer="BicubicResize")
STMedianFilter(3, 3, 1, 1 )
MergeChroma(blur(1.5))
MergeLuma(blur(0.1))

ScriptClip(" nf = YDifferenceToNext()" +chr(13)+ "unfilter( -(fmin(round(nf)*2, 100)), -(fmin(round(nf)*2, 100)) ).TemporalSoften( fmin( round(2/nf), 6), round(1/nf) , round(3/nf) , 1, 1) ")

function fmin( int f1, int f2) {
return ( f1<f2 ) ? f1 : f2
}

But now Virtualdub says it can't open MPEG2 source.
Even if I set another KDVD video or even AVI?
What do I need to do now?

What am I missing.
Also if I was to set that it is to crop to let's say 544*480 what would that mean quality wise, size wise,..?

Also would it be better to use LanczosResize instead of bicubic?

Also I still don't know whether to turn ausio to MP2 or AC3. Which is better bang for the buck-size?

Dialhot · #27 05-23-2004, 05:45 PM

Quote:

Originally Posted by Bleeder

LoadPlugin("C:\Program Files\AviSynth 2.5\plugins\MPEG2Dec3.dll")
LoadPlugin("C:\Program Files\AviSynth 2.5\plugins\GripFit_YV12.dll")
LoadPlugin("C:\Program Files\AviSynth 2.5\plugins\STMedianFilter.dll")
LoadPlugin("C:\Program Files\AviSynth 2.5\plugins\asharp.dll")
LoadPlugin("C:\Program Files\AviSynth 2.5\plugins\unfilter.dll")
LoadPlugin("C:\Program Files\AviSynth 2.5\plugins\undot.dll")

Note : with avs2.5 you can remove all these lines. As soon as the dll is in the "plugins" directory, it will be automatically loaded.

Quote:

GripCrop(720, 576)
GripSize(resizer="BicubicResize")

YOu don't need these 2 lines as you don't have to resize the source.
BTW you missed the third line taht can be removed appart form these two ones : Gripborders !

Else, your script is okay according to me.

Quote:

But now Virtualdub says it can't open MPEG2 source.
Even if I set another KDVD video or even AVI?
What do I need to do now?

Your srouce is not a mpeg2 source but a DV file !
You must use "Directshowsource("nameofthefile.avi").killaudio() " insteed of mpeg2source.

Quote:

Also if I was to set that it is to crop to let's say 544*480 what would that mean quality wise, size wise,..?

You will obtain something taht is not DVD compliant. Start to make compliants ones and then you will ask yourself about the opportunity to play with the standard.

Quote:

Also would it be better to use LanczosResize instead of bicubic?

Not in your case. Lanczos gives sharper picture but bigger file; In your script you already have "asharp" that deals with the sharpness. No need to do it twice.

Quote:

Also I still don't know whether to turn ausio to MP2 or AC3. Which is better bang for the buck-size?

All people aren't english natives so if you can use standard words to ask your question I'll probably be abble to understand it

Bleeder · #28 05-24-2004, 02:02 AM

OK

I ment which of the two (AC3 or MP2) is a better choice considering quality/size? And compatibility.

Another noob question. Do I use the script to open an AVI file or a m2v file previously created by TMPGenc with a template?

Do I open the script in Virtualdub or TMPGenc?
I am now running it in TMPGenc while using KWAGs KDVD template. Is that OK?

So in short:
What is the order of things

1
2
.
.
.

Dialhot · #29 05-24-2004, 04:06 AM

Quote:

Originally Posted by Bleeder

I ment which of the two (AC3 or MP2) is a better choice considering quality/size? And compatibility.

if you need to do 5.1, use ac3. Even if MP2 can be done in 5.1 this not recommanded for quality and compatibility.

For 2.0 the two are equivalent IMHO but there are few good ac3 encoder while software to produce great mp2 are free...

Quote:

Another noob question. Do I use the script to open an AVI file or a m2v file previously created by TMPGenc with a template?

It seem you missed something : the script is the source you will provide to tmpgenc. You use it to open the DV avi, do some filtering (like deinterlace) and resize if needed. Then tmpgenc takes the output of this script and encode it into mpeg.

Quote:

I am now running it in TMPGenc while using KWAGs KDVD template. Is that OK?

Yes it is.

Bleeder · #30 05-24-2004, 06:25 AM

OK, I ran the same DV video (about 42 minutes long and 8.3GB large and is a dark video of mainly internal footage with my DV camcorder) that in the beginning I ran with the first template (the one that's ot even supposed ti be a KDVD according to KWAG) also with KWAG's KDVD template (only changed 16:9 to 4:3 in TMPGenc).

With the first template (posted on page 1 of this thread) that also covered sound (192kbps) the file size was 976 Mbs.

Afterwards I changed video part in that first template with the data from KWAG's template and changed the audio part to 128 kbps from 192 kbps and size was down to 926 Mb. that seemed a bit strange to me since I expected a bigger difference. It's like only the audio made the difference.
And in both I used noise reduction option in TMPGenc.

Now I ran the following script with KWAGs original unchanged KDVD template (no sound but changed 16:9 to 4:3)

Quote:

Directshowsource("D:\New Folder\Video2").killaudio()

converttoyv12()

undot()
asharp(1, 4)

STMedianFilter(3, 3, 1, 1 )
MergeChroma(blur(1.5))
MergeLuma(blur(0.1))

ScriptClip(" nf = YDifferenceToNext()" +chr(13)+ "unfilter( -(fmin(round(nf)*2, 100)), -(fmin(round(nf)*2, 100)) ).TemporalSoften( fmin( round(2/nf), 6), round(1/nf) , round(3/nf) , 1, 1) ")

function fmin( int f1, int f2) {
return ( f1<f2 ) ? f1 : f2
}

I get a video with 812 Mb and I stripped the WMV from original and turned it to MP2 (128 kbps) with Headach3e and the size is 39 Mb

So combined it is 851Mb. How do I improve this further?

I have not combined video and audio yet. Is this something I do in DVDlab?

Also I wish the image would be a bit sharper. What do I change in the script?

Boulder · #31 05-24-2004, 06:46 AM

Do not do any filtering with TMPGEnc's internal filters!

You'll get a sharper image if you remove the MergeLuma line.

Isn't your source material interlaced? You're treating it as if it were progressive.

You'll have to use the file size prediction methods if you wish to obtain a certain file size. CQMatic is a good tool for that, see it's own forum for more details and guides.

Bleeder · #32 05-24-2004, 07:25 AM

OK, I will remove the Luna part.
My DV is interlaced but the KDVD does not have combing. TMPGenc treats or encodes it as interlaced doesn't it?

If I was to include a deinterlace part wouldn't that only reduce quality?

What do you recommend I do?
The source file is interlaced. it's a DV file in Mainconcept DV 2.04 codec.
When I play this file there is no combing.

And in the KDVD there is also no combing.

What to do?

Boulder · #33 05-24-2004, 07:46 AM

If you got the bitrate to spare, encode as interlaced.

However, if you use that script with interlaced sources, you're asking for trouble.

See this thread for using the MA script with interlaced material (and keeping it interlaced):

http://www.kvcd.net/forum/viewtopic....er=asc&start=0

Deinterlacing doesn't necessarily mean lower quality. Interlaced material is very bitrate-hungry and turns into a blocky monster if it's not fed with enough bits. Progressive material doesn't need that much bits but motion is not as smooth and shapness suffers a bit as well.

EDIT: I also recommend that you change the ScriptClip part to this:

Code:

ScriptClip(" nf = YDifferenceToNext()" +chr(13)+ "unfilter( -(fmin(round(nf)*2, 100)), -(fmin(round(nf)*2, 100)) ).TemporalSoften( fmin( round(2/nf), 6), round(1/nf) , round(3/nf) , 1, 2)  ")

Bleeder · #34 05-24-2004, 09:05 AM

OK,

do you think you could paste the entire script I should use?
Cause there are several variants there and I would like try the best one with deinterlacing but for DV and I don't know what if any of that resizing there applies to me,.. there is talk of VHS source quality there,.. so I'm not sure what to use.

Do you think I could get an entire appropriate script pasted here?

Boulder · #35 05-24-2004, 09:40 AM

This one keeps it interlaced so you should encode as interlaced, bottom field first:

Code:

AVISource("path\clip.avi") # or DirectShowSource, if AVISource doesn't work

GripCrop(Your_GripCrop_Parameters_Here) 

UnFoldFieldsVertical(flip=true) 

undot() 
asharp(1, 4) 
GripSize(resizer="BicubicResize") 
STMedianFilter(3, 3, 1, 1 ) 
MergeChroma(blur(1.5)) 
MergeLuma(blur(0.1)) 

ScriptClip(" nf = YDifferenceToNext()" +chr(13)+ "unfilter( -(fmin(round(nf)*2, 100)), -(fmin(round(nf)*2, 100)) ).TemporalSoften( fmin( round(2/nf), 6), round(1/nf) , round(3/nf) , 1, 2)  ") 

FoldFieldsVertical(flip=true) 

GripBorders() 

function fmin( int f1, int f2) { 
  return ( f1<f2 ) ? f1 : f2 
} 
  
function SetParity(clip c, bool parity) 
{ 
    return parity ? c.AssumeTFF() : c.AssumeBFF() 
} 

function UnfoldFieldsVertical(clip c, bool "flip") 
{ 
    flip = default(flip, false) 
    oldParity = c.GetParity() 
    c = c.AssumeTFF().SeparateFields().AssumeFrameBased() 
    top = c.SelectEven() 
    bottom = c.SelectOdd() 
    c = StackVertical(top, flip ? bottom.FlipVertical() 
    \                           : bottom) 
    return c.SetParity(oldParity) 
} 

function FoldFieldsVertical(clip c, bool "flip") 
{ 
    assert(c.Height() % 2 == 0, "FoldFieldsVertical: unexpected frame height") 
    flip = default(flip, false) 
    oldParity = c.GetParity() 
    originalHeight = c.Height() / 2 
    evens = c.Crop(0, 0, c.Width(), originalHeight) 
    odds = c.Crop(0, originalHeight, c.Width(), originalHeight) 
    odds = flip ? odds.FlipVertical() : odds 
    c = Interleave(evens, odds).AssumeFieldBased().AssumeTFF().Weave() 
    return c.SetParity(oldParity) 
}

This one returns a progressive clip, encode as progressive:

Code:

AVISource("path\clip.avi") # or DirectShowSource if AVISource doesn't work
KernelDeint(order=0,sharp=true,threshold=5)

GripCrop(Your_GripCrop_Parameters_Here) 

undot() 
asharp(1, 4) 
GripSize(resizer="BicubicResize") 
STMedianFilter(3, 3, 1, 1 ) 
MergeChroma(blur(1.5)) 
MergeLuma(blur(0.1)) 

ScriptClip(" nf = YDifferenceToNext()" +chr(13)+ "unfilter( -(fmin(round(nf)*2, 100)), -(fmin(round(nf)*2, 100)) ).TemporalSoften( fmin( round(2/nf), 6), round(1/nf) , round(3/nf) , 1, 2)  ") 

GripBorders() 

function fmin( int f1, int f2) { 
  return ( f1<f2 ) ? f1 : f2 
}

I assume here that your clip is bottom field first, hence the parameter order=0 in KernelDeint.

Remove the MergeLuma line for some more sharpness, you'll also get a slightly sharper video if you use LanczosResize as the GripSize resizer.

Bleeder · #36 05-24-2004, 10:05 AM

My video is bottom first.

But when I try to open either script of the two posted I get :
Undot supports only YUY2 color...

EDIT:

Is this OK?

Quote:

AVISource("D:\New Folder\Untitled.avi")
KernelDeint(order=0,sharp=true,threshold=5)

converttoyv12()

undot()
asharp(1, 4)

STMedianFilter(3, 3, 1, 1 )
MergeChroma(blur(1.5))

ScriptClip(" nf = YDifferenceToNext()" +chr(13)+ "unfilter( -(fmin(round(nf)*2, 100)), -(fmin(round(nf)*2, 100)) ).TemporalSoften( fmin( round(2/nf), 6), round(1/nf) , round(3/nf) , 1, 2) ")

function fmin( int f1, int f2) {
return ( f1<f2 ) ? f1 : f2
}

Also: When you say treat as progressive-what does that mean?
That I set TMPGenc as non-interlace in the TMPGencoder's menues?

Boulder · #37 05-24-2004, 10:23 AM

UnDot should support both YUY2 and YV12 material. You can check the colorspace with this script:

AVISource("path\clip.avi")
Info()

See what the preview window shows in VirtualDub.

Your script looks ok, you're not doing any resizing? You might want to use overscan blocks for better quality.

EDIT: Yes, basically progressive means non-interlaced.

Bleeder · #38 05-24-2004, 10:35 AM

The script I pasted works, it didn't work prior to adding "convert to YV12"

It works now and the file sizes of samples area already smaller than with previous scripts.

Quality seems not to have been harmed.

The weird part is that no matter if I treat it as progressive or not (even switching from non interlace as a source and interlaced as destination and vice versa,.. or both non interlaced or both interlaced) in TMPGenc, there's no visible difference between the videos

Boulder · #39 05-24-2004, 10:38 AM

There's usually no visible difference even if you encode progressive material as interlaced, the file size may get bigger though.

Bleeder · #40 05-24-2004, 10:44 AM

Thanks for your realy extensive help with the scripts. This one seems very good both in terms of sicze and quality. It's also sharp enough.

I will use it as progressive. I also recommend it to others.

I have some questions about merging audio with video though.
With which program can I merge them? m2v and MP2?

The reason is that while I will keep my videos on DVDs made by DVDlab, some friends don't have standalone players and would like to save videos from DVDs to normally watch them from their hard disks so I would also like to be able to burn them in usual format on RWs for them to copy them onto their HDs. Like simple video files.

At the same time-I tried DVDlab a bit earlier and strangely it didn't put any audio into the clips. I imported video and audio in and put both ito that timeline...

What's the deal there? Any special things to set for it to actually put both into a file? I also burnt it onto a DVD+RW and no sound.

Also on the HD in the folder i speccified for DVDlab the TS Audio folder was empty while TS video contained data.