or Connect
AVS › AVS Forum › Display Devices › Flat Panels General and OLED Technology › 24/30/60 fps comparison
New Posts  All Forums:Forum Nav:

24/30/60 fps comparison - Page 3

post #61 of 82
Quote:
Originally Posted by borf View Post

24 fps has problems both at 24 and 60 hz (as they say)
but what if the clips in that article were changed to 60 fps?
i already had these clips (first time shown @ 24, then @ 60fps)
again, the tv should be at default 60hz.



even with camera blur (from being shot at 24fps) the second time around the poker table is quite different.
i noticed james bond sitting there this time. it also felt like a pbs special for a moment.
as always the 120hz / 240hz setting is optional & usually adjustable.


Very, very interesting comparisons... the inner logic in me says that higher frame rate would be superior... but the second version of everything did seem weird to watch... hard to figure out if it's just because we're so used to watching 24fps material... or if the blur actually adds realism to the experience somehow... As someone mentioned, I wonder if I would adjust to 60fps filming and it would actually look better...
post #62 of 82
Quote:
Originally Posted by Nielo TM View Post

On an AM-based display, no

On pulse based displays (or displays with black period), higher cycle will help to reduce/eliminate flicker

am lcd displays with higher refresh rate capabilities typically have lower response times, to that end, am lcd displays flaunting their ability to display high frame rates probably will have better transitions between frames because of this. The usefulness of having a high refresh rate is directly limited by the reponse time.

This is similar to the idea that pulse displays look better at higher frequencies as well, they can redraw and draw images that much faster, but is probably less noticeable.

i think some of the 24p appeal comes from the way films are presented. If we were observers in film scenes, near where the camera would be, we would probably be looking all over, thus having a much wider net viewing area, keeping track of all sorts of things. There would also be very large angles to cover for eyes. In a theater or on a tv the camera guides us to the relevant details, and the screen covers a narrow portion of our vision usually, and importantly most relevant details occur near the center of the screen. Therefore our effective fps were we in the actual film scene would be much lower than it is if we are watching on a screen, and so a very high fps display seems less real.
post #63 of 82
Yes it is true that pixels respond faster at higher refresh rates, but hold-time remains unaffected. And it's virtually impossible to distinguish the difference on low-motion contents.
post #64 of 82
Quote:
Originally Posted by Nielo TM View Post

Yes it is true that pixels respond faster at higher refresh rates,

There're two things we're talking about here. One is that if we have a pixel and are having a high refresh rate then the pixel is being informed of what colour it should turn into at a faster pace, which means it can start making that transition earlier in time,*

The other is the fundamental response time limit for the pixel. No matter how fast the refresh rate is, this is the fastest the lcd can transition to another colour. For example computer LCDs sometimes advertise GTG or green to some other colour, sometimes black, and then back to green time as a certain number of milliseconds. Depending on how this number is calculated, if our refresh rate is 240Hz but I have a 8ms response time, then our real refresh rate is limited to 1000/8 because 1000/240 is < 8. Of course in reality it is a bit more complicated than this, but basically this is why higher refresh rate LCDs should typically* have low response times to match.

Finally bring this back to the topic, the LCDs capable of higher refresh rates are thus more likely to have reduced ghosting, although as you say it is not noticeable except for certain cases of fast-moving content. [For me this has been a real headache for computer monitors and gaming, where CRTs seem so much more fluid.]

* and also if the pixel is being informed of what colour to change to more quickly it could change to a more appropriate colour by averaging out colours if the input signal, if said signal is varying at a faster pace. This is why 240Hz will even help on LCDs with high response times.
post #65 of 82
We have to consider three factors actually, which are RT, hold-time and the source.

If one was to input 24p to a 240Hz LCD, the pixels are updated every 41.7ms. And at 60p, the pixels are updated every 16.7ms because in simple AM mode, pixels do not reset after completing each cycle, which was why MCFI was deployed.

So even an LCD with true 16ms RT across the entire gray-scale is sufficient enough to provide smooth motion.

But atm, the primary factor is hold-time. There no reason to reduce RT without reducing hold-time first and this isn't an issue with low-motion contents as the elevated blur helps with frame transitions.
post #66 of 82
I am currently basing the hold-time directly on the lcd refresh rate. In other words, assuming infinitely low response time, the would hold a colour for 240th of a second and then flip colours. Please let me know if I should include other factors.

This being the case, I disagree, I think that hold-time at 240hz is sufficiently low and that response times are lagging behind (every colour transition cannot be made within 4-5ms, some can, obviously, low motion is no problem), though as was said a lower hold-time could help even with crappy response times (averaging).

Why do the pixels have to change more quickly than 41.67ms for 24p? Simple AM mode can only be deployed for certain framerate factors?

As you say the elevated blur is effective-enough interpolation for low-motion frame transitions, if it weren't considered so then we'd be moving into the territory that others have mentioned of varying framerate or different areas of the frame refreshing at different rates because there's little reason not to shoot at a higher framerate if we demand better interpolation accross the board for low and high motion. We'd seemingly be exiting the discussion of 24p content at this point.

Any desireable blur can be simulated on high refresh rate lcds. A more pleasing blur may be able to be created than with the more slowly refreshing lcds.
post #67 of 82
Quote:
Originally Posted by achalas View Post

I am currently basing the hold-time directly on the lcd refresh rate. In other words, assuming infinitely low response time, the would hold a colour for 240th of a second and then flip colours. Please let me know if I should include other factors.

On AM-LCDs, hold-time is governed by the source and MCFI. If the source is 60p, then regardless of the operating frequency, the pixels only alter grayscale value every 16.7ms. If high-level MCFI is enabled, then 60p can be converted to 120p+, in which case, the pixels must be updated at faster rate due to increase in frames with unique information.

Quote:
Originally Posted by achalas View Post

This being the case, I disagree, I think that hold-time at 240hz is sufficiently low and that response times are lagging behind (every colour transition cannot be made within 4-5ms, some can, obviously, low motion is no problem), though as was said a lower hold-time could help even with crappy response times (averaging).

Why can't the pixel change below 4-5ms?

Over-drive functions by switching to either white or black first then back back-stepping to the required grayscale value.

For an example, the middle value of a 8bit gray-scale is 128. So a pixel don't have to exceed that value. If it has to switch from 16 to 230, all it has to do with switch to 0, then to 255 and to 230, which is far simpler then switching from 16 to 230 in liner form.

With ever evolving LCD, I wouldn't say it's not possible to switch to any grayscale value within 4ms.

Quote:
Originally Posted by achalas View Post

Why do the pixels have to change more quickly than 41.67ms for 24p? Simple AM mode can only be deployed for certain framerate factors?

I don't follow

Quote:
Originally Posted by achalas View Post

As you say the elevated blur is effective-enough interpolation for low-motion frame transitions, if it weren't considered so then we'd be moving into the territory that others have mentioned of varying framerate or different areas of the frame refreshing at different rates because there's little reason not to shoot at a higher framerate if we demand better interpolation accross the board for low and high motion. We'd seemingly be exiting the discussion of 24p content at this point.

Any desireable blur can be simulated on high refresh rate lcds. A more pleasing blur may be able to be created than with the more slowly refreshing lcds.

I would prefer if we move to 48p + blur as it seems to produce the most natural motion and it's also cost effective compared to 60p or higher. It's also highly beneficial to S3D as 24p just doesn't have enough frame to induce fluidic motion.
post #68 of 82
Quote:
Originally Posted by IanD View Post

Why unnatural? Human vision doesn't operate on discrete frames, so surely the higher the framerate the more natural and closer to human vision it should appear.

Quite correct. However, frame-rate higher than 48p offers no true benefit as we humans are subtable to motion blur. So filming in 300p with no blur will appear artificial and unnatural and 300p with motion blur will be identical to 48p with motion blur.

Note: I used 300p as 60p is also too slow to induce truly continues movement. We can actually detect individual frame, which is why blur is added.

Quote:
Originally Posted by IanD View Post

Or are you referring to 60i/p as appearing unnatural compared to 24fps, because we are so used to 24fps as representing the natural world?

That's also another factor.


Quote:
Originally Posted by IanD View Post

I've just upgraded to an LCD after my CRT TV died and I'm appalled at 24fps judder on Bluray movies: it doesn't look natural to me at all (natural as in how I would see the real world). Those 60fps videos look much more natural to me. Even cinemas don't seem to have the same sort of judder I see on LCD at 24fps.

That's due your set.

Some don't seem to process 24p correctly. Thus far, only Sharp LCDs yield smooth 24p motion (akin to Digital Cinemas). Maybe they are suing some sort of blur tech, because it certainly wasn't MCFI.

Quote:
Originally Posted by IanD View Post

Consequently, I would likely find frame interpolation more pleasing to watch if only it didn't come with its own artifacts, due to the infant stages of the technology.

Yes, but there's also another factor, shutter speed. Depending on the content and frame-rate, the moving scenes are captured in certain shutter speed to maintain smooth motion without obscuring too much detail. The kind of blur induced by low shutter speed can be replicated during post-production.

Problem is, when 24p is converted to high-motion, not only motion appears smooth but also at times appear artificial, which is due to the lack of sufficient blur. So it would be helpful if manufacturing start implementing blur control.

Quote:
Originally Posted by IanD View Post

Which leads me to another issue that baffles: DVD and Bluray are based on motion vector compression, so it would theoretically be relatively simple to interpolate those vectors in the player and much more efficient than the display having to work it out from scratch; yet I have not seen any manufacturer offer player-based interpolation. Admittedly this approach would require that displays accept 48p/72p/96p/120p, but I expect this would be easier than performing blind interpolation on an incoming signal. I wonder why the industry has chosen not to adopt this approach.

Currently, such process are limited to WinDVD (2004+) and PowerDVD (2009+). However, both still suffer from the same artifices present in TVs.

Having said that, Panasonic seems to prefect the algorithm to satisfactory level. I could see any obvious image artifacts while I was in Muchien (Panasonic Conference 2010)
post #69 of 82
Quote:
Originally Posted by Nielo TM View Post

On AM-LCDs, hold-time is governed by the source and MCFI. If the source is 60p, then regardless of the operating frequency, the pixels only alter grayscale value every 16.7ms. If high-level MCFI is enabled, then 60p can be converted to 120p+, in which case, the pixels must be updated at faster rate due to increase in frames with unique information.

Right, whatever can be done to the source, motion interpolation and processing and whatnot, but the maximum would be 4.167ms (240hz display), assuming RT is a nonfactor. Even without any interpolation or processing, of course the signal could still be updated every 4.167ms, this would reduce blurring but increase 'motion jumping' (i forget what this is called)
Quote:
Originally Posted by Nielo TM View Post

Why can't the pixel change below 4-5ms?

Oh most definitely it can, eventually. I'm saying with the current state in technology it's not the refresh rate that is lagging behind, but rather the response time, even with such technologies as overdrive. Computer monitor LCDs, noted for their fast pixel response time (because for these images it's better to absolutely minimize blurriness), are still having problems with this.
Quote:
Originally Posted by Nielo TM View Post

I don't follow

you mention a time of 33.3ms, which at first glance seems to be right for 30Hz in your previous post, when mentioning 24p on a 240hz display
Quote:
Originally Posted by Nielo TM View Post

I would prefer if we move to 48p + blur as it seems to produce the most natural motion and it's also cost effective compared to 60p or higher. It's also highly beneficial to S3D as 24p just doesn't have enough frame to induce fluidic motion.

I agree that it's cost effective, and we seem to be trending in this direction anyway (240Hz/480Hz S3D), but it is maybe not the best in general. Referring to my first post, I believe 60p+ becomes important for any non-cinema content. Documentaries would fit this, as well as for gaming and computer use, I think 120p is good. I imagine 48p+blur for cinema as acceptable.
Quote:
Originally Posted by Nielo TM View Post

Quite correct. However, frame-rate higher than 48p offers no true benefit as we humans are subtable to motion blur. So filming in 300p with no blur will appear artificial and unnatural and 300p with motion blur will be identical to 48p with motion blur. Note: I used 300p as 60p is also too slow to induce truly continues movement. We can actually detect individual frame, which is why blur is added.

I don't agree that anything above 48p has no true benefit. Are you referencing some study? In the past, some computer applications, such as games, were regularly run at 120p, as people believed they could interpret some information at that framerate. If we can actually detect individual frames, as you mention at even very high framerates, doesn't this indicate that we are able to process some information about the differing framerate, even at this high level? The idea presented was that if the refresh-rate kept being increased, eventually the mind would not perceive any difference from witnessing actual motion, even the frame-change would not be detectable. For instance, if this limit was 300p, then the image would not appear unnatural, and any blurring would be mental. Note-I gave some other reasons for 24p being an appealing fps for cinema a few posts above. At this stage I believe increasing the resolution while maintaining the refresh rates we have today anyway is a higher priority.
post #70 of 82
Quote:
Originally Posted by achalas View Post

Right, whatever can be done to the source, motion interpolation and processing and whatnot, but the maximum would be 4.167ms (240hz display), assuming RT is a nonfactor. Even without any interpolation or processing, of course the signal could still be updated every 4.167ms, this would reduce blurring but increase 'motion jumping' (i forget what this is called).

It won't reduce blurring because the pixels remain motionless. For an example, the default pull-down simply repeats each frame x number of times. Because AM LCDs do not reset after each cycle, they remain motion less, which was MCFI was developed to created additional frames to reduce hold-time.

Quote:
Originally Posted by achalas View Post

Oh most definitely it can, eventually. I'm saying with the current state in technology it's not the refresh rate that is lagging behind, but rather the response time, even with such technologies as overdrive. Computer monitor LCDs, noted for their fast pixel response time (because for these images it's better to absolutely minimize blurriness), are still having problems with this.

TN based panels only have to scan across maximum of 32 shades of grays when overridden. So they can yield excellent motion performance, but the over-all IQ is poor.

But remember lower RT does not = minimal blur. Hold-time has to be reduced as well (as you're well aware).


Quote:
Originally Posted by achalas View Post

you mention a time of 33.3ms, which at first glance seems to be right for 30Hz in your previous post, when mentioning 24p on a 240hz display

That was my bad. I've corrected the error.





Quote:
Originally Posted by achalas View Post

I don't agree that anything above 48p has no true benefit. Are you referencing some study? In the past, some computer applications, such as games, were regularly run at 120p, as people believed they could interpret some information at that framerate. If we can actually detect individual frames, as you mention at even very high framerates, doesn't this indicate that we are able to process some information about the differing framerate, even at this high level? The idea presented was that if the refresh-rate kept being increased, eventually the mind would not perceive any difference from witnessing actual motion, even the frame-change would not be detectable. For instance, if this limit was 300p, then the image would not appear unnatural, and any blurring would be mental. Note-I gave some other reasons for 24p being an appealing fps for cinema a few posts above. At this stage I believe increasing the resolution while maintaining the refresh rates we have today anyway is a higher priority.

We can detect above 300p fps as I've said above, but we do not perceive it in real life because our eyes are not fixed onto a 2D panel.

The best way I can put it is via the age old wave your hands analogy. If you wave your hand in front of your, you won't be able to see the individual fingers or the skin textures and there's no need for 300p to simulate that. However, film it in 300p at high shutter speed >300 and display it on a CRT, we should be able to detect the individual figures and the texture. The reason being is that when an object moves faster than our eyes ability to track it (and to maintain focus), it become blurry and it becomes even more blurry when we focus on something behind it (or on something else). For an example, wave you hand in front of your keyboard. Now focus on the keyboard, as you do you'll notice that the blur has increased.

Remember, we are not trying to simulate the capabilities of our eyes , but rather how we all perceive motion in real life and the majority of the time, blur plays a big part. So 48+ blur is more than adequate and blur removes our ability to defect individual at that frame rate anyway (akin to 50i/p).
post #71 of 82
while i agree with your statements Nielo TM as they are absolutely correct the thing is much of the motion resolution issue you talk about are pretty much restricted to LCD's

CRT and Plasma Displays have no problem displaying 24fps content with the right amount of sharpness and blur to make them look realistic.

All this nonsense has came about due to LCD's poor motion handling of lower but more common 24 and 30 fps frame rates used in movies/TV shows due to sample and hold induced blur.

this is one of the main reasons why i've recently become a plasma convert for normal TV use. simply put if you plan to use a TV primarily as a TV and you do not have abnormaly bright lighting conditions etc. Plasma is the way to go even with their drawbacks IMO.
post #72 of 82
Quote:
Originally Posted by Nielo TM View Post

It won't reduce blurring because the pixels remain motionless. For an example, the default pull-down simply repeats each frame x number of times. Because AM LCDs do not reset after each cycle, they remain motion less, which was MCFI was developed to created additional frames to reduce hold-time.

I am not interpreting the same way as you do for blur. Hold to me means no blur, because nothing is changing. I am considering blur as the smear that occurs when one pixel shifts from one colour to the next. Because the pixels remain motionless, none of this smearing is occuring, until for instance the final frame, if that is how we choose to process the signal (no motion compensation or processing whatsoever). Then the postion of objects in the prior frame will seem to jump to their new position in the next frame if there is some serious level of motion going on. This is what I'm referring to as 'motion jumping'. On the other extreme of our signal processing choices we have the actual frame as it was filmed only play during one refresh cycle and for the remainder of the cycles we have the interpolated, or blurring frames, as I would have referred to them, between the current frame and the next frame, which again only exists for one refresh cycle. We could of course play around with this by inserting black frames etc. as some displays do.
Quote:
Originally Posted by Nielo TM View Post

TN based panels only have to scan across maximum of 32 shades of grays when overridden. So they can yield excellent motion performance, but the over-all IQ is poor.

yet even for TN panels, they are significantly worse than ancient monitor CRTs regarding handling of motion, though seriously TN panels are of such inferior IQ that they cannot be compared to CRTs.
Quote:
Originally Posted by Nielo TM View Post

But remember lower RT does not = minimal blur. Hold-time has to be reduced as well (as you're well aware).

I am considering the case where HT is held constant. For computer screens it is usually more acceptable to minimize interpolation, as described in the first case above.
Quote:
Originally Posted by Nielo TM View Post

We can detect above 300p fps as I've said above, but we do not perceive it in real life because our eyes are not fixed onto a 2D panel.

The reason we can detect p300 is because we have probably needed to do so at some point. Depending on what is being done, the detection capabilities of the eye probably varies overall, and certain parts of the frame probably have higher rates than other areas. To use your hand-waving example, if we were directly looking and waiting for the hand to appear, and trying to notice some peculiarity about its surface texture, then we would be much more able to do it, than if we were just sitting there and someone waved a hand across our FOV.

Quote:
Originally Posted by Nielo TM View Post

Remember, we are not trying to simulate the capabilities of our eyes , but rather how we all perceive motion in real life and the majority of the time, blur plays a big part.

I shared a similar notion in my first reply in this thread when referring to cinema. In general, this would depend on style of the person who is filming the content. For instance, maybe the texture of the hand-waving is done for special emphasis. 3D will also take care of some of the other aspects of this analogy that you mention. Side note- Some viewers would probably take the exact opposite approach, that particularly because they are watching content on a tiny 2D panel that they specifically want a lack of blur, because watching blur on a panel seems unnatural to them, as it is only supposed to exist if they were really in that situation, and not just sitting down and watching it from afar, a sort of 'detached' position. This is partly why in general though I think it is appropriate to have different framerates for sports, documentaries, movies, etc. For computer monitor usage a high fps is desireable. Also, any such fixed number of fps will vary from person to person.

frito, you don't think LCDs from 2010 will nearly defeat this problem, which will bring some of the other relative advantages of lcd into greater light? (pun intended)
post #73 of 82
Quote:
Originally Posted by achalas View Post

I am not interpreting the same way as you do for blur. Hold to me means no blur, because nothing is changing. I am considering blur as the smear that occurs when one pixel shifts from one colour to the next. Because the pixels remain motionless, none of this smearing is occuring...............

Yes, nothing is changing on the screen. However, because your eye (retina) tracks moving objects in an analog continuous fashion, the longer each frame is displayed, the worse the percieved blur on your retina.

This is the source of hold-type blur. There is a conflict between sequential still images (film, video) and how your eyes track motion. To overcome this issue the time each frame is displayed has to be minimized. This is why CRTs are so good at motion even without MCFI. They display each frame for a small fraction of each frame period (~1ms).

It also answers why just having high refresh in an LCD does nothing for hold-type motion blur since without MCFI the frames are just repeated resulting in no change in how long the source frame is displayed.

Study this graphic as it explains everything I just wrote very well.



Cheers
post #74 of 82
Quote:
Originally Posted by frito View Post

while i agree with your statements Nielo TM as they are absolutely correct the thing is much of the motion resolution issue you talk about are pretty much restricted to LCD's.

CRT and Plasma Displays have no problem displaying 24fps content with the right amount of sharpness and blur to make them look realistic.

All this nonsense has came about due to LCD's poor motion handling of lower but more common 24 and 30 fps frame rates used in movies/TV shows due to sample and hold induced blur.

I agree that plasmas look sufficiently realistic when playing 24p material (with even pulldown). I think that Nielo was describing how the lower frame rate of film actually works to its advantage on LCD sets, with the reason being that motion blur is much harder to detect in low-motion (24p-30i/p) content (even though the hold-time is greater than with high-motion content).
post #75 of 82
Thread Starter 
Quote:
Originally Posted by Nielo TM View Post

48+ blur is more than adequate and blur removes our ability to detect individual frame rate anyway.
If you wave your hand in front of your, you won't be able to see the individual fingers or the skin textures and there's no need for 300p to simulate that. (akin to 50i/p).

i don't believe sample and hold blur (or any blur) is natural for the most part:

when waiving your hand across your stationary gaze, it blurs no doubt. but when tracking your hand it no longer moves across photo receptors and therefore shouldn't blur. yet due to sample and hold, the hand will still blur. this second case might be more important as people complaining of lcd blur can attest to. the popular "added blur argument" fails in the second case. i agree it would be nice to have blur for the first case but not the second.
post #76 of 82
I don't know how we can accomplish that without vastly increasing the man-hour. Also, even during tracking, blur can still appear if the eyes can't maintain continues lock.

For an example, as you start waving your hands at an accelerated rate, it becomes impossible to maintain lock (after certain speed/rate) and we immediately perceive blur. However, if I was waving my hands in front of a pulsating light (say around 60Hz), I can easily detect the individual fingers but at the cost of smooth-continues motion. Even at 100Hz, the motion wouldn't be continues. However, ~300Hz, I can't say for sure as my monitor can go that high LOL.
post #77 of 82
Thread Starter 
Quote:
Originally Posted by Nielo TM View Post

at an accelerated rate, it becomes impossible to maintain lock (after certain speed/rate) and we immediately perceive blur.However, ~300Hz, I can't say for sure as my monitor can go that high LOL.

right, i only refer to talking heads and slow pans...these things blur on my lcd when followed by eye.

Quote:
Originally Posted by Nielo TM View Post

However, ~300Hz, I can't say for sure as my monitor can go that high LOL.

my guess is it would help in both cases but probably not solve the frame transition thing completely. how about moving pixels with no frame transitions. just like real life. i bet some one's working on that.
post #78 of 82
yes, its called frameless rendering. ATM, its limited to graphics. But someday, we may have a camera and display to support it natively.

Simply put, instead of embedding pixels (or captured information) on to a frame and display those frames sequentially , we simply update individual pixel in real-time (randomly). So all the old information will remain as they are and only the new information is updated. It will not only save bandwidth but also removes motion artifacts completely.

you can find more information about it on the links below.

http://www.avsforum.com/avs-vb/showthread.php?t=1236953

http://www.acm.org/crossroads/xrds3-4/ellen.html
post #79 of 82
Quote:
Originally Posted by xrox View Post

Yes, nothing is changing on the screen. However, because your eye (retina) tracks moving objects in an analog continuous fashion, the longer each frame is displayed, the worse the percieved blur on your retina.This is the source of hold-type blur....

thanks, I realized what also Nielo was writing about. It's odd that this hold-and-switch manifests as a blur, which intuitively means pixels changing, which is actually what's not happening.
Quote:
Originally Posted by xrox View Post

It also answers why just having high refresh in an LCD does nothing for hold-type motion blur since without MCFI the frames are just repeated resulting in no change in how long the source frame is displayed.

Yea, higher-refresh rates are enabling for mcfi, and one of the main technical hurdles in putting it in, another being RT. Initially I wondered whether there should just be displayed one frame every 10 refresh cycles, (the rest being black frames), which would then be 1 frame for 4.167ms, as opposed to the 1ms figure you mention. They are not doing this, so i guess that even that 4.167 is too long, hence the replacement of some of those black frames with interpolated ones. Maybe one other possible though less likely reason is i guess that due to the RT being nearly the same as 4.167 currently, without the interpolated frames the pixels may never reach their right value for a long enough time, so the interpolated frames assist in not letting the pixels deviate too far from the next "actually-filmed frame"'s values. The problem of course in this is the creation of a pseudo-48fps. Also the 4.167 value isn't fully accurate because of some other considerations.

Updating pixels should not fully remove motion artifacts. The acm article states that it will induce a random blur. This is not really the same as motion blur which is nonrandom but it of course may yield a better result. During scenes of low motion, motion artifacts are less important obviously. In high motion scenes, ultimately the RT and HT for each pixel remain as limiting factors. The most advantageous situation is when resource utilization can be optimized when a small portion of the screen is undergoing high motion, e.g. a javelin throw. The strobing effect that Nielo mentions also should be really diminished. Also, Video compression codecs already use these ideas, so key portions of the bandwidth savings are realized in that sense. (And, as Nielo and IanD say these codecs can be used to improve rendering results.)

About maintaining a lock on a moving object, even in the case of unpredictably accelerative, or unmanageably fast moving objects, the eye can usually pick up a couple of freeze frames of the object with decent detail, (try this with a fan), which would be not possible with a continuous blur.
post #80 of 82
Quote:
Originally Posted by borf View Post

...
Great thread, very informative, Thanks!

Didn't realized you're the originator, toked 3 clips from and put on youtube Pixel Jogger channel: http://www.youtube.com/user/bwon2k10/videos
post #81 of 82
You just bumped up a very old thread from 2008-2010, so I'm adding some additional relevant information:
There are now web-based motion comparisons created since this thread was created:

(1) Web Animation: Comparision between 15fps versus 30fps versus 60fps (most can tell the diference!)
http://boallen.com/fps-compare.html

(2) Web Animation: Bouncing balls at multiple different simultaneous framerates (configurable)
http://frames-per-second.appspot.com/
- Choose the soccer ball
- Set Motion Blur to "none" for both balls
- Set one ball to 60fps and other ball to 30fps

(3) Other motion tests can be found here.
Quote:
Originally Posted by xrox View Post

It also answers why just having high refresh in an LCD does nothing for hold-type motion blur since without MCFI the frames are just repeated resulting in no change in how long the source frame is displayed.
That's true;
TFT Central has a pretty nice article on newer full-screen stroboscopic backlights (generally superior to scanning backlights), modern ones that have the ability to do massive amounts of motion blur reduction:
TFTCentral: Motion Blur Reduction Backlights

Also, a very good page that also explains hold-type motion blur that even occurs on 0ms technologies:
Why Do Some OLED's Have Motion Blur?

You are probably already aware of these newer articles from elsewhere, just posting them in this thread for completeness' sake.
Edited by Mark Rejhon - 3/21/13 at 5:22pm
post #82 of 82
Thread Starter 
Fyi, some of the older links in this thread, Mediafire deleted due to inactivity.
That interactive link is great. Just choose your frame rate and watch...
New Posts  All Forums:Forum Nav:
  Return Home
AVS › AVS Forum › Display Devices › Flat Panels General and OLED Technology › 24/30/60 fps comparison