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.

 

Go Back   Doom9's Forum > Hardware & Software > Software players

Reply
 
Thread Tools Search this Thread Display Modes
Old 22nd October 2012, 16:00   #14981  |  Link
NicolasRobidoux
Nicolas Robidoux
 
NicolasRobidoux's Avatar
 
Join Date: Mar 2011
Location: Montreal Canada
Posts: 269
Non, mais je parle le joual. http://en.wikipedia.org/wiki/Joual

Last edited by NicolasRobidoux; 22nd October 2012 at 16:08.
NicolasRobidoux is offline   Reply With Quote
Old 22nd October 2012, 17:04   #14982  |  Link
NicolasRobidoux
Nicolas Robidoux
 
NicolasRobidoux's Avatar
 
Join Date: Mar 2011
Location: Montreal Canada
Posts: 269
When I have a bit of time, I'll recompute the "optimal" deblur taking into account that you restrict yourself to a 6x6. My guess is that it will go down just a little.
NicolasRobidoux is offline   Reply With Quote
Old 22nd October 2012, 17:10   #14983  |  Link
NicolasRobidoux
Nicolas Robidoux
 
NicolasRobidoux's Avatar
 
Join Date: Mar 2011
Location: Montreal Canada
Posts: 269
Ah! With the optimization criterion that gives 0.981..., it does not make a difference if you restrict yourself to a 6x6. (It makes a difference with the other criterion.)
So: Just keep 0.981..., the deblur I would guess works best for Jinc3 with AR.
NicolasRobidoux is offline   Reply With Quote
Old 22nd October 2012, 17:22   #14984  |  Link
madshi
Registered Developer
 
Join Date: Sep 2006
Posts: 9,140
Quote:
Originally Posted by NicolasRobidoux View Post
Ah! With the optimization criterion that gives 0.981..., it does not make a difference if you restrict yourself to a 6x6. (It makes a difference with the other criterion.)
So: Just keep 0.981..., the deblur I would guess works best for Jinc3 with AR.
Ok, so all is good - thanks!
madshi is offline   Reply With Quote
Old 22nd October 2012, 17:23   #14985  |  Link
DragonQ
Registered User
 
Join Date: Mar 2007
Posts: 934
madshi, can you figure out why this file won't deinterlace with any field order setting in MadVR? I have a feeling MKV Merge is to blame but I'm not sure. EVR struggles with it too - stepping frame-by-frame shows it deinterlaces properly but plays at 46-48 fps instead of 50 and has a bit of a dodgy graph, resulting in jerkiness.

Sample Clip

The original TS file plays fine using both EVR and MadVR.
__________________
TV Setup: LG OLED55B7V; Onkyo TX-NR515; ODroid N2+; CoreElec 9.2.7
DragonQ is offline   Reply With Quote
Old 22nd October 2012, 17:25   #14986  |  Link
madshi
Registered Developer
 
Join Date: Sep 2006
Posts: 9,140
I'm sorry, but if the original file plays fine then it can't be madVR's fault, can it? I really don't have time to look into how mkvmerge might have screwed things up. It's already more than enough work to only work on my own bugs...
madshi is offline   Reply With Quote
Old 22nd October 2012, 17:33   #14987  |  Link
Xaurus
Registered User
 
Join Date: Jun 2011
Posts: 288
Quote:
Originally Posted by madshi View Post
Every current Panasonic plasma has a *much* better native contrast than any LCD. Of course LCD uses some tricks to level the playing field like e.g. local dimming, but these tricks can introduce artifacts on their own. LCD backlights lose light, just as much as plasmas do. Current Panasonic plasmas are rated to reach half brightness after 100.000 hours of use. If you leave on your plasma 8 hours every day, it will reach half brightness after 34 years of use. That's much better than CRTs used to be. And LCD backlights won't hold any longer than this, either. There are valid and good reasons for choosing LCD over plasma, but lifetime is not one of them.
As a previous owner of the following plasma brands:

LG
Pioneer (kuro)
Panasonic

I bought the 60" LG plasma in the fall of 2010. By the spring of this year (1,5 years) I was unable to get it to display the blacks that it initially had during the first period. It became so bad (black was literally grey) that I simply decided to sell it and got Sony's top-model LCD in 55", and now after half a year I still think it's the best TV I ever owned. Ok, 5" down in screen size but the 65" model was too expensive, he he.

Bottom line: For plasmas, take the manufacturer claimed brightness reduction time with a grain of salt. I know the previous gen (2011) Panasonics also had this problem.
__________________
SETUP: Win 10/MPC-HC/LAV/MadVR
HARDWARE: Fractal Design Node 804 | Xeon E3-1260L v5 | Supermicro X11SSZ-TLN4F | Samsung 2x8GB DDR4 ECC | Samsung 850 EVO 1TB | MSI GTX 1650 Super | EVGA G2 750
Xaurus is offline   Reply With Quote
Old 22nd October 2012, 17:34   #14988  |  Link
DragonQ
Registered User
 
Join Date: Mar 2007
Posts: 934
Alright, I'll ask Mosu.
__________________
TV Setup: LG OLED55B7V; Onkyo TX-NR515; ODroid N2+; CoreElec 9.2.7
DragonQ is offline   Reply With Quote
Old 22nd October 2012, 18:15   #14989  |  Link
6233638
Registered User
 
Join Date: Apr 2009
Posts: 1,019
Quote:
Originally Posted by madshi View Post
I've had a plasma myself and I couldn't see any hint of RBE. Plasmas do not draw images sequentially like DLPs do. 1-chip DLPs draw color by color. Plasmas have red/green/blue subpixels and all 3 are active at the same time. So there's no reason plasmas should have any RBE. What they do have is dithering noise, because plasma pixels can only be on or off, but they can't be half on. So they simulate half on by pulsating very quickly. Why don't you go to a shop and check yourself whether you can see any hint of RBE with plasmas? They really do have the best image quality, IMHO, if you can control ambient light.
Actually, Plasmas do draw their image in quite a similar way to DLP. The pixels can only be turned on or off (similar to DLP) and gradation is built up over several subframes by pulsing the pixels on/off for varying lengths of time to control brightness. Because each phosphor has different response/decay characteristics, this does essentially draw the image in sequential colour, even though all the pixels are technically being addressed at the same time.

This video illustrates how Plasma draws its image: http://www.youtube.com/watch?v=tGpdah32n3c

Quote:
Originally Posted by madshi View Post
Every current Panasonic plasma has a *much* better native contrast than any LCD. Of course LCD uses some tricks to level the playing field like e.g. local dimming, but these tricks can introduce artifacts on their own. LCD backlights lose light, just as much as plasmas do. Current Panasonic plasmas are rated to reach half brightness after 100.000 hours of use. If you leave on your plasma 8 hours every day, it will reach half brightness after 34 years of use. That's much better than CRTs used to be. And LCD backlights won't hold any longer than this, either. There are valid and good reasons for choosing LCD over plasma, but lifetime is not one of them.

...

You should get an LCD instead of a plasma if (and only if, IMHO):

- you want to use the display as a computer monitor, too (word, programming, browsing)
- you have lots of ambient light in your room and no way (or will) to control it
- power consumption is extremely important to you

If none of these apply, and if you truly value image quality, then you should seriously consider giving Plasma another chance. 'nough said. This is my last comment on LCD vs Plasma in this thread.
I would argue that if you value image quality, you would give Plasmas a wide berth. They measure well on native contrast, viewing angle and color accuracy (nowadays) but the way they draw their image due to the heavy use of dithering, motion artefacts, the ABL circuit, flicker, crosstalk, poor gradation capabilities (particularly above 60Hz) all result in decidedly poor image quality in my experience.


For what it's worth, some of the best LCD panels out there are approaching 5,000:1 native contrast, and it is only this year that Panasonic has managed to exceed that.

Edge LED and "local dimming" (ha!) Edge LED sets really aren't any good in my opinion. However, the best local-dimming sets that use a full array backlight with hundreds of zones combined with a high contrast panel (for example LG try it with 1,000:1 IPS panels and it looks terrible) will exceed Plasma contrast easily with very little haloing. Examples would be the Sharp Elite and Sony's UV2A HX900. (I hear the newer Sonys with Samsung panels fare worse)

Even the Kuros with their 30,000:1 CR (when calibrated) weren't nearly enough for me, because they still never look black in a dark room. Local dimming achieves that with ease.

Image retention is still a very real issue with Plasma despite what people claim online. (just try only watching 2.37:1 films for a while, or getting hooked on a game for a week without watching anything else and see how "resistant" they really are...)

Plasma still doesn't handle ambient light very well, though they are getting better with it these days.

The thing that kills Plasma for me though, is the "ABL circuit" which starts to dim the screen when the image gets over a certain brightness. With most plasmas you lose over half their brightness when you go from a small white window pattern and a full white screen. It makes bright daylight scenes in films look dull/washed-out, and is a considerable problem if you play games or want to use the screen with a PC. (not that I recommend either due to image retention) It's not an issue of the image not being able to go bright enough - I only watch at reference levels (100cd/mē) but that to achieve that brightness with "brighter" images, you need to calibrate the panel to 200cd/mē or more, which is far too bright for a dark room, or try and compromise somewhere in-between where the image is too bright in some scenes but not bright enough in others etc. LCD does not have this problem at all, and while it affected CRTs, most were well within 10% of their calibrated brightness, not 50% or more.

Motion handling is moot between the two technologies these days. Now that scanning backlights have been introduced, LCDs can match or exceed Plasma motion handling due to the reduced persistence on the retina. (but it may introduce some degree of flicker if you are sensitive to that)


On the plus side, Plasmas actually have a good image when viewed off to the side, if that is a concern for you (I never sit to the side of my TV) because even the best IPS panels still look terrible when you aren't sitting directly in front of them.

They also update the panel progressively, so fast motion won't skew like it does on LCDs.

I agree with your sentiment on the life of these displays - it isn't something you should be concerned about with either LED sets or Plasma, though Plasmas did have a tendency to die early due to power supply failures. Modern Plasma is considerably more efficient now though, so that should be less of a concern.


Another thing is that Plasmas are significantly cheaper to get a well calibrated, high contrast image than LCDs. Panasonic's Plasmas are very cheap compared to LED backlit LCDs. (not edge LED)

If Plasmas image quality problems or other potential issues are not a concern for you, Panasonic's Plasmas have by far the best price/performance ratio out there today.


Quote:
Originally Posted by madshi View Post
Yeah, if I weren't a projector guy, I'd get one of those 65" Panasonic plasmas, myself, for serious movie watching.
I miss having a good projection setup.

I keep seeing the price of a good DLP projector drop and thinking I should buy one again, but I don't really have the space for one now. (there's no way I would buy another SXRD - they're far too soft, and motion handling sucks, JVC's D-ILA is far too expensive, and has similar problems)

Last edited by 6233638; 22nd October 2012 at 18:17.
6233638 is offline   Reply With Quote
Old 22nd October 2012, 18:33   #14990  |  Link
Qaq
AV heretic
 
Join Date: Nov 2009
Posts: 422
Quote:
Originally Posted by madshi View Post
Every current Panasonic plasma has a *much* better native contrast than any LCD.
Yep, but they don't support RGB 4:4:4 yet. So say goodbye to the "high quality chroma upsampling".
Qaq is offline   Reply With Quote
Old 22nd October 2012, 18:47   #14991  |  Link
madshi
Registered Developer
 
Join Date: Sep 2006
Posts: 9,140
Quote:
Originally Posted by 6233638 View Post
Actually, Plasmas do draw their image in quite a similar way to DLP. The pixels can only be turned on or off (similar to DLP) and gradation is built up over several subframes by pulsing the pixels on/off for varying lengths of time to control brightness.
Red/green/blue are driven independently and simultanously which is a key difference to how single-chip DLP works.

Quote:
Originally Posted by 6233638 View Post
Because each phosphor has different response/decay characteristics, this does essentially draw the image in sequential colour, even though all the pixels are technically being addressed at the same time.

This video illustrates how Plasma draws its image: http://www.youtube.com/watch?v=tGpdah32n3c
This video illustrates how one specific LG plasma model draws its image. If you buy a plasma from LG then that's your own fault. Different plasma manufacturers are using *totally* different panels, different phosphors and different driving schemes. It's quite possible to optimize phosphors to get nearer in response/decay characteristics. And it's also quite possible to modify the driving scheme to reduce artifacts. Now I haven't seen Pioneer Kuro and Panasonic 1200Hz captures, so I can't say for sure whether they have color separation problems or not, but I'd like to first see that before I believe it. I know for a fact that Pioneer was using a very different driving scheme than Panasonic a few years back.

Strange enough, people here are clever enough to know that one bad LCD model doesn't mean all LCD models are bad, and that there are different LCD panels with very different properties and problems. But then, 3 different people now mentioned bad LG plasmas, without really acknowledging that maybe not every plasma is the same.

Quote:
Originally Posted by 6233638 View Post
I would argue that if you value image quality, you would give Plasmas a wide berth. They measure well on native contrast, viewing angle and color accuracy (nowadays) but the way they draw their image due to the heavy use of dithering, motion artefacts, the ABL circuit, flicker, crosstalk, poor gradation capabilities (particularly above 60Hz) all result in decidedly poor image quality in my experience.
Well, I guess we will have to agree to disagree here. Whenever I go into a shop and compare (Panasonic) Plasmas and LCDs, I wonder how anybody could buy an LCD for movie playback. They just don't cut it in comparison, from my point of view.

Anyway, we've now heard enough pro and con views about LCD vs. plasma, I think. So maybe we should really give it a rest now?
madshi is offline   Reply With Quote
Old 22nd October 2012, 19:12   #14992  |  Link
Pat357
Registered User
 
Join Date: Jun 2006
Posts: 452
Quote:
Originally Posted by leeperry View Post
I've got a lot of 25fps SD files, some are DVD encodes, some aren't
Those ones that aren't DVD encodes, what are these ?
Exactly those might end up wrong coulored in MadVR when Madshi would implement EBU for *all* 25fps SD content, even with additional check in WxH.
Quote:
because encoding upscaled pixels is a terrible idea to begin with.
Uh ? Why is that ?
Does this mean we should abandon all (up)scaling from our avisynth scripts if we use these to encode !!?
Strong statements you make here, I must say...
Quote:
All I know is that mastering houses usually have their CRT/CCFL LCD screens.
Mastering houses don't use CCFL backlights anymore since a long time. Almost all switched to LED, because off the much better performance & stability over time.
CCFL has a higher black-point and a much less stable curve at low Y.
From the physical POV, controlling a CCFL at low output is much harder than walking over a cord !
The only reason that CCFL is still in the market is price €/m2 compared to LED and other.

Last edited by Pat357; 22nd October 2012 at 19:49.
Pat357 is offline   Reply With Quote
Old 22nd October 2012, 19:28   #14993  |  Link
6233638
Registered User
 
Join Date: Apr 2009
Posts: 1,019
Quote:
Originally Posted by madshi View Post
This video illustrates how one specific LG plasma model draws its image. If you buy a plasma from LG then that's your own fault. Different plasma manufacturers are using *totally* different panels, different phosphors and different driving schemes. It's quite possible to optimize phosphors to get nearer in response/decay characteristics. And it's also quite possible to modify the driving scheme to reduce artifacts. Now I haven't seen Pioneer Kuro and Panasonic 1200Hz captures, so I can't say for sure whether they have color separation problems or not, but I'd like to first see that before I believe it. I know for a fact that Pioneer was using a very different driving scheme than Panasonic a few years back.
It's true that they use different driving schemes (for example when Panasonic introduced their 3D panels, they switched to displaying the brightest subframe first, rather than the dimmest) but they fundamentally need to build up the image over several subframes, and all suffer from these kind of artefacts. Panasonic have done some interesting things with their "2500Hz" driving scheme (which refers to displaying each frame within 0.4ms, not refreshing at 2500Hz) but they still suffer from it: http://www.youtube.com/watch?v=a-03ENXkcs8

For what it's worth, Pioneer's driving scheme meant that they had very limited gradation, and had to use a lot of dithering. At anything other than 60Hz, gradation was extremely poor.

Blue phosphors are still considerably quicker, and red/green still have mis-matched responses. Here is a chart from Panasonic a couple of years ago when they switched to newly developed phospors to better meet the requirements of 3D:



Quote:
Originally Posted by madshi View Post
Strange enough, people here are clever enough to know that one bad LCD model doesn't mean all LCD models are bad, and that there are different LCD panels with very different properties and problems. But then, 3 different people now mentioned bad LG plasmas, without really acknowledging that maybe not every plasma is the same.
If you can find another slow motion video of a Plasma, it will show similar results, with the image being built up over several subframes, and the mis-matched phosphor response resulting in the image essentially being drawn in sequential colour. It just happened to be an LG model being used in that video on YouTube. (there was another video I found, but it doesn't mention what TV was used)

As someone that has owned, tested and reviewed Plasmas from a number of manufacturers, I can tell you that it affects them all.

Quote:
Originally Posted by madshi View Post
Well, I guess we will have to agree to disagree here. Whenever I go into a shop and compare (Panasonic) Plasmas and LCDs, I wonder how anybody could buy an LCD for movie playback. They just don't cut it in comparison, from my point of view.
I don't blame you, LCDs look terrible at their default store settings. (just like Plasmas look awfully dim in a store in comparison)

With a well-calibrated full array LED set, I think your opinion would change. I don't like edge LED sets at all though, the uniformity is too bad, and they have far too few zones for dimming to be effective. (16-32 at best I think?)
6233638 is offline   Reply With Quote
Old 22nd October 2012, 19:28   #14994  |  Link
egur
QuickSync Decoder author
 
Join Date: Apr 2011
Location: Atlit, Israel
Posts: 916
Quote:
Originally Posted by NicolasRobidoux View Post
Not totally sure this is useful, but here are some news, courtesy of the GIMP dev. mailing list:
Vladimir Yuzhikov has release his SmartDeblur program under the GPL at
https://github.com/Y-Vladimir/SmartDeblur

Yuzhikov worked on two common cases: blurring due to incorrect focus and blurring due to movement. "Each of you knows very well," he points out, they "are very difficult to repair." In fact, the examples used as samples by Yuzhikov are impossible to fix using current commercial technology. You can see these amazing results here:
http://gizmodo.com/5953601/incredible-csi-enhance-technology-fixes-unfocused-photos
Actually not a new idea. Usually referred to as inverse filtering. I've seen nice results 14 years ago in an image processing course.
BTW, all the samples he showed belong to the "flat world" class - focus distance is identical to all pixels. Since (out of) focus blurring can be measured and approximated, finding the inverse filter is doable. It gets really hard when there're multiple areas in the image with different distances.

The CSI style operation is usually done by applying weak deblurring kernels again and again to smooth parts of the image. The trick is when to stop. It never looks really good in end but some details can surface. So the CSI use case is valid but fixing your home photos will not work right.

I've read what you wrote on the Jinc function in your website. Basically an ellipse (2D) shaped Sinc like function (right?). But I didn't see any head to head compares against Lanczos.

Is the Jinc AR algorithm proprietary?
__________________
Eric Gur,
Processor Application Engineer for Overclocking and CPU technologies
Intel QuickSync Decoder author
Intel Corp.
egur is offline   Reply With Quote
Old 22nd October 2012, 19:38   #14995  |  Link
leeperry
Kid for Today
 
Join Date: Aug 2004
Posts: 3,477
Quote:
Originally Posted by Pat357 View Post
those might end up wrongly colored in MadVR when Madshi would implement EBU for *all* 25fps SD content, even with additional check in WxH.
I don't see any better technical choice than to use EBU for 25fps SD files. There's a far higher probability that a 25fps SD file is PAL than anything else IMO..be it a WEB-DL, a DVD encode or home made stuff.

Quote:
Originally Posted by Pat357 View Post
Uh ? Why is that ?
Does this mean we should abandon all (up)scaling from our avisynth scripts if we use these to encode !!?
Well, I was mostly talking about 1024x576 PAL DVD rips. The best way is to remux the untouched 720x576 stream and set a DAR of 16/9 or 4:3.

Avisynth? processed in 8bit? Yes, this will temper with the PQ I'm afraid and encoding upscaled pixels is a waste of bits. We can easily upscale at playback by setting the right DAR.

Last edited by leeperry; 22nd October 2012 at 19:43.
leeperry is offline   Reply With Quote
Old 22nd October 2012, 19:40   #14996  |  Link
NicolasRobidoux
Nicolas Robidoux
 
NicolasRobidoux's Avatar
 
Join Date: Mar 2011
Location: Montreal Canada
Posts: 269
Quote:
Originally Posted by egur View Post
...
I've read what you wrote on the Jinc function in your website. Basically an ellipse (2D) shaped Sinc like function (right?). But I didn't see any head to head compares against Lanczos.

Is the Jinc AR algorithm proprietary?
I have a website? You mean the parent site of http://www.imagemagick.org/Usage/filter/nicolas/?
-----
Basically:
Take the inverse Fourier transform of the characteristic function on a centered interval (a.k.a. the "ideal low pass filter") in 1D -> Sinc
Take the inverse Fourier transform of the characteristic function on a centered disc in 2D -> Jinc.
-----
All my stuff is academic research with no patents or what not. (I am an ex-prof.) This includes the million variants of EWA LanczosSharp = Elliptical Weighted Averaging Jinc-windowed Jinc, the basis of madVR's Jinc method. It builds on ideas of, among other people, Paul Heckbert, Andreas Gustafson (and, I would guess, his thesis advisor Heikki Saikkonen), Craig deForest and Anthony Thyssen.
AR is Mathias Rauen's baby, and I'm not in a position to comment on legalese.

Last edited by NicolasRobidoux; 22nd October 2012 at 20:04.
NicolasRobidoux is offline   Reply With Quote
Old 22nd October 2012, 19:55   #14997  |  Link
NicolasRobidoux
Nicolas Robidoux
 
NicolasRobidoux's Avatar
 
Join Date: Mar 2011
Location: Montreal Canada
Posts: 269
@Eric Gur: I don't have a ready made head to head comparison of plain vanilla tensor Lanczos and EWA LanczosSharp (=madVR's Jinc3 AR minus the AntiRipple, which again, is Mathias' private baby).
But you can easily do one yourself. (Warning: I'm a *n*x snob.)
Install HDRI ImageMagick http://www.imagemagick.org/script/hi...amic-range.php, for otherwise you get a version of tensor Lanczos that has clipping artifacts.
Once it's done:
convert input.png -filter Lanczos -resize 800% output.png
gives you plain tensor Lanczos 3-lobe, while
convert input.png -colorspace RGB -filter Lanczos -resize 800% -colorspace sRGB output.png
gives you gamma corrected results.
To get EWA LanczosSharp:
convert input.png -filter LanczosSharp -distort Resize 800% output.png
convert input.png -colorspace RGB -filter LanczosSharp -distort Resize 800% -colorspace sRGB output.png
If you have a reasonably standard 16-bit (not HDRI) version of ImageMagick, the above code will give you correct EWA LanczosSharp results, so if you generate the Lanczos results with some other piece of software, you can compare.
----
If you want to discuss the EWA Jinc part of this (but not AR, which again is linked to madVR), I suggest you take it to the ImageMagick Forums http://www.imagemagick.org/discourse-server/index.php.
P.S. You may want to use a non-bleeding edge version of ImageMagick: A png conversion bug appeared sometime in the last few releases (2 weeks ago? IM updates about 6 times/month). It affects, among other things, 3-channel greyscale.

Last edited by NicolasRobidoux; 23rd October 2012 at 13:50.
NicolasRobidoux is offline   Reply With Quote
Old 22nd October 2012, 20:00   #14998  |  Link
NicolasRobidoux
Nicolas Robidoux
 
NicolasRobidoux's Avatar
 
Join Date: Mar 2011
Location: Montreal Canada
Posts: 269
The disc becomes an ellipse when you warp. For aspect ratio preserving resize, it's always a perfectly round disc.
Elliptical Weighted Averaging was invented to deal with perspective transformations, and is commonly used for splatting. Using it to resize is a recent development.
To make it competitive required "a few tweaks".

Last edited by NicolasRobidoux; 22nd October 2012 at 20:03.
NicolasRobidoux is offline   Reply With Quote
Old 22nd October 2012, 20:05   #14999  |  Link
madshi
Registered Developer
 
Join Date: Sep 2006
Posts: 9,140
Quote:
Originally Posted by 6233638 View Post
Blue phosphors are still considerably quicker, and red/green still have mis-matched responses. Here is a chart from Panasonic a couple of years ago when they switched to newly developed phospors to better meet the requirements of 3D
They seem to at least have brought red and green nearer to each other. So color separation is reduced. And that was a couple of years ago. They might have improved it further in the meanwhile. What I'm saying is that with 1-chip-DLP, as far as I remember, each color is drawn totally separately, in equal time distances, with black in between. Plasma at least *tries* to draw at the same time. Maybe it doesn't work perfectly, due to different phosphor behaviour, but I don't think you can claim that color separation is anywhere near as bad as with 1-chip-DLP. Even with that LG video, although the colors appear to be shown somewhat sequentually, they are relatively near to each other in time, with longer black phases in between.

Quote:
Originally Posted by 6233638 View Post
As someone that has owned, tested and reviewed Plasmas from a number of manufacturers
That doesn't make you the one and only judge, though. Different reviewers still have different opinions. From those comparison reviews I've read from print magazines I actually trust, usually (Panasonic) Plasma is crowned winner, as long as ambient light is low. Does that mean they are 100% right and nobody in his right mind could possibly prefer LCD? Definitely not. Different people are perceptable to different kind of artifacts, and maybe even "looks".

Quote:
Originally Posted by egur View Post
I've read what you wrote on the Jinc function in your website. Basically an ellipse (2D) shaped Sinc like function (right?). But I didn't see any head to head compares against Lanczos.
If you have a couple of test images you want to see scaled with Lanczos vs. Jinc, I can do that and upload the results. Or you could also do that yourself by using the latest madVR version. The key difference for me is that Jinc has virtually no aliasing, no matter which angle the edges are. However, Lanczos is a bit sharper.

Quote:
Originally Posted by egur View Post
Is the Jinc AR algorithm proprietary?
Jinc itself is available in open source software ImageMagic (they call it "EWA LanczosSharp", IIRC). The madVR AR filter is an add-on and is my private invention. madVR is closed source and I've not fully explained the AR algorithm anywhere in all detail. I'm still constantly tweaking the algorithm, anyway, so it's not final yet. I might be willing to explain how it works in concept, but I don't plan to publish the source/shader code. I don't want to make it too easy for my renderer competitors to simply copy my work with zero effort. FWIW, I've already explained some of the basics in this thread:

http://forum.doom9.org/showthread.php?t=145358
madshi is offline   Reply With Quote
Old 22nd October 2012, 20:07   #15000  |  Link
egur
QuickSync Decoder author
 
Join Date: Apr 2011
Location: Atlit, Israel
Posts: 916
Any more details?
Do you scale using 1 pass in 2D? 2 passes in 1D? 2 passes in 2D?
Does it work well in YCbCr?
Is it any better or worse in linear light?
__________________
Eric Gur,
Processor Application Engineer for Overclocking and CPU technologies
Intel QuickSync Decoder author
Intel Corp.
egur is offline   Reply With Quote
Reply

Tags
direct compute, dithering, error diffusion, madvr, ngu, nnedi3, quality, renderer, scaling, uhd upscaling, upsampling

Thread Tools Search this Thread
Search this Thread:

Advanced Search
Display Modes

Posting Rules
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts

BB code is On
Smilies are On
[IMG] code is On
HTML code is Off

Forum Jump


All times are GMT +1. The time now is 17:06.


Powered by vBulletin® Version 3.8.11
Copyright ©2000 - 2024, vBulletin Solutions Inc.