Welcome to Doom9's Forum, THE in-place to be for everyone interested in DVD conversion. Before you start posting please read the forum rules. By posting to this forum you agree to abide by the rules. |
|
|
Thread Tools | Search this Thread | Display Modes |
23rd October 2016, 05:33 | #921 | Link | |
I'm Siri
Join Date: Oct 2012
Location: void
Posts: 2,633
|
Quote:
why should this time be any different? |
|
23rd October 2016, 10:36 | #922 | Link | |
Registered User
Join Date: Nov 2004
Location: Poland
Posts: 2,843
|
Quote:
The things is that Groucho2004 is not using 1 command with a block size, but recalculate option, which I assume may help with quality. Last edited by kolak; 23rd October 2016 at 11:39. |
|
23rd October 2016, 13:45 | #925 | Link |
Registered User
Join Date: Nov 2004
Location: Poland
Posts: 2,843
|
Thanks, but I have some idea.
Some scenes benefit from search=6 or search=7 (they produce double edges but no artefacts which is better when watched in full motion), so if I use 3 step re-calculate: search=4, then search=6, then search=7 then in theory I should get better interpolation (assuming best vectors are picked properly). No? This is almost like my 2 pass adaptive search/block idea. Last edited by kolak; 23rd October 2016 at 13:48. |
23rd October 2016, 14:01 | #926 | Link | |
I'm Siri
Join Date: Oct 2012
Location: void
Posts: 2,633
|
Quote:
You get search=3 by picking vectors calculated by exhaustive search in different (horizontal or vertical) directions based on SAD and you claimed that search=3 sucked,,, |
|
23rd October 2016, 14:20 | #927 | Link |
Registered User
Join Date: Nov 2004
Location: Poland
Posts: 2,843
|
Yes, search=3,4 or 5 maybe produced best vectors SAD wise, but they introduce artefacts on some scenes (e.g. car going quickly across the frame). In such a case search=7 will do bit of double edges, but this is way better than artefacts from e.g. search=3.
|
23rd October 2016, 14:40 | #929 | Link |
Registered User
Join Date: Nov 2004
Location: Poland
Posts: 2,843
|
I thought this will use search=3 for most frames and than just switch to eg. search=7 only for some frames (or none).
I'm running test now which then I will compare against pure search=3 mode (PSNR wise). |
23rd October 2016, 18:50 | #930 | Link |
Registered User
Join Date: Nov 2004
Location: Poland
Posts: 2,843
|
Looks like this makes most frames about the same as pure search=3, but then some of them (those with artefacts) are different as they are coming from search=7. So it does behave sort of hybrid and gives different result than pure search=3. Speed suffers a lot also
but... some scenes don't look good on TV- bit juddery... DCT=1 is stil best bet for quality and slow speed Last edited by kolak; 23rd October 2016 at 20:04. |
24th October 2016, 01:19 | #931 | Link |
HeartlessS Usurer
Join Date: Dec 2009
Location: Over the rainbow
Posts: 10,980
|
In my humble opinion the Feisty2, is the replacement for Didee [EDIT: time will tell if of worth] , the big dog, on the block.
Unfortunately, he is not that nice. He is what he is, we have to accept that and deal with it, not nice I know, but that is the Feisty2, thing. Me Loves him, but what a pain. (Feisty, had a few beers, Forgive me or die) [EDIT: You might want to specify how you will die, ie tragic accident etc]
__________________
I sometimes post sober. StainlessS@MediaFire ::: AND/OR ::: StainlessS@SendSpace "Some infinities are bigger than other infinities", but how many of them are infinitely bigger ??? Last edited by StainlessS; 24th October 2016 at 02:08. |
24th October 2016, 03:47 | #932 | Link | |
Registered User
Join Date: Feb 2002
Location: California
Posts: 2,691
|
Quote:
He is not entirely wrong about that. And I agree with the Didée comparison: neither of them suffer fools (which, I guess, includes me), but they both have pretty unique ways of looking at video problems and therefore develop very unusual solutions to problems. The only real downside is that very few of us can understand how the scripts work and the presentation is often more cryptic than it really needs to be. This makes it very difficult to build on their work or take it in a new direction. This is my only real complaint: both "open source" and the work done in these forums requires a collegial approach because one person can seldom do everything needed to create great software. Last edited by johnmeyer; 24th October 2016 at 03:50. Reason: added last sentence |
|
25th October 2016, 16:30 | #933 | Link |
Registered User
Join Date: Oct 2016
Posts: 3
|
Thank you for the explanation and Suggestion.
I want to ask about MBlockFps Processing Mode, a lot of people apparently using mode = 0 (the fastest), and I read the documentation, it seems the mode 3 to 8 mentioning masking, are they further increased masking and quailty also? Anyone that try this? |
6th November 2016, 00:49 | #934 | Link | ||
HeartlessS Usurer
Join Date: Dec 2009
Location: Over the rainbow
Posts: 10,980
|
Quote:
Code:
Function foo2(clip source, int "multi") { last = source multi = default(multi, 2) prefilt = DeGrainMedian() # some smoothing superfilt = MSuper(prefilt, hpad=16, vpad=16) # all levels for MAnalyse super = MSuper(source, pel = 2, hpad = 16, vpad = 16, rfilter = 4) # one level is enough for MRecalculate bw_1 = MAnalyse(superfilt, chroma = false, isb = true, blksize = 16, searchparam = 3, plevel = 0, search = 3, badrange = (-24)) fw_1 = MAnalyse(superfilt, chroma = false, isb = false, blksize = 16, searchparam = 3, plevel = 0, search = 3, badrange = (-24)) bw_2 = MRecalculate(super, chroma = false, bw_1, blksize = 8, searchparam = 1, search = 3) fw_2 = MRecalculate(super, chroma = false, fw_1, blksize = 8, searchparam = 1, search = 3) bw_3 = MRecalculate(super, chroma = false, bw_2, blksize = 4, searchparam = 0, search = 3) fw_3 = MRecalculate(super, chroma = false, fw_2, blksize = 4, searchparam = 0, search = 3) MBlockFps(source, super, bw_3, fw_3, FramerateNumerator(Source) * multi, FramerateDenominator(Source), mode = 0, Blend=False) return last } Quote:
Same top and bottom if vpad=0. EDIT: I have not been able to reproduce this during current testing, but have noticed it in the past. during some test comparisons, not using MRecalculate() produced some blockiness that was squashed by a single MRecalculate(blksize = 8). I really dont like Blend=True default, it makes detecting scene changes bloody impossible. This is probably what I will use in future, unless someone can suggest something better. Code:
Function DoubleRate(clip c) { c prefilt = DeGrainMedian() # some smoothing superfilt = MSuper(prefilt, hpad=16, vpad=16, rfilter = 4) # all levels for MAnalyse super = MSuper(pel=2, hpad=16, vpad=16,rfilter=4,levels=1) # one level is enough for MRecalculate bw_1 = MAnalyse(superfilt, chroma = false, isb = true, blksize = 16, searchparam = 3, plevel = 0, search = 3, badrange = (-24)) fw_1 = MAnalyse(superfilt, chroma = false, isb = false, blksize = 16, searchparam = 3, plevel = 0, search = 3, badrange = (-24)) bw_2 = MRecalculate(super, chroma = false, bw_1, blksize = 8, searchparam = 1, search = 3) fw_2 = MRecalculate(super, chroma = false, fw_1, blksize = 8, searchparam = 1, search = 3) MBlockFps(super, bw_2, fw_2, num=0, den=0, mode=0, Blend=false) Return Last } EDIT: Scratch above EDIT, produces weird flashing with Overlap=2 in MRecalculate.
__________________
I sometimes post sober. StainlessS@MediaFire ::: AND/OR ::: StainlessS@SendSpace "Some infinities are bigger than other infinities", but how many of them are infinitely bigger ??? Last edited by StainlessS; 6th November 2016 at 01:08. |
||
6th November 2016, 01:03 | #935 | Link | |
I'm Siri
Join Date: Oct 2012
Location: void
Posts: 2,633
|
Quote:
And then you'll see why small blocks suck at frame interpolation |
|
6th November 2016, 01:27 | #937 | Link |
HeartlessS Usurer
Join Date: Dec 2009
Location: Over the rainbow
Posts: 10,980
|
These two are garbage, I must have screwed up somewhere (weird flashing, EDIT: Green and Pink frames).
Code:
Function DoubleRate3(clip c) { c prefilt = DeGrainMedian() # some smoothing superfilt = MSuper(prefilt, hpad=16, vpad=16, rfilter = 4) # all levels for MAnalyse super = MSuper(pel=2, hpad=16, vpad=16,rfilter=4) bw_1 = MAnalyse(superfilt, chroma = false, isb = true, blksize = 16, OverLap=8, searchparam = 3, plevel = 0, search = 3, badrange = (-24)) fw_1 = MAnalyse(superfilt, chroma = false, isb = false, blksize = 16, OverLap=8, searchparam = 3, plevel = 0, search = 3, badrange = (-24)) MBlockFps(super, bw_1, fw_1, num=0, den=0, mode=0, Blend=false) Return Last } Function DoubleRate4(clip c) { c super = MSuper(pel=2, hpad=16, vpad=16,rfilter=4) bw_1 = MAnalyse(super, chroma = false, isb = true, blksize = 16, OverLap=8, searchparam = 3, plevel = 0, search = 3, badrange = (-24)) fw_1 = MAnalyse(super, chroma = false, isb = false, blksize = 16, OverLap=8, searchparam = 3, plevel = 0, search = 3, badrange = (-24)) MBlockFps(super, bw_1, fw_1, num=0, den=0, mode=0, Blend=false) Return Last } Code:
Function DoubleRate2(clip c) { c prefilt = DeGrainMedian() # some smoothing superfilt = MSuper(prefilt, hpad=16, vpad=16, rfilter = 4) # all levels for MAnalyse super = MSuper(pel=2, hpad=16, vpad=16,rfilter=4,levels=1) # one level is enough for MRecalculate bw_1 = MAnalyse(superfilt, chroma = false, isb = true, blksize = 16, OverLap=2, searchparam = 3, plevel = 0, search = 3, badrange = (-24)) fw_1 = MAnalyse(superfilt, chroma = false, isb = false, blksize = 16, OverLap=2, searchparam = 3, plevel = 0, search = 3, badrange = (-24)) bw_2 = MRecalculate(super, chroma = false, bw_1, blksize = 8, searchparam = 1, search = 3) fw_2 = MRecalculate(super, chroma = false, fw_1, blksize = 8, searchparam = 1, search = 3) MBlockFps(super, bw_2, fw_2, num=FramerateNumerator(c) * 2, den=FramerateDenominator(c), mode=0, Blend=false) Return Last }
__________________
I sometimes post sober. StainlessS@MediaFire ::: AND/OR ::: StainlessS@SendSpace "Some infinities are bigger than other infinities", but how many of them are infinitely bigger ??? Last edited by StainlessS; 6th November 2016 at 03:56. |
6th November 2016, 01:45 | #939 | Link |
HeartlessS Usurer
Join Date: Dec 2009
Location: Over the rainbow
Posts: 10,980
|
The two above Garbage functions still fail with Overlap=8, Green, Pink and Gray frames (sometimes), and jumping backwards and forwards in time.
EDIT: The two Garbage functions above, both work ok if changed to MFlowFps() and removing mode=0 (not an arg to MFlowFps). Looks like something may be broken in latest (I think) mvtools MBlockFps.
__________________
I sometimes post sober. StainlessS@MediaFire ::: AND/OR ::: StainlessS@SendSpace "Some infinities are bigger than other infinities", but how many of them are infinitely bigger ??? Last edited by StainlessS; 6th November 2016 at 02:05. |
6th November 2016, 02:00 | #940 | Link | |
I'm Siri
Join Date: Oct 2012
Location: void
Posts: 2,633
|
Quote:
Append "convettoy8()" at the end of the script and check if luma is okay If it is, then that's probably an mvtools bug |
|
Tags |
48fps, 60fps, framedoubling, interframe, smooth motion |
Thread Tools | Search this Thread |
Display Modes | |
|
|