Improving Quailty Script

This forum is for questions and discussion of all the aspects of handling your footage. If you have questions about capturing/ripping footage, AviSynth, or compression/encoding/converting, look here.

Improving Quailty Script

Postby cycangfx » Fri Mar 02, 2007 4:07 pm

Any script to improve the quailty of the FF7 Advent children movie?
User avatar
cycangfx
 
Joined: 02 Dec 2006

Postby cycangfx » Fri Mar 02, 2007 4:10 pm

opps, this should be in the avisynth section sorry....

[MOD 403: topic moved from other video editing systems]
User avatar
cycangfx
 
Joined: 02 Dec 2006

Postby Kariudo » Fri Mar 02, 2007 4:52 pm

I'll get on it
luckilly, I have FF7:AC, so I don't need a short segment from you
Image
Image
User avatar
Kariudo
Twilight prince
 
Joined: 15 Jul 2005
Location: Los taquitos unidos
Status: 1924 bots banned and counting!

Postby Kariudo » Fri Mar 02, 2007 6:14 pm

alright, here's what I came up with
Code: Select all
TFM(order=-1,mode=5,PP=7,field=-1,slow=2)
removegrain(mode=2)
Deen("a3d",1,4,4)
limitedsharpenfaster(wide=true,edgemode=1, strength=150)


dgindex says it's progressive (98% film), but turns out an interlaced file.
TFM fixes that, with "standard" paramaters thanks to Qyot_27

removegrain doesn't do too much more than slightly improve compressibility
mode 2 is really good at preserving detail

deen takes care of a bit more noise, and temporal smoothing improves compressibility.
this smooths out some detail...but in most cases you won't be able to tell the difference, much less after your editing.
the only big problem is if you want to use the scene when reno is in the helicopter near the beginning. If you want to use that scene, just get rid of Deen for that small segment.

limitedsharpenfaster...uhh...sharpens lines. I haven't been able to compare it to other strong sharpeners like vmtoon though.
edgemode = 1 only applies the sharpening to areas where edges are detected
wide = true just makes the sampling area 5x5 as opposed to 3x3

resizing is probably my weakest area; I'm not sure what to do if this is for the RATM mep.
(note for others, the mep tracks are to be submitted at 640x480 and this is an anamorphic widescreen source, 1.85:1)
Image
Image
User avatar
Kariudo
Twilight prince
 
Joined: 15 Jul 2005
Location: Los taquitos unidos
Status: 1924 bots banned and counting!

Postby Kariudo » Fri Mar 02, 2007 6:52 pm

bah...forgot to include a few assumptions:
I am using the US release (region 1) (also means no subtitles)
dgindex says it's 16:9 but the resolution is 720x480 (no letterboxing)
I'm using AviSynth 2.5.7

http://avisynth.org/LimitedSharpen
that'l get you the script for limitedsharpenfaster() and is nice and gives you direct links to required filters

Deen is avaliable in the amvapp...if you go looking for other versions of deen you'll find it...but not the one that has a3d as a paramater

a pm I wrote to someone else wrote:for removegrain you'll have to know what instruction sets your cpu has. if you don't know, you can find out by getting cpu-z
(scroll down until you see version 1.39
if you can only use SSE, put all the .dll files with SSE in thier names into the avisynth plugins folder (and all those that don't have SSE, SSE2 or SSE3 in thier names)

if you can use that and SSE2, put all the .dll files with SSE2 in thier names into the avisynth plugins folder (and all those that don't have SSE, SSE2 or SSE3 in thier names)

if you can use that and SSE3, put all the .dll files with SSE3 in thier names into the avisynth plugins folder (and all those that don't have SSE, SSE2 or SSE3 in thier names)

whatever the highest number after SSE says is what you need to use. (it's in the middle of the page you should see when you start it up
here is an example of the readout that I'm given.
Image

it shows me that I have SSE and SSE2, so I would need to put all the SSE2 things into my plugins folder.


you need to get both removegrain 0.9 and 1.0

it'l save you lots of time if you put all of these into avisynth's plugin directory.

you may also have to get masktools v1.5.8 avaliable from (among other places) here

TFM is part of the TIVTC package avaliable here

to resize, I suggest using lanzcosresize(int width, int height) (comes with avisynth)
Image
Image
User avatar
Kariudo
Twilight prince
 
Joined: 15 Jul 2005
Location: Los taquitos unidos
Status: 1924 bots banned and counting!

Postby Scintilla » Fri Mar 02, 2007 7:36 pm

Kariudo wrote:alright, here's what I came up with
Code: Select all
TFM(order=-1,mode=5,PP=7,field=-1,slow=2)
removegrain(mode=2)
Deen("a3d",1,4,4)
limitedsharpenfaster(wide=true,edgemode=1, strength=150)

Uh, why don't you use <b>d2v="C:\whatever_the_path_is\FF7AC.d2v"</b> in the TFM call?
ImageImage
:pizza: :pizza: Image :pizza: :pizza:
User avatar
Scintilla
(for EXTREME)
 
Joined: 31 Mar 2003
Location: New Jersey
Status: Quo

Postby Kariudo » Fri Mar 02, 2007 8:12 pm

old habbits I guess
went back and added that and didn't notice anything.
no fixed d2v was made when I used avsp nor when I encoded a random 941 frame sample with vdubmod

is there any other reason to specify the d2v paramater?
Image
Image
User avatar
Kariudo
Twilight prince
 
Joined: 15 Jul 2005
Location: Los taquitos unidos
Status: 1924 bots banned and counting!

Postby Scintilla » Fri Mar 02, 2007 8:51 pm

Kariudo wrote:old habbits I guess
went back and added that and didn't notice anything.
no fixed d2v was made when I used avsp nor when I encoded a random 941 frame sample with vdubmod

is there any other reason to specify the d2v paramater?

Yes, because TFM gets hints on how to best match the fields from the D2V file.
ImageImage
:pizza: :pizza: Image :pizza: :pizza:
User avatar
Scintilla
(for EXTREME)
 
Joined: 31 Mar 2003
Location: New Jersey
Status: Quo


Return to Footage Help

Who is online

Users browsing this forum: No registered users and 1 guest