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

[osgZach]

yeah I was looking over TFM the other day and came across the clip2 thing, kind of getting the gist of your guide, but I'll probably have to read it a couple times :p

[thetoof]

Hi toof,

You shall consider finding how to deinterlance mixes of hardtelecine and DHT, i have a lot of anime with this problem that wait for some kind of miracle.
I lack the knowledge to write a detection function that would apply standard ivtc when 2 subsequent frames or less are interlaced and bob+decimate on sections where there's x more.

Unless using manual tweakage (long and you might miss some), the option I see (that is currently implemented) is applying DHT IVTC to the whole clip, with everything it implies (slow, bobbing of progressive frames, etc etc)

[ShocWave]

Is there a way to get AnimeIVTC to work on a 1080i source? I get this:


Avisynth open failure:
GetFrameBuffer: Returned a VFB with a 0 data pointer!
size=12441664, max=43157568, used=153752832
I think we have run out of memory folks!
(AnimeIVTC_-_v1.06.1.avsi, line 426)
(AnimeIVTC_-_v1.06.1.avsi, line 366)


It works fine on lower resolution sources, or if I resized the source prior to applying the script (which I shouldn't be doing, I know, but it was just to see what the max resolution I can get before the error comes up).
I have 4GB ram running on Windows 7 x64. Even at 720i resolution, Avisynth doesn't seem to use more than a few hundred megabytes of memory. Line 426 seems to be a AA function. Any idea what could be going wrong?

[Codex0nz]

Are you using all of the correct required files?

Are you using the 1.06.1 file from the first post?

[Nightshiver]

That's not the problem. He's using to much memory. Use Setmemorymax() to increase the alloted memory.

[ShocWave]

I forgot to mention that I already tried setmemorymax(), and get the same error message if I set it to 3gb or 4gb.

I'm pretty sure most people will get the same error if you did something like put LanczosResize(1920,1080) before AnimeIVTC().

[Nightshiver]

Why would you resize before using animeivtc? Don't do that.

[ShocWave]

I'm not resizing. If you read the post above, I said you can reproduce the out of memory error by having a source with high resolutions.

My script consists of one line:

AnimeIVTC(Mode=1,aa=4)

[thetoof]

use stabilize=false

unless the edges are flickering/trembling, you don't need it. It should greatly reduce memory usage by disabling mvtools' temporal averaging.

[ShocWave]

use stabilize=false

unless the edges are flickering/trembling, you don't need it. It should greatly reduce memory usage by disabling mvtools' temporal averaging.

Thanks for the tip! But sadly, I still get that error message.
I just bought DG NV Tools, but haven't recieved my license file yet. I wonder if that would help.

[thetoof]

Oh! Looking at those lines, I see where the problem is. Stupid noobish coding. It is fixed in the new version I've been working on and off (mostly very off).

Try replacing this
aa0=dec
aa1=dec.ediaa()
aa2=dec.daa()
aa3=dec.maa()
aa4=dec.sharpaamcmod(aastr,dark,thin,sharp,smooth,stabilize,MT,tradius,aapel,aaov,aablk,aatype)

#Prepare clip for aa comparison
compaa = interleave( aa0.subtitle("aa=0",align=9),
\ aa1.subtitle("aa=1",align=9),
\ aa2.subtitle("aa=2",align=9),
\ aa3.subtitle("aa=3",align=9),
\ aa4.subtitle("aa=4\n"+"\n"+"Tweak aastr, dark, thin, sharp and smooth for better results.\n"+"Other more advanced options are also available.",lsp=1,align=9))

#Apply the desired aa method or output aa comparison
aaed = (aa==0) ? aa0 : (aa==1) ? aa1 : (aa==2) ? aa2 : (aa==3) ? aa3 : (aa==4) ? aa4 : compaa.subtitle("You are now comparing the anti-aliasing modes. \n"+"Please choose the one that looks better with your source.",lsp=1)

by thisaaed = defined(aa)==false ? interleave( dec. subtitle("aa=0",align=9),
\ dec.ediaa(). subtitle("aa=1",align=9),
\ dec.daa(). subtitle("aa=2",align=9),
\ dec.maa(). subtitle("aa=3",align=9),
\ dec.sharpaamcmod(aastr,dark,thin,sharp,smooth,stabilize,tradius,aapel,aaov,aablk,aatype).subtitle("aa=4",align=9)) :
\ aa==0 ? dec aa==1 ? dec.ediaa() : aa==2 ? dec.daa() : aa==3 ? dec.maa() : aa==4 ? dec.sharpaamcmod(aastr,dark,thin,sharp,smooth,stabilize,tradius,aapel,aaov,aablk,aatype) : nop()
and tell me what you get...

[ShocWave]

Avisynth open failure:
Script error: expected ':'
(AnimeIVTC_-_v1.06.1.avsi, line 368, column 16)


The line in question:
\ aa==0 ? dec aa==1 ? dec.ediaa() : aa==2 ? dec.daa() : aa==3 ? dec.maa() : aa==4 ? dec.sharpaamcmod(aastr,dark,thin,sharp,smooth,stabilize,tradius,aapel,aaov,aablk,aatype) : nop()

I added a ":" after \ aa==0 ? dec, and I get this when using AA=4:


Avisynth open failure:
Script error: Invalid arguments to function: 'sharpaamcmod'
(AnimeIVTC_-_v1.06.1.avsi, line 368)


AA=1,2 works! AA=3 gives me out of memory.

[thetoof]

Oh, right. This is because I removed the MT argument in aa=4

Try replacing the whole sharpaamcmod by this:

function SharpAAMCmod(clip orig, int "aastr", float "dark", int "thin", int "sharp", int "smooth", bool "stabilize", int "tradius", int "aapel", int "aaov", int "aablk", string "aatype")
{
aastr = default ( aastr, 255 ) # antialiasing strength
dark = default ( dark, 0.2 ) # strokes darkening strength
thin = default ( thin, 20 ) # Presharpening
sharp = default ( sharp, 150 ) # Postsharpening
smooth = default ( smooth, -1 ) # Postsmoothing
stabilize= default ( stabilize,false ) # Use post stabilization with Motion Compensation
Tradius = default ( Tradius, 2 ) # 2 = MDegrain2 / 3 = MDegrain3
aapel = default ( aapel, 1 ) # accuracy of the motion estimation (Value can only be 1, 2 or 4. 1 means a precision to the pixel. 2 means a precision to half a pixel, 4 means a precision to quarter a pixel, produced by spatial interpolation (better but slower).)
aaov = default ( aaov, 4 ) # block overlap value (horizontal). Must be even and less than block size. (Higher = more precise & slower)
aablk = default ( aablk, 8 ) # Size of a block (horizontal). It's either 4, 8 or 16 ( default is 8 ). Larger blocks are less sensitive to noise, are faster, but also less accurate.
aatype = default ( aatype, "Sangnom" ) # Use Sangnom() or EEDI2() for anti-aliasing

w=width(orig)
h=height(orig)
m=mt_logic( orig.DEdgeMask(0,255,0,255,"5 10 5 0 0 0 -5 -10 -5", divisor=4,Y=3,U=3,V=3)
\ ,orig.DEdgeMask(0,255,0,255,"5 0 -5 10 0 -10 5 0 -5", divisor=4,Y=3,U=3,V=3)
\ ,"max").mt_lut("x 128 / 0.86 ^ 255 *")
preaa=(thin==0 && dark==0) ? orig : (thin==0) ? orig.Toon(dark) : (dark==0) ? orig.Warpsharp(depth=thin) : orig.Toon(dark).Warpsharp(depth=thin)
aa= (aatype=="Sangnom") ? preaa.spline36resize(w*2,h*2).TurnLeft().SangNom(aa=aastr).TurnRight().SangNom(aa=aastr).spline36resize(w,h) : (aatype=="EEDI2") ? preaa.ediaa() : blankclip(pixel_type="YV12").subtitle("Please use Sangnom or EEDI2 for aatype")
postsh=(sharp==0 && smooth==0) ? aa : aa.LimitedSharpenFaster(edgemode=1,strength=sharp,overshoot=1,soft=smooth)
merged=mt_merge(orig,postsh,m,Y=3,U=3,V=3)

sD=mt_makediff(orig,merged)

origsuper= orig.MSuper(pel=aapel)
sDsuper = sD. MSuper(pel=aapel, levels=1)


fv3 = tradius==3 ? origsuper.MAnalyse(isb=false,delta=3,overlap=aaov,blksize=aablk) : nop()
fv2 = tradius>=2 ? origsuper.MAnalyse(isb=false,delta=2,overlap=aaov,blksize=aablk) : nop()
fv1 = tradius>=1 ? origsuper.MAnalyse(isb=false,delta=1,overlap=aaov,blksize=aablk) : nop()
bv1 = tradius>=1 ? origsuper.MAnalyse(isb=true, delta=1,overlap=aaov,blksize=aablk) : nop()
bv2 = tradius>=2 ? origsuper.MAnalyse(isb=true, delta=2,overlap=aaov,blksize=aablk) : nop()
bv3 = tradius==3 ? origsuper.MAnalyse(isb=true, delta=3,overlap=aaov,blksize=aablk) : nop()
fiz = tradius==1 ? sD.MDegrain1(sDsuper,bv1,fv1) : tradius==2 ? sD.MDegrain2(sDsuper,bv1,fv1,bv2,fv2) : sD.MDegrain3(sDsuper,bv1,fv1,bv2,fv2,bv3,fv3)

reduc = 0.4
sDD = mt_lutxy(sD,sDD,"x 128 - abs y 128 - abs < x y ?").mergeluma(sDD,1.0-reduc)

return stabilize ? orig.mt_makediff(sDD,U=2,V=2) : merged
}

[ShocWave]

I get this error:


Avisynth open failure:
I don't know what 'sDD' means.
(AnimeIVTC_-_v1.06.1.avsi, line 480)
(AnimeIVTC_-_v1.06.1.avsi, line 368)

[thetoof]

replace fiz by sDD

[ShocWave]

Like this?

sDD = tradius==1 ? sD.MDegrain1(sDsuper,bv1,fv1) : tradius==2 ? sD.MDegrain2(sDsuper,bv1,fv1,bv2,fv2) : sD.MDegrain3(sDsuper,bv1,fv1,bv2,fv2,bv3,fv3)


It appears to have worked :)

Just a few questions though, what exactly was changed?
Would it be okay to leave the script like that for encoding SD resolution clips? Or should I use this only for HD and use the regular script for SD? Any difference in picture quality?

[thetoof]

What has changed is simply better scripting that prevents wasting memory.

Also, aa=4's stabilize is set to false by default (use true only if you have time to waste or unstable edges)

[ShocWave]

Thanks for the help. The encode runs at a wonderful 0.1fps, but atleast it works :)

[Undead Sega]

hi everyone, im having trouble trying to IVTC a video which i am experimenting with from youtube, its a trailer from an anime with fieldblending, and dont worry this isnt encoded by some lame ass who ripped it themsleves etc, this was from the official channel, i am only using their video because i 'assume' it was 24p pulldown to 29.97 and fieldblended at that framerate, the closest i can get to without seeing further frames get dropped is with this:

AnimeIVTC(mode=2, aa=1, precision=3, edimode="nnedi2", iregion="NTSC", oregion="NTSC", postconv="NTSC", decimate=4, cdeblendmthr=1.6, dupthr=20)

and i dont know how to go on further, maybe i am wrong or is that usually the case?

If anyone's interested:

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

I hope im not doing anything wrong here either way...

[ShocWave]

Hm... Mode 2 doesn't work on 1080i resolutions.
I get this error:

AnimeIVTC(mode=2,aa=4)


Avisynth open failure:
vinverse: malloc failure (dlut)!
(AnimeIVTC_-_v1.06.1.avsi, line 348)


Sometimes, I get a slightly different error:


Avisynth open failure:
vinverse: malloc failure (pb3[b] pb6[b])!
(AnimeIVTC_-_v1.06.1.avsi, line 348)

[thetoof]

Woaaah Suzumiya Haruhi!! looking forward to it :p Undead Sega, this is actually blend telecining, not field blending (see how the full frame is blended and not only a field + there is a regular 3:2 pattern... well, not that obvious because of the many duplicates, but it's there). If you ever find out it is 24t 30p hybrid... let me know and I'll add an option, but I've never encountered this before.
So, it's animeivtc(mode=1,blend=true) or srestore(mode="pp3",cache=10).tdecimate()

Both of you, hang in there. I just finished writing v2 (in the very last few minutes actually ^^) and it's being tested as we speak.

I'll be updating the docs until tomorrow to get this out before school starts again.

[Undead Sega]

yes indeed, it is the one and only Haruhi Suzuzmiya! :D Finally they are doing something right, unless if that was their plan all along with their stupid Endless Eight episodes, what was the deal??? Anyways that is interesting, ive never heard of blend telecining, may i ask how is this process done? can it be reversed?

Reason why i assumed its 24fps is because thats wat anime is traditionally animated on, and a spped-up or pulldown is applied for comsumer viewing formats, however of course some frames might look the same with others with little movement etc but those are tips and trades in getting animation done efficiently, so i will say 24fps, but the actual youtube video reads at 29.97fps which could explain why there maybe either frame duplicates or blends?

Maybe something like SRestore to remove the blends and then IVTC would resolve this or something as a last resort?

[Stephen R. Savage]

"Blend telecine" is a misnomer. There are actually two things going on: telecine and blend deinterlacing. First your standard 24p source has 3:2 pulldown applied. Then, at some later point, we use blend deinterlacing because we're lazy.

This can be reversed quickly as long as compression artifacts aren't too heavy, the 3:2 pattern can be found, and the blending method is known. Consider five telecined frames, A B C D E and the corresponding four film frames W, X, Y, and Z. Frames C and D are blend deinterlaced (A, B, and E can be as well, but that just makes them blurrier). Frames A, B, and E correspond directly to frames W, X, and Z. Frame C contains half of X and half of Y. Frame D contains half of Y and half of Z. Normally, we recover Y by just remapping fields and removing the duplicates. However, in the case of blend deinterlacing, the fields have been averaged together, so Y can not be extracted using the trivial method. However, all is not lost. Since we already know about X and Z, we can subtract some weight of them from frames C and D to recover the fields of Y from them. After this, the result (discounting compression and rounding) is equivalent to a normal IVTC. In practice, assuming the blend detection succeeds, the end result is blurrier than the original, and one in four frames of it has very heavy artifacts (usually chroma noise, ringing, and ghosting). The implementation in SRestore probably does not match my description exactly, since the result must be decimated in the case of SRestore.

Actually, blend deinterlacing is itself reversible, even if it doesn't fall into a simpler case like the one above. Since none of the information in the original fields has been lost (only relocated), it is possible to find the original fields by subtraction and division. However, this must be done as a system of equations, with one equation for each line of video, so it is not practical.


ABCDE
=====
WXXYZ
WXYZZ

[thetoof]

Maybe something like SRestore to remove the blends and then IVTC would resolve this or something as a last resort?

Solution was in the previous post; srestore has a dedicated mode for this situation.

So, it's animeivtc(mode=1,blend=true) or srestore(mode="pp3",cache=10).tdecimate()

[Undead Sega]

"Blend telecine" is a misnomer. There are actually two things going on: telecine and blend deinterlacing. First your standard 24p source has 3:2 pulldown applied. Then, at some later point, we use blend deinterlacing because we're lazy.

This is what i thought exactly and i was right on the post ive made concerning this, anime is 24p with a 3:2 pulldown which was then 29.97 blend deinterlaced and i wondered if it was reversible or if one was able to recover from this despite the process done with the resolution and compression having a heavy impact.

Ive tried out:

animeivtc(mode=1,blend=true)

but i get a script error saying that it does not have a named argument called "blend", am i missing something?

[thetoof]

It is simply a new option of the version that'll be released in a few hours, which is why I wrote how to do it externally with srestore and tdecimate.

What srestore basically does is average the two blended frames to recover a clear one. Here's the section is srestore's docos where it is better explained:
To enable the double-blend-removal you have to set a string for omode. This string decides about the
postprocessing mode that is used on the restored frames:
pp0 -> the fastest mode, no postprocessing
pp1 -> use difference masking, higher quality and still good speed
pp2 -> use a special blurring mask on luma and chroma that reduces artefacts
pp3 -> combines postprocessing 1 and 2 -> slowest
You can add also your own postprocessing to these modes. Just add your function like this:
omode="pp3.blur(1)" -> pp3 + blurring
If you want to add a function with its own inputstrings use three quotation marks:
omode="""pp2.deen("a2d")"""

Important Note: Also the double-blend-removal does no decimation. It's created only for double-blends
caused by blenddeinterlacing of telecined sources. You have to add an decimation

Yes, it is indeed "blenddeinterlacing of a telecined source" and the reason I used "blend telecine" is that I saw it being used before as a shorter way to refer to what we have here. tbh, I find it simply convenient :p

[Undead Sega]

I see then, so i was right frmo the start except i described it as 'field blending' dont know if that is a correct erm anyways :D

So which would would i go for? AnimeIVTC or SRestore? hmmmm :D

[Undead Sega]

okay, i had a go with Srestore, and while it seems to have done a good job, there is infact alot of jittering, the blended frames seem to have turned into corrupted frames (visually) and is what might cause the unfluid motion of the video:

http://www.megaupload.com/?d=BT9EKYK9

What do you think?

[thetoof]

I don't get any of this.

- update to 2.7e
- get the latest versions of all the plugins srestore is depending on
- try cache=10 or something
- What is your script?

[Stephen R. Savage]

SRestore's handling of blended pulldown only works if the blend weight is 0.5:0.5. If the blend is weighted more towards one frame or the other, you will get very odd results. Also, there will generally be lots of artifacts regardless because compression distorts the frames. You might not have enough good information to repair this, if the sample came from some streaming site.

[Undead Sega]

Srestore script is:

LoadPlugin("C:\Downloads\VirtualDub-1.8.6\Requirements\MaskTools-v1.5.8\MaskTools-v1.5.8\MaskTools.dll")
LoadPlugin("C:\Downloads\VirtualDub-1.8.6\Requirements\masktools-v2.0a35\mt_masktools-26.dll")
LoadPlugin("C:\Downloads\VirtualDub-1.8.6\Requirements\TIVTCv105\TIVTC\TIVTC.dll")
LoadPlugin("C:\Downloads\VirtualDub-1.8.6\Requirements\RemoveGrain-1.0\RemoveGrainSSE3.dll")
LoadPlugin("C:\Downloads\VirtualDub-1.8.6\Requirements\RemoveGrain-1.0\RepairSSE3.dll")
LoadPlugin("C:\Downloads\VirtualDub-1.8.6\Requirements\RemoveGrain-1.0\RSharpenSSE3.dll")
LoadPlugin("C:\Downloads\VirtualDub-1.8.6\Requirements\Average_21Oct05\Average_21Oct05.dll")
Import("C:\Downloads\VirtualDub-1.8.6\Requirements\R_pack\srestore.avsi")
LoadPlugin("C:\Downloads\dgavcdec109\DGAVCDecode.dll")
AVCSource("C:\Temp\The Vanishment of Haruhi Suzumiya\video[1].dga")
Lanczos4Resize(856,480)
ConvertToYV12()
srestore(omode="pp3",cache=10).tdecimate()

in regards to the lack of information in my source, in theory should i clean it up before processing an IVTC? also this is the highest quality i could rip off from youtube, there being different formats and etc. But with the artifacts on the process of the blends, is there a way to fix that or remove them? The only way i could think of is to remove those frames individually, everyhting else seems fine.

[thetoof]

SRestore's handling of blended pulldown only works if the blend weight is 0.5:0.5. If the blend is weighted more towards one frame or the other, you will get very odd results. Also, there will generally be lots of artifacts regardless because compression distorts the frames. You might not have enough good information to repair this, if the sample came from some streaming site.
Sure sure, but as I mentioned in my earlier post (and correct me if you get different results), srestore outputs an excellent output for me. The recovered frames can hardly be spotted even when comparing to neighbors.

Maybe it's the resize, I don't know, but using this script I get no artifact at all (though I can see what you're talking about by looking at your video1 srestore.264 file, but I do not get these results on my computer).

ffvideosource("video[1].mkv")
srestore(omode="pp3",cache=10)
tdecimate()

[Undead Sega]

thats really stranged, ive even gone ahead and resized the video before putting it in the script for Srestore, but i still get the same results :(

Srestore script:

LoadPlugin("C:\Downloads\VirtualDub-1.8.6\Requirements\MaskTools-v1.5.8\MaskTools-v1.5.8\MaskTools.dll")
LoadPlugin("C:\Downloads\VirtualDub-1.8.6\Requirements\masktools-v2.0a35\mt_masktools-26.dll")
LoadPlugin("C:\Downloads\VirtualDub-1.8.6\Requirements\TIVTCv105\TIVTC\TIVTC.dll")
LoadPlugin("C:\Downloads\VirtualDub-1.8.6\Requirements\RemoveGrain-1.0\RemoveGrainSSE3.dll")
LoadPlugin("C:\Downloads\VirtualDub-1.8.6\Requirements\RemoveGrain-1.0\RepairSSE3.dll")
LoadPlugin("C:\Downloads\VirtualDub-1.8.6\Requirements\RemoveGrain-1.0\RSharpenSSE3.dll")
LoadPlugin("C:\Downloads\VirtualDub-1.8.6\Requirements\Average_21Oct05\Average_21Oct05.dll")
Import("C:\Downloads\VirtualDub-1.8.6\Requirements\R_pack\srestore.avsi")
LoadPlugin("C:\Downloads\dgavcdec109\DGAVCDecode.dll")
AViSource("C:\Temp\The Vanishment of Haruhi Suzumiya\The Vanishment of Haruhi Suzumiya lossless sharpened Dehaloed resized.avi")
ConvertToYV12()
srestore(omode="pp3",cache=10).tdecimate()

What could be wrong?

[ShocWave]

Wow. I'm very suprised to see a v2 seeing how the last major update was over a year ago. Keep up the good work :)
By the way, will v2 fix mode=2 for high resolutions?

[thetoof]

Undead Sega, I don't know. Try putting srestore right after loading your original source or ask MOmonster in his another restore function thread...

ShocWave, we'll know quite soon!

[Undead Sega]

i did try doing that, but because of the resolution size, it wouldnt accept it, therefore i had to do somesort of resizing, only by a few pixels and i still got the same results, how does ur script look like?

Also, is there any chance u could upload the output of ur IVTC of this youtube trailer? otherwise Haruhi will go ballistix! :D

[thetoof]

I posted my script before...

http://www.sendspace.com/file/91v6bv

[Undead Sega]

I must say sir, that does look absolutely wonderful which just adds mroe confusion to me as i dont think im doing anything wrong (provided my scripts above), did u use Srestore or ur new AnimeIVTC?

Probably because u put yours in a mkv container (although i dont think that would make any difference)?

[Undead Sega]

okay, i got ahold of ffmpegsource and contained the .264 file into an mkv and followed ur script and it still claims that YV12 video must have a width divisible by 4.

:( ???

[thetoof]

srestore... which is used anyways in AnimeIVTCv2

I used mkv because I can't read .264 directly, but it shouldn't change anything.

Yes, I resized before srestore in the end and it didn't cause any problem.

[Undead Sega]

I see then, and so it cant be the resize then either. Something here is terribly wrong, its just a shame i cant even get the same results when doing the same thing, its like im gonna ask you to do everything for me or something, hahaha :(

Have you looked at my scripts and probably anything outdated in it?

Also, is it possible just to use AnimeIVTC for Antialiasing only and not actually IVTCing the footage?

[thetoof]

LoadPlugin("C:\Downloads\VirtualDub-1.8.6\Requirements\masktools-v2.0a35\mt_masktools-26.dll")

Unless you use avisynth 2.6, this is wrong + the latest is a36. use mt_masktools-25.dll for avs 2.5.x

[Undead Sega]

i'll give that a go and let you know if it makes any difference, however as i said before, is it possible just to use AnimeIVTC for Antialiasing only and not actually IVTCing the footage?

[thetoof]

If you look in the docs, you'll see that besides every aa mode there is a "use x to call externally"

1 - ediaa()
2 - daa()
3 - maa()
4 - sharpaamcmod()

[Undead Sega]

i was looking in the docs of course, however it requests me to always use somesort of mode to be used.

[Nightshiver]

Then you're doing it wrong. For the external's, don't use the animeivtc() line. Just sharpaamcmod() or whatever you need.

[Undead Sega]

ohhh wait, i got it! :D

[Undead Sega]

Question: How do i use the other functions made for aa=4 a.k.a. sharpaamcmod() ?

[Nightshiver]

Read docs. They explain it all.

[Undead Sega]

i have but when i tried it out it tells me that (besides for using aatype) there isnt a function or argument called ''whatever function ive typed'' and etc. For example like:

SharpAAMCmod(aatype="EEDI2", thin=85, stabilize=true)

unless im doing it totally wrong?