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

[Nightshiver]

Then why would you be using mode=1? That's only if you can see a 3:2 pattern. I used mode=2 because it looked double hard telecined.

[RsVolcom]

Because mode=2 looked so bad. I was just trying different modes!

[egrimisu]

Hi,

Here's the sample i was talking about, i analysed the source and t seems' that are the paterns i 3:2 but sometimes ther 3 progesive frames and right after 3 interlanced frames, probably that's the reason on frames remains interlanced.

http://www.megaupload.com/?d=1K6O5KFC

I finished creating the timecodes.txt for the source we talked a few days ago. Now what's next? i supose i need now to encode to get the mkv file. What script to use and how many passes?
I did what u said about ttempgause_beta but i still get an error, http://img148.imageshack.us/my.php?image=111zx5.jpg

Thanks again, Bye

[Nightshiver]

Probably wrong version of MVtools.

[egrimisu]

Probably wrong version of MVtools.

Latest 1.9.6.2-1 SSE3 found of doom9.

[josey_wells]

Latest version is V1.9.7.0

[Nightshiver]

You should probably keep a favorite tab or bookmark the mvtools thread, as josey updates rather frequently.

[Benus_MinO]

Hey guys!

Why I have 19.18 fps in this configuration after encoding:


DGDecode_mpeg2source("C:\DVD_VIDEO\VIDEO_TS\VTS_01_1.demuxed.d2v")
animeivtc(mode=1,aa=2)
crop( 4, 0, -4, 0)
LanczosResize(640,480)
LimitedSharpenFaster()
Undot()

I wanna 23,976 fps.

[egrimisu]

Got latest mvtools, and everything is working just fine, THANKS

You should probably keep a favorite tab or bookmark the mvtools thread, as josey updates rather frequently.

[egrimisu]

Question 1 : after doing the analysis pass and the rendering i got a 25gb file that contains dubble frames from the original.Now i simply encode that 25gb file with x264 and i use the timecodes in the mkv and thats all? shall i use dup ore something to delete some of those double frames or this is the way is should be?
Question 2 : is there a way to find full 29.97 progrsive frames in a 95% 3:2 movie? i have a mode 6 source. i tried parsing the d2v file but all the frames are reported as 2 (witch probably is progresive since i haven't found no manual that say what it is) and only a few are * (3:2) but the source is 95% 3:2 and the rest progessive.

OFF Topic 1 : limitedsharpenfaster sould be used before or after cleaning, i often use it after cleaning but beside the great sharp image i get some grains, is it recomended to use some lite degrainer like dedot after LSF? or LSF it is more recomended as a realtime filter when playback.
OFF Topic 2 : from your experience is mvdegrain a good detail preserve?

[Adub]

OFF Topic 1: It is recommended to use all denoising techniques before a sharpener. However, feel free to insert one afterwards, if it works for you. Some times a real light pass after sharpening calms things down nicely.

OFF Topic 2: Yes, MVdegrain is a good detail preserver. However, like everything else, it depends on your settings.

[thetoof]

Q1 : The dups should've been decimated... I'm currently trying to improve the vfr decimation of animeivtc, since all I had learned about it was basic stuff, so it's not optimal yet. Can you post the content of your timecode file as well as the animeivtc line of your script?
Q2 : In a mode 6 source ((24t + 30p) + 30i on top of it), you have to identify the 30i section(s) like in any mode with pureint=x and the rest should be done automatically. If it's not... post a sample! :p However, what you are describing (3:2 with 30p) looks like simply hybrid (no 30i on top of it), so mode=5 should do the trick.
For the OTs, Merlin7777 provided the same answer I would have. I'd only add that soothe is a great filter to use after your sharpener.

[egrimisu]

it's mode 6 actually, the credits are fully interlanced;) . Ok i'll give it a try tomorow.
For the othersource that was 29progrsive and a portion 30i ( probably u remember the script you gave me a several post back, i did that analyse pass and renderless pas BUT the source had aprox 40000 frames and 1800 were 30i, NOW the output of the renderless has 60000 frames. I think that the output should have 40000 + 1800(from the full bobing) frames. Am i wrong?

I just love animeIVTC:D, great job, you have made our work a lot easier.

[thetoof]

This must be some kind of error during decimation... However, I need to see the content of the timecode file to confirm that (even if I'm almost sure that there's a bug with how I call vfr decimation).

A good compromise for now would be to simply create a CFR clip and cut half of your temporal resolution for the credits...
#Fill the info for the start/end + the name of your d2v and the rest is done by the script
estart1 = 0
eend1 = frame before credits
istart = 1st frame of the credits
iend = last frame of the credits
estart2 = frame after the credits
eend2 = last frame you want to keep in the episode

s=mpeg2source("whatever.d2v")
a=s.trim(estart1,eend1)
b=s.tempgaussmc_beta2(edimode="nnedi").trim(istart*2,iend*2).selecteven()
c=s.trim(estart2,eend2)
a+b+c
Yes, I'm afraid you'll have to redo your encode...
I just love animeIVTC:D, great job, you have made our work a lot easier.
This in only the beginning ;) There is room for a lot more improvement than the current script!

[Comatose]

it's mode 6 actually, the credits are fully interlanced;) . Ok i'll give it a try tomorow.
For the othersource that was 29progrsive and a portion 30i ( probably u remember the script you gave me a several post back, i did that analyse pass and renderless pas BUT the source had aprox 40000 frames and 1800 were 30i, NOW the output of the renderless has 60000 frames. I think that the output should have 40000 + 1800(from the full bobing) frames. Am i wrong?

I just love animeIVTC:D, great job, you have made our work a lot easier.
Is this the new Negima OVA?

[egrimisu]

No it's Hanbun no Tsuki ga Noboru Sora dvd's R2 japanese

Is this the new Negima OVA?

[egrimisu]

Sorry i have deleted the timecodes.txt and i choosed the 29fps half full resolution deinterlancing metod that you have writen me a few weeks ago. I tried to undelete the timecodes.txt but no luk since i have moved the folder where it was stocked and now it can't be found anywhere. I'll pickup antoher encode like this in the near futere and i'l post that. Thanks

This must be some kind of error during decimation... However, I need to see the content of the timecode file to confirm that (even if I'm almost sure that there's a bug with how I call vfr decimation).

A good compromise for now would be to simply create a CFR clip and cut half of your temporal resolution for the credits...
#Fill the info for the start/end + the name of your d2v and the rest is done by the script
estart1 = 0
eend1 = frame before credits
istart = 1st frame of the credits
iend = last frame of the credits
estart2 = frame after the credits
eend2 = last frame you want to keep in the episode

s=mpeg2source("whatever.d2v")
a=s.trim(estart1,eend1)
b=s.tempgaussmc_beta2(edimode="nnedi").trim(istart*2,iend*2).selecteven()
c=s.trim(estart2,eend2)
a+b+c
Yes, I'm afraid you'll have to redo your encode...

This in only the beginning ;) There is room for a lot more improvement than the current script!

[sno0py]

Hey
I'm getting some very weird effects when I try to use AnimeIVTC.
http://img99.imageshack.us/my.php?image=aivtcissuehr2.jpg
I would like to note that i DON'T have the mask enabled, but also i get those diagonal lines. I should probably also mention that I have seen the VERY SAME diagonal distortions when i tried to use MCBob a while back. Toof had the suggestion that it might be the limits of my system, and based upon this result with two different scripts, he might be right. On the other hand, I know that both use alot of the same prerequisite scripts, and maybe it's a setting that I don't understand. Also, one other note... the footage looks awesome when its not f'n up.. AIVTC ftwthx.... one other question tho... what is a preferred de-rainbower... derainbow isn't really doing the job...

A few PC specs...
CoreDuo E6750 2.67GHz
2.0GB DDR2 800
eVGA nVidia 7900 GT KO
74GB 10k WD Raptor
X-Fi Champion
but just Reglur XP pro.. not 64bit..... yet....

AVS settings:
DGDecode_mpeg2source("H:\1_hp.d2v",cpu=6,cpu2="xxxxxx",info=3)
ColorMatrix(hints=true,interlaced=true)

DeRainbow(25)

AnimeIVTC(mode=1,aa=3)

crop( 8, 0, -6, -2)

Convolution3D(0,3,5,3,5,4,0)
Lanczos4Resize(720,480) # Lanczos4 (Sharp)

Limitedsharpen(ss_x=1.0,ss_y=1.0,smode=3,strength=35)
Undot() # Minimal Noise


megui settings:
program --pass 2 --bitrate 1657 --stats ".stats" --deadzone-inter 18 --deadzone-intra 10 --ref 5 --mixed-refs --no-fast-pskip --bframes 3 --b-pyramid --b-rdo --bime --weightb --direct auto --filter -1,-1 --subme 7 --partitions all --8x8dct --ratetol 2.0 --me umh --merange 32 --threads auto --thread-input --cqm "jvt" --progress --no-psnr --no-ssim --output "output" "input"

[thetoof]

Did you replace all your plugins with the ones in the requirements archive?
Did you check if you get this bug when using only animeivtc?
Care to share a sample?
TComb is a very good dot crawl and rainbowing removal filter, though sometimes I prefer using fft3dfilter on plane=3 with high sigmas to kill chroma noise + rainbowing.
What is your memory usage with the script you posted (check in the task manager)?

[sno0py]

Did you replace all your plugins with the ones in the requirements archive?
---yup.. i systematicly replaced every single one with the ones in the archive
Did you check if you get this bug when using only animeivtc?
---nope. I get this bug also when I use mcbob()
Care to share a sample?
---what's a good free-post service? i'll throw a 10sec clip up if it'll help.
TComb is a very good dot crawl and rainbowing removal filter, though sometimes I prefer using fft3dfilter on plane=3 with high sigmas to kill chroma noise + rainbowing.
---sweet.. tcomb did better, but I haven't tried fft3dfilter yet.. I'll give that a try
What is your memory usage with the script you posted (check in the task manager)?
---When i'm previewing the script in megui, i run ~600MB (of 2GB) ram and 50% CPU load (it maxes out one of my 2 cores).

[sno0py]

here: http://www.megaupload.com/?d=A3LWH2VP

[thetoof]

By AnimeIVTC only, I mean that, in the current script, if you disable all the filters exept aivtc, do you still get the artifact?
Mediafire.com
2 cores? Do you have a setmtmode call somewhere in there? using this + ivtc can lead to weird results...

[sno0py]

By AnimeIVTC only, I mean that, in the current script, if you disable all the filters exept aivtc, do you still get the artifact?
---I removed all the other scripts... and It still happens, but it takes longer.. It all looks so good until i let it run for a little while, and then it all goes to crap. so maybe you're right about the pc limitations...
2 cores? Do you have a setmtmode call somewhere in there? using this + ivtc can lead to weird results...
---Ya... 2 cores, but i dont know what setmtmode is... so i'm not using it... in megui I just set the cpus to auto, so it aims for 1.5 i think is what they say.

If it's a memory thing, I'm planning on upgrading my box with a new better vid card, 8 gb of ram, and xp x64. If it really seems like that's the holdup, then I can put this project on the back burner, and just wait to see how the HW upgrades affect things..

[thetoof]

Oh, I now I think I get it... running x264 + all your avs filters may be simply too much for your comp, so I'd recommend doing a lossless rendering pass of your complete avs script only:
Download and install Lagarith
Load your avs in vdub
video-compression-lagarith (YV12 + multithread enabled)
video-fast recompress
file-save as
wait
new script: avisource("your new lossless filtered clip".avi)
load it in megui
encode
be happy or post again :p

[egrimisu]

Can megui be set to make the renderless pass using lagarith instead huffy?

[Adub]

No. And it's rendering pass.

[RedRider]

I am trying to convert my Simpsons season 1 DVD`s to x.264
they are Double hard telecined
i tried AnimeIVTC 1.02 (mode=2,aa=0) everything is good but sometimes it is trying to blend some frames made from fields of previous and next frame
if they differs a lot,result is bad ,maybe there is a solution fo for it, i have tried different values of decimate but there is no big difference

original video
original1 (http://img87.imageshack.us/img87/9892/original0sx1.jpg)
original2 (http://img87.imageshack.us/img87/1207/original1su1.jpg)
original3 (http://img87.imageshack.us/img87/3330/original2hm2.jpg)
original4 (http://img87.imageshack.us/img87/3330/original2hm2.jpg)

IVTCed
IVTCed1 (http://img76.imageshack.us/img76/1962/ivtced0of8.jpg)
IVTCed2 (http://img126.imageshack.us/img126/3429/ivtced1wn3.jpg)
IVTCed3 (http://img126.imageshack.us/img126/8104/ivtced2ab1.jpg)
IVTCed4 (http://img71.imageshack.us/img71/7977/ivtced3ho7.jpg)

[thetoof]

Weird, can I get a video sample for that section?
load your vob in dgindex, select the section, demux video with honor pulldown flag, upload it to mediafire, post your link.

[RedRider]

sample video (http://www.mediafire.com/?sharekey=b4730766fe6b7bbbd2db6fb9a8902bda)

[thetoof]

From the sample I saw, your clip is badly telecined, but not dht... there'll be an update in the docs about that.

#ivtc, appropriate aa mode for sources with "line breaks"
animeivtc(mode=1,aa=2)
#followed by multiple calls of the aa function to further reduce the issue, use as many as you want. Frame # 226 (after ivtc) is a good reference to see the effect of aa on tricky frames... some adaptative aa is on the way, but for now you'll have to apply it to the whole stream
daa()
daa()
#following it with mvdegraining helps reducing artifacts caused by "over-anti-aliasing" + scene change issues
a=last
vec=a.mvanalysemulti(refframes=x, overlap=4, idx=1, other settings)
a.mvdegrainmulti(vec,idx=1)

[Nightshiver]

When can we expect this :)

[thetoof]

Working on it as we speak... and I have to go for a while in 3h, so unless I'm interrupted by something, today! ^_^

[tw0face]

hello, im a new guy in here and i have a question :)
im ripping avatar the last airbender
it took me a few days to figgure out how to work with .avs script and what options should i use with animeivtc, at least i hope so :)
im using gordian knot, where i set everything, pick none field options, and then just add AnimeIVTC(mode=1, aa=2) in .avs file

first episodes finished great, awsome quality, really something i hoped for, coz previous rips, without animeivtc, didnt get me such good quality, BUT one episode just keeps crashing in second pass and i dont have a clue why
first pass is ok, and at 0% in 2nd virtualdub crashes :(

any idea why? or what do u need from me to figgure out, thanx for ur help :)

[lansing]

i think i need help with my source.
the first 30 seconds or so of the video was 100% ntsc, followed by 100% film to the end.

sample (http://www.mediafire.com/?jxdjdg9rtdi)

My script is

AnimeIVTC (mode=2, precision=1, aa=2, decimate=4)

Lanczos4Resize(640,480)# Lanczo4 (Sharp)

The problem is that there are many thin lines in the images that didn't seen to deinterlace correctly. Another problem is the video didn't run smoothly even after I put in the decimate function.

[Nightshiver]

We're all just waiting for toofy to update.....

[thetoof]

Yup, I'm all done. I only need to get home to zip the latest requirements.7z archive and v1.03 is good to go! Give me 1 hour or so ;)
The next post will also have all the answers to the questions/comments/suggestions that were posted during the 2 months when I was gone.

[thetoof]

@neuron2
The first reason why I assumed that a hard pulldown could create some problem was that I saw this clip (http://www.mediafire.com/download.php?vblqwwgftmd) (check frame # 32 before IVTC and 25 after), but it seems like that frame is only an exeption and that my claim was, in fact, unsupported. I'll be more careful in the future.

How can i find out at what frame it end the op witch is interlanced and when start the progresive part and where stops and start the ed witch is interlanced again? Is there something for this?
I'm afraid you have to check frame by frame... those are the most important sequences to check:
- opening credits
- ending credits
- high motion sequences
You have to check for : interlaced, progressive or telecined credits, pulldown pattern (hard or double hard telecined) and 30p sequences.
If you see double hard telecine somewhere, consider your source DHT (not hard telecine, so forget mode 1 and 3) because if it appeared once, it can at any high motion scene and it'd be a pain to check them all, so just to be safe, use the DHT modes.
You say I should use it after everything else; do you mean I should IVTC normally, run filters, run a lossless pass, and then do the mode=7 after that?
Forget that, I changed how animeivtc deals with vfr
Also, is AA really neccessary for a hard telecined source (IE mode=1)? I don't really see any jaggies...
As it was said by many, it is needed when it is needed... compare and choose the best looking aa method
What pureint is and what is the dirence between 1-7?
"pureint" has been changed to "mix"
It is to tell animeivtc which type of correction to apply to which part of the clip. Read the doc and if it's not clear enough, tell me what is it that you don't understand and I'll improve it.
Opening (IVTC 23.976 CFR) + Middle (Hybrid 24+30fps) + Ending (Bobbed at 60 fps) + IVTC (23.976fps CFR)
Can I have a sample of the hybrid sequence?
i have a few scenes where i have pure interlancing.
I've never seen telecined + interlaced for an episode (the credits, yes, but not the episode), so it's most likely double hard telecined (looks pure interlaced because there is no pattern)... try a mode for DHT, give me some news and we'll see then.
mode=7 seems to be also flawed. mode=7, pass=1 does not make a tfm file, and that file is needed for pass=2 (tdecimate(mode=5)).
From the changelog:
01/19/2006 v0.9.11.8
- don't require input and tfmIn files for mode 5
From TDecimate docs:
As of version 0.9.11.8 of TDecimate, an input and tfmIn file are no longer required for mode 5. If an input file is not specified then a value of ULLONG_MAX-1 (the maximum value tdecimate has) is assigned to all metrics. Likewise, if a tfmIn file is not specified then all matches are assigned as 'c', not combed, and not a d2v duplicate.
I tried to intc heidi dvd witch dgindex recognized it as 77% video that makes it mode 7 that is soft telecined and hard telecined.
No no no! #7 is under "Source type", meaning that you need to use mode= "1 : Use for case #3 & 7 (hard telecine)". Also, make sure to check if it's really hard and not double hard.

If I missed a question or a suggestion, please repost and it'll be a pleasure to help.

[thetoof]

AnimeIVTC() 1.03 Released 2008-09-16 by thetoof

* Fixed bug when splicing clips at different framerates
* Fixed bug where frames were lost to decimation in mode=6
* Fixed bug causing the "tfm.txt" file to be overwritten every time the script was loaded
* Fixed idx values in the MVAnalyse and MVFlowFPS calls
* Use of 2 different pelclips if prefiltered clip is used before motion vectors calculation to prevent b0rked output
* Removed omode=3 for optimal vfr decimation
* Less memory usage by using a named clip for all the blankclips
* Changed syntax to be compatible with the newest version of mrestore
* Faster blends and duplicates detection for decimate=5 by using dclip=i.bob().reduceflicker(strength=1) in mrestore
* Replaced nnediaa with ediaa for better aliasing removal (aa=1)
* Added functions in the script to fix "There is no function named iseven/isodd" bug
* Added "MT" parameter to use with the multithreaded version of MVTools
* Added support for fieldblended PAL & norm-conversion restoration
* Added support for 30p credits on top of telecined background
* New check function to get the frame numbers of the 30p sections (check=0)
* New check function to see the prefiltered clip of the 30p sections (check=9)
* Argument "pureint" changed to "mix"
* 15 new "mix" modes for all the possibilities for 30p and 30i credits
* Replaced decimate=0 with pass=0
* New AA function by Mystery Keeper (aa=4)
* Re-added the possibility to use killcomb=0, but default remains 1
* Changed "precision"'s default to 3 for all modes
* Included TempGaussMC_beta1mod in the requirements package to make bob=3 work with the new "MT" parameter
* Docs update + formatting improvement, reformulation here and there...

[[MA]Debito]

Getting the following Error message:

Script Error: There is no function named "Toon"

Referencing Line 151:
aa4=dec.sharpaamcmod(aastr=aastr, ds=dark, shpre=thin, shpost=sharp, smpost=smooth, tradius=tradius, mt=mt, MC=stabilize, pel=aapel, ov=aaov, blk=aablk)


and line 438:
b=a.Toon(ds).Warpsharp(depth=ShPre)

I've loaded the mfToon and vmToon functions into my script as well with no change.

Here's the AVS line:
AnimeIVTC(mode=2,aa=0,precision=3,iregion="NTSC",oregion="NTSC",decimate=4,MT=true)

Thanks.

[Nightshiver]

It's not talking about either mftoon or vmtoon. You need the dll called "Toon". Get it Here (http://mf.creations.nl/avs/filters/Toon-v1.0.dll)
Also, THANKS toofy! :)

Edit: maybe I spoke to soon. I get an error saying "I don't know what fpsconv means"

[[MA]Debito]

Still getting the no function called "Toon" even when it is autoloaded and also will not work if I use LoadPlugin

Thanks

[thetoof]

im ripping avatar the last airbender

I never used GN, so I can't help with that, but knowing your source, I can assure you it's a hybrid one. Check out the new docs for more info and post again if something's not clear.
animeivtc(mode=5, other settings)
@ lansing
First section = truly interlaced
Second section = hard telecined
So, mode=3
The scripts for the sample you gave me could look like this:
mpeg2source("sample.demuxed.d2v")
animeivtc(3,aa=0,omode=2,edimode="eedi2",pass=0,mix=1,istart1=0,iend1=1509,estart1=604,eend1=639)
Run a lossless rendering pass
avisource("your processed file.avi")
animeivtc(mode=7, pass=1, other settings)
Run video analysis pass
avisource("your processed file.avi")
sharpaamcmod(aastr=99999, other settings)
animeivtc(mode=7, pass=2, settings)
Encode or do another lossless rendering pass. The reason why I split the process is that aa=4 + smart-bobbing for the credits is just too much for the computers of our day and age :p
Debito;1184665']Still getting the no function called "Toon" even when it is autoloaded and also will not work if I use LoadPlugin
Did you put all the plugins + autoload scripts of the requirements archive in the plugins folder of avisynth? If you did, I don't have the slightest idea of what it could be... everything's autoloading perfectly for me.
Edit: maybe I spoke to soon. I get an error saying "I don't know what fpsconv means"
Heh, sry 'bout that... should've done more testing before posting... I'm currently updating the links with a new version.

Changelog:
* Fixed check=3 and check=10 to be compatible with the new add-ons
* Corrected typo in the error message asking to specify the omode
* Disabled aa comparison when using one of the check=x
* nnedipel's default is now false
* Changed the order of the filters so that all the stream (not only the ivtced sections) can be processed with killcomb and aa
* Fixed syntax errors with mode=7

[Nightshiver]

Aight, it works now :)

[Ittipol02]

Debito;1184665']Still getting the no function called "Toon" even when it is autoloaded and also will not work if I use LoadPlugin

Thanks

http://forum.doom9.org/showthread.php?p=1062109#post1062109

[lansing]

@ thetoof

after I updated the filter and applied your first script, I got an error message

Script error: animeivtc does not have a named argument "edimode"

[Adub]

Hmm. Yeah, maybe because it doesn't. My guess is that you had some old settings stored away that you were using with the older version, and you copied and pasted them into your new script thinking it would work.

Okay, that's fine, but read your own error message first. It is telling you exactly what is wrong. Read the documentation.

[thetoof]

Actually... that's a miss by me: I uploaded an "older" 1.04 without the argument.

AnimeIVTC() 1.05 Released 2008-09-18 by thetoof
* Added string "edimode"
* Added int "degrain"
* Fixed typos in the docs
* Removed ineffective error message about the external audio file
* Added the plugins needed for aa=4 + bob/precision=3 in the "requirements" archive & list

[lansing]

I think the function "reduceflicker" is also missing

Script error: there is no function named "reduceflicker" (?, line 151)

[Nightshiver]

Nope, it's in the latest script/dll package that toofy uploaded to mediafire.

[lansing]

nvm, I did find out that it was the missing Avscursion.dll that's causing the error