OK, I have seen this mentioned, but thought I would create a thread to discuss the AV science behind it.

Everyone here seesm to be really picky about PQ and AQ, and it appears that when there is motion on the screen, those with LCD displays, are losing 44.45% of their resolution anyway. Heck, why didn't I just stick with SD?! What good are HD DVD and BR when you lose 1/2 the picture?

I am not sure if my 1080p LCD projector falls in that catagory, but I do have a 50" plasma...to bad its only 768p.

"This year the HD Guru discovers that there is a significant loss in resolution when there is motion -- you know like when the camera pans or a player runs down the field. What he discovered is very interesting, but not surprising; Plasmas fared the best with a loss of 18.52%, microdisplays were next at 27.78% and the LCDs were last at 44.45%. "

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

I think the effect is not that big in reality. Im sure their testfilms showed this. But films are shot wit 1/48 shutterspeed. Wich create motionblur.

The motionblur takes away detail from everything in motion.

LCD's #1 problem is blur due to retina retention. That's why the new LCDs are combining 120Hz and motion interpolation. Next gen will strobe LED BLUs to include "flicker" like CRTs and plasma to reduce blur.

I'm having a hard time using this to recommend SD DVDs. Anyway, just say no to HDM pricing. (I'm a self proclaimed DVD fanboy.)

LCD's #1 problem is blur due to retina retention. That's why the new LCDs are combining 120Hz and motion interpolation. Next gen will strobe LED BLUs to include "flicker" like CRTs and plasma to reduce blur.

I'm having a hard time using this to recommend SD DVDs. Anyway, just say no to HDM pricing. (I'm a self proclaimed DVD fanboy.)

Are you talking TV or projector? I think when you going on upwards to a 80" to even 100" screen, most "DVD fanboys" start to realize DVD upconverted just ain't the same deal.

I think the effect is not that big in reality. Im sure their testfilms showed this. But films are shot wit 1/48 shutterspeed. Wich create motionblur.

The motionblur takes away detail from everything in motion.

I can see that being the 18% that plasma loses, but then LCD loses 26% more than that.

Um...because I want the disks to have the best possible PQ & AQ so that when I upgrade my TV to a full 1080P (I currently have an earlier SXRD that only accepts a 1080i input) there is something to look forward to?

I plan to continue to upgrade my HT, and I sure as heck don't want my "next gen" disks to be the weak link in the chain at this point.

I can see that being the 18% that plasma loses, but then LCD loses 26% more than that.

I think the test were done without motionblur so the plasma loss is still 18%

But most action scenes lose so muc resolution anyway that it shouldnt matter that much.

Look at the image, the camera tracks the policeofficer when he moves so hes very static but everything else is in movement.

http://medlem.spray.se/x17pic/bilder/Image17.jpg

maybe this logic doesnt work, but i still think it looks better b/c your losing that 44% of picture potential on SD movies too, making them look that much crappier, but since HD movies have a much higher picture quality it still looks worlds better than SD.

It's not the disks, it's the displays. And on my 1080P Toshiba 65" they both l;ook excellent. If it's loosing details. I don't really care. They look fantastic and certainly will not be going back to dvd, ever!

maybe this logic doesnt work, but i still think it looks better b/c your losing that 44% of picture potential on SD movies too

Bulls Eye!!!

Well you dont really lose so much because the source also lose res.

Man, next thing you know people are going to be going into the debug consoles on their TVs and setting com_maxfps 125, overclocking, and switching to vertex lighting mode just for pure FPS!

The upcoming published measurement list for various display models (OP) covers vertical resolution loss that's basic to most interlaced or progressive display. Similar drops in vertical resolution, percentage-wise, have always applied to 480i to other formats. (The exception might be progressive non-sampled images originating only from computers, not involving image sampling as with most video signals.)

For both 1080i and 720p broadcast HD, the ATSC approval committee anticipated such resolution losses (both horizontal and vertical) before their mid-'90s testing (see 2.1 Resolution (http://www.atsc.org/news_information/papers/1995_acats/tsreport.pdf) intro paragraph, target and measured resolutions in 2.2 and 2.3, or this simplified table (http://archive.avsforum.com/avs-vb/showthread.php?postid=326565#post326565) of results.)

Commenting on Merson's welcome measurements elsewhere (http://www.avsforum.com/avs-vb/showpost.php?p=11723222&postcount=2), I hoped for similar measurements of program material instead of test patterns (requiring spectrum analysis for horizontal rez). Patterns often aren't sampled signals, so while your setup might show perfect 1920X1080 resolution from test patterns, a movie disc might only deliver ~1000X800 effective resolution (http://archive2.avsforum.com/avs-vb/showthread.php?p=5667245&&#post5667245) (equivalent maximum resolvable detail) from certain scenes, and even less vertical resolution when movement is involved. -- John

Don't our eyes perceive much less detail when dealing with fast action scenes? or perhaps just the hardwired part of the brain cannot keep up with the amount of information of fast motion?

Just wondering.

Don't our eyes perceive much less detail when dealing with fast action scenes? or perhaps just the hardwired part of the brain cannot keep up with the amount of information of fast motion?

Just wondering.

Not in this forum. People here claim they can see individual pixels from ten feet away. Also, they can hear a flea fart at 1000 yards.

i am so glad i picked plasma over lcd for my friend's new display.

where does dlp and lcos rear proj tv's fit into this?

i am so glad i picked plasma over lcd for my friend's new display.

where does dlp and lcos rear proj tv's fit into this?

is there any plasma with matte screen now? I can't stand those glossy screens.

Everyone here seesm to be really picky about PQ and AQ,

Not really.
Bandwidth is an easy target. 30Mbps is 50% more than 20Mbps.

is there any plasma with matte screen now? I can't stand those glossy screens.

Plasma screens are all glass. That is why they are glossy. Some have anti-glare coatings that help.

Not really.
Bandwidth is an easy target. 30Mbps is 50% more than 20Mbps.

Its just funny to see people talking about bandwidth and how they want the BEST PQ and find out they are using an LCD TV.

Are you talking TV or projector? I think when you going on upwards to a 80" to even 100" screen, most "DVD fanboys" start to realize DVD upconverted just ain't the same deal.

Hell, at 42" it ain't the same deal :eek:

Plasma screens are all glass. That is why they are glossy. Some have anti-glare coatings that help.

that shouldn't be an issue. I've mounted various wall-sized pictures in matt-glass frames. How about those anti-glare plasmas? are they still glossy but just not as much?

Having trouble wrapping my mind around the mechanism that creates reduced resolution, motion or not... anyone 'splain this one? The article seems to blame the de-interlacing process, but what about de-interlacing would drop resolution?

OK, I have seen this mentioned, but thought I would create a thread to discuss the AV science behind it.

Everyone here seesm to be really picky about PQ and AQ, and it appears that when there is motion on the screen, those with LCD displays, are losing 44.45% of their resolution anyway. Heck, why didn't I just stick with SD?! What good are HD DVD and BR when you lose 1/2 the picture?

I am not sure if my 1080p LCD projector falls in that catagory, but I do have a 50" plasma...to bad its only 768p.

"This year the HD Guru discovers that there is a significant loss in resolution when there is motion -- you know like when the camera pans or a player runs down the field. What he discovered is very interesting, but not surprising; Plasmas fared the best with a loss of 18.52%, microdisplays were next at 27.78% and the LCDs were last at 44.45%. "

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

Would you like a roughly 20% loss of 720×480 (sd) or 20% of 1920x1080? I'm sure you would want neither but stick with SD? Get real.

This was one of the biggest reasons I chose Plasma over LCD. I jusy couldn't get past the motion blur.

Having trouble wrapping my mind around the mechanism that creates reduced resolution, motion or not... anyone 'splain this one? The article seems to blame the de-interlacing process, but what about de-interlacing would drop resolution?
With deinterlacing, as Merson outlines in his earlier tests/articles (http://archive2.avsforum.com/avs-vb/showthread.php?p=9178202#post9178202), it's using TV fields (half-frames) instead of frames to create progressive frames, (so-called bobbing). Overall, it's the vertical sampling process, with accurate sampling necessarily decreasing as vertical motion increases during capture and playback. Resolution drop is proportional to the rate of motion. There's the Kell factor (~0.7 X line count) that's outlined, along with mention of additional vertical filtering needed for interlaced capture to avoid line 'twitter' of fine details, in Greg Rogers' text associated with his resolution tables (http://www.cybertheater.com/Tech_Reports/HD_Projectors/hd_projector.html) (~1/2 down). Kell reductions apply to interlace capture as well as progressive capture displayed as video. Reportedly, although they're still way too costly for consumer use, motion-compensated deinterlacing (versus today's consumer motion-adaptive deinterlacing), used for pro equipment, greatly reduces vertical resolution loss. Some vertical filtering could be avoided if interlacing wasn't a capture/transmission/viewing factor (CRTs). Lots of other associated factors--enough for texts on both horiz. and vert. resolution. -- John

i am so glad i picked plasma over lcd for my friend's new display.

where does dlp and lcos rear proj tv's fit into this?

Hi,

DLP and LCOS are microdisplay rear projection HDTVs and were apparently bundled in that category's overall results. I'm guessing that, signal processing issues aside, Plasma HDTVs have inherently better motion picture response times because they don't rely on a backlight that causes retina retention issues.

I asked Gary that question over at the Are you getting all the resolution you expect (2007)? (http://www.avsforum.com/avs-vb/showpost.php?p=11760643&postcount=5) thread.

Larry

Jumping back a bit to the original question...

When I read Merson's articles (3, now -- although the third one has not been fully-published, yet -- due out in the November issue of Home Theater Magazine), I was dismayed Perhaps for a different reason than some of you.

I'm a cheapskate in some respects.

I started off thinking I simply had to buy an A/V Receiver to enjoy decent TV -- both due to PQ and AQ. But, I wanted to understand the value of buying an external Video Processor -- be it in the form of an A/V Receiver or a stand-alone VP. I started off my thinking wondering how good the internal VPs in all of these sets were. Not very -- according to Merson -- except in certain brands that passed his tests.

Does one gain anything by buying a higher-dollar set (such as a Pioneer) with a decent VP? I think, ONLY if you have no intentions of buying an A/V Receiver or stand-alone VP. Otherwise, you're buying another box that, basically, duplicates and displaces the built-in VP within the set. Yes?

That's where my cheap or impractical side kicks in. I've read postings here where Owners boast about 3 or more components that seem to duplicate several of the core features of the other components, and the TV -- then turn those features OFF so that one takes control. That seems wasteful, to me.

As in "I spent a bucket of money on this HDTV because it processes the picture so well -- but then I hooked up my beautiful Vdeo Processor because, actually, it does a better job." -- Say what?

Does one gain more by ignoring the Merson testing and simply buying an external VP source? Possibly. Unless I misunderstand their operation.

But it begs the questions:
1) Have these external sources ever been tested to see if they really do pass the full stated signal quality on to the Display?
2) Do the deficiencies within the sets that Merson points out DEFEAT outside source quality improvement by not being able to display the full PQ -- no matter what you do?

As asked in the original question, I'm also shocked a bit at how poor the speakers are on many of these $2,000+ sets. Criminy, my old 21" Sony CRT was plenty loud enough for 15 years. Why is my 37" Sharp such a disappointment? What the hell is wrong with these Manufacturers? If you're going to include crappy speakers on a $2,000 TV -- at least make them removable or replaceable. But maybe that's just the cheapskate in me, again.

Not in this forum. People here claim they can see individual pixels from ten feet away. Also, they can hear a flea fart at 1000 yards.I think that I remember Amir saying that the max bit rate of HD-DVD was good enough to have audio allowing one to hear the flea fart at 1000 yards. He said the higher max bit rate of Blu-ray was unnecessary and a waste of band width.:):)

1) Have these external sources ever been tested to see if they really do pass the full stated signal quality on to the Display?
You raise good points about the value of bypassing internal processes. Greg Rogers in September's Widescreen Review reports on the higher-end Crystalio II processors ($4.5 and $5.5K). Sure hope at these prices they outperform built-in processors. Higher-end FPs (likely other displays) sometimes have the same newer processor chips used in similarly price external processors, but the external devices have more bells and whistles.
2) Do the deficiencies within the sets that Merson points out DEFEAT outside source quality improvement by not being able to display the full PQ -- no matter what you do?
Need testing to confirm, but suspect bypassing set processors and using a good external one would eliminate major problems such as bobbing TV fields (half-frames) into frames instead of using two fields--often sampling other fields--to form each frame. Lots of complexities, though, IMO: If your costly external processor can spit out 24p or other even multiples such as 96 fps, then you'd want to ensure your display can handle that rate. All the permutations of options almost demands buying your own special test DVDs and checking features. -- John

Would you like a roughly 20% loss of 720×480 (sd) or 20% of 1920x1080? I'm sure you would want neither but stick with SD? Get real.

You mean a 44% loss of 1920x1280 or a 0% loss with a 720x480 CRT (SD).
Is there motion blur with a CRT?

With deinterlacing, as Merson outlines in his earlier tests/articles (http://archive2.avsforum.com/avs-vb/showthread.php?p=9178202#post9178202), it's using TV fields (half-frames) instead of frames to create progressive frames, (so-called bobbing). Overall, it's the vertical sampling process, with accurate sampling necessarily decreasing as vertical motion increases during capture and playback. Resolution drop is proportional to the rate of motion. There's the Kell factor (~0.7 X line count) that's outlined, along with mention of additional vertical filtering needed for interlaced capture to avoid line 'twitter' of fine details, in Greg Rogers' text associated with his resolution tables (http://www.cybertheater.com/Tech_Reports/HD_Projectors/hd_projector.html) (~1/2 down). Kell reductions apply to interlace capture as well as progressive capture (film). Reportedly, although they're still way too costly for consumer use, motion-compensated deinterlacing (versus today's consumer motion-adaptive deinterlacing), used for pro equipment, greatly reduces vertical resolution loss. Some vertical filtering could be avoided if interlacing wasn't a capture/transmission/viewing factor (CRTs). Lots of other associated factors--enough for texts on both horiz. and vert. resolution. -- John
Would vertical resolution loss be confined to every other line (only the 'missing' lines which get during the field->frame conversion)? Or is the active field information modified during the conversion to a full frame? Hope I asked that in a technically competent yet clear fashion.

Also, is vertical resolution loss less noticable than horizontal? I seem to recall coming across something in the past that seemed to indicate such.

Finally - why would the TYPE of display matter in the resolution loss? It sounds as if the deinterlacing process is the root of all evil here, why would it be worse on LCD vs. Plasma? Unless it's just coincident that the LCD's user a different type of deinterlacer?

Just trying to load a few technical bullets in the old knowledge clip...

MEC2

Its all about the processing.


My Fujitsu 50" plasma which is 768p looks better then almost every 1080p 50" i have seen.

Differences may be able to be seen when sitting very close, or on displays/screens @ 60" +, but again it largely comes down to processing not extra resolution when comparing 768p to 1080p IMO.

Save yourself the trouble and get yourself a good DLP ;)

Save yourself the trouble and get yourself a good DLP ;)

I spent a ton of $ on a DLP project with a 5x color wheel. I had t sell it because my wife saw rainbows fashing all the time and it gave her a headche!
It did ive me a reason though to switch to 1080p.;)

Would vertical resolution loss be confined to every other line (only the 'missing' lines which get during the field->frame conversion)? Or is the active field information modified during the conversion to a full frame? Hope I asked that in a technically competent yet clear fashion.When vertical filtering is deliberately applied for interlaced delivery, lines for each field (half-frame) are 'smeared' vertically for each frame. As I mentioned, if a purely progressive mode, start to finish, was used, this filtering could be bypassed. Also, again, there's other vertical resolution losses taking place, sometimes with erroneously bobbing fields into frames, for example.

Deinterlacing the prefiltered field lines depends hugely on the effectiveness of the algorithms used, with pixel-by-pixel motion-adaptive techniques the most sophisticated in consumer gear, to using blocks of pixels, to even less sophisticated, resolution-loss techniques. At the highest end, still confined to pro broadcasting gear, are motion-compensation methods, requiring very complex processing to predict motion between TV fields.

Also, is vertical resolution loss less noticable than horizontal? I seem to recall coming across something in the past that seemed to indicate such.IMO, drastic horiz. rez loss is more noticeable. Problem is, there's seldom A-B examples available to spot what's missing. Clearly soft images, achieved with camera filters, might be a director's intent. Noticed the missing horizon rez immediately, for most programming, when I updated my cable converter a few years back from a SA8000HD, limited to 890 lines (http://www.avsforum.com/avs-vb/showthread.php?postid=5175424#post5175424) (per HD picture width!), to a SA8300HD (1290 lines maximum, still considerably short of the 1920 lines a test pattern--but not programming--should provide on a full-rez display).

Finally - why would the TYPE of display matter in the resolution loss? It sounds as if the deinterlacing process is the root of all evil here, why would it be worse on LCD vs. Plasma? Unless it's just coincident that the LCD's user a different type of deinterlacer? That's a surprising test result and not sure why. Cheaper deinterlacing techniques (bobbing TV fields to frames) might account for it, but suspect 540p bobbing crosses display types. Perhaps slower LCD response times could account for it. -- John

I spent a ton of $ on a DLP project with a 5x color wheel. I had t sell it because my wife saw rainbows fashing all the time and it gave her a headche!
It did ive me a reason though to switch to 1080p.;)

+3 (my wife, my daughter and myself)

Save yourself the trouble and get yourself a good DLP ;)

The thing is the article states microdisplays are also subject to it, and since it appears to be a matter of the deinterlacer, a DLP is of no help. I do, matter of fact, have the 71" Samsung DLP from a year ago... not that I've noticed a resolution loss, but who can really know without such testing?

Technical question:

It would seem that deinterlacing can, or should, happen in 2 ways - one, for material that is natively 1080i - that is, each 540 line field is captured every 60th of a second - and another for material that is 30 or less fps - a movie, for instance, but each frame sent in two constituted interlaced fields. In the latter case, one would expect there to be 30 framesps of content per second, not 60 interlaced fieldsps. Make any sense? Since there is no "motion" between the interlaced fields of any 30 fps frame, I would expect construction of such a frame to be less prone to errors (in fact, conceptually error free since motion between motion between frames is not motion between fields) than one that is constituted from frames where each field is from a different sample in time.

Lord, I hope that makes sense...

MEC2

It would seem that deinterlacing can, or should, happen in 2 ways - one, for material that is natively 1080i - that is, each 540 line field is captured every 60th of a second - and another for material that is 30 or less fps - a movie, for instance, but each frame sent in two constituted interlaced fields. In the latter case, one would expect there to be 30 framesps of content per second, not 60 interlaced fieldsps. Make any sense? Since there is no "motion" between the interlaced fields of any 30 fps frame, I would expect construction of such a frame to be less prone to errors (in fact, conceptually error free since motion between motion between frames is not motion between fields) than one that is constituted from frames where each field is from a different sample in time.

Yes, most detailed descriptions of deinterlacing point out that converting interlaced video from TV cameras is different (and much more complex) than from 24p-originated video with complete frames originally. Yet, referring to HD discs, when interlaced output is used for non-progressive CRT display, vertical prefiltering is required. So, using 24p outputs from disc players, with simple fully accurate internal reconstitution of original frames, shouldn't involve vertical resolution loss--providing vertical prefiltering wasn't applied initially. (Yet, believe I recall member Glimmie writing that telecine hardware does perform prefiltering.)

Seems possible that, in part, judder might account for variations of vertical resolution seen on displays: Models with even-multiple display of 24 fps (48, 72, 96, 120 fps) images after inverse 2-3 pulldown wouldn't involve the sometimes subtle judder blurring on display models that instead show 24p frames at 60 fps (with added frame pulldown), which is more suitable for deinterlaced 480/60i and1080/60i sourced from TV cameras capturing at those field rates. -- John