Any Suggestions for Better Uprez to HD? — Page 2

here's some more results using solo 21's script for

upscaling to 720p...

princess leia:

=============================================

gout 8

-------

gout 8 crop

-------

gout 8 => simple resize 720p

-------

solo21 -> gout 8 upscaling 720p

-------

adywan -> gout 8 1080p->720p
-------

here's some more results using solo 21's script for

upscaling to 720p...

darth vader:

=============================================

gout 9

-------

gout 9 crop

-------

gout 9 => simple resize 720p

-------

solo21 -> gout 9 upscaling 720p

-------

adywan -> gout 9 1080p->720p
-------

later

-1

Octorox said:

This thread makes me think that the best way to get an HD OT would be to use Adywan's color corrected SE and splice in elements from the stabilized/upscaled GOUT. Is Adywan's HD version significantly compressed?

 i've thought about that too..

but there are so many differences in color/textures/effects... i'm not sure

how this would be accomplished, and if the differences would be too much..

any ideas?

later

-1

solo21 , or oldschooljedi,

don't know if there any updates to this script..

(OSJ have you tried it yet?)..

anyways, i'm getting some interlacing effects when i use the

script, so i'll try adding some deinterlace filters..

later

-1

Octorox said:

This thread makes me think that the best way to get an HD OT would be to use Adywan's color corrected SE and splice in elements from the stabilized/upscaled GOUT. 

I agree. I think there are a couple of those projects in the works. 

getting back to rescaling images..

there's a photoshop plugin with batch capability that looks impressive:

==================================================

http://www.alienskin.com/blowup/index.aspx

Image Quality: Blow Up® 2 is a Photoshop® plug-in that produces enlargements dramatically sharper than any other software, especially Photoshop! Blow Up uses an innovative algorithm that temporarily converts pixels to a vector representation which results in perfectly smooth, crisp edges. Simulated texture and film grain keep photos looking natural even in extreme enlargements. Blow Up can even clean up JPEG compression artifacts when a client insists on using a low quality source image from the Web.

Image Quality
Image quality is improved in many ways. Results are sharper, both during upsizing and downsizing. Noisy images are handled better. Research from our Exposure product was used to improve simulated grain. You now have control over simulated natural texture.

examples at:

http://www.alienskin.com/blowup/blowup_examples.aspx

check it out...

later

-1

solo21 said:

I've not touched this in a while, but I'll dig up some results for you all.

I was using this script for a prequel trilogy edit, so my original source was the PAL DVD (deleted scenes), upscaling to match the WookieeGroomer 1080p rips.  It's working well for me, but obviously the original source has a lot less grain etc than any OT footage.

 

awesome,

thanks for checking back, i've been having a lot of fun with your script..

yes, i've done some 1080p rescaling, but it takes a lot longer, and its

hard to show here......

if you have any other ideas (about interlace artifacts, improvements), please

let us know..

later

-1

Hi,

Sorry to turn this into a bit of an AVISynth Help Club, but I seem to get the following error with Solo21's script:

MVAnalyse does not have a named argument 'overlap'

I know it's going to be something obvious, but I can't get my head around which bit I need to debug here:

function MVDegrain2i(clip "source", int "overlap", int "dct", int "idx")

{

overlap=default(overlap,0) # overlap value (0 to 4 for blksize=8)

dct=default(dct,0) # use dct=1 for clip with light flicker

idx=default(idx,1) # use various idx for different sources in same script

fields=source.SeparateFields() # separate by fields

backward_vec2 = fields.MVAnalyse(isb = true, delta = 2, pel = 2, overlap=overlap, idx = idx,dct=dct)

forward_vec2 = fields.MVAnalyse(isb = false, delta = 2, pel = 2, overlap=overlap, idx = idx,dct=dct)

backward_vec4 = fields.MVAnalyse(isb = true, delta = 4, pel = 2, overlap=overlap, idx = idx,dct=dct)

forward_vec4 = fields.MVAnalyse(isb = false, delta = 4, pel = 2, overlap=overlap, idx = idx,dct=dct)

fields.MVDegrain2(backward_vec2,forward_vec2,backward_vec4,forward_vec4,thSAD=400,idx=idx)

Weave()

}

Can anyone help? What does it all mean Basil?

It probably means you have an older version of MVTools. The latest is (as far as I know) v2.4.2 and can be found here.

Thanks Mother, I tried the new install (mvtools2.dll) and AVISynth then complains that it can't find the function MVAnalyse, which appears to be now MAnalyse (?) in which case the usage also seems to have changed.

For some reason I was using a really old version, and it looks like once I've upgraded to the last of the 1.x branch it might have fixed the problem.

And I've finally got it working and then I get an Assertion Error and AVISynth explodes...

Might give up on this one...

The MPEG-2 video from the GOUT DVD is YV12, and I'm pretty sure that the stabilization script does not change the colourspace. Therefore, your lossless AVI should also be YV12, unless you changed it to RGB by accident when you saved it.

You can easily find out with this script:
AviSource("C:\filename.avi")
Info()

Just had a closer look at the script, and I have serious doubts about its suitability. Can anyone explain the logic behind using SeparateFields() and Weave() steps for a non-interlaced source?