About DAA()

[zanaitoryoushi]

....................

[manono]

The Repair.dll comes included as part of the RemoveGrain package.

[zanaitoryoushi]

.................................

[manono]

can you evaluate if this is good enough?

No, no one can evaluate a script without seeing the video to which it is being applied.

[zanaitoryoushi]

...........................

[sanlyn]

LoadPlugin("C:\Documents and Settings\Administrator\Desktop\Tools\AviSynth plugins\SmoothLevels\ICL\avs-2.6.x\x86\SmoothAdjust.dll")
video=Smoothlevels(video,chroma=100, preset="tv2pc")

LoadPlugin("C:\Documents and Settings\Administrator\Desktop\Tools\AviSynth plugins\dfttest\dfttest.dll")
video=dfttest(video, tbsize=2,sigma=1.2,tmode=1)

LoadPlugin("C:\Documents and Settings\Administrator\Desktop\Tools\AviSynth plugins\Sharpeners\UnFilter.dll\Unfilter.dll")
video=Unfilter(video, 70, 70)

LoadPlugin("C:\Documents and Settings\Administrator\Desktop\Tools\AviSynth plugins\Spatio-Temporal Filter\deen.dll")
video=Deen(video, mode="a3d", rad=3, thrY=7, thrUV=9, min=0.3)

#Colors
video=Tweak(video,hue=0,sat=1,bright=5,cont=1)

can you evaluate if this is good enough? please point out any flaws that you see, i appreciate criticism, i'm fairly new with this..

Can't advise in detail. Screen captures don't tell how the original video looks. Colors look pretty correct, but your script might be easier to read and maintain if more conventionally arranged:

LoadPlugin("C:\Documents and Settings\Administrator\Desktop\Tools\AviSynth plugins\SmoothLevels\ICL\avs-2.6.x\x86\SmoothAdjust.dll")
LoadPlugin("C:\Documents and Settings\Administrator\Desktop\Tools\AviSynth plugins\dfttest\dfttest.dll")
LoadPlugin("C:\Documents and Settings\Administrator\Desktop\Tools\AviSynth plugins\Sharpeners\UnFilter.dll\Unfilter.dll")
LoadPlugin("C:\Documents and Settings\Administrator\Desktop\Tools\AviSynth plugins\Spatio-Temporal Filter\deen.dll")

Smoothlevels(chroma=100, preset="tv2pc")
Unfilter(70, 70)
dfttest(tbsize=2,sigma=1.2,tmode=1)
Deen(mode="a3d", rad=3, thrY=7, thrUV=9, min=0.3)
#Colors
Tweak(hue=0,sat=1,bright=5,cont=1)

However, this statement

Smoothlevels(chroma=100, preset="tv2pc")

crushes the darkest colors.


[Attachment 13491 - Click to enlarge]

Note how the darks are smashed against the left-hand side of the histogram. You won't see any detail there. Looks OK on a PC, might not look so great to an encoder or TV. Depends on whether your intent is PC display or TV display. In this case, however, we're looking at a jpg capture, not at the original video frame.

[zanaitoryoushi]

...................................

[sanlyn]

Actually, I thought the images showed good work. It does take some experimentation to work with UTube video, some of the material they mount can look awful and requires heavy-duty work. As for processing that slows down systems -- all processing is a slow-down situation, some worse than others. IF you were working with real problems, you'd use strong filters that run at less than 2 or 4 fps, even on a "fast" machine.

It's difficult to judge the effect of scripts or the filters required without a sample of the original. That's not always so easy (at work, anyway).

[poisondeathray]

speed tips:

consider using daa3 instead of daa (same thing, but it uses nnedi3 which is faster and marginally better than nnedi2) - But why are you using daa at all? is there aliasing in the source?

--merange 32 is overkill for psp or SD footage (it's probably overkill for HD footage as well). This will reduce your processing speed with no benefit . Even on "placebo" settings, merange is 24 - this tells you how useless --merange 32 is.

[zanaitoryoushi]

.................................

[poisondeathray]

i'm actually in the office

Stark Tower? are you "Jarvis" ?

i just saw that randomly on the net and decided i use that, so should I eliminate -- merange & -- me esa?
will the end result suffer any loss in detail at all without that?

You won't notice the difference. Only the room will be hotter

[manono]

all i did was to make the colors look more vivid...

Not if you used that script, you didn't. All the Tweak settings do is raise the brightness (black levels) a little bit. Neither the saturation nor the contrast settings do anything. They're both the default settings.

[zanaitoryoushi]

.......................................

[zanaitoryoushi]

all i did was to make the colors look more vivid...

Not if you used that script, you didn't. All the Tweak settings do is raise the brightness (black levels) a little bit. Neither the saturation nor the contrast settings do anything. They're both the default settings.

oh sorry yeah well i guess that's the case, i thought using tv2pc made it for me, the original encode without the plugins are a little on the light side

[sanlyn]

Computer monitors and tv display the contrast and color range differently. A video made "to spec" for tv will usually look a bit more dull at the extremes than it will on a PC. That's why I asked earlier, whether you wanted PC or TV playback.

[zanaitoryoushi]

Computer monitors and tv display the contrast and color range differently. A video made "to spec" for tv will usually look a bit more dull at the extremes than it will on a PC. That's why I asked earlier, whether you wanted PC or TV playback.

yes i understand, you are right . i think a "pc-like" display is much better to look at on my tiny psp screen that's why I deliberately used the tv2pc function.

[poisondeathray]

also is aq-strength still necessary considering the plugins i use?

honestly this was my first attempt in using plugins/filters(whatever they're called), i've never done this before, haha i decided to learn since just setting it to CQ 16 and a sharpen(.3) plus that aq-strength 1.2 to remove unnecessary blocking (macroblocks or so i read somewhere) is not enough to deliver the quality i'm looking for.

You won't notice the difference between 1.0 and 1.2. If you go frame by frame and zoom in you might notice a slight difference in dark, flat shadow areas , but the bitrate will be different because you are using CQ mode (CRF). Strictly speaking, it's not really a valid comparison unless you compare videos with the same bitrate.

If there's blocking in the source (not caused by your encode), then you should prefilter that. If it's caused by your enocoding, then you need to make some adjustments

Just use your eyes, and if you think it looks good enough that's all that matters

[zanaitoryoushi]

.....................................