AnimeIVTC() - All-in-one solution for interlacing and blending - v2.00 [2010-01-07]

[Gavino]

Second, I've change variable call ordering to remove "I don't know what "l01" means" error:

l10 = jmp ? l01 : l_v
l01 = jmp ? l12 : l_v
l12 = l_v
to

l12 = l_v
l01 = jmp ? l12 : l_v
l10 = jmp ? l01 : l_v
But that sets everything to l_v, regardless of the value of jmp, so will give incorrect results. When jmp is true, l01/l12 should be defined from the previous frame, and copied to l10/l01 respectively. Similarly for c10, etc. So you (or thetoof) need to find out why the "I don't know what "l01" means" error actually occurs.

[thanhtu5013]

Hi.
At first sorry for my poor english because I'm french.
I'm looking for antialiasing filters for Avisynth and I've found this thread.
Unfortunately, I don't know how to use AnimeIVTC. I'm interested only in anitialiasing function.
I made some captures videos of GTA IV and there has a lot of aliasing.
I know how to create a script but my mines are usually very simple and short. Long scripts are so hard for me.

Someone could help me ?

Thanks by advance.

[Leinad4Mind]

you have a function I have made that simplify that.

Check http://forum.doom9.org/showthread.php?t=153835

;)

[thanhtu5013]

you have a function I have made that simplify that.

Check http://forum.doom9.org/showthread.php?t=153835

;)
Thanks a lot Leinad4Mind. :)

I watched your script and I have some questions:

- Your script is very long. Does all sentences are wrote ONLY for anti-aliasing and improve it ?
- My script is the following :


AVISOURCE("C:\GTA IV_ENB Series 0.079 SORA - 02 - It'S Your Call_1080p.avi)
ConvertToYV12(matrix="PC.709")


I suppose I have just to add your script below my mine and deleting sentences with "#", right ?

- On your thread, you posted a pack with reguired plugins but there has one guy complaining about missing x64 plugins...
I have Windows 7 64 bits so does your pack will work with me ?

Thanks again for this amazing huge work. :)

[thanhtu5013]

@Leinad4Mind.

I made my script and it works on VirtualDub and MeGui but unfortunately there has no effects in my video... :confused:
There has still a lot of aliasing... :eek:

Watch my screens.

- Original AVI :
http://img697.imageshack.us/img697/1253/screenoriginal.png

- My script :
http://img21.imageshack.us/img21/636/screenretouched.png


My code:

AVISOURCE("C:\GTA IV_ENB Series 0.079 SORA - 02 - It'S Your Call_1080p.avi")

FUNCTION LSharpAAF(clip orig, int "StrDrk", int "ShPre", int "ShPost", int "SmPost", string "aatype"){
StrDrk = default(StrDrk,38) #FastLineDarkenMOD
ShPre = default(ShPre,80) #Presharpening
ShPost = default(ShPost,300) #Postsharpening
SmPost = default(SmPost,100) #Postsmoothing
aatype = default(aatype,"EEDI2") # Use EEDI2() or Sangnom() or maa() or daa() or ediaa() or aaa() or saa() or splinaa() or supaa() or naa() or shaarp() or nnedi() or nnedi2() or antialiasing() for anti-aliasing.
a=orig.AssumeTFF()
w=width(a)
h=height(a)
m=logic(a.DEdgeMask(0,255,0,255,"5 10 5 0 0 0 -5 -10 -5", divisor=4,Y=3,U=3,V=3)
\ ,a.DEdgeMask(0,255,0,255,"5 0 -5 10 0 -10 5 0 -5", divisor=4,Y=3,U=3,V=3)
\ ,"max").greyscale().levels(0,0.8,128,0,255,false)
b=a.fastlinedarkenmod(thinning=0, strength=StrDrk)
b=(aatype=="EEDI2") ? b.TurnLeft().EEDI2().TurnRight().EEDI2().spline64resize(w,h,0.5,-0.5,2*w+.001,2*h+.001) : (aatype=="Sangnom") ? b.spline36resize(w*2,h*2).TurnLeft().SangNom(aa=255).TurnRight().SangNom(aa=255).spline36resize(w,h) : (aatype=="maa") ? b.maa() : (aatype=="daa") ? b.daa() : (aatype=="ediaa") ? b.ediaa() : (aatype=="aaa") ? b.aaa() : (aatype=="saa") ? b.SAA() : (aatype=="splinaa") ? b.splinaa() : (aatype=="supaa") ? b.supaa() : (aatype=="naa") ? b.naa() : (aatype=="shaarp") ? b.shaarp() : (aatype=="nnedi") ? b.NNEDI(dh=true,field=1).TurnRight().NNEDI(dh=true,field=1).TurnLeft().Spline36Resize(w,h) : (aatype=="nnedi2") ? b.nnedi2(dh=true,field=1) : (aatype=="antialiasing") ? b.antialiasing() : blankclip(pixel_type="YV12").subtitle("Please use Sangnom, EEDI2, maa, daa, ediaa, aaa, ssa, splinaa, supaa, naa, shaarp, nnedi, nnedi2 or antialiasing for aatype")
b=b.LSFMod(edgemode=1,strength=ShPost,overshoot=1,soft=SmPost)
mt_merge(a,b,m,Y=3,U=3,V=3)
sD=mt_makediff(a,last)
bv1 = a.MVAnalyse(isb=true, delta=1,idx=13)
bv2 = a.MVAnalyse(isb=true, delta=2,idx=13)
fv2 = a.MVAnalyse(isb=false,delta=2,idx=13)
fv1 = a.MVAnalyse(isb=false,delta=1,idx=13)
sDD = sD.MVDegrain2(bv1,fv1,bv2,fv2,idx=14)
reduc = 0.4
sDD = mt_lutxy(sD,sDD,"x 128 - abs y 128 - abs < x y ?").mergeluma(sDD,1.0-reduc)
a.mt_makediff(sDD,U=2,V=2)
return last
}

FUNCTION ediaa(clip a) {
return a.EEDI2(field=1).TurnRight().EEDI2(field=1).TurnLeft().spline36resize(a.width,a.height,-0.5,-0.5) }

FUNCTION daa(clip c) {
nn = c.nnedi2(field=-2)
dbl = mt_average(selecteven(nn),selectodd(nn),U=3,V=3)
dblD = mt_makediff(c,dbl,U=3,V=3)
shrpD = mt_makediff(dbl,dbl.removegrain((width(c)>1100) ? 20 : 11),U=3,V=3)
DD = shrpD.repair(dblD,13)
return dbl.mt_adddiff(DD,U=3,V=3) }

FUNCTION saa(Clip Clp,Int"SS",Bool"CP") {
OX = Clp.Width
OY = Clp.Height
SS = Default(SS,2)
CP = Default(CP,True)
Clp = Clp.IsYV12() ? Clp : Clp.ConvertToYV12()
Clp.PointResize(OX*SS,OY*SS).SangNom().TurnRight()
\ .SangNom().TurnLeft().BilinearResize(OX,OY)
CP ? Last : MergeChroma(Clp)
Return(Last) }

FUNCTION maa(clip input, int "mask") {
mask = input.mt_edge("sobel",7,7,5,5).mt_inflate()
aa_clip=input.spline36Resize(width(input)*2,height(input)*2).TurnLeft().SangNom().TurnRight().SangNom().spline36Resize(width(input),height(input)).MergeChroma(input)
return mt_merge(input,aa_clip,mask) }

FUNCTION AAA(clip clp, int "Xres", int "Yres", int "Xshrp", int "Yshrp",
\ int "US", int "DS", bool "chroma")
{
clp = clp.isYV12() ? clp : clp.ConvertToYV12()
ox = clp.width
oy = clp.height
Xres = default(Xres, ox)
Yres = default(Yres, oy)
us = default(us, 1)
ds = default(ds, 2)
Xshrp = default(Xshrp, 15)
Yshrp = default(Yshrp, 15)
chroma = default(chroma, false)

us==0 ? clp.PointResize(ox*2,oy*2) : clp.LanczosResize(ox*2,oy*2)

TurnLeft()
SangNom()

TurnRight()
SangNom()

ds==0 ? BilinearResize(Xres,Yres) :
\ ds==1 ? BicubicResize(Xres,Yres) :
\ LanczosResize(Xres,Yres)

Unfilter(Xshrp,Yshrp)

chroma ? MergeChroma(clp.Lanczosresize(Xres,Yres)) : last }

FUNCTION antialiasing(clip orig,int "th_luma",int "th_croma",string "type",int "aath") {

th_luma = Default(th_luma, 20)
th_croma = Default(th_croma, 20)
type = Default(type, "sobel")
aath = Default(aath, 48)
ox = orig.width
oy = orig.height
dx = orig.width * 2
dy = orig.height * 2
clp = orig.IsYV12() ? orig : orig.ConvertToYV12()

a=clp
b=clp.Lanczos4Resize(dx,dy).TurnLeft().SangNom(aa=aath).TurnRight().SangNom(aa=aath) \
.LanczosResize(ox,oy)
c=clp.EdgeMask(th_luma,th_luma,th_croma,th_croma,type)
MaskedMerge(a,b,c) }

FUNCTION splinaa(Clip Clp,Int"SS",Bool"CP") {
OX = Clp.Width
OY = Clp.Height
SS = Default(SS,2)
CP = Default(CP,True)
Clp = Clp.IsYV12() ? Clp : Clp.ConvertToYV12()
Clp.PointResize(OX*SS,OY*SS).SangNom().TurnRight()
\ .SangNom().TurnLeft().Spline36Resize(OX,OY)
CP ? Last : MergeChroma(Clp)

Return(Last) }

FUNCTION supaa(Clip Clp,Int"SS",Bool"CP") {
OX = Clp.Width
OY = Clp.Height
SS = Default(SS,2)
CP = Default(CP,True)
Clp = Clp.IsYV12() ? Clp : Clp.ConvertToYV12()
Clp.nnedi2_rpow2(rfactor=SS,cshift="spline36resize",qual=3).SangNom().TurnRight()
\ .SangNom().TurnLeft().Spline36Resize(OX,OY)
CP ? Last : MergeChroma(Clp)

Return(Last) }

FUNCTION naa(Clip Clp,Int"SS",Bool"CP") {
OX = Clp.Width
OY = Clp.Height
SS = Default(SS,2)
CP = Default(CP,True)
Clp = Clp.IsYV12() ? Clp : Clp.ConvertToYV12()
Clp.nnedi2_rpow2(rfactor=SS,cshift="spline36resize",qual=3).nnedi2(qual=3).TurnRight()
\ .nnedi2(qual=3).TurnLeft().Spline36Resize(OX,OY)
CP ? Last : MergeChroma(Clp)

Return(Last) }

FUNCTION shaarp(clip input, int "mask", int "type") {
mask = Default(mask,1)
type = Default(type,1)
Assert(mask == 1 || mask == 2, "Please use mask = 1 or 2")
Assert(type == 1 || type == 2, "Please use type = 1 or 2")
aa_clip = (type == 1) ? input.spline36Resize(width(input)*2,height(input)*2) : input.nnedi2_rpow2(rfactor=2,qual=3)
aa_clip = aa_clip.TurnLeft().nnedi2().TurnRight().nnedi2().spline36Resize(width(input),height(input)).MergeChroma(input)
mask = (mask==1) ? input.mt_edge("sobel",7,7,5,5).mt_inflate() : input.mt_edge("roberts",0,4,0,4).mt_inflate()

return mt_merge(input,aa_clip,mask) }


P.S: Maybe we should continue our conversation in your thread instead of here. :)

[Ghostlamer]

Hello.
Could you explain how to determine Hard Telecine, Double Hard telecine and others..
Thanks.

Somewhere was a manual for this, but i can not find.

[Nightshiver]

You have not tried to find it at all. Go to the first page.

[boxXx]

i keep getting this error message i don't know how to fix it, it's really bugging me, its working in AvsP but when i try to qeue it in Megui it gives me this : http://www.imagebam.com/image/7366a379845674

here's my script :
LoadPlugin("C:\Program Files\AviSynth 2.5\plugins\TIVTC.dll")
LoadPlugin("C:\Program Files\AviSynth 2.5\plugins\RemoveGrain.dll")
Import("C:\Program Files\AviSynth 2.5\plugins\SRestore.avs")
LoadPlugin("C:\Program Files\MeGUI\tools\dgindex\DGDecode.dll")
DgDecode_mpeg2source("G:\ShinChan\D1\1\VideoFile.d2v", info=3)
AnimeIVTC(mode=2,aa=2)
crop( 0, 0, -8, -2)
Spline36Resize(656,480)
trim(20000,22000)

[Vitaliy Gorbatenko]

It is possible to automatically find and fix antialiased interlaced frames?
On lastest DVD it's a big problem =(

[Leinad4Mind]

It seems to me that u have fix this.

And I don't think that this task is possible. And impossible tasks... is for someone called didée LOOOOL

But after deinterlace if u got some interlace on some frames u can try to apply vinverse() after the telecine.

Like this: "tfm(order=-1,clip2=res,pp=7,slow=2,mode=2,chroma=false,display=false).tdecimate(hybrid=1,mode=0,vfrDec=0).vinverse() #Inverse Telecine + Vinverse post-processing"

Don't know if this is really your problem...

[Vitaliy Gorbatenko]

I'm don't fix this!!! It's a different (next) frame.
Problem of there and it is quite serious.

[Leinad4Mind]

Send me the d2v+m2v+script using, I will take a look.

[Vitaliy Gorbatenko]

Script is simple:

# SOURCE
path_to_d2v="S:\..\t_01.d2v"
# SOURCE
mpeg2source(path_to_d2v, cpu2="oooooo", info=3)
colormatrix(interlaced=true, hints=true)
#
animeivtc(mode=1,aa=0) #the upcoming aa=4 also leads to very good results

Samples here: http://www.mediafire.com/?jwjm4kym4wk

[Vitaliy Gorbatenko]

Sample 2 HAS problemm. 1 broken frame.
http://img63.imageshack.us/img63/8002/0100124575.png
Anime: To love ru TV
I don't have the source from which taken the first example.
Unfortunately such moments there are many recent R2 DVD. And we must somehow learn to find and bypass / repair this.

[Ghostlamer]

You have not tried to find it at all. Go to the first page.
Detailed manual about anime telecining(like in v1) is not there.

http://img143.imageshack.us/img143/8207/82465690.th.png (http://img143.imageshack.us/i/82465690.png/)

[Nightshiver]

The only difference is in the operations, not in how you determine the source. Use the v1 to learn how to tell the difference. Yes, so clever using a screenshot from the v2 faq. Heck, just use the search function for this forum and you'll get several different but accurate ways to learn sources, it's not that hard. If you need the v1 faq, say so and I'll upload it.

[LaughingMan]

When I'm using chrfix to filter defect which you can see in pictures

without chrtfix:
http://img100.imageshack.us/img100/1106/moo02b.png

with chrtfix:
http://img63.imageshack.us/img63/3723/moo02a.png

in this script:
AnimeIVTC(mode=1, chrfix=true, aa=1, omode=1, credconv="blend", e1=40368, i1=42723, e2=43469)
crop(8, 0, -8, 0)

#resize
#denoise

I get this message in final encode:
http://img202.imageshack.us/img202/3331/idontknow.jpg

So could anyone help me? Thanks.

[Vitaliy Gorbatenko]

LaughingMan: I'm have a same error/message.

Error on Credit's :
http://www.mediafire.com/?ztmjlznwjwn

script:
mpeg2source(path_to_d2v, cpu2="oooooo", info=3)
animeivtc(mode=1,aa=0)#, killcomb=0, ifade=true)

How to fix it ?

[Great Dragon]

When I'm using chrfix to filter defect which you can see in pictures
chrfix is broken. It has been discussed on previous page.
http://forum.doom9.org/showthread.php?p=1392244#post1392244

animeivtc(mode=1,aa=0)#, killcomb=0, ifade=true)
remove )#

[Vitaliy Gorbatenko]

Great Dragon: I'm test this options - no hope =(
PS: I'm don't think that the script is broken - becouse video is rendered. I'm think it is a some kind of warning..

[Nightshiver]

You are misunderstanding the point, Vitaliy. The script, AnimeIVTC itself is not broken. The function within the script, chrfix, is. By broken, I don't mean "not going to work and give you error screens", I mean, will give you a result, just not the right one. It currently can't be fixed because thetoof is on hiatus somewhere.

[Vitaliy Gorbatenko]

Nightshiver: Tnx =) But i'm more care about http://forum.doom9.org/showthread.php?p=1401056#post1401056 . For fix chroma exist a special plugins =)

[Great Dragon]

@Vitaliy Gorbatenko: What kind of error do you have exactly?
For fix chroma exist a special plugins
And what is those?

[Nightshiver]

I think the "special plugin" he might be referring to is SRestore, as I can only gather that from the link you gave to the previous pages: http://forum.doom9.org/showthread.php?p=1392244#post1392244

[Vitaliy Gorbatenko]

Chroma correction: http://avisynth.org/mediawiki/External_filters#Chroma_correction
For my issue no responce T_T

[Great Dragon]

Chroma correction: http://avisynth.org/mediawiki/Extern...oma_correction
None of those can deal with chroma blending or chroma swap between fields.

@Nightshiver: In my case (http://forum.doom9.org/showthread.php?p=1391241#post1391241) Srestore isn't work :( And even chroma_rep() (http://forum.doom9.org/showthread.php?p=1268074#post1268074) can't get rid of it. I've tried various combinations of filters (bobers and teleciners) with no results.

[ghscreenki]

Hello again!

I have trouble fixing this R2 source. This source has a lot of bad areas like interlaced frames that are immune to AnimeIVTC's deinterlacing. It can be fixed by deleting that bad frame and replace it with a previous good frame but only works on static scenes though. On scenes with movement, it's a different story.

Here's the source:

http://img229.imageshack.us/img229/2127/source.png

And this is the result of this scene using "mode=1":

http://img717.imageshack.us/img717/2535/sample1r.png

I really don't know how to fix this. I tried using mode=4 to see how it does and this is the result (i1=xxx or p1=xxx):

http://img339.imageshack.us/img339/7515/sample2o.png

mode=4 produced a better result but the problem is still there.

Help!

[Nightshiver]

As always, post a sample.

[ghscreenki]

As always, post a sample.
Whoops. Kept on forgetting. Here's the clip:

http://www.megaupload.com/?d=GXKKQ8ZT

[manono]

Since both fields of the frames involved show that stuff, nothing's ever going to fix it. Is that very short scene the only place where that kind of thing occurs?

[ghscreenki]

Since both fields of the frames involved show that stuff, nothing's ever going to fix it. Is that very short scene the only place where that kind of thing occurs?
It's a short scene around 2 seconds and so far, that's the only scene with that kind of problem.

[manono]

Well, you could use FreezeFrame to replace the bad frames with good frames from either side, which will make it look nice but play slightly jerky, or you could use the BadFrames filter to replace those really nasty ones with blends of the nearby good frames (the way I'd probably do it). That'll give you blends, but nicer looking blends. As far as I know, there's no way to entirely fix the problem.

[ghscreenki]

Don't want any jerkiness so I tried BadFrames() and surprisingly, it's better.

http://img690.imageshack.us/img690/7942/sample3q.png

Didn't fix everything but it's better than the other attempts. =) Thanks!

[clone83]

Hi guys, I'm hoping that you will be able to help me? when i try and load my script i get this error:

AviSynth script error:
Script error: Invalid arguments to function "ediaa"
(C:\AviSynth 2.5\plugins\AnimeIVTC\AnimeIVTC 2.00.avsi, line 427)

The script im trying to use is:

animeivtc(mode=5, omode=2, aa=1, precision=1)

please help...

[clone83]

if you need more info just say so and i will post it up... sort of need this fixed asap. thx guys

[Gavino]

animeivtc(mode=5, omode=2, aa=1, precision=1)
According to the documentation, mode must be between 1 and 4.
Pity that the code does not validate this, leading to such an obscure error.

[clone83]

from what i have found out so far its all to do with the aa settings... any number i put in, it doesnt work.

[Gavino]

from what i have found out so far its all to do with the aa settings... any number i put in, it doesnt work.
Do you mean you have tried with a mode other than 5?
What combinations have you tried? Do you always get the same error?

[clone83]

i have tried it without aa and mode 5 works but as soon as i add any number into aa (1-4) i get errors for each one as it tries to call the external plugin but it doesnt work.

[Gavino]

i have tried it without aa and mode 5 works
It only 'works' by accident (in the sense that no error message is produced) and even then I'm not sure if the results are meaningful.
As I said, there is no mode 5, only modes 1-4 are defined.

Unfortunately, the function code does not check for a valid mode and anything outside the expected range will produce undefined results (including spurious errors as various variables do not have their correct value).
This could have been prevented if the author had included a simple validity check.

[ghscreenki]

I have a very bad frame right before the scene changes and I want to get rid of it but I can't because I'll be clipping off a part of the animation (movement). I don't want to use BadFrames() with "blend=true" because it will blend in the next frame (which is a scene change) and I can't use "blend=false" either because it's just like DeleteFrame and DuplicateFrame - copies the previous frame. Is there a way for me to duplicate a specific frame and put it somewhere in the source other than putting the copied frame next to the original frame?

[Gavino]

Is there a way for me to duplicate a specific frame and put it somewhere in the source other than putting the copied frame next to the original frame?
FreezeFrame (http://avisynth.org/mediawiki/FreezeFrame)(bad, bad, good)
where 'bad' is the number of the frame you want to replace, and 'good' is that of its replacement.

[ghscreenki]

FreezeFrame (http://avisynth.org/mediawiki/FreezeFrame)(bad, bad, good)
where 'bad' is the number of the frame you want to replace, and 'good' is that of its replacement.

Woah! It worked like a charm! :thanks:

[Overdrive80]

One question, with this filter, would i do convert 30i to 30p?

[manono]

One question, with this filter, would i do convert 30i to 30p?
You don't need AnimeIVTC for that, just a deinterlacer.

[boxXx]

Could someone help me out getting the right settings for this sample ?
i've tried multiple ways of de-interlacing this but it never seems to be correct ( Using AIVTC ofcourse :P ).

http://www.megaupload.com/?d=PNTXT2N2

Thanks alot for the help !

[boxXx]

anybody ? :P

[manono]

The file you are trying to access is temporarily unavailable.
It said the same yesterday.

And 90 MB is quite a lot. Usually a well chosen 10 seconds is plenty. Try MediaFire or Sendspace.

[boxXx]

Sorry for not cutting it down to 10 secs, i really have to leave now, here is the Mediafire link :

http://www.mediafire.com/download.php?5d2mtmdltmwytz4

[boxXx]

once again, please ? :P

[Nightshiver]

Rather odd source. Most methods work, but keep some sort of chroma blend/ghost. What worked best for me was

Yadif(mode=1, order=-1)
SRestore()

And then you'll have to manually use FreezeFrame(bad frame#,bad frame#, good frame#) to get rid of the odd blend. Not even VFR worked.

[Emulgator]

Just found a small variable declaration mismatch, easy to fix:
aatype is declared as int, the html mentions int as well , but in the actual 2.00 script a string is expected.
Just open AnimeIVTC 2.00.avsi and change line 14:
, float "dark", int "thin", int "sharp", int "smooth", bool "stabilize", int "tradius", int "aapel", int "aaov", int "aablk", int "aatype"){
into
, float "dark", int "thin", int "sharp", int "smooth", bool "stabilize", int "tradius", int "aapel", int "aaov", int "aablk", string "aatype"){

[aegisofrime]

I'm wondering, could somebody mirror the script and guide on a file hosting service that doesn't suck? I have tried Sendspace for one whole day now, and it won't let me download. Probably because every ISP in my country uses a transparent proxy for their customers. A mediafire mirror would be much appreciated. Thanks!

[Nightshiver]

Here: http://www.mediafire.com/?3ct9wxzdjacj8xw

[aegisofrime]

Here: http://www.mediafire.com/?3ct9wxzdjacj8xw

Thank you very much! That was a lot less painful than sendspace!

[RsVolcom]

Here: http://www.mediafire.com/?3ct9wxzdjacj8xwThanks for this. FYI, It is smaller than the sendspace link, which would not extract.

[Nightshiver]

That's because the sendspace link contains the TGMCMod.

[Shinji-kun]

Hello thetoof

Thread here (http://forum.doom9.org/showthread.php?t=157051). I am trying to restore a berserk trailer that AnimeWorks made (http://www.mediafire.com/?ionx747yy9ch4k0). It has been suggested that the telecine process of the trailer is somewhat irregular. So I tried using your AnimeIVTC() function.

I said it in that thread but, what mode should I use for the trailer? I think it's either 2 or 3 but I still see interlaced frames ("after image" effect in certain frames, don't know if that is interlacing or blending) using both modes. Could you take a look at the sample and give me some help with which mode to use and which parameters to use that might fix the interlacing?

Thanks

[Nightshiver]

The error is not with AAA. "Run out of memory" Well, what do you think it is? It's that Avisynth ran out of processing memory. Use Setmemorymax(number) to correct this. Hopefully you have at least 1gig of RAM in your system, ideally more. (This is for your other post)

It's field-blended. Sometimes there is a 3:2 pattern, other times, there isn't any pattern. You are also using omode and credconv wrong. You don't use those 2 functions with mode=2 or 3. To help you in the future of determining what the source is, I suggest downloading the "old" faq in the original package.

A simple Animeivtc(mode=2,aa=0) works just fine for me on the sample provided. I also would not use AAA or the other filters you currently have chained to this.

Animeivtc(mode=2,aa=0)
source = last
vectors = source.MVAnalyseMulti(refframes=2, pel=2, overlap=4, sharp=1, idx=1)
source.MVDegrainMulti(vectors,thSAD=600,idx=1)
LimitedSharpenFaster(ss_x=1.8, ss_y=1.8, smode=4, strength=180, overshoot=1, undershoot=1)

Works well, and if you want a little more sharpness/thinning, use aa=4.

[Shinji-kun]

which MVTools version did you use?

[Nightshiver]

This one. (http://www.mediafire.com/?1q5yxmmdjqrbklq)

[Shinji-kun]

This one. (http://www.mediafire.com/?1q5yxmmdjqrbklq)

Never mind. 1.9.7. works.

Also, there was still substantial aliasing, grain and banding in the video when I loaded the script you suggested. I think I will use AAA and the others.

I saw interlacing as well.

[Nightshiver]

You claim substantial aliasing,grain and banding while I did not. What are you using to view the output and are you using the sample you provided? There is no such thing as substantial grain after what I did to it, and there is definitely no banding even with grain removal. And don't use the forum for pictures, it's to damn slow. Just upload to tinypic and post a link. I saw no interlacing myself, so please give an exact frame number. You also have dot crawl on the sides, so Checkmate() is needed.

Pics for proof:


Your version: Click (http://img836.imageshack.us/img836/4186/berserk822a.png)
My version: Click (http://img718.imageshack.us/img718/7543/berserk822b.png)
Your version with AAA: Click (http://img715.imageshack.us/img715/690/berserk822awithaaa.png)
My version with LSF: Click (http://img693.imageshack.us/img693/6339/berserk822bwithlsf.png)
Original: Click (http://img269.imageshack.us/img269/9042/berserk822c.png)

Although AAA isn't even a sharpening filter, so it can't really be compared to LSF.

[Shinji-kun]

The images you posted only deal with the grain. There is still aliasing and some interlaced frames when I use your script.

Aliasing

Source
http://img834.imageshack.us/img834/6778/bersrktraileraliasingso.png


Animeivtc(mode=2,aa=0)
source = last
vectors = source.MVAnalyseMulti(refframes=2, pel=2, overlap=4, sharp=1, idx=1)
source.MVDegrainMulti(vectors,thSAD=600,idx=1)
LimitedSharpenFaster(ss_x=1.8, ss_y=1.8, smode=4,strength=180, overshoot=1, undershoot=1)

http://img408.imageshack.us/img408/6086/berserktrailersuggestio.png


AAA()

http://img529.imageshack.us/img529/5968/berserktraileraaagradfu.png (http://img529.imageshack.us/i/berserktraileraaagradfu.png/)


Interlacing

Source

http://img714.imageshack.us/img714/2224/berserktrailerinterlace.png (http://img714.imageshack.us/i/berserktrailerinterlace.png/)


AnimeIVTC(mode=2,aa=0)

http://img689.imageshack.us/img689/1200/berserktraileranimeivtc.png (http://img689.imageshack.us/i/berserktraileranimeivtc.png/)

Uploaded with ImageShack.us (http://imageshack.us)


I viewed these script in vdub 1.9.10.

[Nightshiver]

Give me frame numbers please. And the aa= function of animeivtc is what that's for, aa. I liked aa=2 or 4 for this particular source.

EDIT: Actually, after finding the frame's in question and after testing and thinking about it, those are the only frames that are like that. There are no other frames like that throughout the video, which leads me to believe that what this small part (there are a few other parts) is actually not blended, but it is a video effect the producers/animators put in. To support this, Gut's face is getting an extreme close up at high speed, and in movies and some animation, there are actually blurring/blending effects that are put in to make it look like it's going even faster or to just enhance the effect. I believe that is what the case is here.

Why? Because after thoroughly testing all options of AnimeIVTC, TMPGMCbeta2 and mod, yadif, Srestore, and VFR, those same sections are always like that. I think that this "trailer" is just a special case for this. Where is the source from this? The DVD set from a couple years ago or the "Remastered" set, which is more recent?

[Shinji-kun]

Where is the source from this? The DVD set from a couple years ago or the "Remastered" set, which is more recent?

Giant Robo: Complete Collection. The Remastered is more recent. It's easily obtainable info.

What about the anti-aliasing? I tried the various aa functions in AnimeIVTC and none worked as good as AAA(). The AnimeIVTC aa options always left at least some form of jagged edges in the red parts around frame 1272 (filtered).

[Nightshiver]

Well, if AAA works better for you, then use it. Just out of curiosity, why do you just want the trailer encoded?

[Shinji-kun]

Could you give a link to the old AnimeIVTC guide you mentioned?

[Nightshiver]

Click this. (http://www.mediafire.com/?cnecwea5vm643z1)

[Shinji-kun]

The encode failed again. It gave some error about not being able to cache enough frames and stopped encoding when there was 40 seconds left of the video. This only seems to happen when I use AAA().

About the anti-aliasing, would aa=2 or 4 remove the aliasing in the final encode even if it does not look satisfactory when previewing the script? I think there might be a problem with my video card or MPC-HC because no matter what aa filters I use I can still see aliasing in the encoded X264 MKV files when I view them in mmpc-hc.

I also have a question about an Ed Edd n Eddy episode but I'll post a sample later.

[MatLz]

Hum...I can't see aliasing on the pics...
There is a chroma problem on the blood edges
r=1 #or 2, but not more
deen("a2d",r,1,30,1,1,1)should work.

[Vitaliy Gorbatenko]

Fix chroma problem:
MergeChroma(aWarpSharp2(chroma=4).RemoveGrain(12))

[Shinji-kun]

^ Seems to work. As I mentioned before I have also had a similar problem with Ed Edd N Eddy footage, as well as mpc-hc. The jagged chroma edges are NOT visible when I view the E, E n E episode avs script (or even the source .VOB) in virtualdub, however when I try to view the same avs script in mpc-hc the jagged edges of red/yellow objects ARE visible. This includes when I view the encoded X264/MKV files in mpc-hc, even when using overlay mixer as the renderer, which doesn't even convert the colorspace from YUV IIRC.

So why, I ask, does this chroma problem keep occurring and how can I fix it completely so I won't keep having to add MergeChroma commands to every avs script?

[Nightshiver]

I use AvsP for previewing scripts and VLC to play my files. Did you try the mergechroma on your other sources? You might have to use it everytime, because it might be that each source has a different chroma problem. With mpc-hc, it might be a problem with some of the ffdshow filters. What version of ffdshow are you using? Download the latest one from here (http://sourceforge.net/projects/ffdshow-tryout/files/SVN%20builds%20by%20clsid/generic%20builds/ffdshow_rev3572_20100913_clsid.exe/download). Also, I would suggest downloading CCCP for using things with MPC. So download that from here (http://www.cccp-project.net/) and when you go to install it, un check the box where it wants to install ffdshow, as the version with CCCP is old. And use the MPC version that is included with CCCP, not yours.

[Shinji-kun]

I use AvsP for previewing scripts and VLC to play my files. Did you try the mergechroma on your other sources? You might have to use it everytime, because it might be that each source has a different chroma problem. With mpc-hc, it might be a problem with some of the ffdshow filters. What version of ffdshow are you using? Download the latest one from here (http://sourceforge.net/projects/ffdshow-tryout/files/SVN%20builds%20by%20clsid/generic%20builds/ffdshow_rev3572_20100913_clsid.exe/download). Also, I would suggest downloading CCCP for using things with MPC. So download that from here (http://www.cccp-project.net/) and when you go to install it, un check the box where it wants to install ffdshow, as the version with CCCP is old. And use the MPC version that is included with CCCP, not yours.

I do have the CCCP installed and am using their version of MPC. Although I do have the ffdshow that comes with CCCP checked.

Another problem. I encoded with this:

setmemorymax(2000)
DGDecode_mpeg2source("C:\Users\dvd rip\Berserk Trailer.d2v")
LoadPlugin("C:\program files (x86)\avisynth 2.5\plugins\removegrain.dll")
LoadPlugin("C:\program files (x86)\avisynth 2.5\plugins\yadifmod.dll")
LoadPlugin("C:\program files (x86)\avisynth 2.5\plugins\mvtools.dll")
Checkmate()
AnimeIVTC(mode=2,aa=4)
source = last
vectors = source.MVAnalyseMulti(refframes=2, pel=2, overlap=4, sharp=1, idx=1)
source.MVDegrainMulti(vectors,thSAD=600,idx=1)
LimitedSharpenFaster(ss_x=1.8, ss_y=1.8, smode=4, strength=180, overshoot=1, undershoot=1)
gradfunkmirror()
MergeChroma(aWarpSharp2(chroma=4).RemoveGrain(12))

and I recieved this error:

http://img97.imageshack.us/img97/5165/grab01846.jpg (http://img97.imageshack.us/i/grab01846.jpg/)

What is the problem?

and when you go to install it, un check the box where it wants to install ffdshow, as the version with CCCP is old.

Is there any way to do this without having to reinstall CCCP?

[Nightshiver]

I told you before, you need to use Setmemorymax(). Avisynth is running out of memory to use. How much onboard ram do you have? Divide the total amount of memory (in MB) by 4 and enter that number in-between the brackets. If you only have 1gb of RAM, you have problems, and we'll have to remove some functions/change them.

As for removing ffdshow, all you should have to do is go to the Add/Remove Programs in Windows and just remove it.

[Usedocne]

Filtering in multiply stages via HuffYUV might be helpful too, if he's maxing out his memory?

e.g.
VOB(s) -> First half of filters -> HuffYUV
HuffYUV -> Second half of filters -> HuffYUV

[Shinji-kun]

I have an Ed Edd n Eddy clip that I think might be DHT. There is not constant pattern of 3 progressive frames followed by 2 interlaced. Sometimes there are no interlaced frames for hundrends of frames and other times there will suddenly be a 1:5 pattern. The second thing I see starts when you first see the lamp in Edd's room. The frames look interlaced (fields are present) but there is no combing. I wonder if this is a poor telecine or just a motin=on effect that the animators added. It is clear though that it looks like interlacing even though it is not constant. What is this problem? What AnimeIVTC() mode should I use?

Sample 1 (http://www.mediafire.com/?ahhimxkzf8cumm6)

Sample 2: http://www.mediafire.com/?q999whi40kazmkw

Sample 1 has the part I mentioned above (starts with the lamp) where it looks like there is interlacing but may be an intentional motion enhancing effect originally drawn into the cels.

[Nightshiver]

Another odd source with strange blend-like frames. Mode=3 (DHT) does work, but keeps the strange ghosting images. Even with VFR and trying SRestore, they are still there. If you want more luck, try asking MoMonster in his thread about SRestore.

[Shinji-kun]

Couldn't the ghosting just be an intentionally added effect that the animators put in the original production? Like the Berserk zooming effect.

[Nightshiver]

It could be. E,E, and E would be a little notorious for that, as the creators intentionally made the cartoon pencil work, strange colors, shimmers, etc. So yes, it could just be an added effect.

[Shinji-kun]

Can someone confirm if that's the case?

[egrimisu]

why encodes freezes at some point but only on some sources when ifade=true? using megui(huffy lossles) freeze at some point, using x264 at another point.does anybody else encoutered this problem?
what is the latest version with fixes of animeivtc?might that be the problem...

[Nightshiver]

Read through the thread. The ifade function is broken and there is no fix because toof isn't around anymore.

[egrimisu]

Read through the thread. The ifade function is broken and there is no fix because toof isn't around anymore.

damn,
can i run smartfade manualy(outside animeivtc) i read something that i must run it befor telecide and decimate, how can this be done?


strange i've benn encoding 50 eps from an anime in the past mounth and i had this problem at only one ep. I update mvtools and masktools meanwhile, is smartfade dependent on them or other scripts?

[Nightshiver]

Yes. Smartfade relies on the function daa() within AnimeIVTC and daa() relies on MVtools and masktools.

[egrimisu]

Yes. Smartfade relies on the function daa() within AnimeIVTC and daa() relies on MVtools and masktools.

Hi, i tried with diferent version of mvtools but same thing.
After that i realised that in the avisynth plugins folder i had masktools v 2.a23 and v 2.a45, i've deleted the 2.a23 and now it's working.

PS v2.a23 worked with ifade=true when aa=2 (daa) strange...

Do i need in the plugins folder v1 and v2 of masktools or it is ok if i leave only v2 there?? i'm not sure if masktools v2 is compatible over v1 for older functions/scripts. Thanks

[Nightshiver]

I personally keep both versions in my plugin folder.

[kaishakunin]

Can someone please look at this and give me some hints on how can I IVTC it correctly:
http://rapidshare.com/files/428730305/VTS_01_1.demuxed_cut.m2v
(this is just a 3-second cut from the DVD, size=3Mb)
All methods I tried give unsatisfactory results, the decimated frames are damaged.

[ianken]

Have fun killing all the popus and browser hijacks from the download site. Damn. Even popup blocker failed to stop it. Had to task kill Firefox.

[egrimisu]

Since chrfix=ture does not work when mode=2, does anybody know a workaround to use chrfix? i realy like what it does.

[tasslehoff44]

Hi,

I tried to make a good encoding of a japan anime TS hd file.
I use hardcoded subtitle, trim, resize and some restoring filter.
My encoder gui is staxrip.

I have to deal with vfr and i have some dificulty.

Can you help me ?

[Nightshiver]

Post a sample.

[tasslehoff44]

I don't know how to make a ts sample but if I have the right,
this is here: http://elitistfags.oversubs.org/packs/

on #elitistfags@rizon.net

i make a try on one piece 475

[maki]

I have this m2ts file and I want to know if it's a VFR file or what not.

I demuxed the file and loaded to DGAVCindex. I cut the file short since the file itself is like 4G (DGindex by default has ignore pulldown flags, so I left it as it is and demuxed the file giving me a portion of the h.264 file). This file includes a portion of the beginning and OP of the anime.

http://www.megaupload.com/?d=9255TWVZ

[Nightshiver]

tasslehoff44, you need to upload it to somewhere like Megaupload. I don't have IRC and it's not really worth it for me to install it for the sake of grabbing that 1 file.

Maki, I don't know what the heck you're talking about. Your source is from a BluRay, and it's perfectly 1080 progressive. There's nothing you need to do.

[tasslehoff44]

I cut the opening with FameRing Smart Cutter for DV and DVB v1.3.2, tell me if there is another better tools.
Sample (http://www.megaupload.com/?d=AV21MD50)

[Nightshiver]

Tass, your video isn't VFR and it isn't interlaced. While it does show flags of being 29.97, it's actually progressive. You don't need to do any kind of decimating, IVTC, or de-interlacing.

[tasslehoff44]

Aparently, there is no interlaced frame in the opening, but it's realy an hybrid source, see the frame capture (http://img526.imageshack.us/img526/8753/475w0.jpg).
Mediainfo and other tool report bad info about fps, but media info report that is an interlaced frame.

I do some test and it appear that the opening don't have variable frame rate, but the episode have it.
I upload the encoded opening and episode with the timecode but non muxed: Episode (http://www.megaupload.com/?d=RL75LZ8Z), Opening (http://www.megaupload.com/?d=UPMTI7NN)

[Overdrive80]

Well other case that i have to use this filter. Post update filter for which have problems with errors merge windows. ^^

[Overdrive80]

Hi of new ^^. I update code of this script for use QTMPC and Needi3. I attachment for testing. ^^

[Overdrive80]

Hi if i use animeivtc with hibryd material ntsc with these codes:

AnimeIVTC(mode=4,aa=0,omode=2)

And after use:

AnimeIVTC(mode=4,aa=0,omode=2,i1=3147,e1=40159,i2=42855,e2=44218,pass=2)

The timecodes generated are wrong:

# timecode format v1
Assume 29.970030
# TDecimate v1.0.3 by tritical
# Mode 5 - Auto-generated mkv timecodes file
0,1271,23.976024
1272,1274,17.982018
1275,2430,23.976024
2431,2433,17.982018
2434,22441,23.976024
22442,22444,17.982018
22445,29596,23.976024
29597,29599,17.982018
29600,31327,23.976024
31328,31330,17.982018
31331,35368,23.976024
# vfr stats: 100.00% film 00.00% video
# vfr stats: 44218 - film 0 - video 44218 - total
# vfr stats: longest vid section - 0 frames
# vfr stats: # of detected vid sections - 0

But if i used this script and delete i1=3147 the timecodes are correct, i think.

AnimeIVTC(mode=4,aa=0,omode=2,e1=40159,i1=42855,e2=44218,pass=2)

Timecode generated is:

# timecode format v1
Assume 29.970030
# TDecimate v1.0.3 by tritical
# Mode 5 - Auto-generated mkv timecodes file

# vfr stats: 96.05% film 03.95% video
# vfr stats: 42473 - film 1745 - video 44218 - total
# vfr stats: longest vid section - 100 frames
# vfr stats: # of detected vid sections - 287

Somebody have any idea?

[Nightshiver]

Post the whole video please. And use Megaupload. Meaning, post the whole .vob file you are using to try and encode.

[Overdrive80]

Ok but i think that the code this script is wrong.

eDIT: I go to upload m2v files that is the same because i dont know as splitter vob files.

eDIT: If you do Vfr two passes with TIVTC, the timecodes create is different with respect the timecodes create for animeivtc in mode=4, pass= 2 and specifying interlacing parts.

1 pass

tfm(output="tfm.txt")

tdecimate(mode=4, output="stats.txt")

2 pass

tfm(input="tfm.txt")

tdecimate(mode=5, hybrid=2, vfrdec=1, tfmIn="tfm.txt", input="stats.txt", mkvout="timecodes.txt")

With this script the timecodes says that the material is film in 98.33 %.

But if i use as 2 pass:

animeivtc(mode=4, killcomb=3, aa=0, omode=2, e1=40159,i1=42855,e2=44218, pass=2)

The timecodes create is:

# timecode format v1
Assume 29.970030
# TDecimate v1.0.3 by tritical
# Mode 5 - Auto-generated mkv timecodes file

# vfr stats: 78.21% film 21.79% video
# vfr stats: 34583 - film 9635 - video 44218 - total
# vfr stats: longest vid section - 2190 frames
# vfr stats: # of detected vid sections - 552

[Overdrive80]

Night, I leave de attachment. ^^ Thanks

http://www.megaupload.com/?d=NAYHZB3V

[egrimisu]

Well other case that i have to use this filter. Post update filter for which have problems with errors merge windows. ^^

what probles that this fix solve?

[Overdrive80]

what probles that this fix solve?

Solve the problems that people said, because the script in first post is the original y hasnt modified code.

[egrimisu]

Solve the problems that people said, because the script in first post is the original y hasnt modified code.

does it solve the problem with chrfix=true when using mode=2, the problem that i writed about a few weeks ago. Thanks in advance.

[Overdrive80]

I hope that not, i´m sorry.

[Nightshiver]

I think I've found out your problem, Overdrive. In this incarnation of AnimeIVTC, you are not supposed to use the e1,i1,e2,i2, etc functions for mode=4. Since you are using them, it is screwing up the way VFR detects the interlaced frames.

animeivtc(mode=4,aa=0,omode=2,pass=1)

That should be the first line for using Mode=4. You MUST use pass=1 first. Open MeGUI, load the script, and do a "Analysis Pass". Then, after it finishes, edit the script so that it's:

animeivtc(mode=4,aa=0,omode=2,pass=2)

Reload the changed script into MeGUI and then it will create the timecodes.txt file. When I did this, everything was nice and progressive with no interlaced frames to be found. Here's a little sample of the timecodes.txt that I got when I did it:

# vfr stats: 99.05% film 00.95% video
# vfr stats: 43798 - film 420 - video 44218 - total
# vfr stats: longest vid section - 15 frames
# vfr stats: # of detected vid sections - 68

[Overdrive80]

This option you value, it logical and correct, as i probe at the time. However, you can check yourself if you dont tell to animeivtc the interlacing zones, the ending keep interlacing.

If you can, try pass= 2 with this code:

animeivtc(mode=4, killcomb=3, aa=0, omode=2, e1=40159,i1=42855,e2=44218, pass=2)

The ending that you will see is deinterlacing nice. The timecodes generated is:

# vfr stats: 78.21% film 21.79% video
# vfr stats: 34583 - film 9635 - video 44218 - total
# vfr stats: longest vid section - 2190 frames
# vfr stats: # of detected vid sections - 552

I will upload a result video with this script. Thanks ^^

EDIT: I assumed i1+e1+i2+e2 for the video, and when I indicate to animeivtc i1 then is when the timescodes obtain indicate 100% film, which I thought was a bug of the code. However, if I assumed e1+i1+e2, everything seems normal.

[Overdrive80]

Here i attachment file http://www.megaupload.com/?d=UCDGYPCM

[gnol009]

http://image.hdvnbits.org/graphic/images/2010/December/20/2EA1_4D0F430C.jpg

This is my prob.
How can i fix it?

[Overdrive80]

http://image.hdvnbits.org/graphic/images/2010/December/20/2EA1_4D0F430C.jpg

This is my prob.
How can i fix it?

You no specificied any mode, there are 4 modes.

mode=1, hard telecide
mode=2, field blended
mode=3, double hard telecide
mode=4, material hybrid.

According to the digital material that will work, you must apply a mode or another.

Read documentation, General parameters.

[Gavino]

http://image.hdvnbits.org/graphic/images/2010/December/20/2EA1_4D0F430C.jpg

This is my prob.
How can i fix it?
Read it - I can't imagine a more precise message than that one.

[Ghostlamer]

Hello.
I can not properly remove interlace.

Used "AnimeIVTC(mode=2,ifade=true,aa=0) and AnimeIVTC(mode=3,ifade=true,aa=0, pattern=2)", pattern=2 as the most optimal.

http://screenshotcomparison.com/comparison/15346

Sample: http://www.mediafire.com/?0al7025i98yya
Can be better?, or i dumb?.:)

[manono]

Yadif(Mode=1)#or the bobber of your choice
SRestore(Frate=23.976)

Seems to work OK.

[Overdrive80]

Hello.
I can not properly remove interlace.

Used "AnimeIVTC(mode=2,ifade=true,aa=0) and AnimeIVTC(mode=3,ifade=true,aa=0, pattern=2)", pattern=2 as the most optimal.

http://screenshotcomparison.com/comparison/15346

Sample: http://www.mediafire.com/?0al7025i98yya
Can be better?, or i dumb?.:)
You could try, too.

Tdeint(mode=1).Selecteven().tdecimate()

[Dogway]

Since nobody replied to my question (http://forum.doom9.org/showthread.php?p=1466791#post1466791) I finally decided to use SharpAAMCmod, the result is good enough and easy to use plus it uses both of "my list" filters, toon1.1 and awarpsharp2.
There were a few things I didnt like, so I added eedi3 as an option vs sagnom and also added an option to resize in the same step, so you avoid one resize interpolation in case your final resolution is different.

PD: this is extremely slow, its aimed to quality, not speed.

[Ghostlamer]

You could try, too.

Tdeint(mode=1).Selecteven().tdecimate()
Yadif(Mode=1)#or the bobber of your choice
SRestore(Frate=23.976)

Seems to work OK.
Thanks.
But old AnimeIVTC(mode 2) 1.06 gives much better results with my video than AnimeIVTC 2.00(to many blends(modes 2,3)), and other deinterlacers.

[egrimisu]

Thanks.
But old AnimeIVTC(mode 2) 1.06 gives much better results with my video than AnimeIVTC 2.00(to many blends(modes 2,3)), and other deinterlacers.

Hi, i have checked myself animeivtc 1.06.1 vs 2.00 on your source and yes 1.06.1 does a greater job, but the diference between 1.06.1 and 2 in removing blends is that 1.06.1 uses morestore and 2.00 srestore.

If i know right srestore is the new version of morestore.
i'll try to tweak animeivtc 1.06.1 to use nned3 /eedi3 and qtgmc and no tempgauss_mc.

[egrimisu]

The wierd thing is that:
animeivtc 1.06.1 mode=2 removes some rainbowing without creating bleending(as chubbyrain2,smartssiq,derainbow did on my source) and some dotcrawl and there is no deinterlacing flicker. Mode=2 for animeivtc using bbob=3(default) removes more cleaner the blended frames but leaves a few blended wher 1.06.1 removes those but leaves otherr (for my source).animeivtc 1.06.1 mode=2 makes the image a bit smoother than 2.00

Could someone explain why the in the old version there is no flicker, why are rainbows removed, and why removes better a few blends that in 2.00.

Thanks guys.

[YnK]

I think I'll ask one question... Not that I'm looking for a complete solution :), but rather for some ideas.

I first ran into this little problem when processing an anime of the type "mostly telecined with a little bit of interlaced video" (i.e. 24fps animation with a few completely randomly placed 60fps pans :) ). One panning scene which I could swear wasn't interlaced came out of TIVTC still containing combed frames. I examined the scene and noticed that it has an odd combing pattern which definetely wasn't 3:2 pulldown. I tried to filter it through AnimeIVTC mode 3 (since it would sertainly kill all the combing), but the result was noticeably jerky, as though it missed some random frames.

On close examination, it turned out the thing was actually a botched 30fps pan -- so no wonder it became jerky. So I had to fix it with a bob+selectevery instead. (I didn't simply decide it should be 30fps -- I was just picking the correct frames from a bobbed clip while dropping duplicates, and it ended up at ntsc video rate, so it was intended to be 30fps by the animators.)

So here's my question: is there a way to write a script that could detect and handle such oddball video properly (without unneeded decimation or combing leftovers)? I'm pretty sure there could be other anime dvds that have similar botched 30fps sequences, and on panning scenes, the removal of extra frames becomes very noticeable.

(Incidentally, on some sources, sequences you would probably identify as "double-hard telecine" also become jerky when decimated... it appears that leaving all of the unique frames intact could result in actual 48fps animation, so no surprise here... So does another odd kind of interlacing that is 42fps animation - I saw it used for panning sequences as well. But the botched 30fps pans are worse, because decimation removes frames in such a way you can tell it wasn't supposed to be 24fps.)

[egrimisu]

i look trough the animeivtc script but i couldn't understand when smartfademod() runs , before the ivtc or after? can some one help me? i just realizes that animeivtc 1.06.1 is better for some sources and i wanted to use smartfademod for 1.06.1 manually. thanks

[byme]

chrfix is broken. It has been discussed on previous page.
http://forum.doom9.org/showthread.php?p=1392244#post1392244


remove )#

I downloaded AnimeIVTC dev copy.rar but continues to say there is no function named "TCMCmod"


When I'm using chrfix to filter defect which you can see in pictures

without chrtfix:
http://img100.imageshack.us/img100/1106/moo02b.png

with chrtfix:
http://img63.imageshack.us/img63/3723/moo02a.png

in this script:
AnimeIVTC(mode=1, chrfix=true, aa=1, omode=1, credconv="blend", e1=40368, i1=42723, e2=43469)
crop(8, 0, -8, 0)

#resize
#denoise


how did you run chrfix?

[egrimisu]

nimeIVTC(mode=1, chrfix=true)


I downloaded AnimeIVTC dev copy.rar but continues to say there is no function named "TCMCmod"

alose rename all the tcmcmod with QTGMC in the animeivtc.avsi file.
qtgmc can be downloaded here http://www.mediafire.com/?e5zo2ld67xhdy9b


how did you run chrfix?

Chrfix does not work with mode=2, somekind of confilct with srestore.

[byme]

thanks egrimisu

but I have two errors
if I put the eedi3 in the script:
http://file2.net/110203img2904vg.png.html

without the eedi3 does not find the edeint:
http://file2.net/110203img6593dt.png.html

-----

other two requests:

1) AnimeIVTC not accept the video to 50fps... how can I do?

2) I only need the chrfix, with deinterlaced video. How do I disable all functions of AnimeIVTC, except that the chrfix?

[egrimisu]

put the eedi3.dll in the plugins dir
if your source is 50fps i don't think you need animeivtc, better try qtgmc
unofrtunetly i haven't figured out how to run externally chrfix, but that issue is like dot crawl, so probably you are dealing with dot crawl, try checkmate().

thanks egrimisu

but I have two errors
if I put the eedi3 in the script:
http://file2.net/110203img2904vg.png.html

without the eedi3 does not find the edeint:
http://file2.net/110203img6593dt.png.html

-----

other two requests:

1) AnimeIVTC not accept the video to 50fps... how can I do?

2) I only need the chrfix, with deinterlaced video. How do I disable all functions of AnimeIVTC, except that the chrfix?

[byme]

put the eedi3.dll in the plugins dir
if your source is 50fps i don't think you need animeivtc, better try qtgmc
unofrtunetly i haven't figured out how to run externally chrfix, but that issue is like dot crawl, so probably you are dealing with dot crawl, try checkmate().
I put eedi3.dll in C:\Program Files (x86)\AviSynth 2.5\plugins
but the error remains

checkmate() is not sufficient... and double checkmate() do not like me

qtgmc no need to deinterlace?.. what purpose to use it?

[Nightshiver]

Please post an unprocessed sample.

@egrimisu

chrfix is not a function itself, but is actually a wrapper "function" containing more functions inside of it. From what's in the AnimeIVTC script, chrfix uses a bob mode first, to bob the clip that is about to have its chroma fixed, using either blendbob or clearbob depending on the source. Where the actual "function" of chrfix is per se would be here, in the function chroma_rep:

function chroma_rep(clip source, int "cache")
{
###### parameters & other necessary vars ######
cache = default(cache,-1)


###### source preparation & lut ######
global out1 = source
global out2 = mergechroma(source,mt_adddiff(source.trim(1,0),mt_makediff(source.trim(1,0),source.trim(2,0),y=1,u=3,v=3),y=1,u=3,v=3).blur(1))


###### initialise variables ######
global alfr = -100


###### evaluation call & output calculation ######
scriptclip( source, """

### preparation ###
cfr = current_frame
jmp = alfr+1==cfr

global alfr = cfr

l_v = LumaDifference(out1.trim(1,0),out1.trim(2,0))
c_v = ChromaUDifference(out1.trim(1,0),out1.trim(2,0)) + ChromaVDifference(out1.trim(1,0),out1.trim(2,0))

## luma value shifting ##
l12 = l_v
l01 = jmp ? l12 : l_v
l10 = jmp ? l01 : l_v

## chroma value shifting ##
c12 = c_v
c01 = jmp ? c12 : c_v
c10 = jmp ? c01 : c_v

### condition ###
change = l12*3<l01 && c01*1.5<c12
chango = l01*3<l10 && c10*1.5<c01

change ? out2 : chango ? out1.trim(1,0) : out1

""")


###### final decimation & caching ######
cache<0 ? last : last.RequestLinear(8, cache, 5, false, false)

return last
}

This is the brain of chrfix and of how it actually fixes the chroma fields.

[egrimisu]

Thanks Nightsilver,

shall the function be called before or after ivtc? and what do you recomend for bobbing?

Please post an unprocessed sample.

@egrimisu

chrfix is not a function itself, but is actually a wrapper "function" containing more functions inside of it. From what's in the AnimeIVTC script, chrfix uses a bob mode first, to bob the clip that is about to have its chroma fixed, using either blendbob or clearbob depending on the source. Where the actual "function" of chrfix is per se would be here, in the function chroma_rep:



This is the brain of chrfix and of how it actually fixes the chroma fields.

[Nightshiver]

Well, to see what's being used in AnimeIVTC, just look at the script. For what I just posted, this is being used:

#########
#Bobbing#
#########

#Assign clip to prevent double computation
edeint= i.nnedi2(-2)
emask = i.tmm(1)

#For later NOP() use if blendbob is not needed
usebbob = chrfix && blend || mode==2


#Use non-motion-compensated bobber for blended clips since blends have a negative impact on motion detection (if blends are averaged with frames, they can be harder to detect/remove) (mode=2 and mode=1 with blend and chrfix because blend telecine... well has a lot of blends)
blendbob = usebbob ? bbob<=0 ? i.tdeint(1) :
\ bbob==1 ? i.tdeint(1, edeint=edeint ) :
\ bbob==2 ? i.tdeint(1, edeint=edeint,emask=emask) :
\ bbob==3 ? i.yadifmod(mode=1, edeint=edeint) :
\ i.TGMCmod(edimode=edimode,tr2=degrain,edeint=edeint) : NOP()

#For later NOP() use if clearbob is not needed
usecbob = mode==3 || defined(i1) || chrfix && blend==false

#Allow usage of mo-comp bobber (TempGaussMC) when source is not blended. Highly recommended for stability. (mode=3, interlaced credits, used in chrfix because impact is minimal compared to bob flicker on the whole clip)
clearbob = usecbob ? cbob<=0 ? i.tdeint(1) :
\ cbob==1 ? i.tdeint(1, edeint=edeint) :
\ cbob==2 ? i.tdeint(1, edeint=edeint,emask=emask) :
\ cbob==3 ? i.yadifmod(mode=1, edeint=edeint) :
\ i.TGMCmod(edimode=edimode,tr2=degrain,edeint=edeint) : NOP()


And this is the code for chrfix:


############################
#Chroma swap between fields# chrfix
############################

#Repair double chroma blending if requested by user using xbob depending on source (known by the mode)
chrdeb = chrfix ? blend || mode==2 ? blendbob.chroma_rep(cache) : clearbob.chroma_rep(cache) : nop()

#Remove duplicates (like a safe selecteven/odd) for further processing before final decimation (hybrid decimation (except blend decimation to prevent having 2 calls of tdecimate) or blend telecine IVTC)
#Once bobbed, in mode=1 (tdecimate(1,3,5), in mode=1 + blend = back to 30fps for further deblending, in mode=2 srestore will do the job, in mode=3 tdecimate(1,3,5), in mode 4 VFR needs 30, CFR can do it (personal notes)
debprep = chrfix ? mode==1 && blend || mode==4 && omode==2 ? chrdeb.tdecimate(1,1,2) : chrdeb : i

From the script, chrfix is applied after the IVTC. As you can also see from the script, the functions blendbob and clearbob are also wrappers that contain calls to several different bobbers. Naturally the highest quality bobber would be TGMC, but it is also the slowest. Of course, you could also use QTGMC instead of the regular TGMC for more speed.

[byme]

Please post an unprocessed sample.


http://file2.net/110204img4321pi.png.html

other sample are in this trhead
http://forum.doom9.org/showthread.php?t=159340

with chroma_rep nothing changes
#Chroma swap between fields# chrfix how to use?

[Nightshiver]

I was not talking to you when I mentioned chroma_rep :) If you read my post, that was for egrimisu :) And when I say a sample, I don't mean a picture. That should be common sense. I want VIDEO. Unprocessed video.

[byme]

sorry, but I'm Italian and I know very little English

sample video in YV12
http://www.mediafire.com/?f57gbmaoz6c4kz4

[Nightshiver]

That's not a sample of the anime you posted about above....

[lych_necross]

@Nightshiver

How could one use chrfix by itself (without animeivtc)? Could it be turned into a standalone "wrapper" function?

[egrimisu]

any function that is build in any script that has the avsi extension can be called externally.
For example ediaa is a function built in animeivtc and it stands for aa=1 but can be called externally with ediaa().

@Nightshiver

How could one use chrfix by itself (without animeivtc)? Could it be turned into a standalone "wrapper" function?

[byme]

That's not a sample of the anime you posted about above....

I do not anime, I captured video from TV

but it seemed that the chrfix could do in my case

[Didée]

Rather not. Chrfix is for removing of dotcrawl, with dotcrawl being a very special artifact (crosstalk between luma & chroma carriers in an analog transmission).

The various examples you posted suffer from many different kinds of illness/misfiltering. But they don't show dotcrawl.

[byme]

Rather not. Chrfix is for removing of dotcrawl, with dotcrawl being a very special artifact (crosstalk between luma & chroma carriers in an analog transmission).

The various examples you posted suffer from many different kinds of illness/misfiltering. But they don't show dotcrawl.

that filters I then use for my problem?

:script::thanks:

[darktyron]

I'm trying to deal with Kiki's delivering Service DVD R2, but after AnimeIVTC, result is weird... It's like a ghost residual or something like that.

http://img560.imageshack.us/i/shot1.png/
http://img11.imageshack.us/i/shot2lj.png/
http://img265.imageshack.us/i/shot3i.png/

Mpeg2Source("kiki delivering service.d2v")
AnimeIVTC(mode=1, aa=2)

Source is NTSC (DgIndex told me that) and it's 3:2 pattern for sure (I think).

How do I get rid of that ghost effect? Maybe it's from source?

:thanks:

What's that artifact name? Maybe it's Blending?

---
Edit: I've looked at the source and the artifact is in progressive frames... It's from source... How can I repair it?

[byme]

out of curiosity... you could create a script with only Chrfix? (even if the master Didée says he does not need in my case)

[Mottodj]

Hi everybody, how should i use AnimeIVTC with http://www.mediafire.com/?89dwu9qfa433905 ? R2DVD but making lots of blended frames with any mode or options.

[egrimisu]

Hi everybody, how should i use AnimeIVTC with http://www.mediafire.com/?89dwu9qfa433905 ? R2DVD but making lots of blended frames with any mode or options.

i'm a fraid the transfer was made several years ago and the equipment used weren't that good, this is typical for some old R2 japaneze anime, take maison ikkoku or reukin kenshin, or other. Not all old transfer are like this one, for example magical ema :) best way is to use mode=2, i'll stick to that if i were you.

You could try animeivtc 1.06, it uses morestore instead of srestore and some blends are better removed but other left behind.

[Emulgator]

Maybe somebody has already modified AIVTC to decimate from 25i to 24p ?
If nobody has done so already, I would then have to take up the task of modifying AIVTC.

The source: clips from 24p film stock, hardtelecined with a Rota-Prism scanner
and on-the-fly encoded to MPEG-2 PAL 25i.

AnimeIVTC already did a great job in restoring these particular clips.
and it seemed to be able to get out all frames.
But after comparing the result to a properly preared master indeed roughly every 1000th frame is missing.

Of course I can try to find the holes and interpolate these missing frames,
but a better solution is preferred.

After IVTC I applied Assumefps(25).
Encoded as PAL 25p over 25i.
(Playback as PAL speedup (+25/24 speed = +4,16%) is intended.)

In AIVTC I have to set normconv=true to decimate a PAL source.
This leads to AIVTC always decimating to 23.976fps as output framerate.
I can not specify 24fps.
So the missing frames seem to have an explanation here.

I would be thankful for suggestions where to modify the AIVTC script
before I attempt to try and probably invent the wheel again...

[Didée]

It's all about Srestore, obviously.

This line in AnimeIVTC
frate = Ofrate==1 ? 23.976 : 25
changed to
frate = Ofrate==1 ? 24.000: 25
might do it.

It's line 41 in the AIVTC script I have, though I'm not sure if I have the latest one. (Since I never need or use it.)

[Emulgator]

Many thanks, Didée !
This helped so fast, missing frames are back.
And I was about introducing normconv as float and fitting this into the script... later.

[Boulder]

Is there any good way to repair this video: http://www.mediafire.com/?8dtrwfauw3o63ot ?

Using AnimeIVTC(mode=2) produces a better looking clip but still there are artefacts. Looking at the source, the pattern is not constant so it's difficult to determine what to do with it.

[r42]

http://www.mediafire.com/?w85n8a1ygzybt7a

How AnimeIVTC can help me with this?

[ianken]

Well, I got my MT/SetMTMode issues sorted out...

To test mode 4 (hybrid) I dusted off a known hybrid title (avatar- last airbender) and ran it through a very simple script. The result was a file that was un-smooth and a timecodes file that was essentially empty with the exception of a single entry for a single film-rate segment. A plain TFM/TDecimate run results in a much larger timecodes file and much smoother video playback.

What am I not getting right here?

[K.A.S]

I have some trouble with AnimeIVTC

this is my source:
http://demo.ovh.com/download/c0cb5792d006ac6945c5426510461e3f/Toriko%27s%20END.ts

and my script is:
Import("C:\Program Files (x86)\AviSynth 2.5\plugins\srestore.avs")
DGDecode_Mpeg2Source("C:\KAS Files\Toriko\ep16\Toriko-16.d2v")
trim(43793,45590)

AnimeIVTC (1,aa=4,precision=3,killcomb=2,ifade=true,chrfix=true,blend=true,dark=1,thin=10,sharp=150,smooth=-1,stabilize=true,tradius=3,aapel=4,aaov=8,aablk=16)
Spline36Resize(1280,720,2.5,1.5,-2.5,-1)

help me if something is incorrect

[Leinad4Mind]

How can I deinterlace this kind of TVRip Source?

http://www.mediafire.com/?jseydgmb6t4qcom

I've tried:

tfm(clip2=nnedi3)

edeintted = nnedi3(field=-2)
TDeint(edeint=edeintted,order=-1,mode=2,sharp=true,mtnmode=3,full=false,tryWeave=false,type=1,emask=TMM(mode=1))

interp = nnedi3(field=1)
tdeint(mode=0,order=1,field=1,edeint=interp,slow=2,emask=TMM(mode=0,order=1,field=1))


QTGMC( Preset="Slower")
#With and Without SRestore...
SRestore(25).AssumeFPS(25000,1000)

I haven't try with AnimeIVTC...

If anyone could deinterlace this properly, I would appreciate.

Thanks!

[manono]

The first three scripts won't work because it's field-blended. The fourth is peculiar because you're having SRestore unblend it to 25fps before slowing it to film speed, but since it was converted from 24fps, that's incorrect also. Try the default:

QTGMC( Preset="Slower")
SRestore()

There is still leftover blending, and maybe someone else has a fix for that, but the blends are often in consecutive fields and I don't know how to fix that entirely. I doubt AnimeIVTC can fix it, but never having used it myself I can't be sure of that.

[aNToK]

Is there anywhere to download Set's rewrite of awarpsharp2? The link in the package isn't working and I haven't been able to find it anywhere

[cretindesalpes]

aWarpSharp2 (http://www.mediafire.com/?6d2bw7xd0ehzchy)

[aNToK]

aWarpSharp2 (http://www.mediafire.com/?6d2bw7xd0ehzchy)

Ah, thanks so much cretindesalpes! I've been looking forward to trying this on an old 70's anime dvd set I bought that I want to improve to put up on my widescreen!

[aNToK]

Hmm... I've got all the requirements in place but while trying to run a first pass I get this error:

I don't know what "sfthr" means
(AnimeIVTC 2.00.avsi, line 151)

The series reads as hybrid (mostly film) in megui's analyzer and mediainfo puts it as 29.97 ntsc TFF interlaced. Any thoughts on what I'm missing?


Script: DGDecode_mpeg2source("E:\gatchaman workspace\GATCHAMAN_VOL_5\VIDEO_TS\Ep 25.d2v", info=3)

animeivtc(mode=4,ifade=true,killcomb=3,aa=4,smooth=-1,stabilize=true,aapel=2,omode=1,credconv="mocomp",search=4,nnedi2pel=true,pass=1)

[Nightshiver]

Please upload your copies of AnimeIVTC and the filters required to run animeivtc. I think you may have the wrong version of certain filters and/or AnimeIVTC v2.00\

sfthr is built into the function smartfade inside of AnimeIVTC, and smartfade has been activated in your script because you are using mode=4 and ifade. Please also post a short, unprocessed video sample.

[aNToK]

Will do. I went through both the 1.06 and the 2.0 versions with different plugin folders and neither worked. I'll up them when I get back from Christmas with the family. How long a clip would be needed?

[Gavino]

sfthr is built into the function smartfade inside of AnimeIVTC, and smartfade has been activated in your script because you are using mode=4 and ifade.
However, sfthr is passed into smartfade the first time it is called, without having been previously set:
usefm = mode==1 || (mode==4 && (pass==2 || creds || omode==1))
...
fm = usefm ? ifade ? fm.smartfademod(sfthr, sfshow, cache) : fm : nop()
(from AnimeIVTC 2.00.avsi)

[Undead Sega]

Hi guys, hope you've had a Merry Christmas and a Happy New Year.

Just a quick question, can someone remind me how to do a 2:3 pulldown removal? Meaning that 24p source with added pulldown (2:3 if I'm correct?) to convert it to 25p.

[manono]

Does this have anything to do with AnimeIVTC? Anyway, it's not 2:3 pulldown you want to remove. 2:3 (or 3:2) pulldown takes a progressively encoded 23.976fps video and adds the TFF/RFF flags so that it outputs interlaced 29.97fps. If you're sure your source is 24fps converted to 25fps by the addition of two duplicate fields every second, then one way (using TIVTC) to remove the dupe fields is:

TFM()
TDecimate(Mode=0,Cycle=25,CycleR=1)

After TFM does the field matching, TDecimate removes that (now) one duplicate frame in every 25 frames. If that doesn't work for some reason, please make a short sample available.

And in case I've misunderstood the question, if it really is an NTSC source with hard pulldown, then:

TFM().TDecimate()

returns it to 23.976fps. To speed it up to 25fps (why?) add:

AssumeFPS(25)

Or, if making this for DVD and you want to keep the 24fps within a PAL DVD, after encoding for progressive 23.976fps, run the resulting MPV through DGPulldown with the custom box checked and 23.976->25.

[Undead Sega]

Well, just to be on the safe side, here is a sample of what I was given (looks familiar? :D):

http://www.filesonic.com/file/OHGPOiH

I'm pretty sure it is what I've described, cause it was from a PAL VHS tape, and he animation is of course made at 24 frames per second, plus I'm quite certain as well that this is hard telecined too :D

[manono]

Yes, it's as I said in my first script. TFM() used alone does the field matching after which you'll find one duplicate frame in every 25. After adding TDecimate(Mode=0,Cycle=25,CycleR=1), all the dupe frames are removed, leaving you with a 24fps video.

Oh, since it's animation you may want to use TDecimate(Mode=1,Cycle=25,CycleR=1) for those cases when the animation is drawn at less than 24fps. The TDecimate doc explains more about Mode=1.

[Gerry62]

However, sfthr is passed into smartfade the first time it is called, without having been previously set:
usefm = mode==1 || (mode==4 && (pass==2 || creds || omode==1))
...
fm = usefm ? ifade ? fm.smartfademod(sfthr, sfshow, cache) : fm : nop()
(from AnimeIVTC 2.00.avsi)

Ive just comeup with the same issue, unzipped every dll from the pack and started afresh, still same error this chap has, has anyone solved it yet?


Oh I may be stupid but does this process keep the interlace? I know im creating a 23.976 Progressive file in the end, BUT does it blend or something the Feilds, I need both feilds you see, More definition you see..!

[Gerry62]

Well, just to be on the safe side, here is a sample of what I was given (looks familiar? :D):

http://www.filesonic.com/file/OHGPOiH

I'm pretty sure it is what I've described, cause it was from a PAL VHS tape, and he animation is of course made at 24 frames per second, plus I'm quite certain as well that this is hard telecined too :D

Yes it does look familiar!
I'm currently Restoring it for Mr Gilchrist, Cleaning By hand, Frame by frame!

[Gerry62]

function smartfademod(clip c,
As you may have noticed, this option somehow ended up untested after being modified and I was unable to do so until now. Sorry for the trouble.

To eplain why i dredged this up..
If you want to use 'ifade=True' option then you need to have a look over some posts over at page 40ish for some corrected code and change your avsi script, after doing that a couple of times it then comes with the error

I Don't Know What "C" Means
Line 504 abs(avg) > sfthr ? interleave(sep.selecteven(),sep.selectodd()).weave().daa() : last")
Line 153 fm = usefm ? ifade ? fm.smartfademod(sfthr, sfshow, cache) : fm : nop()


So I take it we can't use smartfade then?
Anyone found a fix for the smartfade sfthr prob yet?

[Gavino]

Anyone found a fix for the smartfade sfthr prob yet?
I don't use AnimeIVTC, so haven't tested this, but from looking at the code of function 'smartfademod', I believe it should be changed to this:
function smartfademod(clip c, float "sfthr", bool "sfshow", int "cache") {

global sfthr = default ( sfthr, 0.4 ) # Threshold for fade detection.
show = default ( sfshow, false ) # Displays luma difference between fields without processing anything.
cache = default ( cache, 10 ) # Cache the frames to help seeking and other functions requesting many frames

show ? scriptclip(c, "subtitle(string(abs( separatefields().selectodd().averageluma() \
- separatefields().selecteven().averageluma())))") : \
scriptclip(c, "sep = separatefields() \
...

EDIT: In fact, I see that's exactly what thetoof was saying in posts #880 and #882.

[Gerry62]

I don't use AnimeIVTC, so haven't tested this,

EDIT: In fact, I see that's exactly what thetoof was saying in posts #880 and #882.

Opps, Missed that one, thetoof might want to make a Ver2.1!

Thanks for the help! and out of interest what would you use?

[Gavino]

out of interest what would you use?
The reason I don't use it is simply that I rarely work with anime or sources that require IVTC.
So I'm not in a position to say if there is anything better.

[parent5446]

Is there a script similar to this that deals with non-anime sources?

[Emulgator]

AIVTC will handle other content too.

[parent5446]

Awesome. I was kind of thinking it could be used for other stuff considering the filter is fairly generic enough, but I'm nowhere near an expert enough at Avisynth to know that for sure. One more quick question: what will happen if you use AnimeIVTC on a pure interlaced source? Will it hurt the quality of the source since technically it's not telecined at all?

[Emulgator]

It depends on the nature (generation, conversion) of your source and what you need to achieve (final delivery format).

As the name AnimeIVTC suggests, inverse telecine is the goal.
You want to have the progressive frames back as they were shot.

Generic interlaced footage coming unaltered from a video camera will not profit from IVTCing.

A proper Film 23.976p on NTSC 29.97i hard pulldown can easily be reverted without AIVTC.
SelectEvery() is your friend.

But if the pulldown pattern changes, shifts, capturing missed something, normconversion was underneath,
then AIVTC comes in handy. And in hardest cases, YATTA...

And there is a PAL species of pure interlaced material that came from progressive frames.
24p Film that had been scanned with a Rota-prism scanner on PAL25i is such a case.
This can be IVTCed with AIVTC.

If 24p on 25i hard Euro-pulldowned with fixed pattern shall go back to 24p,
then SelectEvery(...) may be better.

No matter if NTSC or PAL: Find out the nature of your source.

AssumeTFF().SeparateFields().Spline36Resize(width, height)
AssumeBFF().SeparateFields().Spline36Resize(width, height)

Step through the fields. Don't give up.
Start, end, scenecuts
Do this in every place, even if it takes hours.
(Sometimes I had so scrambled material to restore,
it took more than just hours to assess field cadence throughout the clips.)

[Rheffera]

I have read through the readme but cant get this script to work. all i get is Aviscript error:"Mistake in i1,e1 parameters..."

AnimeIVTC(mode=4,omode=2,pass=1,killcomb=1,e1=43369,i1=45328)

as far as i know this is correct syntax.

even if i set e1=1,i1=2 it still gives me the error. Any idea?

i havent touched the script. all i have done is load dependencies. Im using the latest stable avisynth.

The clip in question is definetley film and then interlaced for the credits. DecomIVTC mad the content look good, but the credits a jerky interlaced mess. Plain yadif made the credits look great, but the content jerky. Its clear to me that it is hybrid.

[frenshprince]

Hi there,

First message on this forum, but longtime reader, hello everyone :)

I'm finally posting today, beacuse I dont understand what I'm doing wrong.

I have a telecine NTSC video 29.97, that I would like to IVTC.
In order to obtain a PAL 720*576 23.976 vidéo.

I used the TIVTC plugins, with one of this script
tfm(order=0).tdecimate(mode=1)
or
tfm(order=0).tdecimate(mode=1,hybrid=1)
And the result is almost perfect.

Almost, because there are still some frames whiche are not deinterlaced.
And its not good to see.

Here's some pics of what I mean
http://desmond.imageshack.us/Himg407/scaled.php?server=407&filename=loki1.jpg&res=medium
http://desmond.imageshack.us/Himg72/scaled.php?server=72&filename=loki2.jpg&res=medium
http://desmond.imageshack.us/Himg651/scaled.php?server=651&filename=loki3.jpg&res=medium


Ive just tried anime ivtc, and noting is working but a simple script as "AnimeIVTC(mode=4,omode=1)

I could post a part of the video, if lt's needed.




Your help would be really appreciated :o

Thanks

[manono]

Got a small sample that shows the problem when IVTC'd?

[Bloax]

Yeah, I know I'm useful as hell.

But that reminds me a lot of what happens if you tried to deinterlace Polavision (http://forum.doom9.org/showthread.php?t=163574).

[frenshprince]

I tried with ifade= true, and it fixed the issue :D

But now I'm trying the mode 2
AnimeIVTC(mode=2,bbob=4,normconv=true)

and I have this error "there is no function named "srestore".

What I a m doing wrong ?

[Emulgator]

Download srestore and put it in your plugins folder.

[frenshprince]

It works !

Thanks very much ;)

[Undead Sega]

Yes it does look familiar!
I'm currently Restoring it for Mr Gilchrist, Cleaning By hand, Frame by frame!

That means you're Ade Gregg right?!?!? :D

[ENAJonas]

Can someone please weigh in on my post http://forum.doom9.org/showthread.php?t=164267 ? Is this all I need to do?

[TheFluff]

animeivtc is most likely much too slow to run in realtime, so it won't be of much use to you.

[horrormaster34]

anybody know what the requirements are for just SharpAAMCmod?

[GMJCZP]

anybody know what the requirements are for just SharpAAMCmod?

horrormaster34, you should seek the help that is included in the package animeivtc (see first post).

Now I tell them my question:

In help file of AnimeIVTC recommend avoiding SetMTMode instead recommend something that I modified as follows:

SetMemoryMax (512)
# SetMTMode (5)
mpeg2source ("Cap.d2v", cpu = 4, moderate_h = 20, moderate_v = 40)
# SetMTMode (2)

s = AnimeIVTC (mode = 2, bbob = 4). YLevelsC (0, 0.75, 245, 0, 255)
threads = 2
partition = int (framecount (s) / threads)
a = trim (s, 0, partition)
b = trim (s, partition +1,0)
a # load this script in vdub and save as rendering1, Then change a to b and save as rendering2, ideally on a different HDD ... etc etc

Then I put the two pieces in VirtualDubMod simultaneously to process, this would not damage my hard drive?
I have a single hard disk with two partitions.

[Mr. Keska]

thanks . . .


^_^

[GMJCZP]

Then I put the two pieces in VirtualDubMod simultaneously to process, this would not damage my hard drive?
I have a single hard disk with two partitions.

There's someone I can clarify my doubt?

[Boulder]

Why would it damage your HD? In normal use, there are reads and writes all the time even though you may not notice it. Having one HD and performing multiple HD-intensive operations simultaneously only means that they get slower as the drive heads need to seek more.

[GMJCZP]

Thank you, Boulder

[eriler]

Hi,
I'm having several problems with using AnimeIVTC - It simply just doesn't allow some parameters to be set.
Firstly, when trying to set the i1, e1, p1, etc. parameter it doesn't allow me to set anything if i don't have e1 somewher ein the script. i.e. "p1=40407, i1=43279, p2=44096" doesn't work, however "p1=40407, e1=43279, p2=44096" works. Same if you replace the p's with i's. I get the error message "Mistake in entering i1, e1..... parameters. Please check."

Secondly, "ifade=true" does not work, i get the error "I don't know what "sfthr" means". I made the changes to AnimeIVTC 2.00.avsi as explained in post #1169.

These are just some errors which i get when i change settings around. The current avs line i have now is as follows (This one works):
AnimeIVTC (mode=1, ifade=false, killcomb=3, aa=0, p1=40407, e1=43279, p2=44096, omode=2, pass=1)

The video im encoding has progressive video for the intro, main part and for the preview (Anime), and i think it has interlaced for the ending song (Video of real people).

[Nightshiver]

Hi,
I'm having several problems with using AnimeIVTC - It simply just doesn't allow some parameters to be set.
Firstly, when trying to set the i1, e1, p1, etc. parameter it doesn't allow me to set anything if i don't have e1 somewher ein the script. i.e. "p1=40407, i1=43279, p2=44096" doesn't work, however "p1=40407, e1=43279, p2=44096" works. Same if you replace the p's with i's. I get the error message "Mistake in entering i1, e1..... parameters. Please check."


Well yeah, that's how it's supposed to work. You can't just pick and choose, it has to be a complete section. Can't have e2 without an e1 first and vice versa.

[eriler]

Well yeah, that's how it's supposed to work. You can't just pick and choose, it has to be a complete section. Can't have e2 without an e1 first and vice versa.

I think you've misunderstood; I don't want to have an e1 or e2 at all. I want it to be p1, i1, p2. But that doesn't work.

[raido]

I installed animeivtc 2.00 and it keeps asking for an output mode. I'm assuming I want omode=1 for most purposes? What is omode=2 for? When would I need vfr?

[x265]

Should i use AnimeIVTC or TIVTC on anime DVD's?

[Guest]

There's no universal answer for all videos. You have to base your decision in each case on the character of the specific video.

[x265]

http://img46.imageshack.us/img46/2214/capturelpr.png << What have i done wrong?

[vdcrim]

Off-topic, but you can copy the error as text in Edit -> Miscellaneous -> Copy AviSynth error to clipboard.

[Overdrive80]

eedi3 not is a option in official version, in mod version yes, but this version has issues.

[x265]

Ok, so which anti-aliasing script should i use for anime(SD)?

[Overdrive80]

LSharpAAF (http://forum.doom9.org/showthread.php?t=153835) agglutinates enough anti-aliasings. You should use that offer you better look according your source.

[x265]

Will there be any future updates for this filter?

[Guest]

@x265

You seem to like posting this question. It would be much more valuable for everybody if you would explain what you think is required that would justify an update. If it ain't broke, don't fix it.

[sadsack]

any chance of a new upload containing all the required plugins? The upload on the first page expired.

[Max_Cady]

Hi guys,

I'm dealing with my claymore Japanese dvds.Looks like it's most of hard-telecined and ending credits pure interlaced.

Sample:
http://www.multiupload.nl/V7ETLCMK3Q

Can't find proper solution.I'd appreciate any feedback.TIA!

[Overdrive80]

any chance of a new upload containing all the required plugins? The upload on the first page expired.

Mirror: http://db.tt/gdrf1hRP

[mp3dom]

Hi guys,
I'm dealing with my claymore Japanese dvds.Looks like it's most of hard-telecined and ending credits pure interlaced.


Claymore is VFR.
Opening/Endings, excluding credits texts, are a mixture of 23.98p and 59.94i. The anime itself is animated at 30p.

[Max_Cady]

-Nevermind

[faxmactor]

I don't know what "usemocompi" means (AnimeIVTC 2.00avsi, line 218)

Somehow lines got swapped. Use this: http://www.mediafire.com/?z0mdtwyoefn

I also get this, but MF link is now (long) gone.

I tried to move up usemocompi definition in the script, but the script drops:

Invalid arguments to function "nnedi2_rpow2"

error.

Could anyone help?

[real.finder]

I don't know what "usemocompi" means (AnimeIVTC 2.00avsi, line 218)



I also get this, but MF link is now (long) gone.

I tried to move up usemocompi definition in the script, but the script drops:

Invalid arguments to function "nnedi2_rpow2"

error.

Could anyone help?

try use this link 1 (http://www.mediafire.com/download/0wrsyae0wn9tyf9/AnimeIVTC+2+mod.avs) or link 2 (http://forum.doom9.org/showthread.php?p=1673928#post1673928)

but use it with nnedi3pel instead of nnedi2pel

--------


edit: now chrfix work in mode 2 + other updates mentions in script :)

edit2: some note about iFade and Anti-aliasing and other remnants Interlaced, here (http://forum.doom9.org/showthread.php?p=1639679), so try not to use iFade and aa in AnimeIVTC

for Anti-aliasing use maa2 (http://forum.doom9.org/showthread.php?t=168315) or maa2mod (http://pastebin.com/fqR4CMbV)

and now you can use an External bobber (http://pastebin.com/UaSEtX8i), Details inside the script, EX:- AnimeIVTC(2, bbob=5, extbob=MC_A_bob, mode22=true)

[osgZach]

Does anyone have an active link for the Requirements package?

MediaFire seems to have removed it for "Terms of Service Violation" or some other such nonsense..

[lansing]

Does anyone have an active link for the Requirements package?

MediaFire seems to have removed it for "Terms of Service Violation" or some other such nonsense..

Here (http://www.mediafire.com/download/go4earhv7ju3btc/AnimeIVTC_Requeriments_v2.7z)

I uploaded into my mediafire account, it should stay available as long as I'm still active.

[osgZach]

Thanks lansing !

[IRCUSAGE]

wuzzup all
does anyone have the images mentioned in this section with the settings used on them

# Examples:
Double Hard Telecine - At first, your source may look like a typical Hard Telecined one, but after a more thorough inspection, you'll see that it's actually a DHT one: same source, different high motion scene.
Interlaced on top of telecined (can be hard or double hard) - In this case, it was a hard telecined source. Image
Obviously, since there's no pattern in DHT and interlacing on every frame with pure interlaced material, regular IVTC will output blends and jerkiness.

[StainlessS]

The images that IRCUSAGE is referring to are linked in the opening post, but are 404, not found.

EDIT: Quick look on WayBack archive found nothing.

[IRCUSAGE]

The images that IRCUSAGE is referring to are linked in the opening post, but are 404, not found.

Correct!!!! :D

[bruno321]

I've an NTSC source which has field blends, where QTGMC().srestore(23.976) works generally well, but there are some sequences of frames where srestore doesn't quite get it right and doesn't select the correct frames, so a frame is skipped and instead there's a sequence "good frame - blended frame".

I've tried AnimeIVTC(mode=2,bbob=4) and it seems to be working quite great with it. I haven't found such problems thus far.

My question is, what is that AnimeIVTC line doing that's different from the above QTGMC+srestore? I tried to look at the code but it's too advanced for me. I understand QTGMC is called (from bbob=4) and srestore is used as well, but somehow it's doing something different, I'd like to understand what's going on.

[real.finder]

I've an NTSC source which has field blends, where QTGMC().srestore(23.976) works generally well, but there are some sequences of frames where srestore doesn't quite get it right and doesn't select the correct frames, so a frame is skipped and instead there's a sequence "good frame - blended frame".

I've tried AnimeIVTC(mode=2,bbob=4) and it seems to be working quite great with it. I haven't found such problems thus far.

My question is, what is that AnimeIVTC line doing that's different from the above QTGMC+srestore? I tried to look at the code but it's too advanced for me. I understand QTGMC is called (from bbob=4) and srestore is used as well, but somehow it's doing something different, I'd like to understand what's going on.

it use srestore with ", dclip=i.bob(-0.2,0.6).reduceflicker(strength=1)"

[AdamCarter]

[QUOTE=real.finder;1926150]

Hi real.finder. Been using your animeivtc it's awesome on 95% of stuff i throw at it.

I am having an issue with pole position UK. a USA broadcast on PAL UK dvd.

I have chrfix=true which worked on the first 2 episodes, but not the third. I have a few areas where the chroma of one frame shows up as a faded version on the next frame so it looks like a blend.


this is script
AnimeIVTC(mode=2, bbob=4, chrfix=true, normconv=true)
crop(10,4,-16,-4)
including a link to the source .vob

https://www.swisstransfer.com/d/009c1169-b66f-4fb0-aef2-6f1c21663179

Thank you.

[real.finder]

Hi real.finder. Been using your animeivtc it's awesome on 95% of stuff i throw at it.

I am having an issue with pole position UK. a USA broadcast on PAL UK dvd.

I have chrfix=true which worked on the first 2 episodes, but not the third. I have a few areas where the chroma of one frame shows up as a faded version on the next frame so it looks like a blend.


this is script
AnimeIVTC(mode=2, bbob=4, chrfix=true, normconv=true)
crop(10,4,-16,-4)
including a link to the source .vob

https://www.swisstransfer.com/d/009c1169-b66f-4fb0-aef2-6f1c21663179

Thank you.

your source dont need chrfix from what I saw
maybe you can try it with mode22 like this
AnimeIVTC(mode=2, bbob=4, normconv=true, mode22=1)