![]() |
Quote:
Keep up the good work guys :D . Racer99 |
@Kwag,
My test was Drumline DVD, 118 minutes, full screen, low action movie, and the target filesize was one 80 minute CD. Tok file prediction was very low for 704x480 (CQ 15), 528x480 (CQ 19) and 352x480 (CQ 30). I used the current Optimal Script for v2.5x and default settings for VCD in Tok. 1) ToK CQ predictions at any resolution very low :cry: 2) Encoding time veeeeeeerrrrry slooooooooow :cry: 3) Filesize bigger than using avisynth 2.08 :cry: 4) Picture quality pooooor (due to low CQ prediction) :cry: I've got to wonder if this process is better than using avisynth 2.08+ :roll: -bp |
8O Something is WAY wrong bp :!:
My encoding with the current script and 2.52 is much faster than with 2.08. Have you tried deinstalling 2.08 completely, and then installing 2.52 :idea: The quality I have achieve now has surpassed any previous encoding I have ever done before. There's just no comparison :D -kwag |
@Kwag,
Kwag wrote: Quote:
errors, but Tok file predictions are much lower than 2.08 for the same resolutions. Encoding times are 15+ hours at 352x240 instead of 5 to 6 with v2.08. Filesize is also gaining weight with v2.5.2. I don't know what magic you're using but I can't seem to duplicate anything close to the picture quality you've getting :? LoadPlugin("E:\DVD BACKUP\1 - DVD2KVCD\3 - MOVIESTACKER\Filters25\MPEG2Dec3.dll") LoadPlugin("E:\DVD BACKUP\1 - DVD2KVCD\3 - MOVIESTACKER\Filters25\Unfilter.dll") LoadPlugin("E:\DVD BACKUP\1 - DVD2KVCD\3 - MOVIESTACKER\Filters25\STMedianFilter.dll") ## Defined Variables and Constants ## # MaxTreshold = 1.58 scd_trigger = 15 # Scene change trigger value. cf = 0 # Current frame. lf = 0 # Last frame val = 0 # Dynamic value applied to filters # #### ## Main section and static filters ### # Mpeg2Source("D:\Temp\movie.d2v") Limiter() UnFilter(50, 50) BicubicResize(528,480) STMedianFilter(8, 32, 0, 0 ) TemporalSoften(2,7,7,3,2) # Experimental! MergeChroma(blur(1.50)) MergeLuma(blur( 0.1)) # #### ## Dynamic Linear Adaptive filtering and Scene Change Detection ## # FrameEvaluate("cf = YDifferenceToNext()") FrameEvaluate("val = (cf / 16) + 0.1") ScriptClip("((cf - lf) < scd_trigger) ? (val < MaxTreshold) ? MergeLuma(blur(val)) : MergeLuma(blur(MaxTreshold)) : MergeLuma(blur(1.5))") # #### #LetterBox( Your_Values_Here ) # Depends on situation. Use MovieStacker! #AddBorders( Your_Values_Here ) # Depends on situation. Use MovieStacker! Limiter() FrameEvaluate("lf = YDifferenceToNext()") # #### -bp |
Hi bp,
Could this be a processor related issue :?: I'm on a P4 @1.6Ghz. Are you on Intel or AMD :?: -kwag |
Bug Squashed !!!
New script posted :!:
:imstupid: There was an almost invisible problem on the previous posted script. It was very hard do see, but it was present if you looked at your .avs and moved frame by frame with VirtualDub. The third frame after every scene change was slightly blurred, in comparison to the adjacent frames. It's been squashed :!: :gun: I also added a "Version" to the script, so we can all keep track of the last date/time modified. I hope this script can now create a basic framework so we all can grow it, and apply different filters (if nedded!) for different filtering strategies. -kwag |
Im using the newest script posted, but while the movie is encoding I get this green bar along the right side in TMPGenc. Does anyone know what causes this?..Here is my script from TOK
## DLL Section ## # LoadPlugin("C:\Filters25\MPEG2Dec3.dll") LoadPlugin("C:\Filters25\STMedianFilter.dll") LoadPlugin("C:\Filters25\UnFilter.dll") # #### ## Defined Variables and Constants ## # MaxTreshold = 1.58 scd_trigger = 15 # Scene change trigger value. nf = 0 # Current frame. lf = 0 # Last frame val = 0 # Dynamic value applied to filters # #### ## Main section and static filters ### # Mpeg2Source("C:\DVD\matrix.d2v") Limiter() UnFilter(50, 50) BicubicResize( 352,240,0,0.6,8,0,740,480 ) STMedianFilter(8, 32, 0, 0 ) TemporalSoften(2,7,7,3,2) # Experimental! MergeChroma(blur(1.50)) MergeLuma(blur( 0.1)) # #### ## Dynamic Linear Adaptive filtering and Scene Change Detection ## # FrameEvaluate("lf = YDifferenceToNext()") FrameEvaluate("nf = YDifferenceToNext()") FrameEvaluate("val = (lf / 16) + 0.1") ScriptClip("((nf - lf) < scd_trigger) ? (val < MaxTreshold) ? \ MergeLuma(blur(val)) : MergeLuma(blur(MaxTreshold)) : \ ( (sign(nf-lf)) != -1) && (val < MaxTreshold) ? MergeLuma(blur(val)) : MergeLuma(blur(.75)) ") # #### #LetterBox( 16,16,16,16 ) # Depends on situation. Use MovieStacker! #AddBorders( 16,16,16,16 ) # Depends on situation. Use MovieStacker! Limiter() # #### AssumeFPS(23.976) LoadPlugin("C:\TOK\ToK_EXTRAS\Sampler\Sampler-2.5.dll") oldfps = framerate interval = round((FrameCount/24)/59.940)/10 nFrames = round(24) SelectRangeEvery( (round(framecount/interval)),nFrames) David |
Quote:
-kwag |
Quote:
Ignore me, Im just thinking "out loud" :) Quote:
Anyway you've changed the script now so it doesn't matter :) |
Re: Bug Squashed !!!
Quote:
|
Re: Bug Squashed !!!
Quote:
Right now, the script blurs only the frame after a scene change, using mergeluma(blur(MaxTreshold)). -kwag |
Problem with mergeluma now corrected.
I've found what's possibly a bug in AviSynth and/or mergeluma.
After doing a test encode, I see some scenes where there's no scene change at all, and one frame slightly increases brightness. It looks like a small "flash". I've traced the problem to be when mergeluma blur value makes a transition from around ~0.180 to ~0.205. I've corrected this by changing the minimum value to be 0.2. So now mergeluma fluctuates between min=0.2 to max=1.58. I have been able to reproduce the problem consistently. Here's a very short mpeg file which shows the effect: www.kvcd.net/flash.mpg And here's the same clip with the corrected min of .2 showing a clean clip: www.kvcd.net/no-flash.mpg The current script posted corrected this. -kwag |
Quote:
|
Re: Problem with mergeluma now corrected.
Quote:
|
Re: Problem with mergeluma now corrected.
Quote:
I just PM'ed sh0dan about it :) -kwag |
Just a quick (and stupid :oops: ) question...
Where in the script are you supposed to put the lines: FieldDeinterlace() or Telecide() Decimate() Am I correct to presume I should put them directly after the first Limiter() line? Newbie warning, I know... Sorry! :roll: |
Quote:
-kwag |
Thanks kwag! :wink: I just wanted to be sure to get it right.
You're the Man! 8) |
I need a simple script to reproduce this.
Does this happend with or without mergeluma? (When you just blur)? |
Hi Kwag,
In testing the new script I have noticed that on scenes with very little action the image seems to shift from left to right.... The Scene I'm looking at is from the movie "Meet the Parents" when they are at the dinner table. You can see the glass wine move from left to right... I have just tried this a couple of minutes ago with the last script you posted. I don't notice this effect in scene with action/movement... Have you notice the same? |
Quote:
You can reproduce the effect with the current script, which is this: Code:
## DLL Section ##That will make the value for mergeluma to go from ~0 to 1.58 -kwag |
Quote:
Try removing that filter and encode that scene. I don't think it's related at all to the current script, which basically changes luma values. The other filters are applied statically and evenly throughout all frames. So I don't think luma variations, specially low variations, would make any shifting of pixels. -kwag |
@kwag: "Simple"???
Anyway - it would seem like it is a blur() issue. It's coded by someone who isn't any longer in touch with us, so it might be hard to track down. For now - you seem to do fine with values > 0.2! |
Quote:
doom9: green pixels in leftmost column, but only when using Virtualdub to decode XviD fourCC |
:( that kinda sucks that we have to stick with Blur values >0.2 ...
I still think this blurs the picture a lot, even at 704x480/576... Well I guess we have to live with it... |
Quote:
Are you interested in seing a small sample? Follow this link: http://www.geocities.com/dorvalcs/meettheparents.zip The file is less than 700k... But it is very noticeable! I'm cureous to see what you think... As usual! :wink: |
Quote:
We'll find a fix or a workaround in a few "centons" :mrgreen: -kwag |
Quote:
-kwa |
Quote:
Thanks for taking the time to look at it!! |
The 0.2 luma problem fixed!
Get rid of STMedianFilter NOW :!: :twisted:
It screwes up the luma somehow :!: That was the problem. I just found that out, and I just updated the script. Now the luma varies from 0 to MaxTreshold again, for an extra "crispy" movie without "flashes" Edit: This sucks :!: The STMFilter definitely causes a shift in color balance at some scenes, but I'm still getting some flashes when I lowered the value to MIN on luma. So I guess we'll have to stick to 0.2 for a while until this is further debugged :cry: :evil: -kwag |
Quote:
Are you positive this doesn't show, even minimal, if you encode without filters :?: -kwag |
Quote:
Mmmmmm.... I should have done this before :oops: ... I've been encoding all morning with various setting to see how I could get rid of this :evil: After removing ALL the filters... The image is still shifting :x Sooooooo.... I put my DVD in my player.... And guess what??? The shifting is part of the source :oops: :oops: :oops: Oh boy!!! Sorry for the trouble Kwag... Next time I'll know what to look at first :wink: |
Quote:
-kwag |
Kwag,
i test all the last scripts from 2 days ago at the last.(really new) i ever see that less filters is better, is a question of choice of arguments like you did...and you was great :!: the last script is faster and give more details. :wink: i can't find words to write sometimes and the opinions of girv are great to me. i can speak like he posted. he always see the same as me and he wrote with big details. go ahead friends, is really better. :wink: thanks. |
Would some kind soul care to test this please :roll:
I hope I'm finally hitting the nail correctly, and hope that we can all have a stable script that we can start to depend on, without changing every hour :!: Hopefully, this one won't "blink" :!: This one also blurs the frames adjacent to a scene change. That is, the frame before the scene change, AND the frame after the scene change. Not visibly detectable, but softer to the encoder and final file size. Encode a small clip, and then load it up in Vdub. Move slowly through your clip, and notice the scene changes. See if you can see the small blur between scene changes and if there are any frames with any color bleed or flashes. THERE SHOULDN'T. I'm trying to reproduce the problems, but right now I can't. It seems that it's working. I added the STM filter again, because apparently it was some interaction with STM and the way scene detection was working. I really can't pin point what was happening. But the way the script is now, it seems that the anomalies don't show anymore 8O I won't update the current script until I see some reports about this script that it works flawlessly :!: Code:
## DLL Section ## |
Well man i think is kicking now i just did the same clip that i been doing now 4 2 friking days and the the change between Blur and sharp scenes it much much better so i think we cooking now we almost if not done. :wink:
PS: i just love this crop :lol: |
@ Girv
I don't know if you have already fixed your problem. But i got also a green bar at the bottom when i forgot to change the resize and addborders values when i started a new project. When i used the correct values (from moviestacker) the bar was gone. @ Kwag I used the script for avs 2.52 yesterday and everything was ok. But today I wanted to do some tests and the script doesn't work corect anymore. The color is messed up, so know i have blue people. Quite funny but it's not what I want. Do you have any idea what this can be? I am using your optimal script. Thanks |
Quote:
It's either a mathematical problem ( floating point error or something deep in the blur code ) of some other logical software control flow error. -kwag |
@ovg64 and all,
Make sure this line reads: FrameEvaluate("val = (lf / 16) + 0.05") I forgot the "+" sign on the script above. I edited the script. -kwag |
Yeah its Kool now, using the old adaptine script there was a part where actually the face was sharp and the mouth was a blur 8O :lol: , no more of that.
|
Site design, images and content © 2002-2026 The Digital FAQ, www.digitalFAQ.com
Forum Software by vBulletin · Copyright © 2026 Jelsoft Enterprises Ltd.