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. Domains: forum.doom9.org / forum.doom9.net / forum.doom9.se |
26th March 2008, 14:03
|
#2 | Link |
|
Registered User
Join Date: Jan 2005
Location: cz
Posts: 704
|
deinterlace:
http://image.diku.dk/sunebio/KellerLauzeNielsen.pdf |
|
|
|
17th July 2008, 11:52
|
#4 | Link |
|
Registered User
Join Date: Apr 2002
Location: Germany
Posts: 5,407
|
Nice finding.
 Note on the last one: It's really not very hard to implement something quite similar with a script. The following should give the idea (example is actually not-working, it's just to show a possible general flow) Code:
dark = last bright = dark .levels(...) # or Tweak(), ColorYUV(), HDRAGC() ... or whatever suits you diff = mt_makediff( dark, bright, U=3,V=3 ) bvec1 = dark .MVAnalyse( isb=true, idx=1 ) fvec1 = dark .MVAnalyse( isb=false,idx=1 ) diff_temp = diff1.MVDegrain1( bvec,fvec, idx=2 ) # or use MVDegrain2/3 with more vectors diff_spat = diff1. MinBlur(2) # or use tbilateral; MinBlur is the fast+lazy shortcut err_bw = dark.MVMask( bvec1, kind=1, ml=50 ) # tweak ml, perhaps also gamma err_fw = dark.MVMask( fvec1, kind=1, ml=50 ) # (kind=1 -> motion-compensated , kind=0 -> motion adaptive) err = Merge( err_bw, err_fw, 0.5 ) diff_spatemp = diff_temp .mt_Merge(diff_spat, err, U=3,V=3) result = dark .mt_makediff( diff_spatemp, U=3,V=3) After a quick glance through the paper, there are a few points ... they don't use mo-comp at all, the method is 'only' motion-adaptive. Temporal-Bilateral and Volumetric-Bilateral ... well, those we don't have in Avisynth. At least Temporal-Bilateral could be self-cooked with ttempsmooth, via pfclip and according masks. Reminds me, one feature that's missing in MVTools: create vectors on clip A, then exchange compensation with clip B (MVChangeCompensate) and re-derivate SADs from clip B accordingly. Like "MVChangeSAD(vectors, newclip)". Occasionally, this might become handy ... e.g. for stuff like this here. 
__________________
- We´re at the beginning of the end of mankind´s childhood - My little flickr gallery. (Yes indeed, I do have hobbies other than digital video!) |
|
|
|
|
5th August 2009, 09:50
|
#8 | Link |
|
Registered User
Join Date: Jan 2005
Location: cz
Posts: 704
|
http://www.nashruddin.com/phase-corr...v.html#results
http://www.wedesoft.demon.co.uk/horn...ation-txt.html http://www.google.cz/url?sa=t&source...lEHfeyoXz8PSOA there are many articles about phase correlation, could this be improvement when implemented to mvtools? Depan already uses phase-shift. Last edited by Terka; 5th August 2009 at 10:18. |
|
|
|
|
1st January 2010, 13:01
|
#9 | Link |
|
Registered User
Join Date: Jan 2005
Location: cz
Posts: 704
|
http://www.faculty.idc.ac.il/toky/Pu...l/superRes.pdf
http://www.utia.cas.cz/files/Soutez_...i_help_v12.pdf the matlab application is free for non commercial application |
|
|
|
 |
| Thread Tools | Search this Thread |
| Display Modes | |
|
|
Linear Mode
