logo Sign In

Info: Messing with ROTJ 1986 Special Collection - Suggestions wanted

Author
Time
 (Edited)

Hi there all I’ve been messing around with the 1986 ROTJ Laserdisk and just recently did a capture of side 1 through my LD-V4400 and i was messing around and created this script

loadplugin(“DGDecode.dll”)
loadplugin(“SangNom.dll”)
loadplugin(“FDecimate.dll”)
loadplugin(“ColorMatrix.dll”)
import(“deblockqed.avis”)
loadplugin(“MaskTools.dll”)
loadplugin(“UnDot.dll”)
loadplugin(“mt_masktools.dll”)
loadplugin(“deen.dll”)
loadplugin(“hqdn3d.dll”)
loadplugin(“GrainOpt.dll”)
loadplugin(“GuavaComb.dll”)
loadplugin(“Decomb.dll”)
loadplugin(“cnr2.dll”)
loadplugin(“TDeint.dll”)
loadplugin(“mvtools.dll”)
import(“SeeSaw.avs”)
loadplugin(“RemoveGrainSSE3.dll”)
loadplugin(“RepairSSE3.dll”)
loadplugin(“RSharpenSSE3.dll”)
import(“mvbob.avs”)
import(“MCBob_v03u.avs”)
loadplugin(“EEDI2.dll”)
loadplugin(“nnedi.dll”)
loadplugin(“GuavaComb.dll”)
loadplugin(“BruteIVTC.dll”)

BlindDeHalo - v0.1

This script removes the light & dark halos from too strong “Edge Enhancement”

Adaption of the parameters has to be done manually for each source, the function is not adaptive.

Needed Plugins: MaskTools.dll >= 1.4.15

#                 undot.dll

function BlindDeHalo( clip clp, int “strength”, int “radius_x”, int “radius_y”, float “darkamount”, string “mode”, float “maskblur”, bool

“fullframe” )
       {
         strength =   default( strength,  80)
         radius_x =   default( radius_x,   2)
         radius_y =   default( radius_y,   2)
         darkamount = default( darkamount,1.0)
         mode =       default( mode,  “soft”)
         maskblur =   default( maskblur, 1.0)
         fullframe =  default( fullframe, false)
         modus = mode==“hard” ? “hardlight” : “softlight”

ox = clp.width
         oy = clp.height
         rx = sqrt(radius_x)
         ry = sqrt(radius_y)

clp
         edge=Dedgemask(0,255,0,255,“6 10 6 10 -64 10 6 10 6”,setdivisor=true,divisor=64).levels(2,0.5,6,0,255,false).undot#.coloryuv

(analyze=true)#

blur1 = bicubicresize(int((ox/rx)/4 +.5)*4,int((oy/ry)/4 +.5 )*4,-1.0,1.0).lanczosresize(ox,oy)
         blur2 = bicubicresize(int((ox/rx)/4 +.5)*4,int((oy/ry)/4 +.5 )*4, 1.0,0.0).lanczosresize(ox,oy)

diff = YV12subtract(blur1,blur2)

overlay(diff,diff,mode=“hardlight”).levels(0,1.0,255,88,168,false)

darkamount == 1.0 ? NOP : overlay(last,last.levels(0,1.0,128,int(128-128*darkamount),128,false),mode=“lighten”)
         overlay(last,last.levels(0,1.0,255,128+strength,128-strength,false)) # levels reversed: no more “invert” in below overlay()

overlay(clp,last.greyscale.blur(maskblur),mode=modus)
         fullframe ? NOP : MaskedMerge(clp,last,edge.deflate.expand.blur(1.58).inflate)

return(last)
       }

#Load Sources and trim frames to match
a=avisource(“i:\Composite 2_20050702_204306.avi”).trim(96,50228).guavacomb(“NTSC”)
b=avisource(“i:\Composite 2_20050702_183413.avi”).trim(207,50339).guavacomb(“NTSC”)
c=avisource(“i:\Composite 2_20050703_001628.avi”).trim(378,50510).guavacomb(“NTSC”)
d=avisource(“i:\Composite 2_20050702_180458.avi”).trim(202,50334).guavacomb(“NTSC”)
e=avisource(“i:\Composite 2_20050702_234755.avi”).trim(200,50332).guavacomb(“NTSC”)

#Merging sources
f=merge(merge(merge(a,b),merge(c,d)),e,0.2)

#IVTC/Decimation
g=f.converttoyv12(interlaced=true).bruteivtc().tdecimate(hybrid=1).crop(0,52,0,-140).NNEDI(dh=true,field=1)

h=g.lanczos4resize(720,288)

#Degrain

backward_vec2 = h.MVAnalyse(blksize=16,isb = true, delta = 2, pel = 2, overlap=8, sharp=2, idx = 1)
forward_vec2 = h.MVAnalyse(blksize=16,isb = false, delta = 2, pel = 2, overlap=8, sharp=2, idx = 1)

i=h.MVDegrain1(backward_vec2,forward_vec2,thSAD=400,idx=1)#.limitedsharpenfaster(smode=4,dest_x=1280,dest_y=544,wide=true,strength=250

#Derainbow

j=i.fft3dfilter(sigma=3,sigma2=3,sigma3=12,sigma4=3,bt=3,plane=1).fft3dfilter(sigma=3,sigma2=3,sigma3=12,sigma4=3,bt=3,plane=2)

#Upscale to 856x364

k=j.limitedsharpenfaster(smode=4,dest_x=856,dest_y=364,wide=true,strength=250)

return k

And here were my results so far

And i was curious what else i can do to this.  And on the last 2 frames i was wondering what the weird lines and blocks were and i was also curious about what you guys use for IVTCing and how to fix the strange halos?

Author
Time
 (Edited)

What a coincidence, I captured the same LD (and also Star Wars and Empire). But I think your sharpness settings on your capture device are set too high (I've had the same problem but Moth3r showed me the way). If you look at the first picture you clearly see ghosting effects around Luke. Setting the sharpness lower will solve this problem. Ofcourse the image will look a little bit softer but that's just what it looks like. If the sharpness is set too high you only get "fake" detail and ghosting effects. Did you try different settings? Here is a useful thread: http://originaltrilogy.com/forum/topic.cfm/Setting-sharpness-for-analogue-capture/topic/9566/

What capture device did you use?

Did you try the derainbow before the inverse telecine step? I think derainbowing works better before inverse telecine.

I also tried derainbowing but I noticed I got some black and white frames just before a scene change. Did you have the same problem?

For IVTC I used virtualdub in manual mode.

Fez: I am so excited about Star Whores.
Hyde: Fezzy, man, it's Star Wars.
Author
Time

looks good,

 

i'll play around with it..

 

should this thread be moved to the Technical area?

 

later

-1

[no GOUT in CED?-> GOUT CED]

Author
Time

I wasn't sure if this thread should be in the technical section either lol.

And the card i am using is this http://www.newegg.com/Product/Product.aspx?Item=N82E16815100120 the drivers for the card say Conexant 2388x and I am capturing using a program called Christv cause those retarded WMD drivers or whatever didn't let me capture above 320x240 lol.

My Sharpness setting is at 0 and i cannot set to negative numbers so i am unsure what to do and i have not tried derainbow before ivtc yet but i did learn that doing derainbow after upsizing to 856x364 has a lot less effect.

Also i was wondering if you meant VirtualdubMod or Virtualdub and where would these manual settings be.

 

Thx

Author
Time

Oh, that's a pitty you can't set the sharpness to negative values.

I use virtualdubMpeg, but I think it can be done with any version of virtualdub. You have to go to video > framerate. At the bottom you can select the manual IVTC. But IVTCing the special collection is a lot of work. The pulldown changes from time to time so the settings have to be adjusted for specific pieces of video. Personally I just processed every side 5 times (offset 0 through 4) and I cut and pasted the correct IVTCed pieces together.

Fez: I am so excited about Star Whores.
Hyde: Fezzy, man, it's Star Wars.
Author
Time

Is there anything I can do to get around the sharpness problem or should i look for a new capture card lol?

Author
Time

I don't know if this makes a difference or not with captuer cards, but could changing your TV settings help?  Again, I'm not sure - in fact, I think it probably doesn't.  Just throwing it out there though.

A Goon in a Gaggle of 'em

Author
Time

There's a lot of ghosting/haloing in those caps- is that from the LDs themselves or is something in the script causing it?

Author
Time
Mielr said:

There's a lot of ghosting/haloing in those caps- is that from the LDs themselves or is something in the script causing it?

 

I think it has to do with my capture card

Author
Time
 (Edited)
Mielr said:

There's a lot of ghosting/haloing in those caps- is that from the LDs themselves or is something in the script causing it?

 

 It's not on the LDs themselves. It happens when the sharpness level of the capture device is too high.

 

inurenegade said:

Is there anything I can do to get around the sharpness problem or should i look for a new capture card lol?

 

 The only thing I can think of is a capture card that allows you to change the sharpness settings. I'll post a screenshot from my capture later for comparison.

Since you used dehalo, are the ghosting effects much worse in the raw capture?

Fez: I am so excited about Star Whores.
Hyde: Fezzy, man, it's Star Wars.
Author
Time
 (Edited)

I didn't call for dehalo in the script its just a function that i was thinkin about using. And surprisingly after watching the LD through my TV and then recording the LDs through my dvd recorder i see some halos LOL. So now im really curious to see what your source has.

Author
Time
 (Edited)

Here's a screenshot from my raw capture:

http://img219.imageshack.us/img219/6643/xx0000mj7.jpg

 

You can still see some ghosting, but I thought it was acceptable. If I set the sharpness even lower, the image was too soft in my opinion. So I choose this setting, but it's just a personal preference.

Fez: I am so excited about Star Whores.
Hyde: Fezzy, man, it's Star Wars.
Author
Time
 (Edited)

hmm thats about what my dvd recorder and tv had and my capture is worse then that.

I'm curious what capture card are you using and how much does it cost approximately?

Btw heres my raw capture of the same image

Image Hosted by ImageShack.us

Im starting to think that my capture card has a comb filter in it...

Author
Time

I'm using a sweetspot video processor (aka PDI deluxe). It was about 150-160 euros.

I don't think a combfilter has anything to do with the ghosting...

I was wondering, in the last picture, did you cut off something on the sides, or is it just captured like that?

Fez: I am so excited about Star Whores.
Hyde: Fezzy, man, it's Star Wars.
Author
Time

Capture card did that and i meant that i think my capture card has a comb filter cause of all the rainbowing lol. I guess i get what i pay for with a 17 dollar capture card lol.

Author
Time

A comb filter removes the rainbowing.

There's also rainbowing/chroma noise in my capture. It's what bugs me the most.

Fez: I am so excited about Star Whores.
Hyde: Fezzy, man, it's Star Wars.
Author
Time
inurenegade said:

i was also curious about what you guys use for IVTCing and how to fix the strange halos?

The most reliable method of IVTC is using:

doubleweave
pulldown(x,y)

as described here.

Automatic methods may not correctly detect the pattern in some cases - the initial starfield pan down being a good example, as there is very little for the detection algorithm to lock on to.

However, as Arnie correctly pointed out, if there are numerous cadence changes then it does become rather time consuming to go through manually and identify each change point.

Halos - if you can't eliminate them with the sharpness setting, you should try and look for an alternative driver for your card, or maybe the 2388x reference driver if you can find it.

Also, this thread at doom9 may help.

Guidelines for post content and general behaviour: read announcement here

Max. allowable image sizes in signatures: reminder here

Author
Time
 (Edited)
Arnie.d said:

A comb filter removes the rainbowing.

No, a comb filter separates the luminance (y) and chrominance (c) elements of a composite video signal.

Since video is stored on laserdisc in analogue composite form, there must be a comb filter somewhere in the chain.

A poor comb filter will lead to dot crawl and "rainbows" (aka cross-colour effects). A good quality comb filter will reduce the appearance of these effects.

Since the player the OP used does not (I think) have an s-video output, he would have hooked it up with a composite video cable, therefore the (cheap?) comb filter in the capture card will be used.

If the card has an s-video input, then a potential improvement could be obtained by connecting the LD player to the DVD recorder with a composite cable, then the DVD recorder to the capture card with an s-video cable. This uses the comb filter inside the DVD recorder, which is likely to give higher quality results.

 

Guidelines for post content and general behaviour: read announcement here

Max. allowable image sizes in signatures: reminder here

Author
Time
Moth3r said:
Arnie.d said:

A comb filter removes the rainbowing.

No, a comb filter separates the luminance (y) and chrominance (c) elements of a composite video signal.

Since video is stored on laserdisc in analogue composite form, there must be a comb filter somewhere in the chain.

A poor comb filter will lead to dot crawl and "rainbows" (aka cross-colour effects). A good quality comb filter will reduce the appearance of these effects.

Since the player the OP used does not (I think) have an s-video output, he would have hooked it up with a composite video cable, therefore the (cheap?) comb filter in the capture card will be used.

If the card has an s-video input, then a potential improvement could be obtained by connecting the LD player to the DVD recorder with a composite cable, the the DVD recorder to the capture card with an s-video cable. This uses the comb filter inside the DVD recorder, which is likely to give higher quality results.

 

Whoops...  I knew that.

 

Fez: I am so excited about Star Whores.
Hyde: Fezzy, man, it's Star Wars.
Author
Time
 (Edited)

Just so ya know i went through my closet and found my older kworld card which has yeilded much better results and has negative sharpness settings ! Also it doesn't crop off the image like my other card so im gonna do a couple of captures of side 1 and see how the video turns out.  :D

 

Heres a comparison of my old capture and my new capture

old one

new one

i realized though that i need to go back and fix the capture cards colors :/

Author
Time
 (Edited)

Hey guys i was curious what this line is on the side of the video and are the lines all around the video combing artifacts?

Also i wanted to show you guys the difference between my DVL-700 via S-Video  and my LD-V4400 via Composite

DVL

V4400

Author
Time
 (Edited)

The DVL-700 via S-Video is clearly superior. The rainbowing on the V4400 indicate a low-quality comb filter in the capture card.

As for the first image, can't see what you mean on the left hand side of the video (would help not to obscure it with a huge red line!).

The horizontal lines you see are not interlacing artefacts (trying to be careful with definitions so as not to confuse). Nor do I think they are caused by imperfect y/c separation in a low quality comb filter. I'm not sure what causes this effect, but I have seen it before. In that particular case it was related to the sharpness setting.

Guidelines for post content and general behaviour: read announcement here

Max. allowable image sizes in signatures: reminder here

Author
Time

Here's mine:

http://img399.imageshack.us/img399/8007/x0000ie8.th.jpghttp://img399.imageshack.us/images/thpix.gif

Fez: I am so excited about Star Whores.
Hyde: Fezzy, man, it's Star Wars.
Author
Time
 (Edited)

I cant wait to see your version finished Arnie.d its lookin great!

Moth3r said:

The DVL-700 via S-Video is clearly superior. The rainbowing on the V4400 indicate a low-quality comb filter in the capture card.

As for the first image, can't see what you mean on the left hand side of the video (would help not to obscure it with a huge red line!).

The horizontal lines you see are not interlacing artefacts (trying to be careful with definitions so as not to confuse). Nor do I think they are caused by imperfect y/c separation in a low quality comb filter. I'm not sure what causes this effect, but I have seen it before. In that particular case it was related to the sharpness setting.

 

Yeah i think the dvl has somewhat of an upper hand to the ld-v4400 but it stinks that it applies DNR. I wish there was a way to just have the 3d adaptive comb filter and the time base correction active and the dnr deactivated.

Author
Time
 (Edited)

I have hooked up my old dvd recorder to use it as a 3D comb filter for the LD-V4400 it got rid of almost all the dot crawl in the opening crawl. It looks really amazing. But after hooking it up and making 5 captures i have run into something strange.... When i line up all the captures to merge them the frames are different.. Like I match up the begining frame and end frame for all 5 videos but when i do ShowFiveVersions(a,b,c,d,e) after like a scene change or a subtitle the frames arent matched in the merge and the video is messed up.

I have to say all of you who have worked with capturing analog sources for preservation deserve a lot of credit for all the hard work you have had to go through of fixing and cleaning up your caps!

 

Here a picture of what im talking about

Free Image Hosting at www.ImageShack.us

they are all on the same frame but something got out of whack