http://hdguru.com/?p=187

The results of this test (3:2 pulldown) were disappointing with only 14 out of 75 sets (just 18.66%) properly handling the signal!

Static and Motion Resolution

A particular HDTV may resolve a stationary test signal at full bandwidth, but what happens when motion is introduced? The answer, the on-screen resolution drops. This can have a significant affect on your viewing experience, especially if you tastes tend toward sports and action.

...three distinct groups emerged from this test of the twenty 1080p displays. All displays in the top group were plasma HDTVs. They all had a static resolution of 1080 lines and a measured motion resolution of 830-880 lines, depending on the specific display. The next group consisted of microdisplay rear projectors, static measured 1050-1080 (depending on the display) while motion resolution ranged of 610-780 lines. The bottom group were all the LCD flat panels, with a static resolution of 400 (one panel) to 1080 lines and motion rez coming in at 360 lines (one set tested) to 600 lines. Three of the LCDs tested were 120 Hz models (one was the Sony KDL-46XBR4 reviewed here), all 120 Hz models had 600 lines of motion resolution.

Does not have the actual list of sets up yet :(

Looking forward to the results.
Thanks!

Yeah, me too, ready to pull the trigger but now I wanna see what's up.

Glad this is coming out now. I'm just about to purchase a new set and keep flip-flopping between plasma and lcd.

Glad this is coming out now. I'm just about to purchase a new set and keep flip-flopping between plasma and lcd.

No kidding, me too.

Anyone that wants lack of resolution with motion is crazy!

I agree artwood. Why would anybody want bad off angle viewing and motion blur, even with 120hz? And pay a lot more for it.

The following Panasonic press release pretty much confirms that there was a 600 line limit with moving images on LCD. It looks like Panasonic has improved this, although it doesn't say how much more than 600 lines. Looks like HDGURU was spot on with his results. And it looks like the 1st gen implementations of 120hz were not done correctly.

http://panasonic.co.jp/corp/news/official.data/data.dir/en070809-7/en070809-7.html

"The new 37-inch LCD TV, TH-37LZ75, features the Full-HD IPS Alpha Panel, offering consistently clear pictures even when viewed off-center. The panel displays crisp fast-moving images, with a moving-picture resolution of over 600 lines, thanks to the new Full-HD W Speed that achieves a refresh rate of 120 frames per second, twice as fast as the typical rate. The new Full-HD W Contrast AI also gives third-dimensional appearance to images with a high 7,000:1 contrast ratio"

I agree artwood. Why would anybody want bad off angle viewing and motion blur, even with 120hz? And pay a lot more for it.Off angle viewing, for most, is not the big a deal. Most tier one sets are more than adequate.

IMHO resoultion and motion blur are two different creatures. Motion blur does exist but is no way near the deal killer you would infer. "The latest LCDs are not devoid of motion blur completely as plasma is, but they're fast enough to watch even the fastest of sporting events with few desultory effects." http://www.hometheatermag.com/advic.../407plasmavlcd/ .

BTW, the plasma sets tested also suffered a loss in dynamic (motion) resolution apparently not as bad as the tested LCDs. So, if a loss of dynamic resolution and motion blur are the same, wouldn't this mean that plasmas also suffer from motion blur?

As for costs, they are all over the place. A Sharp AQUOS LC-52D62U 52" LCD costs less than a Panasonic TH-50PZ700U 50" Plasma TV http://electronics.pricegrabber.com/plasma-lcd-televisions/p/197/ Yet the LCD is larger. I'm sure you can go to other sites and find the reverse.

This article is interesting. It attempts to explain the cost difference. "When it comes to the market price of a TV set, the positions of PDP and LCD TVs are inverted with pricing for PDP TVs rising above that for LCD TVs. Average market price for a 50-inch full HD PDP TV was as high as $4,453 as of the 1Q of 2007, compared with a relevant LCD TV's $3,709, according to DisplaySearch.

As reasons for the PDP TV's inverted pricing state, DisplaySearch cited factors including the high component costs needed to assemble a PDP to a TV set. PDP TVs have no other choice but using optical filters to reduce EMI and prevent reflection, as well as expensive power supply circuits that support their 180 to 200 V high drive voltage, which further result in high costs to assemble a set, said DisplaySearch.

An optical filter used in 50-inch full HD resolution models, for example, cost $104.2 (as of the 1Q of 2007). LCD TVs do not need any optical filters. A power supply circuit cost $61.5, about three times more than the LCD TV's $21.1.

Reflecting the high voltage, components around the power supply circuit in a PDP TV were also about twice more than those in an LCD TV. When building up a TV set, a PDP TV required labor costs of $203.2 for assemblies including the installation of an optical filter, more than $30 higher than that required for an LCD TV." http://techon.nikkeibp.co.jp/english/NEWS_EN/20070925/139644/

I agree artwood. Why would anybody want bad off angle viewing and motion blur, even with 120hz? And pay a lot more for it.

ME!!! Have you seen the Samsung 81 series? Flaws and all this is TV is un friggin believable.

Off angle viewing, for most, is not the big a deal. Most tier one sets are more than adequate.

IMHO resoultion and motion blur are two different creatures. Motion blur does exist but is no way near the deal killer you would infer.

Of course not. :rolleyes:
Just like reflective LCD's and LCD burn in. ;)
If your panel starts macroblocking and/or blurring when motion gets faster, you are losing resolution.
How can you say it is unrelated?

My panel does not start macroblocking and/or bluring with motion. I'm sorry if yours does.

Never said 'unrelated' I said they are two different things.......and they are.

How much 'blur' do you notice with your pdp? They lose resolution with movement as well.

My panel does not start macroblocking and/or bluring with motion. I'm sorry if yours does.

Never said 'unrelated' I said they are two different things.......and they are.

How much 'blur' do you notice with your pdp? They lose resolution with movement as well.

I don't notice ANY on mine....but every single LCD Tv I see, there it is. ;)
Face it, HDguru hit's it on the head and LCD pushers just get offended and defensive and then pull the BIAS card.

To me off center viewing is a huge detriment to owning an LCD. As the center piece to a home theater system, good off center viewing is an absolute must. In Sound&Vision Mags latest issue they test the new Toshiba 52LX177. In the minus attribute section they say "typical limited LCD viewing angle". That kind of statement tells you that this shortcoming is expected by the pro reviewers. Expected yes, but not acceptable.

I don't notice ANY on mine....but every single LCD Tv I see, there it is. ;)
Face it, HDguru hit's it on the head and LCD pushers just get offended and defensive and then pull the BIAS card.Who pulled a bias card? I must have missed reading those posts.

Macroblocking is a signal/compression issue, not related to display type. You must not have understood this. As for blur related to resolution, if this is true, pdps must blur as well since HDGuru reported that pdps also suffered resolution loss with motion. ;)

Anyone that wants lack of resolution with motion is crazy!

Thx U......I posted this article in the Plasma section on Mon. I guess I should of posted it in here b/c my thread basically went dead. But I was trying to explain to one gentleman that the motion resolution flaw is not only unique to display devices alone. That other pieces of equipment along the chain suffer from this as well.

My panel does not start macroblocking and/or bluring with motion. I'm sorry if yours does.

Never said 'unrelated' I said they are two different things.......and they are.

How much 'blur' do you notice with your pdp? They lose resolution with movement as well.

Well I doubt alot of ppl will notice this issue b/c at the present moment there is very little 1080p live action sporting events being broadcasted. And you wont notice this flaw while watching a film b/c most films move at rate of around 24 fps.

Good point

To me off center viewing is a huge detriment to owning an LCD. As the center piece to a home theater system, good off center viewing is an absolute must. In Sound&Vision Mags latest issue they test the new Toshiba 52LX177. In the minus attribute section the say "typical limited LCD viewing angle". That kind of statement tells you that this shortcoming is expected by the pro reviewers. Expected yes, but not acceptable. As the "center piece to a home theater system" most people I know try to to sit pretty much in front of, or within a few degrees off center from, the screen. Here is a review on an LCD from HDTV Solutions on a JVC. "Also, the viewing angle on the LT-46FN97 was one of the best that we have seen on LCDs that we have reviewed. Most pictures on LCDs will wash out as you move off center. This JVC maintains its image quality even at fairly severe viewing angles. " http://www.hdtvsolutions.com/JVC_LT-46FN97_Review.htm I believe their chart indicates a 150 degree viewing angle.

Who pulled a bias card? I must have missed reading those posts.

Macroblocking is a signal/compression issue, not related to display type. You must not have understood this. As for blur related to resolution, if this is true, pdps must blur as well since HDGuru reported that pdps also suffered resolution loss with motion. ;)

Yes...just like you seem to be missing that LCD motion blur and off axis viewing. ;)
I used the term macroblocking because it is similar to what I have seen LCD's do with fast motion but I am sure there is a more accurate term.
If you read the HDGuru article, you would know what he said...and that was the best LCD's were 120hz and even those just were not in the running vs the plasma's.
I have a 720p (768p) plasma, so it's no wonder I don't see any loss. :cool:
I guess we could say my lowly plasma resolution beats your 1080p LCD in motion resolution scenes. :D

Degrees of viewing angle means little to the average viewer. I know that my main viewing room is 27x15. With a couch on one side wall and a chair on the other, placed about 10' from the several 52" LCD displays I've tried, the PQ is noticeably lessened. The same goes if I lay on the floor. I haven't tried the JVC. Even the article you mention states that "MOST"pictures on LCD's will wash out as you move off center. That's a really bad thing when you can say that about most LCD's. Why would you want to buy one for your main display? Sure, maybe you can squeeze two or three people in the sweet spot but what if you have more people than that? Why bother, just get a plasma if you want a family size flat panel.

Yes...just like you seem to be missing that LCD motion blur and off axis viewing. ;)
I used the term macroblocking because it is similar to what I have seen LCD's do with fast motion but I am sure there is a more accurate term.
If you read the HDGuru article, you would know what he said...and that was the best LCD's were 120hz and even those just were not in the running vs the plasma's.
I have a 720p (768p) plasma, so it's no wonder I don't see any loss. :cool:
I guess we could say my lowly plasma resolution beats your 1080p LCD in motion resolution scenes. :DYou could but you would be incorrect. ;)

Degrees of viewing angle means little to the average viewer. I know that my main viewing room is 27x15. With a couch on one side wall and a chair on the other, placed about 10' from the several 52" LCD displays I've tried, the PQ is noticeably lessened. The same goes if I lay on the floor. I haven't tried the JVC. Even the article you mention states that "MOST"pictures on LCD's will wash out as you move off center. That's a really bad thing when you can say that about most LCD's. Why would you want to buy one for your main display?Overall quality.

Overall quality of what? Do they use a better grade of plastic? Certainly not picture quality.

How much 'blur' do you notice with your pdp? They lose resolution with movement as well.


i see 'blur' and macroblocking with motion on my plasma and all plasmas at bb.

i also see it on my lcd with motion.


it looks more pronounced on my lcd (i assume) because the lcd is generally sharper than plasma ie: if there is something bad in the signal the sharper set will show it more.

neither is perfect. pick your poison

i see 'blur' and macroblocking with motion on my plasma and all plasmas at bb.

i also see it on my lcd with motion.


it looks more pronounced on my lcd (i assume) because the lcd is generally sharper than plasma ie: if there is something bad in the signal the sharper set will show it more.

neither is perfect. pick your poison

It sounds like you have a poor source. Try FIoS Tv. :D
LCD vs plasma noise.....they are just different.
I would hardly call LCD sharper. Have you seen the Aquos noise up close?
I did not see any claims of perfection but please, ignoring what LCD's are doing to objects in motion is just plain silly.
Off axis angles, motion blur with resolution loss....these are real problems dogging LCD even now.

Face it folks--we're still not getting 1920X1080 every single pixel in all of its glory all the time. Real HD is not perfect yet. We should all be after the least bad or the least imperfect.

That happens to be Plasma at the current time. Why do people have such a problem with accepting the truth?

Sounds like LCD needs to rev up to about 240hz to get a significant increase in motion resolution.

It sounds like you have a poor source. Try FIoS Tv. :D
LCD vs plasma noise.....they are just different.
I would hardly call LCD sharper. Have you seen the Aquos noise up close?
I did not see any claims of perfection but please, ignoring what LCD's are doing to objects in motion is just plain silly.
Off axis angles, motion blur with resolution loss....these are real problems dogging LCD even now.


absolutely. poor sources will do it every time no matter what you have.

i don't notice any blurring on the lcd with hd dvd or sd dvd.

hd ota and cable is another story and some channels are hideous.

anybody claiming perfection is a fanboy in my book.

if you or anyone can cite a sequence from a movie on hd dvd (chapter time) that you claim to see blurring on lcd but not plasma, i will check it out on my sony xbr2 (yeah yeah, old stuff but i love it) and will be 100% honest in my response.

doesn't have to be hd dvd. any sd material will do. i can check it out on both the lcd and plasma.

no fanboy here.

Is DLP at least better than LCD?

The greatest Mitsubishi is.

Is DLP at least better than LCD?

Maybe, if you're willing to sit directly in front of the screen and at the perfect height for the picture to look good.

If you move two feet in any direction away from the sweet spot, then No.

At least LCD's don't have SSE. :D

Seems like there is a vital piece(s) missing to this test. What sources are available for sports content at 1080p? There is plenty of sports programming at 1080i. Since there is no telecine scheme for this interlaced content, there is no direct process to create a progressive image w/o inherently losing vertical resolution under motion. That has always been the case, and will likely be for some time, as long as true interlaced programming exists.

So the part that is missing from the test is something to evaluate how much resolution is lost under motion from a traditional 60 Hz interlaced test pattern. Should it be any surprise to anyone here that this process alone is enough to drop native resolution by as much as 50%? In that respect, the relative performance between any of these displays (regardless of the the technology) are just not going to be the limiting factor from any sports event you will be pulling off the air, from a satellite, or by cable.

Now factor in the amount of compression effects that occur in the typical digital sports broadcast. The technical resolution of the broadcast may be "1080", but the effective resolution of the content will fall far short of that, in direct response to the amount of on-screen motion. The more motion, the more blocking...the more blocking, you won't be in "1080-land". Effective resolution can drop by a factor of 8x once the blocking occurs in a scene (essentially, the 1080 image is degrading into 8x8 blocks).

So, the resolution loss under motion of these various displays may seem harsh, but rest assured, the performance of the programming you would feed these displays is even more compromised (for the time being). The bottleneck isn't at the display.

If the author of the article wanted to really make a valid point, he would not be pointing to sports broadcasts as a scenario where his tests might become relevant. He should have pointed to hd gaming from a console or computer as a potential source to exercise these limits.

The issue of Home Theater Magazine with these tests should be hitting the stands now.

Barnes and Noble is supposed to have it already.

So the part that is missing from the test is something to evaluate how much resolution is lost under motion from a traditional 60 Hz interlaced test pattern.

I think he did this with the "deinterlace test" using the Silicon Optix HQV HD - that is if that test covers 1080i60 non-telicine signals (i.e non 24p based).

So, the resolution loss under motion of these various displays may seem harsh, but rest assured, the performance of the programming you would feed these displays is even more compromised (for the time being). The bottleneck isn't at the display.

Agee with you point about compression but the display is still important though. Whether 50% resolution loss occurs during deinterlacing depends on the TV (HQV) (http://cache.gizmodo.com/assets/resources/2007/06/SILOPT_HD_HQV_Benchm%5B6%5D.pdf) and failure to detect 3:2 can result in an equell amount of resolution loss as well as it's ability to resolve bandwidth down to the pixel level. A display could be conveying only 1/4 the resolution. As far as 1080p24, this article (http://www.hometheatermag.com/hookmeup/1106hook/) suggests all resolution is recoverable depending on the TV (though i would argue with them your point about compression artifacts).

He should have pointed to hd gaming from a console or computer as a potential source to exercise these limits.

yeah, maybe using a CRT monitor as a reference.

Home Theater Magazine's latest issue (got it today) covers most of these issues and a few more. They tested 75 displays and rated them on 3:2 pull down ability, motion blur, de-interlacing ability and a few other things. Pretty interesting read.

I think he did this with the "deinterlace test" using the Silicon Optix HQV HD - that is if that test covers 1080i60 non-telicine signals (i.e non 24p based).

I think this was just a basic test to determine if interlace was handled properly. It doesn't evaluate the quality of the resolution under various motion scenarios. Afaict, the article does not bother to specify how much resolution is lost under motion interlace from a reference display (to delineate how much resolution is lost by virtue of the deinterlace process, itself) vs. the various test displays, either. That would have been the most valid test for the article from the standpoint of how these displays would respond to real and potentially most demanding broadcast material (1080i60).

Agee with you point about compression but the display is still important though. Whether 50% resolution loss occurs during deinterlacing depends on the TV (HQV) (http://cache.gizmodo.com/assets/resources/2007/06/SILOPT_HD_HQV_Benchm%5B6%5D.pdf) and failure to detect 3:2 can result in an equell amount of resolution loss as well as it's ability to resolve bandwidth down to the pixel level.

This is true, but then we are evaluating motion performance in the 24p realm, not 60i realm...which implies a movie of some sort...which implies typical cinema motion blur will be in effect for any motion scene, anyway. Again, the display will not be the bottleneck, as the program will contain its own motion blur just by virtue of its medium and style. Even in the extremely unusual case that there is zero motion blur in the program, we are still talking about a framerate of 24p, which the modern display should have no problem keeping up with if it has any hope of doing 60i reasonably well. The viewer will be at odds with far more insidious image issues with 24p under motion, as well. There is such a spatial jump from frame to frame at 24p, it's not like the viewer will be able to concentrate on the nth degree of detail, when the shear slowness of the framerate is not able to render genuinely smooth motion in the first place. ;)

So, like I was saying, the real relevant part of this test centers around 1080i60 performance, as 1080p24 has its own set of issues that preclude it from really stressing the limits of a particular display, when it comes to motion and resolution (and even 1080i60 isn't an "ideal" source, hence my mention of high fps computer-sourced material).

The Home Theater Magazine URL link in Post #9 is BAD (abbreviated) - when clicked, you only get the home page.

here is the correct URL (for the actual page) -
http://hometheatermag.com/advicefromtheexperts/407plasmavlcd/

I see your point better now. Between compression artifacts, 24p judder and camera blur (anything else?) it's amazing even 800 lines of resolution remain during motion. Then again it's not known if these artifacts were present on the "FPD Benchmark Software" used to test motion - it might be similar to a computer source in which case it should be accurate -
A true motion test would use a computer source to test the panel itself AND something like the HQV disk to test the processing (3:2 detection, deinterlacing ect). If the FPD software was a "computer-like" source then both bases were covered in the article. This is inconclusive for now afaik. Of course none of this detracts from the fact that many motion artifacts will be inherent in the source and can't ever be removed! BTW doesn't 60i broadcast use camera blur too? or just 24p content.
The article was a bit disappointing for me as there is only one 120hz set that i see (sharp D92). He certainly doesn't say how well any of them did in the HQV interlaced motion test - only pass or fail.

I think there is some native motion blur in video cam shots, but it is much less than the deliberate effects that are put in "film" movies. I think even then it is pretty well overwhelmed by the effect of digital compression in high motion scenes in your typical cable/satellite/OTA broadcast, however. (this blurring effect occurs under motion, but isn't the same as "motion blurring", as it doesn't correlate to the vector of the motion so much as simply existing/clouding in the areas of the screen that are undergoing some kind of motion)

If the author of the article wanted to really make a valid point, he would not be pointing to sports broadcasts as a scenario where his tests might become relevant. He should have pointed to hd gaming from a console or computer as a potential source to exercise these limits.

Excellent points but even a 720p sports broadcast is going to really show LCD's coming up short. I can only imagine those LCD owners who run that 720p through a 1080i/p conversion to their 1080p panels and wonder WTH! :eek:

...but then you have given up resolution with respect to 1080i, anyway, so it's really a moot point at that scenario. You still have to put up with macroblock artifacts in heavy motion broadcasts, so aren't even getting a genuine 720p level video, anyway.

In every case where the display has a weak point, the quality and format of the broadcast is even weaker. The only thing at this time that transcends these limitations is if you intend to put computer/console gaming into the display.

...but then you have given up resolution with respect to 1080i, anyway, so it's really a moot point at that scenario. You still have to put up with macroblock artifacts in heavy motion broadcasts, so aren't even getting a genuine 720p level video, anyway.

In every case where the display has a weak point, the quality and format of the broadcast is even weaker. The only thing at this time that transcends these limitations is if you intend to put computer/console gaming into the display.

Yes..but at less than 50" panel sizes....I think 1080 is really wasted anyway (not that I think it's a bad thing to have 1080p).
I have to say that my direct coax to cablecard 720p is really void of any noticable artifacts. With FIoS as my source soon..it should even get better.
I do agree though....our weak content sources are the biggest problems by far.

other than hqv, is there any other benchmark disks to use? i have only have avia (standard dvd) but thats for calibration.

Yes..but at less than 50" panel sizes....I think 1080 is really wasted anyway (not that I think it's a bad thing to have 1080p).

Never wasted - all depends on your sitting distance from the display; benefits of 1080 resolution is easily noticed (compared to a 720p panel) if sitting between 4 to 6 feet from the display, or 6 to 9 feet from a 60"+ size 1080p plasma/LCD.

Don't discount 1080p in small sizes. Earlier this year Geoffrey Morrison of Home Theater Magazine mentioned a forthcoming product - 9" LCD HDTV with 1080p resolution, and said something like: "don't laugh - how would you like to have this unit in front of your eyes while flying in an airplane?"

1080p is always more beneficial than 720p when viewing at distances where your eyes can obviously see the difference in resolution (compared to 720p). It should, since 1080p contains TWICE the number of pixels!

other than hqv, is there any other benchmark disks to use? i have only have avia (standard dvd) but thats for calibration. I don't know of a benchmark disk but your question raised some questions in my mind. Are you looking for a benchmark disk to use at home or for shopping? I could see benefit if you could use one while shopping - if you're allowed to use it in a store. I do not see much benefit for use at home.

My queston is would you trade out a set you already owned, even one that looks good to your eyes - pleases your family, because it had lower scores than another set you tested? Would knowing your set had high numbers make your viewing experience any better?

I must be misunderstanding something. If the signal is the bottle neck (I agree) and is not sending all the res it should how does that invalidate the test that where done? It seems to me that it would only make it more important that the display did not loss even more res.

If the signal is losing x and the display is losing y then it seems that the end result is you are losing x+y. You can't control x but you can control y by getting a set that losses the least amount of res.

My queston is would you trade out a set you already owned, even one that looks good to your eyes - pleases your family, because it had lower scores than another set you tested? Would knowing your set had high numbers make your viewing experience any better?

I have :D I really hate this place sometimes but also love it too as I am getting better and better TVs each time. :D

I still have my first HDTV (use it in my office) it amazes me that I thought it looked great when I first got it.

I must be misunderstanding something. If the signal is the bottle neck (I agree) and is not sending all the res it should how does that invalidate the test that where done?

It doesn't invalidate the tests done in the article, but it does make them less relevant (perhaps, even irrelevant) to what you will see in real use of the displays using real material.

If the signal is losing x and the display is losing y then it seems that the end result is you are losing x+y. You can't control x but you can control y by getting a set that losses the least amount of res.

In the case of real usage scenarios, the "x" is much larger than the "y", hence a more dominant effect. What you will be seeing is the effect of "x", by far, so the contribution of "y" becomes insignificant. Even if you were able to eliminate "y", you won't be able to see any more resolution of the original program, because that resolution has already been lost by "x".

Once "x" has occurred to the signal chain, you cannot go backwards in quality, regardless of how much you can eliminate "y". The best you could do is eliminate "y" to the point where you simply see the problems that "x" created more clearly/accurately. Since "x" has already corrupted the original signal, I'm not sure that really is a benefit, outside of if you are in a diagnostic/R&D stage of the product.

I don't know of a benchmark disk but your question raised some questions in my mind. Are you looking for a benchmark disk to use at home or for shopping? I could see benefit if you could use one while shopping - if you're allowed to use it in a store. I do not see much benefit for use at home.

My queston is would you trade out a set you already owned, even one that looks good to your eyes - pleases your family, because it had lower scores than another set you tested? Would knowing your set had high numbers make your viewing experience any better?

just for home. ive already bought my displays, im just curious to see how they perform. i can give a rats ass what anyone thinks. i just want to see how good they are at certain areas

In the case of real usage scenarios, the "x" is much larger than the "y", hence a more dominant effect.

Agreed

Even if you were able to eliminate "y", you won't be able to see any more resolution of the original program, because that resolution has already been lost by "x".

If y is contributing to the loss (even if its not the dominant factor) why would you not get more back if you eliminate the y loss?

Once "x" has occurred to the signal chain, you cannot go backwards in quality, regardless of how much you can eliminate "y". The best you could do is eliminate "y" to the point where you simply see the problems that "x" created more clearly/accurately. Since "x" has already corrupted the original signal, I'm not sure that really is a benefit, outside of if you are in a diagnostic/R&D stage of the product.

This I do not understand unless you are saying that y is only eliminating problems in the signal but I doubt that is the case. I am sure y eliminates some of the problems x created but y is also eliminating some clear information as well.

If there is already loss why would you want more?

"x" is the dominant effect and occurs prior to "y". So it matters little whether "y" is mitigated or not, since the real detail information has already been lost at the stage "x" occurs.

The only way to circumvent "x", for the time being, is by feeding a high fps, progressive source that has no data compression, such as a PC/console game.

Put it another way, both "x" and "y" are taking losses from the same resolution band (the highest). Once that resolution has been lost by "x", then "y" will simply lose the same resolution (but that resolution has already been lost, so it doesn't much matter if "y" happens or not). It doesn't necessarily lose "more" resolution on top of whatever "x" does.

So you are saying that x and y are losing the same data in this case resolution in movement so it does not matter.

If that is what you are saying and that only a uncompressed, high fps, progressive source would not have the x loss then why when using a test disk are the results different of different TVs should they not be all the same? As the disk is not a uncompressed, high fps, progressive source.

Are you talking about the bluray test disc? If they are using a player worth its beans (let alone test purposes), that should indeed be a progressive test signal. Ideally, the moving content has been prepared for utmost detail performance using the least amount of compression possible. It is not clear from the article if the moving content is 24p or 60p (or possibly even 60i, for that matter), but since 60p isn't even supported (afaik), I would imagine this was a 24p test- essentially what the "perfect" film transfer would deliver that had zero motion blur (a very atypical scenario).

If that is what you are saying and that only a uncompressed, high fps, progressive source would not have the x loss...

Ok, I guess this is a good time to be even more specific in my wording. It's not entirely impossible that a "compressed" source could provide a good test. It really relies on the nature of the compression (...and color space, framerate, yadda-yadda-yadda...). The best way to test would still be to sidestep those variables altogether using an "uncompressed, high fps, progressive source" such as direct CGI material. That's really the point I was trying to make.

Mind you- that would be the most relevant kind of test for the scenario pc/console gaming as the source. If the evaluation is to target the losses you would see only for movies/broadcast, then it is irrelevant (as we have discussed further up).

Once that resolution has been lost by "x", then "y" will simply lose the same resolution


I would just point out if we are talking about interlaced content (where half the resolution is lost at the get go), motion-adaptive deinterlacing is designed to replace what is lost (i.e. theoretically "y" isn't limited by "x")

Thank you Mr. Hanky I now understand you point I may not completely agree with it but thanks for taking the time to explain your point to me.

Borf

Interlaced does NOT mean half the res is lost it means that each frame is broken into 2 fields (oversimplified but I am in a hurry) in the case of film based it can be completely restored in the case of video it can come close to getting a full frame.

I would just point out if we are talking about interlaced content (where half the resolution is lost at the get go), motion-adaptive deinterlacing is designed to replace what is lost (i.e. theoretically "y" isn't limited by "x")

It would be yet another variable, that the tests in the article do not cover well. It is one thing that motion-adaptive deinterlacing exists, and it is another to correlate that any one display will have it, let alone every display having the same "flavor" of it, so that the result is as "correct" as is technically possible. Ultimately, the article just doesn't cover deinterlace performance to the level that it should if it wants to correlate the performance you see to the the type of broadcast material you will have access to. It just does a pass/fail as if it is being done "correctly" (which could simply mean it weaves when it is suppose to weave and bobs when it is suppose to bob, with no qualification to resolution lost/preserved in either state). Obviously, the quality of the "pass" will determine how much deinterlace resolution you can eek out. Maintaining full resolution under static interlace conditions is easy enough. Retaining some fraction of that full resolution under moving interlace conditions is the tricky part (theoretically it isn't possible to achieve a full frame because each field is inherently a different place in time under motion...we only achieve something that "looks" about right via various blending schemes or interpolation...it isn't the full resolution of that point in time, because such a frame never existed, in the first place)

Additionally, the compression artifacts issue has still to be accounted if the interlaced broadcast comes over a digital service. So you could have the "perfect" deinterlace going on, but you still have the compression artifacts under motion issue to deal with. So "y" is still ultimately limited by "x", because "x" comprises a number issues which include interlace/deinterlace performance.

Does anyone know where this loss of motion lines of resolution happens?

Given my understanding of the video signal path at HD resolution, from a media player to a tv:

VOB ---> MPEG stream ---> decoded RGB values in framebuffer ---> HDMI cable ---> tv framebuffer ---> DSP effects (adjusting colorspace, brightness etc) ---> panel buffer ---> panel driver ---> pixels

If it's 800 lines of motion that would suggest that the panel driver itself is only capable of instantaneously processing and releasing updates on (800 lines * 1920 pixels/line) pixels.

In which case it would seem that it's the panel driver chosen, not any intrinsic technology difference between LCD displays and PDPs that determine motion handling.

Am I wrong? This is sort of a half-educated guess. Can someone explain this properly?

Interlaced does NOT mean half the res is lost

From my understanding of 1080i60 (non-24p) based content the fields are temporaly different and can't be merged, meaning 1/2 resolution is lost. With 24p based content every field pair represent a single point in time and resolution can be restored by merging fields.

Never wasted - all depends on your sitting distance from the display; benefits of 1080 resolution is easily noticed (compared to a 720p panel) if sitting between 4 to 6 feet from the display, or 6 to 9 feet from a 60"+ size 1080p plasma/LCD.

Don't discount 1080p in small sizes. Earlier this year Geoffrey Morrison of Home Theater Magazine mentioned a forthcoming product - 9" LCD HDTV with 1080p resolution, and said something like: "don't laugh - how would you like to have this unit in from your eyes while flying in an airplane?"

1080p is always more beneficial than 720p when viewing at distances where your eyes can obviously see the difference in resolution (compared to 720p). It should, since 1080p contains TWICE the number of pixels!

Well of course. That's no arguement. Buy 1080p sit closer. The bigger the screen the more obvious the advantage of 1080p is.

From my understanding of 1080i60 (non-24p) based content the fields are temporaly different and can't be merged, meaning 1/2 resolution is lost. With 24p based content every field pair represent a single point in time and resolution can be restored by merging fields.

You loss some but not 1/2. There are a few was to do it some work better then others but the only one that you would loss 1/2 would be BOB deinterlaceing where one field is resized to the native res of the screen.

Try this

http://en.wikipedia.org/wiki/Deinterlacing

Its a little basic but is a good starting point

It's a crapshoot, at that point. The advanced/adaptive methods of deinterlacing can achieve something very close to what an original progressive frame would have been, but it is all a guess (albeit, an intelligent guess). If the type of motion cannot be identified by the particular algorithms implemented, then it regresses to classic deinterlacing techniques which are limited to about 1/2 resolution in motion scenarios.

Ironically, it just adds another variable to the whole "x" and "y" situation. Now we'll have various levels of classic deinterlacing, motion blur, digital compression, display smearing, AND potentially deinterlacing ghosting added to the mix. If you are able to eliminate any one of these factors, you still have all the other factors to mar the end result. Mitigating all of the factors might be possible, but only under a very concerted effort and rethinking some very fundamental concepts of video transmission.

You loss some but not 1/2. There are a few was to do it some work better then others but the only one that you would loss 1/2 would be BOB deinterlaceing where one field is resized to the native res of the screen.

Try this

http://en.wikipedia.org/wiki/Deinterlacing

Its a little basic but is a good starting point

Why thank you tower101 for your input as i said this exact same thing a couple posts ago. Make up your mind on the 1/2 thing and let's not go in circles please.

motion-adaptive deinterlacing is designed to replace what is lost

Plowed through some posts here, and commented on Merson's column generalizing about his test results earlier in another thread. Haven't seen the published article yet. Maybe the published magazine article clears up what's meant by motion resolution.

In the online pre-publication article it indicates the monoscope pattern moves left to right for motion tests. That side-to-side motion, it seems, would give a rough indication of horizontal resolution loss (using the converging vertical lines). This motion loss was measured in the mid-90s, presumably using interlaced or progressive CRT displays, and the results for 1080i/720p appear in this simplified table (http://archive.avsforum.com/avs-vb/showthread.php?postid=326565#post326565). Table 2.3 (http://www.atsc.org/news_information/papers/1995_acats/tsreport.pdf), the source for the simplified table data, compares what a panel of video experts measured with what they'd anticipated (table 2.2 and 2.1 Resolution introduction paragraph). Since they anticipated ~20% horizontal resolution loss from standard filtering, the ATSC-approval tests apparently involved sampling resolution patterns (as virtually all HD programming also samples).

In contrast, non-sampled computer or pattern-generator test patterns needn't undergo the same type of filtering (Nyquist) before viewing. So a HD display, ideally, might resolve 1920 alternating B&W vertical lines and 1080 alternating horizontal lines from non-sampled test-pattern sources, yet TV cameras or telecines sampling images at ~74 MHz would show less equivalent motion- or static-video details. The ATSC test committee used a rotating test pattern for motion testing (see Note 3, table 2.3). Such rotating patterns are available with commercial software (higher-end products).

The online prepublication testing report (http://hdguru.com/?p=187), indicating steadily declining motion "resolution" from plasmas, to microdisplays, to LCDs, seems to imply everything refers to vertical resolution (since horizontal resolution, which would involve the vertical converging lines, isn't mentioned AFAIK).

Typically, resolution as a measured parameter refers to horizontal resolution, and is given as TV lines or lines per picture height in professional use and in lines per picture width in consumer use and within AVS. Greg Rogers' resolution tables (http://www.cybertheater.com/Tech_Reports/HD_Projectors/hd_projector.html) (about 1/2 down), still applicable to fixed-pixel displays, provide both. (All of Merson's motion resolution test results could logically represent consumer-type horizontal resolution, which for 16X9 HD displays only involves multiplying by 1.78 to convert lines/PH to lines/PW. But the text only seems to discuss vertical resolution.)

The "resolution" numbers in the prepublication summary imply gradually falling vertical resolution from one type of display to another. It's unclear whether variances in deinterlacing effectiveness or display type caused the significant differences. LCD switching times, it seems, might account for some of the widely varying resolutions, and presumably any type of display model still using 540p bobbing with motion could more than halve vertical resolution.

For sampled images, the Kell factor (~0.7 X line count) reduces vertical resolution (see text near Rogers' table linked above). And for interlaced 1080i additional vertical filtering--'smearing' lines together--is used to minimize the 'twittering' of fine details. The Kell factor applies whether images are sampled progressively or interlaced. The sublinked authoritative article outlining 720p effective resolution (http://archive2.avsforum.com/avs-vb/showthread.php?p=5667245&&#post5667245) mentions in the final paragraphs why Kell doesn't apply to computer-based images. The reduced resolutions for 720p outlined in that article, and the ATSC approval test results, however, don't factor in as much when 720p is derived from 1080/24p (master tapes for dramas) downconversions, which can boost 720p effective resolution (resolvable detail) much closer to the format resolution. However, 24p-captured dramas still appear limited to 800--1100 lines effective resolution (equivalent detail) (see quotes/sublinks (http://www.avsforum.com/avs-vb/showthread.php?p=9314235&&#post9314235))--unless newer 4k telecines or 4k downconversions used for HD discs boost those results after similar spectrum analysis. -- John

Why thank you tower101 for your input as i said this exact same thing a couple posts ago. Make up your mind on the 1/2 thing and let's not go in circles please.

You know what you right what was I thinking :confused: Ofcourse they just toss out 1/2 the res and here I thought they where using bending and weaving and other methods at the pixel level to retain as much as info as possible.

You know you should let the networks know about this as they could save a lot of money and just send out the 540 and let the TV's make up the rest maybe you could get a percentage.

Wow thanks John very informative I have a lot of reading to do. :)

So a HD display, ideally, might resolve 1920 alternating B&W vertical lines and 1080 alternating horizontal lines from non-sampled test-pattern sources, yet TV cameras or telecines sampling images at ~74 MHz would show less equivalent motion- or static-video details.
-- John

Using vertical lines and horizontal lines is a very poor way of describing resolution. I am not sure why some do this.
Horizontal resolution should not be referring to anything vertical and vice versa, in my view. ;)

Using vertical lines and horizontal lines is a very poor way of describing resolution. I am not sure why some do this.
Horizontal resolution should not be referring to anything vertical and vice versa, in my view. ;)
Agree it can create confusion. But as the monoscope pattern illustrating Merson's online article shows, converging wedges formed by lines are used for measurements. Zoom in on the lines enough and you'll see individual picture elements (pixels), which in turn can differ between sampled and computer-generated images. You need vertical lines for horizontal resolution because a horizontal line would just appear as a line--and zooming in on alternating B&W pixels wouldn't be easy to see. -- John

Typically, resolution as a measured parameter refers to horizontal resolution, and is given as TV lines or lines per picture height in professional use and in lines per picture width in consumer use and within AVS. Greg Rogers' resolution

John, to accurately portray resolution, you need both horizontal and vertical resolution. But I do agree there is a tendency for many publications and articles to express it simply as '1080i' or '720p' without regard for the horizontal numbers. We need both.

Agree it can create confusion. But as the monoscope pattern illustrating Merson's online article shows, converging wedges formed by lines are used for measurements. Zoom in on the lines enough and you'll see individual picture elements (pixels), which in turn can differ between sampled and computer-generated images. You need vertical lines for horizontal resolution because a horizontal line would just appear as a line--and zooming in on alternating B&W pixels wouldn't be easy to see. -- John

Thanks for the info.
Interesting post.
I noticed in this table, you used 720 and 1080 vertical lines. Is that a typo?

http://archive.avsforum.com/avs-vb/showthread.php?postid=326565#post326565

I noticed in this table, you used 720 and 1080 vertical lines. Is that a typo?

http://archive.avsforum.com/avs-vb/showthread.php?postid=326565#post326565
Been quite a while since I posted that table, but recall I went with the terms the video experts used in table 2.1 in that next link: (http://www.atsc.org/news_information/papers/1995_acats/tsreport.pdf) (scanning formats as opposed to measurements, where they use cycles per active picture height). That's the various H/DTV video formats. 720 and 1080 lines there refers to stacked (vertically) horizontal lines, and listing the horizontal pixels/line just beneath defines the two HD scanning formats. But yes, as mentioned earlier above, such terminology sometimes can be confusing. -- John

You know what you right what was I thinking :confused: Ofcourse they just toss out 1/2 the res and here I thought they where using bending and weaving and other methods at the pixel level to retain as much as info as possible.

You know you should let the networks know about this as they could save a lot of money and just send out the 540 and let the TV's make up the rest maybe you could get a percentage.

"They" - the broadcast company DO send out 540 lines per field. PS what this and i have been saying refers generally to 30fps content not 24p which has temporaly identical fields. The TV can recover the resolution with 24p like i said before and a bit less with 30fps.
if the video contains moving objects, resolution is necessarily
lost; it was lost at the time of the recording.
A good video processor needs to distinguish between objects in motion or objects that
are not in motion. Doing so ensures that all of the resolution is preserved. If a video
processor assumes that a non-moving object is, in fact, moving, as much as half of the
useful resolution is being discarded.

If a given display is properly processing all the
lines, the pattern’s alternating horizontal lines should remain intact. If the display is only
displaying single 540 line fields, (all odd or all even numbered lines) and upconverting to
the TV’s native resolution, the boxes will strobe all black and then all white. If this occurs
at any time, the display is listed as “fail”, because the display will only be working with
half of the available information as it strobes.

Do CRT (all) & LCD projectors have the same loss of resolution?
Thanks.

And if the TV does not fail you do not loss 1/2, like you said it does. You do know that not all the TVs failed right.

Originally Posted by borf
I would just point out if we are talking about interlaced content (where half the resolution is lost at the get go)

Originally Posted by borf
From my understanding of 1080i60 (non-24p) based content the fields are temporaly different and can't be merged, meaning 1/2 resolution is lost.

If you want to believe that you are only seeing 540 when 1080i(v) is being sent to your TV then great, just like you believe that 5:5 would make a TV unwatchable. :rolleyes:

Interesting post :-D

"If you want to believe that you are only seeing 540 when 1080i(v) is being sent to your TV then great..."

When things are nearly still, then yes, the 2 fields can be combined to achieve a full 1080p's worth of information w. little to no ill effect. Under moderate to high motion, the fields cannot be combined directly because each field will show the subject in a slightly different place in time. The fields will be mismatched. So the classic approach dictates that only half the resolution of a full frame will be available at any 1 point in time. The only thing you can do at that point is to upscale a single field (with some edge blending, perhaps, to soften the jaggies) or you could try to synthesize the missing resolution based on preceding and/or adjacent fields. Either way, you still only had 1 legitimate field of information for that point in time. Anything beyond that is an interpretation/intelligent guess of what should be there. The end result may look right, it may not, but it is an entirely different entity than if you were able to capture a full frame of information for a single point in time, in the first place.

Logic would suggest that the more complex the detail of the subject is, the less likely you will be able to synthesize an accurate mating field while under motion. It is that very scenario where you would likely encounter detail utilizing the highest resolution of the implementation (your tv and the broadcast). So the end result could, at best, maintain the performance of 1/2 resolution of a full frame or somewhere above that, but fall decidedly short of mimicking the complex/entropic detail at the highest resolution of the genuine full frame of information that would have existed at that point in time (...and maintaining genuine detail occurring at the highest resolution is the whole goal, right?...the workaround is a good substitute, but it is still no replacement for the real thing).

I just received my latest issue of Home Theater Magazine, and find a contradiction with Gary's test results, vs reviews I've read previously of selected models.

On the magazine's front page is: 74 HDTVs Tested - What Sets PASS & What Sets FAIL?

So, you read this and would think the test results are absolutely "clear cut", without any particular bias affecting the final results - but is this really so?

Example #1...

Gary gives every Panasonic plasma a "P" [Pass] under the "Deinterlacing" column. But, both cnet and H.T.Mag's recent Panasonic plasma reviews of both 720p and 1080p panels have noted that while they usually pass the Video-30fps HD signal deinterlacing test, they FAIL the same for Film-24fps 1080i deinterlacing [as well as also failing to pick up the 3:2 sequence].

This can be confirmed by viewing the reviews , especially the well-known Cnet "Geek Box" as the bottom of each review, where a red POOR is shown for displays that fail the 1080 film-signal deinterlacing test.

So, is this fair to give the Panasonic's a "pass", since they pass one 1080i test, but fail another 1080i test? Remember, the film-deinterlacing test is particularly important because of 1080i signals received from HD-DVD, Blu-ray, and even upconverted 1080i standard DVDs [artificial 1080i]. I don't know what type of processing chip Panasonic uses, but they never promote it's name or processing ability, or if it's one of the top-tier chips (obviously not in my opinion).

Example #2...

Gary states near the end of the article... [I]"JVC's rear projectors and some Sharp LCDs would produce Pass or Fail (as noted), depending on whether the test signal was delivered via HDMI or component video. In these cases, I counted it as a failure in the final tally."

This is even more of a surprise! Gary indeed "fails" the premium Sharp 52" 92U panel, even though Geoffrey Morrison of this same magazine stated in his summer review of the Sharp that it does PASS 1080i deinterlacing [thru the HDMI input, but not component].

Gary has a *disclaimer at the bottom of the results page... "Set must be in user-activated Advanced menu film mode to pass; fails in the out-of-box factory-default setting." Well OK, but I would expect most HDTV owners, from day one, to turn "on" the Film-mode option , which means the Sharp [like all of their premium models] will pass the 1080i deinterlacing test, via the HDMI inputs.

In mentioning this, it kinda reminds of Consumer Reports testing of HDTV's, with one of their test results being how well the displays look with DVD images. In the small print C.R. states their reviews are done with DVDs viewed [I]only thru the component video inputs, and NOT the superior HDMI inputs!

I mean, why not the better emphasis on viewing DVDs etc (or 1080i deinterlacing tests for that matter) USING THE HDMI INPUT as opposed to any other input.

So, in my opinion, the tests that Gary did, while overall helpful, are unfortunately compromised (and not fully accurate for if a display justifiably earns a "Pass" or "Fail"), because of his particular criteria for the settings, which determine the results (as outlined above)... although these are not the complete results, as shown in my examples above.

All I can say is WOW. I know of no TV that does half sizing (tossing out 1/2 the res) the worst do a BOB and that does not toss anything out but scales each field to a full frame you don't loss any thing, though its not the best way to do it.

Here is an older link but does a good job of explaining deinterlacing.

http://www.100fps.com/

I know of no deinterlacer that uses interpolation (making up new info based on existing info). Since you seem to know of them perhaps you could post a link to help me understand. As even my Lumagen does not do that but I would love to get one that does. It does do a Per-pixel video deinterlacing and is very good at detecting 3:2 but I am always interested in upgrading.

My deinterlacer takes a look at each pixel deciedes what pixels are in motion and then applies the best method to use and will use many different ones to keep as much as the original image as possible. Do you end up with 1080 no of-course not but you don't loss 1/2 the res either.

If you could provide a link that shows that any deinterlacer on the market today that uses half-sizing I would love to see it as that is the only form of deinterlacing that tosses out 1/2 the res. Also a link to one that shows one that does interpolation that would be great too as I may upgrade.

Thanks

RUSTY PELICAN

I don't know about the rest but

Example #1...

Gary gives every Panasonic plasma a "P" [Pass] under the "Deinterlacing" column. But, both cnet and H.T.Mag's recent Panasonic plasma reviews of both 720p and 1080p panels have noted that while they usually pass the Video-30fps HD signal deinterlacing test, they FAIL the same for Film-24fps 1080i deinterlacing [as well as also failing to pick up the 3:2 sequence].

As I read my copy it does say P for deinterlace BUT it says F for THREE-TWO, that is the same as the c-net review. The deinterlace test is for video and the THREE-TWO is for film. I agree its weird the way they wrote it out but that is what it means.

Scaling each field to full frame is essentially building the frame from "half" the resolution. It doesn't throw anything out. It simply makes do with 1 field's worth of information, which corresponds to 1/2 of a progressive frame (if that frame actually existed).

I don't have any links to various interpolations in use, but I don't doubt it is possible to some degree or other. I simply offer that scenario to cover that "one person" who will always say such and such is possible if they just do xyz. In a sense, that is much the same concept as what you describe your deinterlacer does, except it uses adjacent and/or preceding fields to determine motion which drives the algorithm that bobs and weaves the image on an area-by-area basis. So you achieve an image that is full resolution in the areas not in motion and an effective half resolution in the areas that are in motion.

It really doesn't change any of the points I have made regarding loss in resolution due to motion in the deinterlacing stage superceding loss in resolution in the display due to image smear of a moving subject. The display will still be capable of full resolution in the static areas of the screen, and the moving areas of the image will probably smear, but it will be smearing a region that was only at an effective half resolution due to motion deinterlacing, anyway. As for if smearing + half-resolution deinterlacing has an additive effect that affects even lower resolution components, well the tests in the article didn't really bother to address that scenario, now did it? ;) (maybe it should have, if it this case is of such importance?)

I don't have any links to various interpolations in use, but I don't doubt it is possible to some degree or other. I simply offer that scenario to cover that "one person" who will always say such and such is possible if they just do xyz.



LOL Ya you have to be careful around here there is always someone that has a friend who's cousin had one.

I understand what you are saying and I agree just makes me upset when someone comes out with you can't deinterlace video and you are only getting 1/2 of the res because that is not the case. You may not/will not get a full frame but you don't loss 1/2.

Some one reads that here takes it at face value then they tell 2 friends and they tell 2 friends and so on LOL.

Although the article in question does say that 35.14% did fail the video test :(

To be fair to Home Theater, the article is a good source of info for most people as the average reader/TV viewer may not be as anul as some of us here. Although we would have like a more in-depth review with better explanations.

I still think the motion loss is a good test not for broadcast but for thouse of us with HD-DVD, BD and HD gaming systems. I can tell you that NINJA garden does loss ress in motion on my PDP and on a friends CTR it does not, although no where as bad as my LCD does. So in that sense it validates the test results.

Thanks for the updates and further comments^^^

I posted the following recently on another thread, about Sony's "interpolation" technology included on their XBR4/5 line...

According to a new XBR4 review (http://www.avsforum.com/avs-vb/showthread.php?p=11862002#post11862002) <link by Home Theater Magazine, the Sony uses interpolation by creating new frames from the content, and not "inserted" black frames [as featured in the 1-year old Sharp D62U series panels, although they didn't advertise it].

Thanks for the updates and further comments^^^

I posted the following recently on another thread, about Sony's "interpolation" technology included on their XBR4/5 line...

According to a new XBR4 review (http://www.avsforum.com/avs-vb/showthread.php?p=11862002#post11862002) <link by Home Theater Magazine, the Sony uses interpolation by creating new frames from the content, and not "inserted" black frames [as featured in the 1-year old Sharp D62U series panels, although they didn't advertise it].

Yup some of the new 120 LCDs do use Interpolation for frame rate but like I said I don't know of any that use it for deinterlacing. I am sure they will in the future as things are always moving forward. When that happens I will have to upgrade again :D

I too read the latest HDGuru tests in the Nov. Home Theater Mag and noticed a discrepancy between what CNet tests with the HQV test disc and what Merson reports with the Sony S3000 series 768p LCDs. I bought one and did my own testing.

CNet says the set scores a "Poor" on film resolution of 1080i/60 sources, while Merson scores it as a "Pass" on 3:2. It should be clarified that these are two different tests. The film resolution test demonstrates high detail in pans of 3:2 material, i.e. pans over an empty stadium. The Sony sets definitely fail, showing minor moire in the upper level seating. However, the Sonys can lock onto 1080i 3:2 pulldown pattern without flicker. This test pattern is on the HQV test disc (the one right before the stadium test).

I am still trying to find a real world example of where 'high' film detail failure would be noticible.

I too read the latest HDGuru tests in the Nov. Home Theater Mag and noticed a discrepancy between what CNet tests with the HQV test disc and what Merson reports with the Sony S3000 series 768p LCDs. I bought one and did my own testing.

CNet says the set scores a "Poor" on film resolution of 1080i/60 sources, while Merson scores it as a "Pass" on 3:2. It should be clarified that these are two different tests. The film resolution test demonstrates high detail in pans of 3:2 material, i.e. pans over an empty stadium. The Sony sets definitely fail, showing minor moire in the upper level seating. However, the Sonys can lock onto 1080i 3:2 pulldown pattern without flicker. This test pattern is on the HQV test disc (the one right before the stadium test).

I am still trying to find a real world example of where 'high' film detail failure would be noticible.

I've read previously [Home Theater Magazine I think] that they use the HD-DVD or Blu-ray edition of MISSION: IMPOSSIBLE 3, chapter 8 or 9, during a scene in the Vatican involving stair steps - stating it takes a very good processor to properly deinterlace/scale this sequence without noticeable artifacts.

Also, you could do some tests using movies with dark (low light) scenes, to see how well the display's processing works. UNDERWORLD EVOLUTION on Blu-ray would be good, also the few very dark scenes in the HD-DVD of WOLF CREEK, would be good examples (among others).

I've read previously [Home Theater Magazine I think] that they use the HD-DVD or Blu-ray edition of MISSION: IMPOSSIBLE 3, chapter 8 or 9, during a scene in the Vatican involving stair steps - stating it takes a very good processor to properly deinterlace/scale this sequence without noticeable artifacts.

Ah yes, I recall that scene on HD DVD. Useful, but still rare IMO.

Also, you could do some tests using movies with dark (low light) scenes, to see how well the display's processing works. UNDERWORLD EVOLUTION on Blu-ray would be good, also the few very dark scenes in the HD-DVD of WOLF CREEK, would be good examples (among others).

I did a fair amount of watching low light scenes (Chronicles of Riddick HD DVD). I refuse to watch the Wolf Creek HD DVD again ;) What can I say? LCD is not the same as CRT. Everything below 20IRE just kinda blends together. There's not much posterization, but detail is flat. I did a careful brightness calibration with HD DVE. Adjusted my set to where the 2% bar just dissapears.

One thing I did have a question on when using HD DVE was the pattern for 8-bit testing. There's a b&w pattern, as well as one with color. The test instructions say you should make out 22 steps, any more and you are running up against the limits of the processor. I don't see any steps, just a smooth gray to white transition. Is this good or bad?

AIUI, sets with constant 540p bobbing (see Merson's three earlier articles (http://archive2.avsforum.com/avs-vb/showthread.php?p=9178202#post9178202)) use each 1/60-sec 540-line TV field from 1080i to create full frames. Newer test DVDs or a standard SMPTE test pattern detect this with flashing white and black screens as each TV field is displayed.

The two 540-line interlaced half-frames are supposed to be shown 1/60-sec apart on interlaced CRT displays, or digitally deinterlaced into full frames for fixed-pixel displays. With my 1080i CRT RPTV, my eyes merge the two 1/60-sec TV fields into 1/30-sec frames. With most fixed-pixel displays a buffer memory helps create 1/60-sec frames, preferably combining both fields/frame rather than creating frames from only 540-line half-frames.

The mid-90's ATSC-approval test was cited earlier above (http://www.avsforum.com/avs-vb/showpost.php?p=11853671&postcount=64). Test results (see simplified table (http://archive.avsforum.com/avs-vb/showthread.php?postid=326565#post326565)) showed that a static 1080i test pattern had 800 lines of effective vertical resolution (resolvable lines), while a dynamic (5-rpm) pattern only 400 lines. Vertical resolution is the number of resolvable B&W horizontal lines from screen top to bottom. This drop in vertical resolution with motion varies directly with the rate of motion, so movement less than ~5 rpm from a rotating test pattern shouldn't cause as much loss of vertical resolution. Also, the vertical resolution loss from movement should only take place where movement occurs. 540p bobbing, however, could diminish vertical resolution over the whole screen.

In the "2.1 Resolution" introduction of the ATSC report (linked above), the video experts wrote: In the case of an interlaced scanning system, for vertical resolution under motion, the picture will take on the characteristics of a progressively scanned system with half the number of scanning lines (i.e., vertical resolution of 1080I becomes 540P under motion).
Again, that's where motion is taking place within the image, and the resolution drop varies with the rate of motion.

Also, as outlined originally above (yesterday), it's possible to use a non-sampled computer or pattern-generator test images and measure the full 1920X1080 resolution of many displays. Test patterns or actual programming sampled at ~74 MHz are filtered to reconstruct images from the samples, resulting in both vertical and horizontal resolution loss. Came across an interesting diagram, Fig. A3, (http://www.ebu.ch/en/technical/trev/trev_308-hdtv.pdf), from 2006 European HD studies, that compares the effective resolutions (resolvable details) of both 1080i and 720p after the Kell factor (~0.7 X line count) and anti-twitter filtering is factored in. Obviously such diagrams involving maximum resolutions expressed as lines correspond to equivalent details in scenes, and maximum resolvable details can be far less--if a film or video camera is heavily filtered, for example, or from many other limitations along the delivery chain. -- John

I know the results from the tests probably can't be posted.

But can anyone post a list of the sets that were tested?

Overall, do the test results mirror opinion on this board, or were their some "sleeper" TVs that did very well in the tests.

Cheers,

Detox

I am still trying to find a real world example of where 'high' film detail failure would be noticible.

This is really the problem.
What is really affecting real world viewing?
Does everything get converted to video over broadcast already?
Does your HD player do all the heavy lifting anyway?

This is really the problem.
What is really affecting real world viewing?
Does everything get converted to video over broadcast already?
Does your HD player do all the heavy lifting anyway?

I'm not sure I understand your (rhetorical?) question about everything getting converted to video over broadcast. Of course the short answer is "no", and we can expand on that,because there is plenty of 1080i/60 film based content available from NBC, CBS, HBO, etc.

As far as my DVD player doing the heavy lifting, the answer there is also no. I was able to do the same HQV tests last night but with my Toshiba HD A1 outputing 720p. The player did terrible! Failed the 1080i motion-adaptive test, as well as the 3:2 pulldown for film sources. But I suppose it is a moot point, as no film-based material would ever be encoded on the disc at 1080i/60. The one test that the player passed at 720p that my Sony 32SL130 failed @1080i input was "Film Mode High Detail" which shows up as moire in the stadium stands.

I guess I need to reconsider my assesment after thinking that through... As long as the HD DVD is 1080p/24, then outputing to 720p appears to preserve all the fine detail when viewing on a 768p set. For the 1st gen Toshiba models, my recommendation for viewing on a 720p set would be: use 720p output for 1080p film based content, and 1080i output for video based HD DVD content (assuming your display can do proper motion adaptive deinterlacing of 1080i sources).

AIUI, sets with constant 540p bobbing (see Merson's three earlier articles (http://archive2.avsforum.com/avs-vb/showthread.php?p=9178202#post9178202)) use each 1/60-sec 540-line TV field from 1080i to create full frames. Newer test DVDs or a standard SMPTE test pattern detect this with flashing white and black screens as each TV field is displayed.

The two 540-line interlaced half-frames are supposed to be shown 1/60-sec apart on interlaced CRT displays, or digitally deinterlaced into full frames for fixed-pixel displays. With my 1080i CRT RPTV, my eyes merge the two 1/60-sec TV fields into 1/30-sec frames. With most fixed-pixel displays a buffer memory helps create 1/60-sec frames, preferably combining both fields/frame rather than creating frames from only 540-line half-frames.

The mid-90's ATSC-approval test was cited earlier above (http://www.avsforum.com/avs-vb/showpost.php?p=11853671&postcount=64). Test results (see simplified table (http://archive.avsforum.com/avs-vb/showthread.php?postid=326565#post326565)) showed that a static 1080i test pattern had 800 lines of effective vertical resolution (resolvable lines), while a dynamic (5-rpm) pattern only 400 lines. Vertical resolution is the number of resolvable B&W horizontal lines from screen top to bottom. This drop in vertical resolution with motion varies directly with the rate of motion, so movement less than ~5 rpm from a rotating test pattern shouldn't cause as much loss of vertical resolution. Also, the vertical resolution loss from movement should only take place where movement occurs. 540p bobbing, however, could diminish vertical resolution over the whole screen.

In the "2.1 Resolution" introduction of the ATSC report (linked above), the video experts wrote:
Again, that's where motion is taking place within the image, and the resolution drop varies with the rate of motion.

Also, as outlined originally above (yesterday), it's possible to use a non-sampled computer or pattern-generator test images and measure the full 1920X1080 resolution of many displays. Test patterns or actual programming sampled at ~74 MHz are filtered to reconstruct images from the samples, resulting in both vertical and horizontal resolution loss. Came across an interesting diagram, Fig. A3, (http://www.ebu.ch/en/technical/trev/trev_308-hdtv.pdf), from 2006 European HD studies, that compares the effective resolutions (resolvable details) of both 1080i and 720p after the Kell factor (~0.7 X line count) and anti-twitter filtering is factored in. Obviously such diagrams involving maximum resolutions expressed as lines correspond to equivalent details in scenes, and maximum resolvable details can be far less--if a film or video camera is heavily filtered, for example, or from many other limitations along the delivery chain. -- John

Thank you for the EBU technical review you posted on here.

I'm not sure I understand your (rhetorical?) question about everything getting converted to video over broadcast. Of course the short answer is "no", and we can expand on that,because there is plenty of 1080i/60 film based content available from NBC, CBS, HBO, etc.

As far as my DVD player doing the heavy lifting, the answer there is also no. I was able to do the same HQV tests last night but with my Toshiba HD A1 outputing 720p. The player did terrible! Failed the 1080i motion-adaptive test, as well as the 3:2 pulldown for film sources. But I suppose it is a moot point, as no film-based material would ever be encoded on the disc at 1080i/60. The one test that the player passed at 720p that my Sony 32SL130 failed @1080i input was "Film Mode High Detail" which shows up as moire in the stadium stands.

I guess I need to reconsider my assesment after thinking that through... As long as the HD DVD is 1080p/24, then outputing to 720p appears to preserve all the fine detail when viewing on a 768p set. For the 1st gen Toshiba models, my recommendation for viewing on a 720p set would be: use 720p output for 1080p film based content, and 1080i output for video based HD DVD content (assuming your display can do proper motion adaptive deinterlacing of 1080i sources).

Interesting...how do you know about broadcast?
What are they doing to the original content and does every broadcaster do the same thing?
The reason I wonder is I just can't see a big difference in PQ between panels that have passed interlace/3:2 tests and panels that failed using broadcast signals.
As far as the HD DvD players....isn't it the same as when people were saying DvD players tested on this panel failed but this panel did not when they were actually just confusing and testing player performance?
If you send DvD 480i...then you are testing the panel, yes?

How do I know about broadcast? I understand the transmission standards that the content providers use to deliver content that originated at 1080p/60 or 1080p/30, and how it is converted to 1080i/60 or 720p/60 for ATSC transmission. If you are hinting that other broadcasters do it differently than one another, or incorrectly, I ask that you be more specific or give examples of what you would like to discuss further.

I can see a difference in video based and film based content on panels that pass the deinterlace/3:2 tests. My Sony 32SL130 appears shaper than a friends 32" Olevia 232 model with the same content.

Yes, you are correct. If I'm sending DVD 480i, I'm definitely testing the panel's ability.

Interesting...how do you know about broadcast?
What are they doing to the original content and does every broadcaster do the same thing?
The reason I wonder is I just can't see a big difference in PQ between panels that have passed interlace/3:2 tests and panels that failed using broadcast signals.
As far as the HD DvD players....isn't it the same as when people were saying DvD players tested on this panel failed but this panel did not when they were actually just confusing and testing player performance?
If you send DvD 480i...then you are testing the panel, yes?

How do I know about broadcast? I understand the transmission standards that the content providers use to deliver content that originated at 1080p/60 or 1080p/30, and how it is converted to 1080i/60 or 720p/60 for ATSC transmission. If you are hinting that other broadcasters do it differently than one another, or incorrectly, I ask that you be more specific or give examples of what you would like to discuss further.

I can see a difference in video based and film based content on panels that pass the deinterlace/3:2 tests. My Sony 32SL130 appears shaper than a friends 32" Olevia 232 model with the same content.

Yes, you are correct. If I'm sending DVD 480i, I'm definitely testing the panel's ability.

Not hinting. I am just asking.
So your Sony and that Olevia are hooked up via cablecard and coax?
Otherwise....you are just testing the STB.
You really need to compare the same source using the same hardware or it's just not valid.

Edit: Some STB's do have pass-thru.

Both were hooked up to OTA digital antenna for HD sources.

Not hinting. I am just asking.
So your Sony and that Olevia are hooked up via cablecard and coax?
Otherwise....you are just testing the STB.
You really need to compare the same source using the same hardware or it's just not valid.

Edit: Some STB's do have pass-thru.

Not hinting. I am just asking.
So your Sony and that Olevia are hooked up via cablecard and coax?
Otherwise....you are just testing the STB.
You really need to compare the same source using the same hardware or it's just not valid.

Edit: Some STB's do have pass-thru.

My STB has pass-thru and my video processor tells me what it is getting including if it is Film or Video. Almost all prime-time shows and almost all movies are Film, sporting events and the news are video. The cable co just passes what is sent to them.

As far as my DVD player doing the heavy lifting, the answer there is also no. I was able to do the same HQV tests last night but with my Toshiba HD A1 outputing 720p. The player did terrible! Failed the 1080i motion-adaptive test, as well as the 3:2 pulldown for film sources. But I suppose it is a moot point, as no film-based material would ever be encoded on the disc at 1080i/60. The one test that the player passed at 720p that my Sony 32SL130 failed @1080i input was "Film Mode High Detail" which shows up as moire in the stadium stands.

I guess I need to reconsider my assesment after thinking that through... As long as the HD DVD is 1080p/24, then outputing to 720p appears to preserve all the fine detail when viewing on a 768p set. For the 1st gen Toshiba models, my recommendation for viewing on a 720p set would be: use 720p output for 1080p film based content, and 1080i output for video based HD DVD content (assuming your display can do proper motion adaptive deinterlacing of 1080i sources).

Sam,

2 questions - Does your Toshiba A1 have the first firmware update? Are you viewing using the HDMI input?

I also own the Toshiba A1 HD-DVD player, and after the 1st firmware update, the 720p quality is excellent if viewing thru HDMI, but poor [somewhat noisy] if using component, although 1080i thru either input is pretty good [but a little soft on component compared to HDMI]. I'm referring to the movie content on HD-DVD discs.

According to these tests involving motion which LCD sets sucked the least?

How far are they behind plasmas?

The article stated that all 120hz sets retained 600 lines of resolution. One set was the Sony KDL-46XBR4. All pdps lost resolution during motion as well. PDPs retained 830-880 lines of 1080. So I guess both technologies blur with motion.

Reading through this thread, I'm not quite sure what the results really mean. I seem to remember reading that the tests don't tell us much unless you report the H & V lines. I also believe the 3:2 and deinterlacing tests are important. I look forward to seeing those results.

The subject of this thread, has changed for me.
Does the article change from the OP?

At first, I thought this was about 1080p displays displaying 1080p.
Is the review only about 1080i & 3:2 being displayed?
Thanks.

Sam,

2 questions - Does your Toshiba A1 have the first firmware update? Are you viewing using the HDMI input?

I also own the Toshiba A1 HD-DVD player, and after the 1st firmware update, the 720p quality is excellent if viewing thru HDMI, but poor [somewhat noisy] if using component, although 1080i thru either input is pretty good [but a little soft on component compared to HDMI]. I'm referring to the movie content on HD-DVD discs.

I'm running firmware version 2.3, and I used HDMI to test the Sony panel. I definitely think that 720p via HDMI is the preferred setting for a 720p-native display assuming the HD DVD content is 1080p/24.

According to these tests involving motion which LCD sets sucked the least?

How far are they behind plasmas?

For the motion tests, the results are all over the board. The Sony LCDs fair the best overall, but others do well too. 65% of the sets tested passed the motion adaptive test, but only 13 of 74 sets pass both the motion and 3:2 film detection.

So according to the test the best PDP retain 880 lines out of 1080. That's an 18.5% drop.

If the same applies for 720p PDP, they would retain 587 out of 720.

The best 1080p LCD retain 600 out of 1080.

So it looks like a 720p PDP has just a bit less motion resolution than a 1080p LCD, correct?

Further, if the best 120hz 1080p LCD are dropping 55% resolution during motion, then a 720p LCD would have 400 lines of motion resolution at best. Odds are it's even less since it's not 120hz.

It would be a huge assumption to think that a 720p(768p) plasma like my 500u would lose the same 18% @ 720p resolutions.
It's losing little if any, according to my eyes. ;)
Eat your heart out, 1080p LCD's as my plasma is pnwning you in motion. :D

Sounds like LCD needs to rev up to about 240hz to get a significant increase in motion resolution.The difference between some of the 60 Hz LCDs and the 120 Hz LCDs was 580 compared to 600. In my opinion the doubling of the refresh rate probably had nothing to do with the difference and it was most likely due to a slightly better response time with the LCD panels that were used to make those 120 Hz LCDs.


So it looks like a 720p PDP has just a bit less motion resolution than a 1080p LCD, correct?Well based on the FPD benchmark test that is true but that test was made by a group of Plasma companies. A fact that for some reason Gary Merson forgot to mention in his description of the FPD benchmark test. The test obviously does test motion resolution but I have the hunch that it was made more to show the largest possible difference between LCDs and Plasmas than to do an objective test on motion resolution. Also considering that the DLPs only got 650 to 700 on that test I am somewhat curious to see what HD CRTs would have gotten.


It would be a huge assumption to think that a 720p(768p) plasma like my 500u would lose the same 18% @ 720p resolutions.Your right, it could be losing more than that ;). Seriously though I am a bit curious to know why no 720p displays were tested for motion resolution. I am also curious to why the Plasma companies that made the FPD benchmark test didn't make it available to the general public.

Last week Panasonic placed a full page add in our local paper comparing motion on LCD and Plasma at both 720p and 1080p. The graph showed the Plasma 720p TVs did not loose any resolution, while the 1080p plasms dropped below 900 lines. This I assume is because 720p is below the 880 line threadhold of plasma motion. For LCD, the 720p & 1080p TVs dropped to a similar 600 lines of motion.

The difference between some of the 60 Hz LCDs and the 120 Hz LCDs was 580 compared to 600. In my opinion the doubling of the refresh rate probably had nothing to do with the difference and it was most likely due to a slightly better response time with the LCD panels that were used to make those 120 Hz LCDs.
To everyone,

This test does not take into account 'hold-type' motion blurring which is the majority contributer to percieved motion blur by the viewer. It only tests response time as the contribution as Richard states above. In this scenario plasma will win hands down and OLED will be even better.

To properly test motion blur perception you need to use a moving camera system to simulate the human retina tracking a moving object on the screen. In that scenario 120Hz LCD will show a huge improvement over 60Hz LCD. But plasma will still be slightly better.

Moving Picture Resolution (http://www.advanced-pdp.jp/fpd/english.html#4)

Cheers

Last week Panasonic placed a full page add in our local paper comparing motion on LCD and Plasma at both 720p and 1080p. The graph showed the Plasma 720p TVs did not loose any resolution, while the 1080p plasms dropped below 900 lines. This I assume is because 720p is below the 880 line threadhold of plasma motion. For LCD, the 720p & 1080p TVs dropped to a similar 600 lines of motion.

Thanks for posting. I had incorrectly assumed each type of display was dropping a certain percentage.

Nice to know 720p LCD are not dropping into the cellar.

The next question is, since our eyes are not instruments, how 'visible' is the difference between 880 and 600 lines of motion resolution? Will different people notice it more than others?

To everyone,

This test does not take into account 'hold-type' motion blurring which is the majority contributer to percieved motion blur by the viewer.Where did you read that from? Also what exactly is the difference between "hold type" motion and response time?


It only tests response time as the contribution as Richard states above. In this scenario plasma will win hands down and OLED will be even better.I would mention that Gary Merson thought little of the motion resolution (http://hdguru.com/?p=195) of the OLED display he watched at CEATEC.


To properly test motion blur perception you need to use a moving camera system to simulate the human retina tracking a moving object on the screen. In that scenario 120Hz LCD will show a huge improvement over 60Hz LCD. But plasma will still be slightly better.

Moving Picture Resolution (http://www.advanced-pdp.jp/fpd/english.html#4)Honestly speaking I don't even see where you are getting that from on that Plasma company website.

Where did you read that from? Also what exactly is the difference between "hold type" motion and response time?When an object is in motion on a video screen our eyes follow (track) the movement continously. However, since video is made up of successive still frames our eyes are still moving while each frame is stationary. The longer the frame is visible the more blur is drawn onto our moving retina. Since LCD is a sample and "hold" technology it is the most susceptible to this phenomenon. And with current LCD tech having fairly good response times the hold-type blurring is the major contributor.

http://scitation.aip.org/getabs/servlet/GetabsServlet?prog=normal&id=DTPSDS000036000001001590000001&idtype=cvips&gifs=yes

And 120Hz LCD technology was devopled to combat this phenomenon by reducing the "hold" time by 50%.

I would mention that Gary Merson thought little of the motion resolution (http://hdguru.com/?p=195) of the OLED display he watched at CEATEC. OLED has the best response times but since it uses active matrix addressing similar to LCDs the pixels are "held" on for the entire frame period just like LCDs. Therefore it will have poor motion resolution compared to plasma. A passive matrix OLED display will have excellent motion resolution.

Honestly speaking I don't even see where you are getting that from on that Plasma company website.MTF and MPRT are becoming the standard measurement systems in the industry. Using a tracking camera is the most accurate method developed to date. Not just Plasma manufacturers are using it.

http://scitation.aip.org/getabs/servlet/GetabsServlet?prog=normal&id=DTPSDS000035000001001266000001&idtype=cvips&gifs=yes

Last week Panasonic placed a full page add in our local paper comparing motion on LCD and Plasma at both 720p and 1080p. The graph showed the Plasma 720p TVs did not loose any resolution, while the 1080p plasms dropped below 900 lines. This I assume is because 720p is below the 880 line threadhold of plasma motion. For LCD, the 720p & 1080p TVs dropped to a similar 600 lines of motion.

Interesting...
I knew it!!!
Hehe....

This test does not take into account 'hold-type' motion blurring.. It only tests response time as the contribution as Richard states above.



Wait, I still have faith in Gary...The test was a subjective eye test. Shouldn't it account for sample and hold blur?

This Blu-ray disc contains a pattern called a monoscope pattern which is made up of a series of four black lines drawn so they gradually come together in a wedge like pattern. There are numbers adjacent to the lines indicating resolution.....the motion pattern with the area where the four lines could still be discerned as separate and not blurred together was noted. These numbers were then compared.

Wait, I still have faith in Gary...The test was a subjective eye test. Shouldn't it account for sample and hold blur?To get an accurate test of sample and hold you need a pursuit camera and a video object moving at a high enough velocity to induce 'hold-type' blurring.

I assume that because Gary's data shows 120Hz to show little improvement that the test fails to capture the sample and hold effect.

Even if Gary test somehow met the proper conditions, it is so subjective because retinal persitence is not the same for every viewer and can vary widely.

The MPRT method is being adopted for a good reason. That being all manufacturers (LCD,OLED,PDP) think it is accurate.

Cheers

So after this thread I went and bought the magazine and read the article. I've been considering a 52" Sharp LCD TV for a while now and was pretty close to the LC-52D82U model. It showed up on the test list and failed the deinterlace and only passed 3:2 using component inputs. That really surprised me, but what was more strange to me was that the LC-46D62U model succeeded in both tests and still kept 600 lines of resolution despite being a 60Hz model. The update to the 62U is out now, LC-52D64U and I'm debating the merit of that compared to the LC-52D82U. 64U deinterlaces and handles 3:2 pulldown very well, is less likely to have Sharps known banding issue, and is cheaper, but runs at 60Hz. 82U fails those tests but runs at 120Hz.

So given the choice between a 120Hz model that failed these tests and a 60Hz model that passes, which would you go with?

So after this thread I went and bought the magazine and read the article. I've been considering a 52" Sharp LCD TV for a while now and was pretty close to the LC-52D82U model. It showed up on the test list and failed the deinterlace and only passed 3:2 using component inputs. That really surprised me, but what was more strange to me was that the LC-46D62U model succeeded in both tests and still kept 600 lines of resolution despite being a 60Hz model. The update to the 62U is out now, LC-52D64U and I'm debating the merit of that compared to the LC-52D82U. 64U deinterlaces and handles 3:2 pulldown very well, is less likely to have Sharps known banding issue, and is cheaper, but runs at 60Hz. 82U fails those tests but runs at 120Hz.

So given the choice between a 120Hz model that failed these tests and a 60Hz model that passes, which would you go with?

IMO the tests performed in the article doesn't apply to majority of todays HD programming. The information in that article is good information to keep in mind for future or if you are deciding whether or not shelling out extra cash for 1080p HD set right now is worth it.

Ther real stoy about all of this is that LCD sucks when it comes to motion.

Plasma is better but isn't perfect.

I wonder how the upcoming in January Sony top of the line SXRD will perform in regards to motion?

Home Theater magazine doesn't believe motion blur is a deal breaker - at least for most: "The latest LCDs, despite faster and faster response times, are not devoid of motion blur as plasma is, but often they're fast enough for most people not to be bothered by it. " http://www.hometheatermag.com/advicefromtheexperts/407plasmavlcd/index.html

Probably useful to consider so-called resolution-pumping blur as well as LC switching-time blur. There must be times when the two combine. Resolution pumping takes place between static images (or image areas) when deinterlacing circuits for video-originated signals (from cameras) simply weave together odd-even TV lines, providing crisp images; then when even slight motion takes place blurring results because the deinterlacing algorithm isn't sophisticated enough to avoid noticeable contrasting blur.

Here are four excerpts (http://www.avsforum.com/avs-vb/showpost.php?p=11843791&postcount=200) from WSR articles about recent higher-end front projectors and how their video processors handle resolution pumping. -- John

Home Theater magazine doesn't believe motion blur is a deal breaker - at least for most: "The latest LCDs, despite faster and faster response times, are not devoid of motion blur as plasma is, but often they're fast enough for most people not to be bothered by it. " http://www.hometheatermag.com/advicefromtheexperts/407plasmavlcd/index.html

Then that magazine needs to have a good talk with all these people, and convince them that they should not be so bothered with the motion blurring problems on their brand new Samsung LCD sets.

Check out the owner's thread on the 71 Samsung LCD models. Read the last couple of pages. Owners are now joining in registering mass complaints with Samsung about the blurring problem. So far, Samsung is telling them that there will be no fix issued, and even advised one person to just return the set.

http://www.avsforum.com/avs-vb/showthread.php?t=893827

To be fair greenland, the issue that people are registering complaints with Samsung on is related to the processing artifacts inherent in the current (immature) implementation of 120hz.

Basically the newest weapon in the fight against motion blur introduces some processing artifacts that get more severe as you crank it up. Additionally, the US models don't allow you to turn the feature off completely as the EU model apparently does.

Well it makes me wonder how much detail the eye can really see when things are moving fast. If the object being tracked doesn't stay still isn't there a natural blur of things anyway?

I wonder if this isn't more obvious just due to the fact that stationary objects are sharper now in HD than they have been in the past.

Mike

Well it makes me wonder how much detail the eye can really see when things are moving fast. If the object being tracked doesn't stay still isn't there a natural blur of things anyway?



An object being tracked should not blur because it isn't moving relative to the eye (it's in effect stationary). The idea you just mentioned is what LCD manufacturers used for years to ignore LCD blur.

when eye tracking is considered, a moving object is rendered stationary with respect to the retina so that temporal frequencies fall to zero. In this case much the same acuity to detail is available despite motion. It astonishes the author that video engineers so often state that softening of moving objects is inevitable and acceptable, when it plainly isn't.

http://www.microsoft.com/taiwan/whdc/archive/convrg.mspx

I think it bears mentioning that the situation where a moving object is perfectly tracked by the viewer or camera to enable full detail for the object would be extremely rare (perhaps, unlikely), with the exception of a deliberately contrived CG effect. There will always be some degree of motion blur, though obviously not as much as when the subject is in full motion past a static viewer.

Think of it this way, if you are out playing around with your digital camera, how likely are you to be able to capture an absolutely focused shot of a chosen subject if you attempt to snap the shot while you walk in a circle around the object while tracking the position of the subject in your viewfinder? It's difficult enough to achieve a perfect capture just standing still, depending on how "trembly" you are. Anyone who takes pictures even as a hobby knows that the best focused and unblurred shots come from when you aren't supporting the camera, at all, on your person...the camera is put on a "dead" tripod.

Hence, as soon as something is "moving" through the shot or as soon as someone has to hold/position/track the camera through the moving shot, logistics dictate that "all bets are off" that the subject can be shot in hard "perfect" focus with zero motion blur down to the last pixel of camera resolution. So if the source material cannot even meet this criteria, it's really a moot point that the display device needs to meet this criteria. Someday it may be subject to such material, but not for the time being with current live production techniques.

I think it bears mentioning that the situation where a moving object is perfectly tracked by the viewer or camera to enable full detail for the object would be extremely rare (perhaps, unlikely), with the exception of a deliberately contrived CG effect. There will always be some degree of motion blur, though obviously not as much as when the subject is in full motion past a static viewer.

Think of it this way, if you are out playing around with your digital camera, how likely are you to be able to capture an absolutely focused shot of a chosen subject if you attempt to snap the shot while you walk in a circle around the object while tracking the position of the subject in your viewfinder? It's difficult enough to achieve a perfect capture just standing still, depending on how "trembly" you are. Anyone who takes pictures even as a hobby knows that the best focused and unblurred shots come from when you aren't supporting the camera, at all, on your person...the camera is put on a "dead" tripod.

Hence, as soon as something is "moving" through the shot or as soon as someone has to hold/position/track the camera through the moving shot, logistics dictate that "all bets are off" that the subject can be shot in hard "perfect" focus with zero motion blur down to the last pixel of camera resolution. So if the source material cannot even meet this criteria, it's really a moot point that the display device needs to meet this criteria. Someday it may be subject to such material, but not for the time being with current live production techniques. I see what you are saying but I don't see the relevance to display induced blur measurement. A camera tracking an object in real life is irrelevant because both are moving continuously. And the display motion blur measurement using a tracking camera is not done by hand. It is a very accurate tracking system.

Regarding display measurement, it is called "hold-type" blurring for a reason. On a display the object being tracked is not moving continuously. It is a series of still images. Yet our eyes are moving continously. Since our eyes are moving while the object is not, there is blur produced on our retina. The more retinal persistence one has the more blur is percieved.

The only way to eliminate subjectivity and make the measurement comparable between manufacturers is to use a tracking camera with set specifications ( tracking speed, shutter speed...etc)

The relevance to display induced blur measurement is that the chances of actual footage material appearing on a broadcast or movie production that would actually stress a display in such a manner are virtually nill. The measurement, itself, is fine from an academic standpoint, but its impact will never manifest in real use with real material because the source, itself, is already compromised to a greater extent.

It's like measuring the existence of valve float at 7000 rpms on an engine. While the problem certainly exists and is certainly measurable, it's impact in real use of that engine is nonexistent, since the production incarnations of that engine in real vehicles are automatically rev limited by the engine computer at 6000 rpms. In the case of LCD panels, motion blur can certainly be induced by specific test patterns, but the impact does not occur in real use with real video material because the material simply does not probe those limits by virtue of the logistics/limitations of shooting objects in motion.

Mind you, I only provided this counterpoint in response to the following statement which occurred in this discussion:

"It astonishes the author that video engineers so often state that softening of moving objects is inevitable and acceptable, when it plainly isn't."

I maintain that by taking into the account of details of such a scenario, the assessment of "inevitable and acceptable" is certainly reasonable for classic capture of motion scenes (no CG trickery ;) ).

The premise of the eye tracking an object in motion is conceptually sound. It's the details of assessing how well it actually does this, that make practice deviate from concept. Tracking an object perfectly down to the very limits of your available resolution is easier said than done.

The relevance to display induced blur measurement is that the chances of actual footage material appearing on a broadcast or movie production that would actually stress a display in such a manner are virtually nill. The signal is not stressing the display? It is a visual perception problem as I've described countless times here. If you want to describe it in terms of stress you would say the display puts stress on our eyes by holding each image frame too long.

The measurement, itself, is fine from an academic standpoint, but its impact will never manifest in real use with real material because the source, itself, is already compromised to a greater extent.You are saying that display induced motion blur can not be percieved because the sourcing induced motion blur has compromised the signal already?? You forget that display induced motion blur has nothing to do with source quality and therefore will be additive to any blur existing in the source itself. Manufacuturers are trying to make the display handle motion better, not improve the signal quality? Why would all the manufacturers, universities, research centres work on this display induced motion blurring problem when in your opinion it is not a problem? The problem is relative.

Competing technologies want the best motion "handling" regardless of signal quality. To do this they need to reduce the hold time of the display.

So if the source material cannot even meet this criteria, it's really a moot point that the display device needs to meet this criteria.

ah, what xrox said. And don't forget video games and similar have no blur from the source. (well most)

Tracking an object perfectly down to the very limits of your available resolution is easier said than done.You miss the point here. The reason camera tracking is being used to evaluate display induced motion blur is that it takes the subjectivity out of the measurement. Whether the tracking is accurate is not very important. What is important is that each manufacturer use the same system and parameters to measure. And the measurement is mathematical eliminating the user from the equation. Even so the tracking is done by computer and motion control so I would say it is pretty accurate.

Regarding our perception of 'hold-type' blur, since a moving object on a video monitor is actually never in motion then "any" amount of tracking by our eyes will induce blur onto our retina. What determines the amount blur induced is retinal velocity, image hold time, and retinal persistence.

I think we are talking toward to different points. When I mention "camera tracking", I'm not disputing its relevance in the context of a motion test.

I am referring to how it is involved in the capture of real footage which in turn goes into some sort of broadcast program or movie media. Motion imagery will always be subject to some degree of imperfection in focus or motion blurring. Hence, the programming that is available for viewing on a particular monitor/display can never transcend that limitation. There's never going to be an instance where an object in motion is perfectly tracked in focus and position in a camera setup involving human operation or manual mechanical operation (that pretty much leaves CG to create a specific special effect). So the "eye perfectly tracking a moving object such that all motion vectors zero out" scenario seems a tad oversimplified to describe what happens in real use. That was the point I was commenting on.

I am referring to how it is involved in the capture of real footage which in turn goes into some sort of broadcast program or movie media. Motion imagery will always be subject to some degree of imperfection in focus or motion blurring. Hence, the programming that is available for viewing on a particular monitor/display can never transcend that limitation. Yes, I agree on this. No display today can improve on the blur already present in the signal.

There's never going to be an instance where an object in motion is perfectly tracked in focus and position in a camera setup involving human operation or manual mechanical operation (that pretty much leaves CG to create a specific special effect). When filming yes, but regarding the MPRT test I might disagree as all parameters can be controlled accurately.

So the "eye perfectly tracking a moving object such that all motion vectors zero out" scenario seems a tad oversimplified to describe what happens in real use. So how would you describe what happens in real use? Remember, I am only talking about "display" induced motion blur.....

So how would you describe what happens in real use? Remember, I am only talking about "display" induced motion blur.....

In real use, the tracking is never "perfect". So if you are viewing footage involving the tracking of an object moving across a background, the camera will still be dealing with small motion vectors and dynamic deviations from correct focus. That is going to be the genesis of some degree of lost detail in the "still" part of the shot that will be there regardless of how much or how little motion blur is occurring from the LCD panel, itself. Similarly, you could have some hypothetical "blur-less" LCD, and that's not going to suddenly make that tracking shot of a football in mid-pass appear in absolute pixel-by-pixel detail, the same as if you could take a static shot of an absolutely motionless football (yes, I'm aware a thrown football has a spin, so it's not a perfect example- so sue me...you get the point). It's just not the same thing.

Mr. Hanky you seem to be suggesting that since source material already has blur in the signal, then if the display adds more then it's ok.

I would disagree. If the source has some blur than I certainly don't want the display to introduce additional blur, lets minimize it.

Also, I enjoy playing games which in most cases don't have any source blur.

In real use, the tracking is never "perfect". So if you are viewing footage involving the tracking of an object moving across a background, the camera will still be dealing with small motion vectors and dynamic deviations from correct focus. That is going to be the genesis of some degree of lost detail in the "still" part of the shot that will be there regardless of how much or how little motion blur is occurring from the LCD panel, itself. I said "display induced motion blur", not signal inherent motion blur. Again, as I've tried to explain, motion blur reduction in displays is not concerned with how much blur is in the signal. They are only concerned with "display" induced blur, and that blur will be additive to any in the signal. I don't see what you are trying to argue here? Why are you focusing on blur inherent to the signal when that is irrelevent to this discussion?

Similarly, you could have some hypothetical "blur-less" LCD, and that's not going to suddenly make that tracking shot of a football in mid-pass appear in absolute pixel-by-pixel detailNo one said it did. Again, a display cannot improve on blur in the signal, but it can make it much worse. The goal is to reduce the motion blur produced by the display regardless of the blur in the signal.

We seem to be going in circles here.......

xrox, how can you tell if the blur is from the signal or induced by the set? I can't see how you can determine this unless you have a known clean signal. Since tracking is never perfect, where would you get such a signal that is clear of motion blur? If you don't know the amount of motion blur in the input signal how can you determine if the signal or set is at fault?

Mr. Hanky you seem to be suggesting that since source material already has blur in the signal, then if the display adds more then it's ok.

It is not clear (no pun intended) that a display can "add" more blur to a pre-blurred signal. If blur has already been applied to the signal, then it is likely that the display will be able to keep up with the degree of moving detail that is left (hence, it will not add more blur to material that is already blurred). Even if it didn't, could you really tell one kind of blur from another? Blur is blur, when it's all said and done.

Once the signal has been motion-blurred from the source, that means high frequency information in moving objects has been reduced/lost. What is left is low(er) frequency components which would not blur on the LCD screen since the transition is slower as it moves past numerous pixels on screen. Each pixel has more time to adjust to the correct level (instead of lagging and hence blurring).

The whole reason motion blurring occurs entirely boils down to the presence of detail in the frequency domain. If the high-frequency components have been removed/attenuated, then there isn't much left to cause more blurring in the display. The display will be able to keep up with the lower frequency components of the detail that is left, even if it measures up to lose half the native resolution under motion.

These motion blur tests are fine and all, but people should remember to take them under the context of the conditions that they test. It is true that the display loses a considerable amount of native resolution under motion. However, if this high frequency detail is already blurred in the source (broadcast or movie disc), then it is pretty much a moot point. The "tests" are testing an envelope of performance that far exceeds anything you would encounter in a broadcast or movie disc. So the test is useful from a design/optimization/QC standpoint, but not so relevant for actual viewing of real program material that consumers have access to (where blur at the source is an inevitability by sheer artistic license or the inherent limitations of the capture device). You could invent a 100% blur-free LCD panel, and you are still going to see the same motion blur in broadcast and movie material, because that blur was/had been always in the source, to begin with.

If the high-frequency components have been removed/attenuated, then there isn't much left to cause more blurring in the display. The display will be able to keep up with the lower frequency components of the detail that is left

Mathematically that looks good but pixels transiton at a frequency of 1/60 seconds period. The frequency attenuation that occured in the source has no bearing on this. A blurry source will blur more if the panel can't keep up with the refresh rate.

The 1/60 sec thing would only apply if the motion scene contained absolutely no motion blur on its own. When motion blur is added/occurs at the source, the amount of change at a pixel from one frame to the next is reduced (because it is now a blended mixture of what was in the adjacent pixel and in that pixel 1/60th sec earlier. There is less change, so the timely transition to the new value can be achieved by the display.

The frequency attenuation is essentially dictating the maximum amount of change a pixel will "see" from one frame to the next and as the values of adjacent pixels moves to another pixel (motion in time and space). As high frequency components are attenuated, the transition in any one pixel will be smaller. This transition has then reached a point where the display can accurately follow the change.

Given that 1/60 sec pertains to a rating of 17 ms, and the current generation of displays seem to be getting down to sub-10 ms ratings and using some variant of the overdrive scheme, I'd say 17 ms performance is pretty well covered.

xrox, how can you tell if the blur is from the signal or induced by the set? I can't see how you can determine this unless you have a known clean signal. Since tracking is never perfect, where would you get such a signal that is clear of motion blur? If you don't know the amount of motion blur in the input signal how can you determine if the signal or set is at fault?Since they are seperate entities and are additive, the only accurate, transferrable, and quantitative way to do so is to use a computer generated test pattern that moves a designated number of pixels per frame along with a tracking camera. This isolates the display "hold-type" motion blur contribution.

As for the additive combination of signal blur and display blur well why do you think that response time, and hold-time are such big issues with LCD? The reason is that consumers complained that LCDs show more blur than CRT or Plasma even thoug they are playing the same signal with the same signal blur. The blur from the display is additive...................

And why do you think that LCD now is moving to LED strobbing, 120Hz, backlight scanning.....etc. The answer is to reduce 'hold-type' motion blur regardless of signal blur.

Mr. Hanky, I doubt that any more debate will change your misconceptions on this topic. Your ideas go against ubiquitous scientific evidence.

It is not clear (no pun intended) that a display can "add" more blur to a pre-blurred signal. Yes it is clear. The problem lies in that you don’t clearly understand why. And it seems clear that I can’t explain it properly to you. Would it be more convincing if I posted numerous documents on the subject?

If blur has already been applied to the signal, then it is likely that the display will be able to keep up with the degree of moving detail that is left (hence, it will not add more blur to material that is already blurred). From this it seems you still do not know what ‘hold-type’ blur is. Like I’ve already said the signal is not stressing the display, the blur is produced on our retina. If my explanations aren’t clear than maybe you should read some documents on hold-type blurring.

Even if it didn't, could you really tell one kind of blur from another? Blur is blur, when it's all said and done. MPRT measure the blur produced by the display only by using a computer generated moving test pattern. As for the consumer, well many people can watch an LCD and Plasma side by side and easily see the increased motion blur on the LCD even though both are playing the same signal. It all depends on one's retinal persistence and sensitivity to temporal artifacts.

Once the signal has been motion-blurred from the source, that means high frequency information in moving objects has been reduced/lost. What is left is low(er) frequency components which would not blur on the LCD screen since the transition is slower as it moves past numerous pixels on screen. Each pixel has more time to adjust to the correct level (instead of lagging and hence blurring).This comment makes in very clear you have no idea what I’m talking about and/or what hold-type blur is. Hint: It has nothing to do with pixel transitions….

You could invent a 100% blur-free LCD panel, and you are still going to see the same motion blur in broadcast and movie material, because that blur was/had been always in the source, to begin with. This is absolutely correct. Problem is there is no such thing as a 100% blur free panel and this is due to the ‘hold-type’ problem of video reproduction.

The 1/60 sec thing would only apply if the motion scene contained absolutely no motion blur on its own. When motion blur is added/occurs at the source, the amount of change at a pixel from one frame to the next is reduced (because it is now a blended mixture of what was in the adjacent pixel and in that pixel 1/60th sec earlier. There is less change, so the timely transition to the new value can be achieved by the display.

The frequency attenuation is essentially dictating the maximum amount of change a pixel will "see" from one frame to the next and as the values of adjacent pixels moves to another pixel (motion in time and space). As high frequency components are attenuated, the transition in any one pixel will be smaller. This transition has then reached a point where the display can accurately follow the change.

Given that 1/60 sec pertains to a rating of 17 ms, and the current generation of displays seem to be getting down to sub-10 ms ratings and using some variant of the overdrive scheme, I'd say 17 ms performance is pretty well covered.Pixel transitions and hold-type blurring are seperate causes of blur. To be more specific, an LCD with sub-1ms ratings will still show lots of hold-type blur.

http://scitation.aip.org/getabs/servlet/GetabsServlet?prog=normal&id=DTPSDS000036000001001590000001&idtype=cvips&gifs=yes

What can I say, the discussion is pretty much over when you give up and simply claim "the other person doesn't know what they are talking about", in lieu of addressing the points presented.

What can I say, the discussion is pretty much over when you give up and simply claim "the other person doesn't know what they are talking about", in lieu of addressing the points presented.That is fine, I tried my best. All I'm left with is the truth and I'll post it whenever it needs to be.

Cheers

As the "center piece to a home theater system" most people I know try to to sit pretty much in front of, or within a few degrees off center from, the screen. .

Sure if it is a tiny display.

Since they are seperate entities and are additive, the only accurate, transferrable, and quantitative way to do so is to use a computer generated test pattern that moves a designated number of pixels per frame along with a tracking camera. This isolates the display "hold-type" motion blur contribution.

As for the additive combination of signal blur and display blur well why do you think that response time, and hold-time are such big issues with LCD? The reason is that consumers complained that LCDs show more blur than CRT or Plasma even thoug they are playing the same signal with the same signal blur. The blur from the display is additive...................

And why do you think that LCD now is moving to LED strobbing, 120Hz, backlight scanning.....etc. The answer is to reduce 'hold-type' motion blur regardless of signal blur.IMHO, response time is no longer an issue with LCDs - the real issue, as you state, is the LCD sample and hold time. 120Hz panels reduce that in half. I recently read of a 180Hz demo'd in Japan.

Although a computer generated test pattern would show differences. I would suggest the identical movie, game, etc. displayed on a plasma, lcd, and crt side by side. I believe a 'real world' test would be more meaningful to the average viewer than a computer generated test pattern.

I know that some viewers have a complaint over LCD motion blur but I don't know if that number is large. If it is, it is not reflected in the market. Even Home Theater Mag states that motion blur on LCDs should not be a problem for "most" viewers. http://www.hometheatermag.com/advice...lcd/index.html

Pixel transitions and hold-type blurring are seperate causes of blur. To be more specific, an LCD with sum-1ms ratings will still show lots of hold-type blur.

http://scitation.aip.org/getabs/serv...cvips&gifs=yes

I sure it will, but it will still be blur amongst many other layers of blur occurring throughout the entire chain from source to your eyes. One must be sure to maintain context and perspective over the entire situation, not just blur in one link. You can concentrate on eliminating blur in one link, but if there is still various other kinds of blur occurring in the other links and in much greater degree, what have you accomplished? Maybe "hold-type blur" is the big bogey for this marketing year? Maybe next year it will be a different kind of blur. After a while, you figure out all this toil is mere marketing manipulation and very little real progress.

Although a computer generated test pattern would show differences. I would suggest the identical movie, game, etc. displayed on a plasma, lcd, and crt side by side. I believe a 'real world' test would be more meaningful to the average viewer than a computer generated test pattern.

I know that some viewers have a complaint over LCD motion blur but I don't know if that number is large. If it is, it is not reflected in the market. Even Home Theater Mag states that motion blur on LCDs should not be a problem for "most" viewers. http://www.hometheatermag.com/advice...lcd/index.html A real world test is too subjective as our eyes have varying amounts of retinal persistence.

And the number of consumer complaints must have been large enough to drive the current technology? Many of the new LCD technologies have been developed for this purpose (Strobbing, scanning, 120Hz)

I sure it will, but it will still be blur amongst many other layers of blur occurring throughout the entire chain from source to your eyes. One must be sure to maintain context and perspective over the entire situation, not just blur in one link. You can concentrate on eliminating blur in one link, but if there is still various other kinds of blur occurring in the other links and in much greater degree, what have you accomplished? Maybe "hold-type blur" is the big bogey for this marketing year? Maybe next year it will be a different kind of blur. After a while, you figure out all this toil is mere marketing manipulation and very little real progress.Try the last 6 years. Manufacturers realized early that pixel response did not solve the problem and that they needed to reduce the hold time to increase motion handling performance.

After a while, you figure out all this toil is mere marketing manipulation and very little real progress.So when manufactures spent 10 years solving the response time issue with LCDs there was no real progress? You make no sense?

Early LCD models were just awful blurry messes thanks to 45ms response times. I'd say that todays LCDs have made pretty good progress.

Minimizing display induced motion blur is a good thing IMO and science and industry seem to agree.

When motion blur is added/occurs at the source, the amount of change at a pixel from one frame to the next is reduced so the timely transition to the new value can be achieved by the display.

I see exactly where you're comming from but this is not what happens AFAIK.
The situation is opposite: grey-to-grey transitions actually take more time than high contrast situations (black/white ect.) because a much lower voltage must be used. This can be seen on any LCD response time graph. Even if response times were quicker with like colors, how would that reduce blur on borders of moving objects?


Given that 1/60 sec pertains to a rating of 17 ms, and the current generation of displays seem to be getting down to sub-10 ms ratings and using some variant of the overdrive scheme, I'd say 17 ms performance is pretty well covered.

Arguably yes but sample and hold blur is now the major source of LCD blur so the display can still degrade the source.

Here is the abstract of a great scientific paper describing how 120Hz does improve motion perception and there is room for more improvement.

Abstract —The ideal frame rate for the highest motion-image quality with respect to blur and jerkiness is presented. In order to determine the requirements for avoiding these impairments, motion images from a high-speed camera and computer graphics were combined with a high-speed display to perform a psychophysical evaluation. The camera, operating at 1000 fps, and image processing were used to simulate various frame rates and shutter speeds, and a 480-Hz CRT display was used to present motion images simulating various frame rates and time characteristics of the display. Subjects were asked to evaluate the difference in quality between motion images at various frame rates. A frame rate of 480 fps was chosen to be an appropriate reference frame rate that, as a first estimation, enables coverage up to the human-dynamic-resolution (HDR) limit based on another experiment using real moving charts. The results show that a frame rate of 120 fps provides good improvement compared to that of 60 fps, and that the maximum improvement beyond which evaluation is saturated is found at about 240 fps for representative standard-resolution natural images.

You can concentrate on eliminating blur in one link, but if there is still various other kinds of blur occurring in the other links and in much greater degree, what have you accomplished? Maybe "hold-type blur" is the big bogey for this marketing year? Maybe next year it will be a different kind of blur. After a while, you figure out all this toil is mere marketing manipulation and very little real progress.

No, it's always been the bad "bogey" for LCD HDTV's. 3-4 years ago it was really noticeable with LCD displays having 22ms and higher response times (even 16ms, unless you sit further away from the screen). But even if you could have an 0ms response time, the basic sample-and-hold technology of LCD can still result in some viewers noticing the "native" blur [as others have stated, because it's subjective to each person's retinas].

I'm glad that LCD panels are being improved in reducing the inherent blur [regardless of response time], as we see experiments with new technology, like 120hz refresh etc (a small step up, but still needs improvement).

I see exactly where you're comming from but this is not what happens AFAIK.
The situation is opposite: grey-to-grey transitions actually take more time than high contrast situations (black/white ect.) because a much lower voltage must be used.

This applies to "classic" LCD designs that are many generations old by now. When you move to an overshooting scheme, you can achieve minimal latency on pretty much any transition, as long as you have range to overshoot beyond the target value. So ironically, this moves the latency problem back to the extremes, rather than grey-to-grey scenarios. By now, they are probably multiple generations on with the overdrive scheme, so I would guess latency across the board is on a tight leash well within the performance requirements for 60 Hz operation.

Here is the abstract of a great scientific paper...

That's great, but achieving such extreme refresh rates won't amount to much for real displays playing real material that is bound at 24 and 60 Hz and varied degrees of motion blur added on top of that to boot. There's point where radical improvement in one area is pointless due to the contribution of all of the other areas which have not been improved.

That's great, but achieving such extreme refresh rates won't amount to much for real displays playing real material that is bound at 24 and 60 Hz and varied degrees of motion blur added on top of that to boot. Your opinion based on misconception.

From the same paper : "Because there is a large difference between the visual resolution limits for the real and the displayed images in Fig. 6, improvement of the temporal resolution of hold-type displays is required for the realization of higher-quality motion images. The perceived MTF of hold-type displays has been extensively studied.5 The decrease of visual resolution of hold-type displays in smooth pursuit even for images taken with a high-speed shutter is caused mainly by retinal slip, which today is well-known. "

There's point where radical improvement in one area is pointless due to the contribution of all of the other areas which have not been improved.Of all the papers I've read over the years I remember one that seemed to partially agree with this regarding signal vs display blur. I'll have to look it up again.

When an object is in motion on a video screen our eyes follow (track) the movement continously. However, since video is made up of successive still frames our eyes are still moving while each frame is stationary. The longer the frame is visible the more blur is drawn onto our moving retina.Interesting explanation, but has that really been proven? I have to point out that a lot of Plasma companies have invested billions of dollars into Plasma production so I have some right to be skeptical when someone says that the "hold type" motion of LCD is a negative. I personally have never heard about that before and I would need to see some prove of it to believe that. Also there are several other "hold type" displays that don't suffer from motion blur at 60 Hz so it seems a bit convenient that only LCD is affected.


Since LCD is a sample and "hold" technology it is the most susceptible to this phenomenon. And with current LCD tech having fairly good response times the hold-type blurring is the major contributor.What about DLP and LCOS? Also it just seems to me that if what you said was true that it would be really noticeable with DLP since we have digital movie theaters that use 3 chip DLP projectors at 24 Hz.


MTF and MPRT are becoming the standard measurement systems in the industry.Okay, but that doesn't prove that their is a noticeable difference between 60 Hz and 120 Hz displays. I have read several reviews in which the reviewer said that there is no noticeable difference unless motion interpolation is used.


So given the choice between a 120Hz model that failed these tests and a 60Hz model that passes, which would you go with?I would recommend a 120 HZ display that passes all of the HQV tests and from what I have read (http://www.engadgethd.com/2007/10/10/engadget-hd-review-ln-t4671f-120hz-hdtv/) the Samsung 71 LCD series has done that.


Home Theater magazine doesn't believe motion blur is a deal breaker - at least for most: "The latest LCDs, despite faster and faster response times, are not devoid of motion blur as plasma is, but often they're fast enough for most people not to be bothered by it. "And from what I have read the Home Theater magazine tends to be pretty harsh when it comes to LCDs. Also just to add to this the reviewer for the Samsung LN-T4665F at Widescreen Review said this when referring to Xbox 360 games: "I never saw any smearing or blurring due to display lag, no matter how fast I moved my characters through their virtual worlds."

Interesting explanation, but has that really been proven? I have to point out that a lot of Plasma companies have invested billions of dollars into Plasma production so I have some right to be skeptical when someone says that the "hold type" motion of LCD is a negative. I personally have never heard about that before and I would need to see some prove of it to believe that. Also there are several other "hold type" displays that don't suffer from motion blur at 60 Hz so it seems a bit convenient that only LCD is affected.


It's even been documented in some LCD manufacturer's publications. Why do you think they have so much effort behind 120hz?

Interesting explanation, but has that really been proven? Proven and measured. It is difficult to measure as it is a human perception issue and we all have varying sensitivity to it. This is why MPRT was implemented. This way a mathematical measurement free of subjectivity can be taken and all manufacturers can compare numbers without bias (assuming no fudging…:))

I have to point out that a lot of Plasma companies have invested billions of dollars into Plasma production so I have some right to be skeptical when someone says that the "hold type" motion of LCD is a negative. I personally have never heard about that before and I would need to see some prove of it to believe that.
Just type in google “LCD hold type blur”. And Plasma is not a sample and hold display.

Also there are several other "hold type" displays that don't suffer from motion blur at 60 Hz so it seems a bit convenient that only LCD is affected. What displays do you refer to? DLP and Plasma are PWM displays. I have no clue how LCOS works and don’t know much about microdisplays either so I can’t comment on those. As for flat panels the only true sample and hold displays are LCD and OLED. If you compare the hold time versus display technology you get a pretty clear indication of the effect.

CRT – 1-2ms hold time
Plasma – 4-6ms hold time
LCD (120Hz) – 8.3ms hold time
OLED (120Hz) - 8.3ms hold time
LCD (60Hz) – 16.7ms hold time

It is no coincidence that CRT is considered the reference for motion handling of video as it has short response times and short hold times (the ideal combination)

Okay, but that doesn't prove that their is a noticeable difference between 60 Hz and 120 Hz displays. I have read several reviews in which the reviewer said that there is no noticeable difference unless motion interpolation is usedBecause the blur is produced on the human retina and not the display, sensitivity to the phenomenon can vary widely.

At least LCD's don't have SSE. :D

what is SSE?

It's even been documented in some LCD manufacturer's publications. Why do you think they have so much effort behind 120hz?Proper display of 24 fps video? Motion interpolation to create smoother video? It just seems to me that their are other reasons for 120 Hz LCD and that the "hold type" nature of LCD is only one of them.


And Plasma is not a sample and hold display.Just to mention this but I already knew that which is why I wonder if Plasma companies might be fanning the flames of this issue.


DLP and Plasma are PWM displays.Okay, but are you sure that means DLP is not a "hold type" display? I have read several articles that have said it is a "hold type" display including this article (http://www.digitalcinemareport.com/issuesin3ddisplay.html) from a person who said many of the things you have said in this thread. I would mention though that Real D Cinema (http://en.wikipedia.org/wiki/Real_D_Cinema) is technically 72 Hz per eye and keep that in mind since he tends to say 144 fps when he refers to it.


I have no clue how LCOS works and don’t know much about microdisplays either so I can’t comment on those.I have read (http://europe.elecdesign.com/Articles/ArticleID/16743/16743.html) that LCD, DLP, and LCOS are "hold type" displays.


Because the blur is produced on the human retina and not the display, sensitivity to the phenomenon can vary widely.Possible, but when I start reading professional reviews that state that the reviewer can't see any difference between 120 Hz and 60 Hz it makes me wonder how many people are really affected by this issue.

If you compare the hold time versus display technology you get a pretty clear indication of the effect.

CRT – 1-2ms hold time
Plasma – 4-6ms hold time
LCD (120Hz) – 8.3ms hold time
OLED (120Hz) - 8.3ms hold time
LCD (60Hz) – 16.7ms hold time

It is no coincidence that CRT is considered the reference for motion handling of video as it has short response times and short hold times (the ideal combination)


Yes, for now, CRT is still the reference standard, both for response time and for depth of blackness. Unfortunately, a 50-inch direct-view CRT HDTV would be very expensive and extremely large and heavy. Compromises must be made.

BTW, what a lot of people see as "motion blur" is actually in the source, such as in a hockey game when the camera pans back and forth to follow the action up and down the ice. The background blurs on a CRT exactly the same as it does on an LCD. Also, the "blocking" or "pixelation" is caused by compression artifacts in the source, not by the display.

Proper display of 24 fps video? Motion interpolation to create smoother video? It just seems to me that their are other reasons for 120 Hz LCD and that the "hold type" nature of LCD is only one of them.


Only a subset of the 120hz displays do 5:5. When you say motion interpolation to create smoother video I view that as motion interpolation to smooth video and avoid motion blur.

Yes, for now, CRT is still the reference standard, both for response time and for depth of blackness. Unfortunately, a 50-inch direct-view CRT HDTV would be very expensive and extremely large and heavy. Very true, you'd have to build your house around that monster CRT :).

BTW, what a lot of people see as "motion blur" is actually in the source, such as in a hockey game when the camera pans back and forth to follow the action up and down the ice. The background blurs on a CRT exactly the same as it does on an LCD. Also, the "blocking" or "pixelation" is caused by compression artifacts in the source, not by the display.Yes, this was explained several times in this thread already. My posts are only about motion blur induced by the display, and with current technology hold-type blur is the main culprit. And since this blur is additive to any in the signal it is still very important to reduce or eliminate it.

Cheers

Can someone explain why LCD's have to drop resolution significantly during movement? It doesn't make sense to me...we dont look at a single still picture there is always movement going on in film thats the whole point of film. So if an LCD can only display 600 lines of resolution during "motion" wouldn't that mean its effective resolution is only around 600p?

Proper display of 24 fps video? Motion interpolation to create smoother video? It just seems to me that their are other reasons for 120 Hz LCD and that the "hold type" nature of LCD is only one of them.I'd say that is true. But I'd also point out that the primary purpose 120Hz refresh LCDs were developed was to reduce "hold-time". Judder improvement is a secondary issue and most 1st generation 120Hz LCDs don't even address Judder at all.

Just to mention this but I already knew that which is why I wonder if Plasma companies might be fanning the flames of this issue. LCD manufacturers were researching 'hold-type' remedies back when it was the least of LCDs problems (viewing angle, light leakage, response time). And as of today there are prototype LCDs with hold-times and MPRT that well surpass Plasma performance.


Okay, but are you sure that means DLP is not a "hold type" display? I have read several articles that have said it is a "hold type" display including Quite possibly as I don't know much about DLP. I assumed it was not hold-type because it used PWM like plasma?

Possible, but when I start reading professional reviews that state that the reviewer can't see any difference between 120 Hz and 60 Hz it makes me wonder how many people are really affected by this issue.I can only go by what I read on the subject. One persons review or one persons thoughts on the matter will not change my opinion based on literally hundreds of scientific articles I've read on the subject. Again, temporal artifact sensitivity varies WIDELY from person to person. And for the record I personally have very little sensitivity to image blur. The only display induced blur I can percieve is on a 60Hz LCD, and only when it is beside a CRT or Plasma for reference.

reviews that state that the reviewer can't see any difference between 120 Hz and 60 Hz it makes me wonder how many people are really affected by this issue.

Check out first hand reports in any of the 120hz tv threads. This posted yesterday:


On the right side of the screen with 120Hz "Off", I could see a big difference in motion blur in all the tests. On the left where the effect is "On", honestly, the car looks ALMOST as clear as the static image. But trust me there was a big difference, not just a difference that I want there to be. A BIG difference.
http://www.avsforum.com/avs-vb/showthread.php?t=852139

So if an LCD can only display 600 lines of resolution during "motion" wouldn't that mean its effective resolution is only around 600p?

Yeah....don't say it too loud. It's a touchy subject with the LCD community these days to know that 720p plasmas have higher resolution during motion. :o

Only a subset of the 120hz displays do 5:5.True, and it is difficult to figure out exactly which 120 Hz displays do 5:5 with 24 fps video.


When you say motion interpolation to create smoother video I view that as motion interpolation to smooth video and avoid motion blur.Using motion interpolation with low frame rate video would help create smoother motion with any type of display. Of course the issue of artifacts caused by motion interpolation does make it a double edged sword.


Can someone explain why LCD's have to drop resolution significantly during movement? It doesn't make sense to me...we dont look at a single still picture there is always movement going on in film thats the whole point of film. So if an LCD can only display 600 lines of resolution during "motion" wouldn't that mean its effective resolution is only around 600p?Well if we take the Plasma companies test on motion resolution at face value, and I don't really think that would be a good idea, the drop with LCD is from 1080 to 600 while the drop with Plasma is 1080 to 880. The test itself was subjective with the people that Gary asked supposedly giving answers +/- 25 compared to his answer. As for how much the test results are related to noticeable motion resolution that is a question I would love to see objectivity answered since I have the hunch that it was made more to show the largest possible difference between LCD and Plasma than for any objective analysis of motion resolution.


I can only go by what I read on the subject. One persons review or one persons thoughts on the matter will not change my opinion based on literally hundreds of scientific articles I've read on the subject.I have never read any scientific articles about "hold type" displays (and certainly not "hundreds" of them ;)) but it seems to me that when most professional reviewers see no difference the next logical question would be how many people would be fine with 60 Hz displays? How many people would be fine with 120 Hz displays?


Check out first hand reports in any of the 120hz tv threads. This posted yesterday:


http://www.avsforum.com/avs-vb/showthread.php?t=852139Note that I was referring to the difference between 60 Hz and 120 Hz excluding the use of motion interpolation and I mentioned that in an earlier post (http://www.avsforum.com/avs-vb/showpost.php?p=11975761&postcount=155).


Yeah....don't say it too loud. It's a touchy subject with the LCD community these days to know that 720p plasmas have higher resolution during motion.Only if you take at face value the results from a motion test made by Plasma companies. Also considering the scaling that is needed for 1080i/p and 720p video signals for them be shown on 1366x768 and 1024x768 Plasma displays I am highly skeptical of the supposed claim (http://www.avsforum.com/avs-vb/showpost.php?p=11940011&postcount=111) that they lose no motion resolution. In fact can anyone name me an independent reviewer (not a Plasma company) that has even used their motion test on a 768p Plasma display?

Only if you take at face value the results from a motion test made by Plasma companies. Also considering the scaling that is needed for 1080i/p and 720p video signals for them be shown on 1366x768 and 1024x768 Plasma displays I am highly skeptical of the supposed claim (http://www.avsforum.com/avs-vb/showpost.php?p=11940011&postcount=111) that they lose no motion resolution. In fact can anyone name me an independent reviewer (not a Plasma company) that has even used their motion test on a 768p Plasma display?

Oh come on now.
All you need is good eyesight to see this effect on LCD's.
It's interesting that the 81 series Samsung, the most plasma like LCD out there, seems to eliminate this problem and has been measured to show plasma like motion performance.

Oh come on now.
All you need is good eyesight to see this effect on LCD's.I was merely pointing out that scaling of the video signals should affect motion resolution on 1366x768 and 1024x768 displays. Also when did I ever say that motion resolution was not an issue with most current LCDs? All I said is that I am skeptical of a motion resolution test that was made by Plasma companies and that is not accessible to the general public.


It's interesting that the 81 series Samsung, the most plasma like LCD out there, seems to eliminate this problem and has been measured to show plasma like motion performance.Well from the sounds of it what the 81 Samsung series does to increase motion resolution on the FPD test is to scan the LEDs top to bottom (http://hdguru.com/hd-guru-exclusive-first-revew-of-samsungs-ln-t4081f-led-driven-lcd-102907/201/). The fact that such a method affects the FPD test but that motion interpolation on 120 Hz displays does not is to say the least a bit odd.

I was merely pointing out that scaling of the video signals should affect motion resolution on 1366x768 and 1024x768 displays. Also when did I ever say that motion resolution was not an issue with most current LCDs? All I said is that I am skeptical of a motion resolution test that was made by Plasma companies and that is not accessible to the general public.



Ah, I see.
I don't see it as a problem both from my own plasma viewing and just going by the amount of data being processed. It's much higher with 1080p.

I was merely pointing out that scaling of the video signals should affect motion resolution on 1366x768 and 1024x768 displays. Also when did I ever say that motion resolution was not an issue with most current LCDs? All I said is that I am skeptical of a motion resolution test that was made by Plasma companies and that is not accessible to the general public.


Well from the sounds of it what the 81 Samsung series does to increase motion resolution on the FPD test is to scan the LEDs top to bottom (http://hdguru.com/hd-guru-exclusive-first-revew-of-samsungs-ln-t4081f-led-driven-lcd-102907/201/). The fact that such a method affects the FPD test but that motion interpolation on 120 Hz displays does not is to say the least a bit odd.

Mr. HDGURU does it. But he also has been called a Plasma fanboy. The 81 LCD has gotten high praise from him. I agree. :)

Last week Panasonic placed a full page add in our local paper comparing motion on LCD and Plasma at both 720p and 1080p. The graph showed the Plasma 720p TVs did not loose any resolution, while the 1080p plasms dropped below 900 lines. This I assume is because 720p is below the 880 line threadhold of plasma motion. For LCD, the 720p & 1080p TVs dropped to a similar 600 lines of motion.I'm guessing that graph is the same one in this pdf report on their site:

ftp://ftp.panasonic.com/pub/Panasonic/consumer_electronics/whitepapers/Future_Looks_Bright_for_Plasma_TVs.pdf

It's mostly about the higher lumen advances in plasma tech but if you look at pages 17-18 it addresses the moving picture resolution and shows the graph of three of their sets, one being LCD. It also mentions the effect the 120hz has on the LCD moving resolution.

I simply believe most people, including many "professional reviewers", simply do not understand that there is a difference between liquid crystal "response time" issues and moving picture resolution issues. It's two completely different things. I own a sony 40" V2500 1080p panel and see no response time problems at all. I am however bothered by the sample & hold issue under certain viewing. I've always called it "full field blur" because that is exactly what happens. Hocky games are the worst as the camera pans back and forth stopping at either end. Very obvious.

BTW, what a lot of people see as "motion blur" is actually in the source, such as in a hockey game when the camera pans back and forth to follow the action up and down the ice. The background blurs on a CRT exactly the same as it does on an LCD.Not true. I own a 36" sony 1080i CRT and have set it up right along side my LCD to test this very thing. While there is most definitely camera (source) induced blur when watching hockey, the sample & hold blur on the LCD simply adds to that making it even worse. My eyes may be susceptible to this but it's not hard to see at all for me. The LCD panel is great for watching many things. But watching hocky isn't one of them.


ron

I was merely pointing out that scaling of the video signals should affect motion resolution on 1366x768 and 1024x768 displays. Also when did I ever say that motion resolution was not an issue with most current LCDs? All I said is that I am skeptical of a motion resolution test that was made by Plasma companies and that is not accessible to the general public. I understand what you are saying but why then are LCD manufacturers using the same test to show motion performance? Correct me if I'm wrong but MPRT and the Plasma method from APDC use the same tracking camera sync'd with a moving test pattern.

Well from the sounds of it what the 81 Samsung series does to increase motion resolution on the FPD test is to scan the LEDs top to bottom (http://hdguru.com/hd-guru-exclusive-first-revew-of-samsungs-ln-t4081f-led-driven-lcd-102907/201/). The fact that such a method affects the FPD test but that motion interpolation on 120 Hz displays does not is to say the least a bit odd.That's because motion interpolation reduces hold time by 50% (if done right) while backlight scanning can reduce hold time much more than 50% depending on the duty cycle per LED row per frame.


Edit: Actually, I just thought about it and one reason that 120Hz may not look better than 60Hz for motion is that motion interpolation and 5:5 pulldown may not be done together. For instance only motion interpolation is done on 60hz signals and 5:5 is only done on 24Hz signals (never 5:5 and motion interpolation).

This would mean that 24Hz signals will look identical (or nearly) at 60hz and 120hz as the hold times are similar. Only with full frame motion interpolation can 5:5 fully reduce hold time by 50%. This also means that 80% of the frames are calculated :yikes: And even worse, backlight scanning would be useless as the hold time is mostly set by the frame repetition.

I understand what you are saying but why then are LCD manufacturers using the same test to show motion performance? Correct me if I'm wrong but MPRT and the Plasma method from APDC use the same tracking camera sync'd with a moving test pattern.Can you provide evidence for that? After all the FPD benchmark test was made by Plasma companies so how do you know that the test pattern on it is the same one that is used by the industry?


That's because motion interpolation reduces hold time by 50% (if done right) while backlight scanning can reduce hold time much more than 50% depending on the duty cycle per LED row per frame.An interesting theory but did the 120 Hz displays that Gary Merson test have any difference in score depending on whether motion interpolation was used? This is one of the reasons I am suspicious about the test since by limiting its access to pro-Plasma reviewers they prevent the test itself from being tested.


Edit: Actually, I just thought about it and one reason that 120Hz may not look better than 60Hz for motion is that motion interpolation and 5:5 pulldown may not be done together.Possible, but without being able to test the FPD benchmark test we won't really know what the answer might be.

This gives us a lot to look forward to when far more perfected displays are introduced over the next 10 years or so. Resolution is just one of many factors that need to be worked on; I can't wait.

Richard

I missed out on getting a nov copy of Home Theater, Bookstore already carrying Dec. Does anyone know if this arcticle with ratings will show up on Home Theater WEB site or HDGURU?
Also trying toi find out if I can get a back copy from Home Theater, bit didn't see anything oin their WEB site and have heard back on email request.
Any info on possible how to get a copy of this article would be appreciated.

I am however bothered by the sample & hold issue under certain viewing. I've always called it "full field blur" because that is exactly what happens. Hocky games are the worst as the camera pans back and forth stopping at either end. Very obvious.

Not true. I own a 36" sony 1080i CRT and have set it up right along side my LCD to test this very thing. While there is most definitely camera (source) induced blur when watching hockey, the sample & hold blur on the LCD simply adds to that making it even worse. My eyes may be susceptible to this but it's not hard to see at all for me. The LCD panel is great for watching many things. But watching hocky isn't one of them.


I feel this same way about video games. Like racing games, I make a minor steering correction and all of the detail in the foreground (like 65% of the screen) goes into this full field blur then snaps back. It drives me crazy.

FYI

The back issue office can be contacted by calling 1-212-915-4160

Motion Resolution: Note that this is a 60Hz LCD. 120Hz would be between the PDP and LCD and 120Hz with BL scanning would be better than the PDP.

http://i222.photobucket.com/albums/dd126/xrox/sampleandhold.jpg

The majority of 120 Hz (LCD) TV's use the faster frame rate to better control, using timing, the pixel change delays. Generally it takes a shorter time to get from white to black or from black to white compared with from gray to gray. Using 120 Hz refresh, a request all the way to black can be done and then the correct gray value requested on the next 1/120'th second interval. Depending on the exact video content, an overshoot may or may not occur, but motion blur getting back up, or getting from dark to light in any situation, is less of a problem.

Resolution pumping, where a large more or less rectangular section of stationary matter goes blurry because something nearby moved, not to mention when the whole screen going blurry because something moved, to me constitutes a "Fail" for the TV.

This is due to inferior de-interlacing.

The current "good" deinterlacing is "pixel by pixel motion adaptive". I am guessing that less than 5% of TV's made in the last half of 2007 have this for HDTV (1080i). In it, all moving subject matter still relies on just the current field, using interpolation so as to have at most half the resolution compared with true woven stationary subject matter. IMHO the fact that processing moving subject matter using just the current field, and interpolation, is less noticeable, not more noticeable, considering the other factors contributing to blurriness notably camera shutter speed related blur at filming time.

The next step up in de-interlacing sophistication is called "motion compensation". The processing tries to guess where moving subject matter went to, in order that material from a non-matching odd field can be woven into an even field and made to match, etc. No affordable HDTV examples exist yet.

Video hints: http://members.aol.com/ajaynejr/viddoubl.htm

Hi Alan,

I would be very interested in your (or anyone else's) opinion of the system my new TV (Philips 47PFL9532 aka 47PFL9732 in the US) uses for smoothing motion.

I am happy with it (so probably should not be examining it too closely:D) but am not in a position to make meaningful comparisons with other manufacturers products.

It supposedly has a 3ms response time (4ms for the US model) and does this http://www.research.philips.com/newscenter/dossier/naturalmotion/judder-free.html

To me, the motion looks as good as or better than the 42" SD plasma it replaced.

Hi Alan,

I would be very interested in your (or anyone else's) opinion of the system my new TV (Philips 47PFL9532 aka 47PFL9732 in the US) uses for smoothing motion.

FYI - That Philips LCD is reviewed in the latest (December) issue of Home Theater Magazine . Another AVS member asked about it recently, and I posted much of the review on another thread - [b]Posts #121 & #123 HERE (http://www.avsforum.com/avs-vb/showthread.php?p=12149707#post12149707) <link

The majority of 120 Hz (LCD) TV's use the faster frame rate to better control, using timing, the pixel change delays. Generally it takes a shorter time to get from white to black or from black to white compared with from gray to gray. Using 120 Hz refresh, a request all the way to black can be done and then the correct gray value requested on the next 1/120'th second interval. Depending on the exact video content, an overshoot may or may not occur, but motion blur getting back up, or getting from dark to light in any situation, is less of a problem.This makes no sense to me? Maybe I'm reading this wrong but it seems that you are relating pixel response and frame rate?

In no way does increasing the frame rate change pixel response or even the perception of pixel response. It is a method to reduce the frame period and improve hold-type blurring and also to give the capability to reduce judder.

There are many resources on this forum (this thread is one) and the web that explain this fully.

Cheers