I have an H79 at the moment and am seriously thinking of upgrading either to a new 1080P DLP 1 chip or a new 3-chip DLP and will be going the CH route with a cinema scope screen but wanted to confirm before purchasing a new PJ what causes motion judder?

I realy HATE motion judder & have seen it in many displays including LCD's & SXRD's so I know it's not a problem of DLP. I see it ALL the time and want to confirm if it's caused from my PJ and all other displays projecting at 1080p/60 and not 1080p/24? Will the problem dissapear if I get a unit the projects 1080p/24 or am I wasting my time cause it's in the transfer of DVD & HD DVD?

Film is typically shot at 24 frames per second. Any time that film is displayed at a frame rate that is not an even multiple of 24, judder will result. This is what happens when film goes through the telecine process and comes out at frame-rates of 30 or 60.

To eliminate judder, the following things are required:

1) A display that will accept 1080p/24 input from HD-DVD/Blu-Ray and that will display it at an even multiple of 24.

and/or

2) A display that will deinterlace film based content and display it at an even multiple of 24.

Below is a good description of the telecine process that Gremmy is talking about which takes film at 24p and converts it to 60i. 60i creates a 3-2 cadence (3 film frames shown followed by 2 film frames) in order to make up for the fact that 60i (30p) isn't an even multiple of the native 24p film rate. 60hz interlaced comes from the ntsc standard which we are stuck with for broadcast sources but as Gremmy points out we can avoid if a projector has a 24p or 48p inpput. A person also needs a source to support 24p or 48p output or a video processor that will take 60i and convert it to 24p or 48p like the DVDO Iscan's do.

http://www.hometheaterhifi.com/volume_7_4/dvd-benchmark-part-5-progressive-10-2000.html

1) A display that will accept 1080p/24 input from HD-DVD/Blu-Ray and that will display it at an even multiple of 24.

and/or

2) A display that will deinterlace film based content and display it at an even multiple of 24. The Sony Pearl does this with 1080i film based content

I was just thinking it would be great to have a running list of PJ's that meet these requirements and maybe how it was confirmed (judder test, test patterns, mfg specs, etc)

This feature is must for my next projector. Maybe a new thread is in order.

Can the upcoming Mits 5000 do 1 of these 2 things?

Can the upcoming Mits 5000 do 1 of these 2 things?

it can do both :)

Nice :)

"Film is typically shot at 24 frames per second. Any time that film is displayed at a frame rate that is not an even multiple of 24, judder will result."

Likely not terribly practical from a processing standpoint but couldn't 24hz film be frame rate converted to 120hz cleanly (5x frame rate) then frame rate converted back to 60hz. (1/2 frame rate)?

Sort of like how with digital audio you can do sampling rate conversion cleanly between 44.1kHz and say 96kHz by first oversampling to a common multiple of 44.1 and 96kz then downsampling to 96kHz.

Shawn

"Film is typically shot at 24 frames per second. Any time that film is displayed at a frame rate that is not an even multiple of 24, judder will result."

Likely not terribly practical from a processing standpoint but couldn't 24hz film be frame rate converted to 120hz cleanly (5x frame rate) then frame rate converted back to 60hz. (1/2 frame rate)?

Sort of like how with digital audio you can do sampling rate conversion cleanly between 44.1kHz and say 96kHz by first oversampling to a common multiple of 44.1 and 96kz then downsampling to 96kHz.

Shawn

I don't think so. At 120hz, each frame is displayed 5 times per second. Slicing that in half (to 60hz) results in each frame being displayed 2.5 times per second, which sounds to me like another 3:2 cadence and resulting judder.

Maybe someone else has a better idea of how this could be possible, but I'm not so sure.

I don't think so. At 120hz, each frame is displayed 5 times per second. Slicing that in half (to 60hz) results in each frame being displayed 2.5 times per second, which sounds to me like another 3:2 cadence and resulting judder.

Maybe someone else has a better idea of how this could be possible, but I'm not so sure.
Correct, you must have even multiples: 24, 48, 72, 96, or 120

Having said that I would like to see displays with a 120Hz (internal) refresh rate. This would work great with 24fps film and 60Hz video. It would also have to do 100Hz for PAL.

Also to be clear, usually when people talk about judder they do mean 3:2 cadence or other problems which indeed can be pretty unpleasant.

But also remember that even with film-rate content at an even multiple of 24hz you still do have the fact that the source is 24fps which is not that high. That "judder" still is present, and you can see that sometimes on really fast pans even when you're at the theater watching film.

Oh Chris you're always such a spoil sport. ;) As a kid at the Saturday matinee I used to wonder how they got the stage coach wheels to turn backward as it slowed down. Remember, 24fps, it's a feature not a bug!

Also to be clear, usually when people talk about judder they do mean 3:2 cadence or other problems which indeed can be pretty unpleasant.

But also remember that even with film-rate content at an even multiple of 24hz you still do have the fact that the source is 24fps which is not that high. That "judder" still is present, and you can see that sometimes on really fast pans even when you're at the theater watching film.

Excellent point. 24fps is inherently a pokey technology that creates a lot of annoying blur. I had an interesting discussion with Wm about this topic and I hope he doesn't mind me saying that he has viewed some 60hz motion compensated HD material that really looks a lot better than native 24fps shown at 48hz. The only problem that I see with it is that it increases storage capacity by 2.5x. This could be the killer app for BD-50 though.

Mark: there is also a featuer on some video cards that interpolate new frames basically increasing the FPS of the source content. I forget what it's called, it's like trivision or dimension or DNM TNM or TVM or something like that. Someone will know what I'm talking about and then I'll look foolish, but.... :)

Excellent point. 24fps is inherently a pokey technology that creates a lot of annoying blur. I had an interesting discussion with Wm about this topic and I hope he doesn't mind me saying that he has viewed some 60hz motion compensated HD material that really looks a lot better than native 24fps shown at 48hz. The only problem that I see with it is that it increases storage capacity by 2.5x. This could be the killer app for BD-50 though.

24 fps doesn't create blur -- film creates blur. When the shutter opens while the film is moving, this creates a blur on the film exposure itself. This can actually help create somewhat natural looking motion. The weasel cam routine from the battles in Gladiator and Saving Private Ryan achieves that effect by using a very short exposure, which avoids bluring with fast camera motion.

That said, the illusion of motion created by 24 fps is very visible, even at a multiple of 24. That said, I think the eye does a better job tricking the brain into believing in the motion of a regular cadence. The 3:2 cadence is harder for the brain to believe in, and is less pleasing to the eye.

60i video has the best illusion of motion currently available (without 'lab' technologies). Unfortunately, shooting film at higher that 24 fps is very expensive, and would require industry wide changes. I believe that this was tried in the 70's and didn't go very far.

When the shutter opens while the film is moving, this creates a blur on the film exposure itself. I would be very surprised if the film moved while the shutter was open. This would create a vertical-only blur. I always assumed that the film is stationary and any blur is due to exposure time of each frame (approaching 1/24 sec). Seems like this could easily obliterate any advantage of 1080p over 720p.

BTW, not that it matters for the purposes of this discussion but NTSC frame rate is actually 0.1 % less than 30 Hz (ie: 29.97 Hz) with corresponding field rate of 59.94 Hz). During telecine conversion, the film frame rate has to be slowed as well (to 23.976 Hz). Not sure if this carries over into HDTV standards.

Brent

While it has been a slower processes than with other mediums we will probably see the end of most use of film in the next decade. The unfortunate problem at this point is that Hollywood is sticking by the legacy 24fps format. They really need to move to at least 30fps in the post film era.

Mark: there is also a featuer on some video cards that interpolate new frames basically increasing the FPS of the source content. I forget what it's called, it's like trivision or dimension or DNM TNM or TVM or something like that. Someone will know what I'm talking about and then I'll look foolish, but.... :)

Interesting, thanks for the heads-up. I hadn't thought that something this sophisticated could be done in real-time. I'm sure more processing helps so it coulld be a YMMV case as far as the technology goes.


I would be very surprised if the film moved while the shutter was open. This would create a vertical-only blur. I always assumed that the film is stationary and any blur is due to exposure time of each frame (approaching 1/24 sec).


I too would be surprised if the film was moving while the shutter was open. It's been years since I last played around with an Ariflex but I thought that the shutter and film were time synchronized and locked together. The film doesn't necessarily have to stop moving for this to happen. If they weren't synhronized then increasing the fps would only add to the vertical blur and not improve it.


BTW, not that it matters for the purposes of this discussion but NTSC frame rate is actually 0.1 % less than 30 Hz (ie: 29.97 Hz) with corresponding field rate of 59.94 Hz). During telecine conversion, the film frame rate has to be slowed as well (to 23.976 Hz). Not sure if this carries over into HDTV standards.


Good point Brent and yes the same 59.94hz and 23.98hz carries over into HDTV standards including things like the HDMI spec. Interestingly 1080p24 timing is actually spec'ed at 1080p @ 23.98hz so the reduction at 24hz doesn't apply just to film. These timings are all available in the CEA-861-D or later spec.

Mark: I decided to stop being lazy and google it, what I was thinking of was Trimension DNM, and has been around for quite some time. The HTPC forum will likely have info about it as some have tried it out and stated it was very very different some like it I guess, others didn't.

But also remember that even with film-rate content at an even multiple of 24hz you still do have the fact that the source is 24fps which is not that high. That "judder" still is present, and you can see that sometimes on really fast pans even when you're at the theater watching film.
Getting a little off topic, but it's worth pointing out that at least one film director (James Cameron) appears to be trying to get studios and theaters to finally break the 24Hz barrier, and 120Hz seems a likely target.

While I applaud any effort in that direction, if/when it finally begins to become a reality, I'll be anticipating an upsurge of motion sickness in the theater.

Mark: I decided to stop being lazy and google it, what I was thinking of was Trimension DNM, and has been around for quite some time. The HTPC forum will likely have info about it as some have tried it out and stated it was very very different some like it I guess, others didn't.

Interesting, I did a search on the HTPC forum and you're right, it pops up everywhere. I haven't been using my HTPC much since I installed my HD-A1 so it's probably getting behind the times as far as the technology in htpc goes. At any rate, this trimension stuff is really interesting and as you mention pretty controversial. Here is also some stuff about it from Philips that includes some downloadable demos.

http://www.trimension.philips.com/index.php?page=home.html


Getting a little off topic, but it's worth pointing out that at least one film director (James Cameron) appears to be trying to get studios and theaters to finally break the 24Hz barrier, and 120Hz seems a likely target.

While I applaud any effort in that direction, if/when it finally begins to become a reality, I'll be anticipating an upsurge of motion sickness in the theater


Even 30hz would be a big improvement and many film cameras support this fps. At 30hz a normal cadence could be used which would improve both motion blur and telecine judder. Plus only 25% more film is consumed.

Don't get me wrong though I'm all for 120hz and I would think that the reduced blurring would result in added realism. Bring on the motion sickness bags!

I agree with Mark that 30fps would likely give a noticable increase in smootheness to the eye, and have minimal impact in the amount of film consumed.

Adding to that, it then fits inside one of the standard broadcast HD containers (60i) with minimal fuss, and eliminates 2:3 judder to boot!.

I've always wondered about film stock and processing costs. I mean for a movie with a double-digit-millions budget, is 25% more film really that big of an impact? I assume the telecine, storage, and FX costs would go up too, but still how much of the overall budget goes to such things?

...Even 30hz would be a big improvement and many film cameras support this fps. At 30hz a normal cadence could be used which would improve both motion blur and telecine judder. Plus only 25% more film is consumed...
Film speed will not be changed because of all the legacy equipment. No one is about to spend large amounts of capital on replacing all the film equipment with the digital change over looming. If this were the case I would also like to ses Vista Vision for constant height in film. ;) The problem is going to get the industry to change the standards to a higher rate for digital "film". I would like to see something like the following to become the standards.


A 4K plus system with cameras and projectors
3840x2160 for 1.78 (drop 1.85)
5080x2160 for 2.35 scope (no lenses or matting)
36 to 42 bit color depth
60fps or 120fps

While 120Hz may be excessive there are a couple of good reasons to use it.

60Hz sounds to common and a consumer standard plus the most important reason 120Hz gives you backwards compatibility with 24fps film.

I would be very surprised if the film moved while the shutter was open. This would create a vertical-only blur. I always assumed that the film is stationary and any blur is due to exposure time of each frame (approaching 1/24 sec). Seems like this could easily obliterate any advantage of 1080p over 720p.

BTW, not that it matters for the purposes of this discussion but NTSC frame rate is actually 0.1 % less than 30 Hz (ie: 29.97 Hz) with corresponding field rate of 59.94 Hz). During telecine conversion, the film frame rate has to be slowed as well (to 23.976 Hz). Not sure if this carries over into HDTV standards.

Brent
Crap -- that was a typo. The film is held motionless. I meant the camera moving while the film was still. The exposure of the film shows this motion as blur.

Edit: you can have registration issues between the shutter and the film frame, and that could cause issues, but that's the exception.

Is there any real advantage to the "multiple of 24" idea with a digital display? Won't the device simply repeat the frame X number of times? I understand that with a CRT this would be a problem, because you would get flicker from the slow refresh, but with digital devices the entire frame is displayed at once. What benefit would you get from 72fps over 24fps?

Is there any real advantage to the "multiple of 24" idea with a digital display? Won't the device simply repeat the frame X number of times? I understand that with a CRT this would be a problem, because you would get flicker from the slow refresh, but with digital devices the entire frame is displayed at once. What benefit would you get from 72fps over 24fps?

24 works fine, the point is it just should be a multiple of 24. If it's not an even multiple (including a multiple of 1, which is 24) then you get judder introduced.

24 works fine, the point is it just should be a multiple of 24. If it's not an even multiple (including a multiple of 1, which is 24) then you get judder introduced.
Okay, yes, I get that. Unfortunately the NTSC/ATSC standards are set and there's nothing we can do about that. I was addressing the fact that when it comes to disc based film source material, that some people are obsessed with 72fps or 120fps. I just wanted to know if there was an advantage to it considering those specific circumstances. I appreciate your reply. Question answered.

Okay, yes, I get that. Unfortunately the NTSC/ATSC standards are set and there's nothing we can do about that. I was addressing the fact that when it comes to disc based film source material, that some people are obsessed with 72fps or 120fps. I just wanted to know if there was an advantage to it considering those specific circumstances. I appreciate your reply. Question answered.
24fps is a old Hollywood standard and is not supported by NTSC. It does have provisions in the ATSC standard but I don't think it has been used yet.

If a digital display refresh rate is at only 24Hz it will flicker the same as film (unless the pane has a super slow decay rate). So you need to up the refresh rate to at least 48Hz to avoid flicker. I can even tell when my display is in 48Hz mode by looking and focusing just off screen. I can detect flicker that I can't see in 60Hz mode.

If a digital display refresh rate is at only 24Hz it will flicker the same as film (unless the pane has a super slow decay rate). So you need to up the refresh rate to at least 48Hz to avoid flicker. I can even tell when my display is in 48Hz mode by looking and focusing just off screen. I can detect flicker that I can't see in 60Hz mode.

There's no inherrent benefit to one multiple over another in terms of flicker, as long as the displays themselves are equal. If you had a "zero decay" display that could instantaneously refresh, it would look exactly the same with 24 hz film content at 24 hz as at 120, or 12000 hz...

(William, I think you're in agreement, I'm just stating htis for other folks on the thread).

If a digital display refresh rate is at only 24Hz it will flicker the same as film

24fps film flickers at 48Hz. With a bowtie shutter, each frame is flashed twice before the next frame is advanced into the gate. (A 2:2:2… cadence)

Dave

Guys, you're confusing framerate with refreshrate/flicker. with a digital display there isn't really any flicker at all. But the framerate means that any inherent judder to the 24fps source will appear the same. When you display film, you have 24fps judder, and then 48hz flicker usually. If you use a CRT, then you introduce flicker issues again as well. For instance I run 72hz refresh on the CRT, but the FPS is still 24, each is just repeated 3 times because of the nature of the way CRTs work. With a digital, you could just use 24hz and it would look fine, if the display supported that, for instance. There's no flicker issues because there really aren't any flicker concerns involved.

Guys, you're confusing framerate with refreshrate/flicker. with a digital display there isn't really any flicker at all. But the framerate means that any inherent judder to the 24fps source will appear the same. When you display film, you have 24fps judder, and then 48hz flicker usually. If you use a CRT, then you introduce flicker issues again as well. For instance I run 72hz refresh on the CRT, but the FPS is still 24, each is just repeated 3 times because of the nature of the way CRTs work. With a digital, you could just use 24hz and it would look fine, if the display supported that, for instance. There's no flicker issues because there really aren't any flicker concerns involved.
If a digital projector is operating at 24Hz doesn't that mean it will display 24 images a second? If so it will have more panel off time than at 72Hz. I can tell 100% of the time if my projector is in 48Hz mode by looking and focusing off screen and using my peripheral vision to see the flicker (can't see it at 60Hz). It doesn't have to be film playing and I can tell much easer with a static Windows background.

I've always assumed that the image frame was double buffered and independent of the refresh rate. In actual practice though, they could be implemented to be synchronized. For example at 24fps a projector might simply show a frame twice to achieve 48fps and then run the refresh rate at the same 48hz. It's probably best however to disassociate what is theoretically possible from actual implementations. For example many people have complained about 48hz in various forms but many of the problems seem to be due to problematic implementations on the source and/or display device which shouldn't be construed to apply to all 48hz implementations.

At any rate, I did a quick web search to see if I could find any DILA or SXRD specifics on how they run their panel refresh rates but I came up empty. I did find an interesting new release from JVC on their LCD flat panels. Apparently they are also getting into the motion compensation business:

http://news.thomasnet.com/companystory/488708

This is the problem with many of the common theory questions. If the manufacturers would just tell us exactly what they are doing, then it would not be an issue, we would know. Every notice how you can get almost zero technical details about the inner workings of projectors? Why so secretive? What EXACTLY happens when you feed a 1080p24 feed to a particular projector? Is there a global refresh/blanking ala CRT on digital projectors in general or on some models specifically? The Pearl claim of 24p in equals 96p out that I have seen mentioned for example. This only makes sense to do if there is blanking between frames, and/or some sort of “phosphor decay” like phenomena going on in the display panel. If there is no blanking/decay, then you may as well just display at 24fps.

…with a digital display there isn't really any flicker at all. …With a digital, you could just use 24hz and it would look fine, if the display supported that, for instance. There's no flicker issues because there really aren't any flicker concerns involved.

Every display or projector I’ve ever looked at (designed to display the illusion of images in motion) with a photo diode flickers at its framerate or a harmonic. The illusion of motion IS a series of still images periodically flashed. They all have had some amount of periodic time the image is blanked before the next is presented. Having no shutter, blanking or reset between frames produces a blurry mess to human perception.

“For instance”, what digitals have you experienced that can sync natively to a 24Hz vsync and not flicker at 24Hz or a harmonic?

Dave

24fps is a old Hollywood standard and is not supported by NTSC. It does have provisions in the ATSC standard but I don't think it has been used yet.

That's what I meant when I said that "I get that." I fully understand jutter caused by telecine/inverse telecine. My point was that we must simply learn to live with it when dealing with broadcast material such as OTA. In my question, I was referring specifically to film sourced material on a disc that had the ability to output 24fps or some multiple thereof.


If a digital display refresh rate is at only 24Hz it will flicker the same as film (unless the pane has a super slow decay rate). So you need to up the refresh rate to at least 48Hz to avoid flicker. I can even tell when my display is in 48Hz mode by looking and focusing just off screen. I can detect flicker that I can't see in 60Hz mode.
I can see this problem with CRT scan lines decaying before the screen refreshes, but I don't see how a digital display can cause this, when the entire screen is displayed and refreshed at once. Is it due to the off time between screen refresh? The off time between refresh should be a set time. If the refresh rate increased then the total off time in a given second would actually increase as well.

For example, if the off time is 1ms (I have no idea what it might be really, I'm just using this number as an example) and the refresh rate was set at 24hz, then the total off time in one second would be 24ms. If the refresh rate were bumped to 48hz, then the total off time would be 48ms in that same second. This may affect overall brightness as well, if this is really how it worked.

The question then is this: Is there really any off time between frames on a digital display, or do the pixels simply transistion from one to the next?

If a digital projector is operating at 24Hz doesn't that mean it will display 24 images a second? If so it will have more panel off time than at 72Hz.*snip*
How will 24Hz have MORE off time in a second? See the second part of my post above.

How will 24Hz have MORE off time in a second? See the second part of my post above.
It may be related to the decay rate which is the time it takes to go from fully on state to fully off state. At a higher rate like 72Hz the end of the decay of one refresh may happen at about the same time as the next on state eliminating or reducing the off time.

Also I would think that if the off time per second was still the same amount of time over all, but divided into smaller increments it would be less perceptible.

Maybe it is neither one of these but I know that there is a deference because I can actually (indirectly) see it.

Most of this is covered here: http://en.wikipedia.org/wiki/Flicker_fusion_threshold

Brent

Thanks for the link, but it still leaves some questions unanswered. The article states, "LCD flat panels do not seem to flicker at all as the backlight of the screen operates at a very high frequency of nearly 200 Hz, and each pixel is either 'on' or 'off'."

This does not apply to front projection, which has a constant light source in a UHP lamp. William, are you able to see this on front projection as well as backlit LCD displays?

Also, does anyone know for sure if there is an "off time" between frames?

Well, note it says "This article or section is in need of attention from an expert on the subject"

This example is talking about flicker of the back light and a static image, for example a white field, not the inter-frame judder.

LCD "response times" are notoriously slow (around 5 - 10 mSec for LCD monitors) compared with a frame time at 60 Hz of 17 mSec. In fact this is why single chip, pulse width modulated, LCD & LCOS displays (with the exception of the now defunct Philips RPTV) are not available. Therefore there would appear to be no possibility of an "off-time" between frames.

OTOH, DLP is a whole different story.

Brent

Also, does anyone know for sure if there is an "off time" between frames?

This is a more complicated question than it looks. In the days of broadcast NTSC and CRT's it always corresponded to the vertical refresh timing. Nowadays it can be as quick as the time it takes for a single pixel to be clocked.

It's best to think of refresh rate and display rate as independent parameters that may be (but not always) synchronized in multiples to one another. In most displays the toggling between frames can be nearly instantaneous (no off time needed) but there are reasons not to do this. One reason is that "tearing" can occur when a frame is shown that is a mix of the two adjacent frames. In severe cases, ghosting can happen if a frame is shown for a different duration than other frames.

In most modern double buffered systems (also called a ping pong buffer), frame data is placed into two distinct areas of memory. When one frame is finished the video controller chip can start displaying the next frame by simply updating it's memory address to point to the other buffer so that the next pixel written is from the next buffer (which in turn contains the next frame). As mentioned above though, updating the image in the middle of a frame looks terrible because it creates a mix of two different frames. So what is typically done is to synchronize the frame update with the device refresh time so that a clean frame is shown. Most PC's for example have a vertical synch setting that allows the frame update to be synchronized to the display rate. In the case of HTPC's judder will occur if this setting is used and the refresh rate isn't a multiple of the frame rate because a frame needs to be dropped periodically to stay in synch with the frame rate. If on the other hand a person disables this setting tearing can occur. For HTPC's of course the best solution is to use a refresh rate that is a multiple of the frame rate (71.9x hz for example).

In the case of display devices, the refresh rate for most technologies is dependent on the physical properties of the device. For CRT's for example it's related to the persistence of the phosphor and the perception of flicker, but for LCD and LCOS devices it's related to the length of time the crystal can be twisted. For all of these technologies, there is a range of variability that can be used which in turn can be exploited by a system designer so that the frame and refresh rates are synchronized and so that tearing and other artifacts are minimized.

All of the above is a description of generalities, actual implementations can vary greatly which is why we have forums like this but also as I said earlier it also explains why the results of some implementations of 48p may differ greatly with others.

...William, are you able to see this on front projection as well as backlit LCD displays?...
I have a JVC HD10K (I have a JVC HD10K LCoS projector that operates at 48Hz, 50Hz, or 60Hz. I don't see flicker while watching directly but can identify it at 48Hz by the method I described above. I can't identify flicker at 60Hz. My projector will also operate at 50Hz but have never tried it. I may try 50Hz just to see.) LCoS projector that operates at 48Hz, 50Hz, or 60Hz. I don't see flicker while watching directly but can identify it at 48Hz by the method I described above. I can't identify flicker at 60Hz. My projector will also operate at 50Hz but have never tried it. I may try 50Hz just to see.

I have a JVC HD10K (I have a JVC HD10K LCoS projector that operates at 48Hz, 50Hz, or 60Hz. I don't see flicker while watching directly but can identify it at 48Hz by the method I described above. I can't identify flicker at 60Hz. My projector will also operate at 50Hz but have never tried it. I may try 50Hz just to see.) LCoS projector that operates at 48Hz, 50Hz, or 60Hz. I don't see flicker while watching directly but can identify it at 48Hz by the method I described above. I can't identify flicker at 60Hz. My projector will also operate at 50Hz but have never tried it. I may try 50Hz just to see.

Hi William, I think you'll find that JVC uses hertz as a euphimism for frame rate rather than actual panel refresh rate. It would be very interesting to see what the actual panel refresh rate corresponds to @ 48p, 50p and 60p frame rates.

... but for LCD and LCOS devices it's related to the length of time the crystal can be twisted.Hmmm, this is getting interested. I didn't realise this was a limitation. However, I can imagine (without bothering to find out how these things really work) that there is a capacitance discharge mechanism at work. So, for a static display there wouldn't be an off-time per se (ie: not a rectangular brightness vs time function) but rather a charge/discharge cycle. The brightness vs time function would be approximately saw tooth (or, since Mark is participating in this thread, I had better say, more correctly represented by interleaved exponentials :) ).

I can imagine that this is what William is seeing at 48 Hz "refresh rate".

Brent

If a digital projector is operating at 24Hz doesn't that mean it will display 24 images a second? If so it will have more panel off time than at 72Hz. I can tell 100% of the time if my projector is in 48Hz mode by looking and focusing off screen and using my peripheral vision to see the flicker (can't see it at 60Hz). It doesn't have to be film playing and I can tell much easer with a static Windows background.

There is no panel "off" time, it just continues displaying the same thing until it is refreshed, at least with LCD and LCOS. DLP is slightly different because it is always active, but the principle is the same. Essentially, if it's 24fps content, if it's displayed with a 24hz refresh or 48hz, etc it would look indistinguishably the same.

There is no panel "off" time, it just continues displaying the same thing until it is refreshed, at least with LCD and LCOS. DLP is slightly different because it is always active, but the principle is the same. Essentially, if it's 24fps content, if it's displayed with a 24hz refresh or 48hz, etc it would look indistinguishably the same.
Then why or how can I tell 48Hz from 60Hz?

Then why or how can I tell 48Hz from 60Hz?

After giving it more thought, I think what people are thinking of panel refresh time is actually panel response time. For DILA I think it's around 10ms (100hz) and for SXRD it's something like 8ms (125hz). As Chris mentions it's continually varying (no refresh off time as opposed to frame off time) but this doesn't mean it couldn't still vary slightly as a multiple of the frame rate (to minimize tearing) so this might be what William is seeing. This is just speculation though, we need an engineer from JVC or Sony to jump in and give us the facts.


or, since Mark is participating in this thread, I had better say, more correctly represented by interleaved exponentials


:D I find it hard to get technical points across without being precise in terminology but it probably comes across as though I'm a total geek. Which is only partially true lol.

Then why or how can I tell 48Hz from 60Hz?Well, if my hunch about the sawtooth refresh is correct, and you happen to have a photodiode/phototransistor hooked up to an oscilloscope, then you could tell pretty easily. :rolleyes:

What *might* work is viewing the screen through a video camera. With a static pattern 60 Hz source should show identical frames whereas 48 Hz should show some kind of frame-to-frame beat pattern (or brightness modulation) that repeats every 5 frames on the video camera. (5 frames at 60 Hz has the same duration as 4 frames at 48 Hz, but this assumes refresh rate is locked to input frame rate). Might need to step through still frames and/or capture frames and analyse in photoshop to detect any beats however. Viewing rapid motion or action sequences from the 48 Hz source should be even more revealing.....

Brent

What other current processors can deconstruct a 1080i60 signal from cable hd and output at 24fps like it has been stated the dvdo scan can? Can the realta or gennum chip?

Then why or how can I tell 48Hz from 60Hz?

Because if you're running at 60hz, the duration is uneven, so you have a 3:2 cadence that one can see, rather than an even cadence where the duration is even.

Because if you're running at 60hz, the duration is uneven, so you have a 3:2 cadence that one can see, rather than an even cadence where the duration is even.

Exactly right. Although in this thead there seems to be confusion about frame rate, refresh rate and panel response time so it's worth clarifying that Chris means the frame rate.

What other current processors can deconstruct a 1080i60 signal from cable hd and output at 24fps like it has been stated the dvdo scan can? Can the realta or gennum chip?

The Realta can do it.

Because if you're running at 60hz, the duration is uneven, so you have a 3:2 cadence that one can see, rather than an even cadence where the duration is even.
On a static Windows background screen (blue is the easiest color) with no program running and no DVD in the drive? I don't understand how 3:2 cadence comes into play? :confused: Please explain this.

BTW: I know and see judder with a HD-DVD or DVD at 60Hz vs 48 HZ (that is why I use 48Hz for film based playback) but I'm talking about flicker.

On a static Windows background screen (blue is the easiest color) with no program running and no DVD in the drive? I don't understand how 3:2 cadence comes into play? :confused: Please explain this.

BTW: I know and see judder with a HD-DVD or DVD at 60Hz vs 48 HZ (that is why I use 48Hz for film based playback) but I'm talking about flicker.

Nono, I was talking about judder. There won't be any flicker involved if it's LCD/LCOS/DLP. And if the image is static of course, then you won't see the judder.

There won't be any flicker involved if it's LCD/LCOS/DLP.I would have thought so too, except William can see flicker in his peripheral vision at 48Hz on his JVC HD10K LCoS. My post #45 above provides a possible reason why this might occur on LCD/LCOS.

To expand on this a little, despite LCD/LCOS being included on the Digital Hi-End Projectors forum, when all is said and done they are analogue display technologies (albeit fixed panel rather than scanned). The brightness of each pixel is proportional to the analogue voltages applied across the liquid crystal cells. The voltages are stored as charges on capacitors and subject to decay over time, hence the need for refresh even for static images.

Brent

Nono, I was talking about judder. There won't be any flicker involved if it's LCD/LCOS/DLP. And if the image is static of course, then you won't see the judder.
...but I can see flicker indirectly by looking and focusing off screen and using my peripheral vision with a static image (non film) at 48Hz but not at 60Hz. :confused:

It is not a problem and I can't see it watching a film. I only notice it when I have a static picture and turn my head away from the screen. It is not something to worry about but there is some deference between 48Hz and 60Hz.

...but I can see flicker indirectly by looking and focusing off screen and using my peripheral vision with a static image (non film) at 48Hz but not at 60Hz. :confused:

It is not a problem and I can't see it watching a film. I only notice it when I have a static picture and turn my head away from the screen. It is not something to worry about but there is some deference between 48Hz and 60Hz.

Huh, well I've never seen anything like that, and I am very susceptible to seeing flicker and things like that. That sounds quite strange to me.

...but I can see flicker indirectly by looking and focusing off screen and using my peripheral vision with a static image (non film) at 48Hz but not at 60Hz. :confused:

It is not a problem and I can't see it watching a film. I only notice it when I have a static picture and turn my head away from the screen. It is not something to worry about but there is some deference between 48Hz and 60Hz.

There are a couple of things that I can think of that would cause this. The first is what I was alluding to earlier in this thread. If the image is double buffered to eliminate tearing (which it is in the case with an HD10K). A new frame is being serially transmitted while the old frame is being buffered and displayed. In this case there is likely a synchronization with the panel refresh timing when the new frame is ready for display. This synchronization could be in the form of varying the panel refresh time (although as I mentioned earlier, I don't know this for a fact). The other thing that could cause this is the source (htpc) itself. There are known problems with Win XP SP2 and renderless VMR9 that cause flicker at 48hz.

Also, as I was saying earlier we're now getting into specific implementation issues of both the source (HTPC) and the display device (HD10K). Issues with specific implementations of 48p shouldn't be taken as a problem with this timing in general. As an example, the HD10K uses an SRAM frame buffer to double buffer the frame, it also uses an analog backplane which contains a MOS transistor to charge and discharge a capacitor that turns the pixel on and off. Another JVC implementation of DILA use a digital backplane and the pixel is driven with a PWM clock. Another manufacturer might replace the SRAM frame buffer with a DRAM buffer to save a few pennies so in all of these cases the refresh timing can change drastically which in turn might impact how a device performs at varying frame rates.

I may have figured out what and/or why I'm seeing flicker. First it is not source or film related. I see it on a Windows Desktop with no software running and no DVD in the drive. I also see it with my HD-DVD STB on the Setup menu with no HD-DVD in the drive.

I set the projector to 48Hz and turned on the internal test patterns and I can't see it with them on. This leads me to believe that I'm some how able to see the actual signal flicker that is inherent with a picture at sent at 48Hz. Or it could be that the internal test patterns are handled in a deferent manor than the input signal.

I also discovered something I had never noticed before in 7 years of projector ownership. On the test patterns I could see heat waves in the lower center of the screen. :confused: At first I thought it was internal but when I moved my head the waves moved. So it was the heat coming off the top of my head distorting the picture. Well just one more thing to worry about. I guess to improve our pictures we should all wear ice packs on our heads. :eek:

I set the projector to 48Hz and turned on the internal test patterns and I can't see it with them on.

Yeah, it's likely that the HD10K uses different source timing when it displays the internal test patterns so that might be why the flicker goes away. Clocking in new frames, storing them into a buffer and then clocking out the buffer to refresh each pixel in the panel is a lot more complicatd than people think. There are a lot of variations on the theme which can cause flicker that is not necessarily related to frame judder even though changing the frame rate may cause variations in the flicker.

Ummmm...

You guys kinda lost me could you ALL please retype & repeat what you all said? Right from the start. Thx.

You guys kinda lost me could you ALL please retype & repeat what you all said? Right from the start. Thx.

done. link here: :) http://www.avsforum.com/avs-vb/showthread.php?t=737604&page=1&pp=30

done. link here: :) http://www.avsforum.com/avs-vb/showthread.php?t=737604&page=1&pp=30

Lol. Thanks for saving us the typing. Also, when a person reads it from the start, please realize that there are many different causes for motion problems. This thead discusses a bunch that need to be kept separate in a readers mind. Off-hand, I can think of the following topics in this thread: telecine judder due to non-equal (1-1,2-2, etc) cadences. Film motion blur (object being viewed is moving relative to a stationary film frame), slow camera frame rates that heighten camera blur, slow refresh rates, refresh rates that are synchronized to frame rates and blanking periods between frames. Did I forget any?

In regards to the question of the thread title: as some have already indicated "judder" doesn't come only from the telecine process; it's also inherent in the limitations of 24ps film techniques and shows up on the movie screen as well. Since many here have been covering the telecine/video process-born version of judder, perhaps I'll blather a bit about the film-born version.

24fps is low horsepower - a very low "sample rate" - for capturing motion and that particular number-per-second makes for some odd artifacts. A classic is the "wagon wheel that starts going backward," - e.g. in westerns as a horse-drawn wagon with large wheels and long spokes begins to slow down, the spokes appear to briefly start going backward, and then reverse back to the proper direction. (It's not hard to figure out why that occurs if you think about it).

But the judder that is problematical to film cameras isn't actually a problem during fast motion or fast pans so much as "medium" paced (or medium slow) pans. A lot of effort goes into dealing with the limitations of 24fps in making movies. If you pan quickly all the film captures is blur, so a fast pan lasting too long will bug out your eyes like you can't focus. But it's ok as long as there is an object for your eyes to focus on in the frame, even if all the rest is a blur. So that's obviously why the actual relative movement of an actor or object within a frame tends to be restrained. In other words, the actor running through a crowd will generally be kept around one area of the frame so that at least he remains in focus even if everything else is blurred by the panning motion (likewise with space-ship flying scenes).

Judder becomes a problem especially with camera movement, e.g. a pan, that becomes medium tempo and is exacerbated by hard-edged, vertical objects - the nightmare scenario being the medium pan across a picket fence outside a house. If the pan is fast enough, the vertical boards of the picket fence will become blurred on each frame of film, creating a blurred-together look as we watch it on screen. This looks fairly natural to us, since objects in motion blur in our vision naturally (just wave your hand in front of your face, especially under low lighting if you haven't noticed it before). But if the pan gets slower, the individual pickets in the fence are captured with less blur, until you can see the more discrete, more defined pickets. But at 24 fps doesn't allow capture (at such a speed) of all the successive positions needed for a smooth motion. So
what you notice are the pickets stuttering across the screen, because the camera has only captured the pickets in position A...D...G...J...M.....instead of all the steps....ABCDEFGHIJKLM
(Or I....I....I....I....I....I instead of a continuos IIIIIIIIIIIIIIIII).

Something like this same phenomena is seen in movies like Private Ryan and Gladiator, which employed a technique sometimes referred to as "skinny shutter" (or 45 degree shutter), in which the shutter angle is changed such that
exposure time is reduced for each frame. So, just like raising the shutter speed on your still camera ensures a sharper image, you end up with sharper discrete images on each film frame. But you also therefore loose the "natural" looking blurring that is typically captured with motion on film. Hence when played back on-screen, the sharp images at 24fps allows our eyes to pick up the discrete
time sampling of the action, giving the image that sharp, but jerky look seen in the gladiator battles.

(And it's why stop-motion animation like Ray Harryhausen's work had that jerky quality. It wasn't so much due to a lack of ability on H's part - obviously he was amazing. But it was a limitation of the technique, because unlike regular real-time motion captured on film, which causes some blur, each frame of a Harryhausen's model's movement was a still motion, making for sharp images each frame and thus the stuttered movement like you see in gladiator. It wasn't until The Empire Strikes Back that this deficiency in stop motion animation was addressed and Phil Tippet created "go-motion," in which the models were actually moved a little with the shutters open, giving the image a natural bit of blur on each frame and restoring fluidity. For just these reasons artificial blur is added to motion in CGI characters/animation, to mimic the blur of "real" objects as captured on film).

But back to the picket fence....if the camera slows down still more, even though the individual pickets are caught with greater clarity, they will move more smoothly through the frame because, obviously, more frames-per-second were able to be devoted to the movement of each picket, making for smoother motion.

In fact if you read a cinematographer manual, such as American Cinematographer 35mm Manual, it has specific recommendations for panning speeds (also taking into account screen brightness, focal length etc), so as to avoid strobing and skipping effects.

This limitation is something that really bothered Douglas Trumbull (Special Effects supervisor for Close Encounters Of The Third Kind, special effects guy on 2001, director of Brainstorm etc). The low 24fps sample rate of film is such a limit on it's fidelity (speed any sampling system up - analog tape, digital what have you - and you increase fidelity), which is one reason why Trumbull created "Showscan," in which his company did many tests for arriving at frame rates that led to a clearer image and smoother motion. It's shot on either 65mm or 70mm film at 60fps with an extremely wide shutter angle of 180 degrees, making for a shutter time of 1/125 second. The very short exposure time makes for very sharp images, and the high 60fps sample rate makes for smooth motion. The combination makes for super-sharp images with natural motion. Too bad it didn't catch on
with the major film companies and was relegated to a theme-park "ride." But if you've ever seen it - wow! - you've never seen film look like that before. It's amazingly 3 dimensional-looking.

But, anyway, whenever I was shooting film I had to be conscious of judder scenarios and the general limitations of trying to capture motion at 24fps.
It still kind of drives me nuts when it occurs watching movies projected in theaters as well.

(Sorry for the length, but given this thread title I s'pose there must be someone interested in the details...and btw earlier there was a discussion wondering if the movement of film in the gates could be responsible for judder. However, each frame of film moving through a camera's gates is momentarily stopped for the film exposure. Obviously this is necessary, otherwise all film would be a blur. So judder and motion artifacts are a matter of exposure time and the low sampling rate of 24fps film).

Cheers,

Rich

(Been a long while since I've shot film at this point, so forgive if my memory is getting creaky).

It appears that the only reduction that can be had in judder for mere mortals with VP’s is to send 24p or a multiple there of to our projectors. Then you get into equipment specifics about what rates and types you can send (vp limitations) and what the particular projector you are sending to will do with those rates (projector limitations). Digital transmission tops out at 1080p60 for current consumer implementations AFAIK. VP limitations and projector limitations will intersect giving you a subset that starts at 48p and goes downward. So for digital we have 24p 24psf and 48p as likely sending rates. Assuming your VP will do those three rates, what should you send? Looks like it will depend on your projector. My recollection is that the only 24p variant you can send to the Ruby is 24psf in gives you 48p out. Pearl is rumored to be 24p in = 96p out. So what happens if you send 24psf or 48p to the Pearl? Will it accept it, and if it does, what will it display it at? 48p? If so, you are better off sending 24p in that case. We might be trending toward 24p outputs from source devices for film based HD content (as seen or promised on HD-DVD and BluRay). That makes sense, it keeps the bandwidth requirements down and hopefully the concept of multiplying input frame rates is simple enough that the display manufacturers can do it correctly – unlike the telecine butchering fiasco we currently have.

Two people have already referred to the fact that LCD/LCoD/SXRD diplays typically have a capacitor that is driven via transistor to drive each individual liquid crystel cell making up a pixel. You can actually see this in the schematic reperesentation in the new JVS DILA whitepaper referenced in another thread.

The electrical characteristics of the panel will determine how long it takes to charge the capacitor, as well as how quickly it will drain, both when unloaded via the transistor, as well as via leakage from the coupled LC cell. Typically, matrices such as these are strobed via row & column address lines. So each capacitor is charged in a fraction of the display refresh time, as during that interval all 1080 rows need to be strobed.

This means that the capacitor then must maintain the charge to bias the LC cell on. Given the decay time of the charge is non-zero, it's likely that some this decay becomes visible if the strobe rate becomes too low. Therefore that's probably why some PJ's actually drive the display at a multiple of the framerate (i.e. 24fps at 96Khz for te Pearl). It also may explain why one poster can see this behavior... that display may be getting toward the bottom end of the range where it can maintain each cell charge without beginning to visibly decay.. at least for senstive viewers.

Hello,

I have a new Panasonic PT-AX100 720p (my first projector), and I have read through the whole discussion on 48Hz LCD flicker - because I have the same problem as William explained.

Please, if you have also an LCD projector connected to HTPC try to do the same as William did - look at the black Windows/MacOS:) screen. I use my HTPC for browsing and working, so on contrary to William - the SMALL FLICKER IS A PROBLEM for me.

On my regular LCD (notebook) I use 80Hz refresh rate and I am fine. On the Panasonic I am not - i tried to setup manually the refresh rate therefor, but I found out that I cannot go above 60Hz (quite ridiculous). And after 2 hours working with all white Excel/Word my eyes are getting really tired.

I have read a lot of reviews on my projectors, and I wonder how come that nobody mentioned this problem?? You just dont realize it in shop when you test the equipment - but after hours of work it is a pain. I would have gone for DLP if I new that.

Please if you have the LCD projector with HTPC, try it, and let me know how you have solved it.

ps. I dont see it directly on 60Hz, but its same as with the monitor, you dont see it, but you just find out that your eyes are tired after some time. With 24p I see it so obviously (dont you know why the PJ does not convert the 24 to 48 at least? - or - where can I see the actual refresh rate employed by PJ?)

PLEASE HELP - I dont want to damage my eyes

In regards to the question of the thread title: as some have already indicated "judder" doesn't come only from the telecine process; it's also inherent in the limitations of 24ps film techniques and shows up on the movie screen as well. Since many here have been covering the telecine/video process-born version of judder, perhaps I'll blather a bit about the film-born version.


Rich, Somehow I missed this when it was first posted and I just read it. It provides a great description of judder and motion blur from a film prospective. Thanks for posting it.

To eliminate judder, the following things are required:

1) A display that will accept 1080p/24 input from HD-DVD/Blu-Ray and that will display it at an even multiple of 24.

and/or

2) A display that will deinterlace film based content and display it at an even multiple of 24. The Sony Pearl does this with 1080i film based content

Or

3) interpolation with motion compensation to 30 or 60 hz.

Hi William, I think you'll find that JVC uses hertz as a euphimism for frame rate rather than actual panel refresh rate. It would be very interesting to see what the actual panel refresh rate corresponds to @ 48p, 50p and 60p frame rates.

The panel refresh rate (as measured with photo diode and o'scope) is 48, 50, and 60hz on the HD10K (and Meridian MF1).

The panel refresh rate (as measured with photo diode and o'scope) is 48, 50, and 60hz on the HD10K (and Meridian MF1).
William, don't know if you read through the whole thread but how am I able to see the deference between 48Hz and 60Hz? I can tell when my HD-10K is in 48Hz mode by looking and focusing just off screen. I can detect flicker that I can't see in 60Hz mode. What am I seeing and how can I tell?

The panel refresh rate (as measured with photo diode and o'scope) is 48, 50, and 60hz on the HD10K (and Meridian MF1).

Interesting. Are you sure that what you're measuring is panel refresh rate and not the frame rate, frame blanking or an artifact of the way the panels are driven? The reason that I ask is that driving LCDs is complicated by the fact that the liquid crystal needs to be periodically relaxed or it degrades and becomes permanently polarized. So each pixel needs an alternating +- DC bias and there are many different refresh schemes that can be used in getting the pixel alternately reversed (refreshed). Some schemes use simple single pixel row/col scans, others are line based, others alternate every other pixel etc. There is also the issue with the electrical parasitics in the gated capacitor which in turn governs how long the bias voltage can be applied to each pixel. Another complication is matching the alternating +- DC bias so that both states yield the same reflectivity.

Without really knowing what's going on inside it's really hard to say what is causing this specific 48, 50 and 60hz flicker. I suppose the simplest design would be one where the frame rate and refresh are coupled and are in fact one in the same. In other words Frame 0 uses negative biases for all the pixels in the frame and when frame 1 is available (at 48, 50 or 60hz) it then uses positive biases for all of the pixels. This would certainly describe what William and others are seeing, but this would be surprising because most designers prefer driving LCD's at > 60hz. It would be disappointing if JVC used a simple scheme like this knowing that it would lead to flicker problems at 48hz. It's possible though that 48p support was added late in the game and the designers went with what they had. It would be really interesting to get a schematic and see exactly what's going on. Anyone have one?

The panel refresh rate (as measured with photo diode and o'scope) is 48, 50, and 60hz on the HD10K (and Meridian MF1).


William, I'd be interested in trying this with my scope and my Pearl, to determine if/when it switches to 96hz mode.

Any recommendations of a phtodiode that has the correct response characteristics for these kind of tests?

Thanks.

Mark,

The observed rate, using a photodiode and o'scope, changes with the input signal. If this were an artifact of the way the panels are driven, why would it change? JVC tells us the HD10K supports 48hz. I don't get what point you are trying to make here. Earlier JVC projectors did not support 48hz, only 60 hz, then later 50 and 60 hz.

A waveform includes both positive and negative swings. I think this might account for the alternating DC Bias requirement. Could it be that you are making this more complicated than it is?

William

Steve,

I used a simple photodiode I picked up at Radio Shack. A 9 V battery and resistor provide suitable bias. I used an op-amp to amplify the signal but I don't think it's really necessary. A photo-darlington would probably work too.

William

William, don't know if you read through the whole thread but how am I able to see the deference between 48Hz and 60Hz? I can tell when my HD-10K is in 48Hz mode by looking and focusing just off screen. I can detect flicker that I can't see in 60Hz mode. What am I seeing and how can I tell?

I read the thread. Not sure I understand this question - you are seeing a difference in the refresh rate. What else can we tell you about this? If you want more information I think you'll need some test equipment.

William, I'd be interested in trying this with my scope and my Pearl, to determine if/when it switches to 96hz mode.

Any recommendations of a phtodiode that has the correct response characteristics for these kind of tests?

Thanks.

Steve, I would very interested in seeing your test methodology and results if you do this. I think this could be extremely interesting!

Mark,

The observed rate, using a photodiode and o'scope, changes with the input signal. If this were an artifact of the way the panels are driven, why would it change? JVC tells us the HD10K supports 48hz. I don't get what point you are trying to make here. Earlier JVC projectors did not support 48hz, only 60 hz, then later 50 and 60 hz.

A waveform includes both positive and negative swings. I think this might account for the alternating DC Bias requirement. Could it be that you are making this more complicated than it is?

William

Wm,

I thought my points were pretty clear, but let me summarize:

The frame rate and the panel refresh rate don't have to be coupled, in fact they don't even need to be driven at integer multiples of one another, although it probably simplifies the design to do so.
JVC says that they support 48hz, 50hz and 60hz. They are clearly talking about 48p, 50p and 60p frame rates. In fact there is no mention of panel refresh in any of the spec sheets that I've seen. This was my original point to William that you seem to be taking exception to for whatever reason.
Using panel refresh rates < 60 hz is problematic because of flicker. So if in fact JVC (and the MF1) are doing this they have basically substituted flicker for judder at 48p. Not the best implementation. This goes back to what I was saying earlier in the thread though that 48p (and 24p, 24PsF, etc.) implementations may not be ideal in some cases and people should keep in mind that the problems that they may encounter are not necessarily related to removing telecine judder but rather implementation problems with those specific frame rates (on their specific hardware).
Yes driving LCDs is complicated and it's not at all like driving a CRT. I'm not making it overly complicated and in fact if you do a web search on LCD Inversion schemes you'll find many if not dozens of commonly used schemes. Each has different flicker properties. Here is an example quote from a link that I will provide below that talks about the basic problem:

" it is very difficult to get exactly the same voltage on the cell in both polarities, so the pixel-cell brightness will tend to flicker to some extent at half the frame-rate. If the polarity of the whole screen were inverted at once then the flicker would be highly objectionable. Instead, it is usual to have the polarity of nearby pixels in anti-phase, thus cancelling out the flicker over areas of any significant size. In this way the flicker can be made imperceptible for most "natural" images. "


You've hooked up an oscope and are seeing a periodic waveform at exactly the frame rate. Without more information, I really don't know what this means. It could be that you're simply measuring the frame rate. It might mean that the panels are refreshed at exactly the frame rate or it might not. I truly don't know. You didn't say for example if you're measuring close to the panel and integrating many pixels or if you are looking at one specific pixel.
There are other ways to test for the refresh scheme used. Below is a link that includes test patterns for no less than 12 common LCD refresch schemes (Pixel dot-inversion, RGB sub-pixel dot-inversion, Line-paired pixel dot-inversion, Line-paired pixel dot-inversion (offset), Line-paired RGB sub-pixel dot-inversion, Line-paired RGB sub-pixel dot-inversion (offset), Row-inversion (also known as line-inversion), Line-paired row-inversion, Line-paired row-inversion (offset), Dot-inversion, Line-paired dot-inversion, Line-paired dot-inversion(offset)).

Interestingly enough, these test patterns work by creating perceived flicker even though the panel refresh rate (and scheme) is left unchanged.

http://www.techmind.org/lcd/



EDIT: One other thing related to your measurement that I don't understand is that you say, "A waveform includes both positive and negative swings. I think this might account for the alternating DC Bias requirement". You can't possibly measure an alternating DC Bias on the panel using a photodiode. This is because the alternating DC bias causes the same twist in the liquid crystal but of opposite direction. The net result is that both biases create the same reflectivity, which means the light measured by the photodiode should be the same. You might be able to detect the transition for a single pixel with a photodiode, but it won't be a positive and negative swing.

Steve,

I used a simple photodiode I picked up at Radio Shack. A 9 V battery and resistor provide suitable bias. I used an op-amp to amplify the signal but I don't think it's really necessary. A photo-darlington would probably work too.

William

Cool, thanks.

Steve, I would very interested in seeing your test methodology and results if you do this. I think this could be extremely interesting!

Yeah, I'll defiinitely share. It may be a bit before I have an opportunity though.. I'm demo'ing my Pearl to a couple of fellow AVS'ers tomorrow night so they can get a real life looksee, and Sunday I've got a football bash going on at my place.

And it's probably been 8 years since I've even turned my O-scope on... hopefully it still works!

Hmmm... Mark said:

" it is very difficult to get exactly the same voltage on the cell in both polarities, so the pixel-cell brightness will tend to flicker to some extent at half the frame-rate. If the polarity of the whole screen were inverted at once then the flicker would be highly objectionable. Instead, it is usual to have the polarity of nearby pixels in anti-phase, thus cancelling out the flicker over areas of any significant size. In this way the flicker can be made imperceptible for most "natural" images. "

(emphasis mine)

Interesting. I wonder if that's why the Pearl displays at 96hz when the input is 24p. that would make the "flicker rate" 48hz, which would likely make it less noticable.

Yeah, I'll defiinitely share. It may be a bit before I have an opportunity though.. I'm demo'ing my Pearl to a couple of fellow AVS'ers tomorrow night so they can get a real life looksee, and Sunday I've got a football bash going on at my place.

And it's probably been 8 years since I've even turned my O-scope on... hopefully it still works!

I think I've got you beat. It's been about 9 years since I last turned on my Tek oscilloscope :) I'll also be interested in hearing what you find. As I was alluding to Wm though, you need to make sure that you're not integrating over multiple pixels so putting the sensor by the screen rather than the by the panel will help. Even if the sensor is confined to one pixel on the screen you will still be measuring the combined results of the three RGB panels! You'll also want to put up a static display so the frame data isn't changing. If the frame data is the same you should be able to detect the rapid decline in light output as the LC is reversed followed by a rapid upswing to the same (or close to the same) bias that it was before. It's really hard to predict exactly what it will look like though because there are many different ways that the panel can be driven. The days of CRT were much more straight forward ;)

Hmmm... Mark said:


(emphasis mine)

Interesting. I wonder if that's why the Pearl displays at 96hz when the input is 24p. that would make the "flicker rate" 48hz, which would likely make it less noticable.

It would be interesting to find out exactly how Sony is driving their panels too. I've also heard that the Pearl is driven at 96hz, but I don't know for a fact that this means that they are refreshing the panel 4 times for each frame. It's likely that this is the case though. 24hz is getting slow and there could also be problems maintaining the sample and hold cell voltage on the pixel for this long. It might even degrade the liquid crystal by maintaing the twist for too long.

People talk about SXRD showing less motion blur in sports than DILA, but another important result is that the fast LC response times probably allow the panel to be refreshed more often which helps to reduce flicker. It also may imply that JVC wasn't able to support 24p properly without redesigning their panels for faster response times.

I hate to use all these "may" and "probably" terms, but as I mentioned there are a lot of possible variations in how panels are driven and I've never heard any specifics on exactly how this is accomplished in either DILA or SXRD. It's likely that this is proprietary and unique to each version of the technology because it's not dependent solely on the LCOS device physics.

Dear gentlemen,

I think you speak about 2 different think, resp. there are 2 different communities of users:
1. cinema (i.e. moving pictures) watchers (who bother often about judder)
2. HTPC (i.e. still picture of computer desktop/application) watchers (who suffer from flicker)

Let me explain:

As I described in my post the problem IS NOT in:
- JVC projector
- 24p/48p signal

Because THE SAME PROBLEM OCCURS on:
- my new Panasonic PT-AX100
- with 50p signal

I see mismatch between you in the following:
- cinema watchers do not experience flicker at 50Hz or lower, just because the flicker is much LESS perceivable at moving pictures or/and colour-changing pictures. In moving pictures, human eye even does not strongly perceive flicker at terribly low 24Hz in old theaters.

- on the contrary William (I, and other HTPC users) do experience the flicker on OUT HTPC because we watch long on the MONOCOLOUR and NON-MOVING pricture (i.e. the blue windows desktop, or white word processor). At this situations eye is much more sensible, up to the 60-70 Hz.


I think that those of you who argue with william HAS NOT TRIED the 50Hz HTPC blank screen. Because if you would, you would have to see the flicker, at 24,48 and 50 (at 60 it is much less evident - but it is, in the night when the screen shines strongly to your eye).

PLEASE PLEASE PLEASE TRY IT ! - at any LCD projector. at 48 or 50 Hz - it doesnt matter. The flicker is there.

Note - i do not know why the projector does not refresh LCD more than the 50Hz which it receives from PC, but it does not, just look)


ps. - many of you discussed the difference between refresh, sampling and other rates. This is however NOT RELEVANT to HTPC screeen STILL PICTURES. There are now moving pictures. There is just the REFRESH RATE that the digital video processor sends to the output. My iMac mini allows full scale of refresh rates up to 100 Hz, but for some reason the compatible refresh rates for my LCD projector on HDMI input is only up to 60Hz (through VGA though I can go up to 100Hz)

PLEASE DO NOT ARGUE WITH ME - TRY THE STILL PICTURE AT 50Hz ON ANY HTPC+LCDprojector combination.

There was a question what was measured by the Osciloscope.

Actually I do not care what does it measure (refresh, frame, etc) , or what couse it, THE ONLY IMPORTANT THING IS that just as the osciloscope "sees" waves of changes 48 times a second, just so DOES SEE it our eye.

And the fact is, that 48 or 50 is a TIRING for our eye (maybe also for the osciloscope:)).

Whether the single cell refresh at 1000Hz I do not care, the overall impression from the picture on the oscilopscope (and my eye) is that it changes only at 48/50 Hz and that is low and tiring.

The question than is, how can William, I, and others CAN:

>>> INCREASE "The What the osciloscope (and our eye) measures" up from 48/50Hz to 70-80Hz.<<<


ps. I found that at my Panasonic the VGA output support various refresh up to 87Hz - here are complete specs:
panasonic.co.jp/pavc/global/projector/products/ax100/specifications.html

But on the preferred HDMI it supports only 48-60Hz. Why? What to do with it? How to increase it.

ps. Guys, you who think the 48/50p is not the "real" LCD refresh. What is the real one you guess? Because to my amateur calculations it has to be less than 100Hz with this slow technology (for ca. 10ms on-off time = at 100Hz there would be no lumen output given the technology as it was explained here)

IIRC, all of the current crop of processing chips (Realta/Reon, VXP, and ABT) can, in theory, output the original 24 frames from a 60i signal. However, apart from the DVDO VP50 (which uses ABT), there are implementation issues.

I have been reliably informed that there is currently no implementation of the Realta in a shipping product which achieves 24p from 60i. Similarly, the Crystalio II currently has issues with the same thing (Pixel Magic Systems are working on a fix).

My understanding (please correct with evidence if wrong) is that neither the Mitsubishi HC5000 (Reon VX chip) nor the Sony Pearl do this. i.e. Both projectors deinterlace 60i to 60p and display the reconstructed frames in the 3:2 telecine pattern. IIRC, the HC5000 will accept 24p and 48p signals, and the Pearl will accept 24p.

Cheers for the informative post Viruskiller, its been hard to nail down the capabilities of the different 1080p scalers on the market in regards to this. I don't think having a 1080p24 display and source is going to necessarily get rid of judder. At the What HiFi show at the weekend, they had the pioneer blu ray player connected up to the pioneer 5000ex (both of which display at 24fps or multiples thereof, and the people there seemed to think both were running this) but there was still judder on the blu ray demo disc. The only perceivable difference I could see from other HD displays doing variations on 1080p/i, was that the judder was more uniform. When viewing HD DVD on the HT5000 next door, I noticed less judder on panning shots. Obviously, the processing in the HT5000 is going to be top-notch but none the less I found this suprising. All I can say, is that the judder on the 50" Pioneer did look different but was more distracting than what I saw on the 12' HT5000! I got to the sim2 demo early and they let me watch about half an hour of MI3, so not just the action sequences, and went straight to the Pioneer booth after. The only answer this leads me to is that the processing plays a more important role in reducing judder than native display capabilities. Anyone got any explanantions for this, or experienced similar?

...I don't think having a 1080p24 display and source is going to necessarily get rid of judder. At the What HiFi show at the weekend, they had the pioneer blu ray player connected up to the pioneer 5000ex (both of which display at 24fps or multiples thereof, and the people there seemed to think both were running this) but there was still judder on the blu ray demo disc. The only perceivable difference I could see from other HD displays doing variations on 1080p/i, was that the judder was more uniform...
I have a VP 50 running at 48Hz and it does get rid of the 60Hz judder problem (small judders followed by a big judder) with film based source. The small judders are inherent in 24fps film.

Can the Pioneer BD player output 1080p24? If it doesn't then it's likely the feed was actually 1080i60 and the 5000ex (which has very good processing) was converting it to 1080p60 and displaying it at 60Hz.

The Sim HT5000 uses a "DigiOptic Image Processor", but I'm not sure what chip is inside it.

IIRC, all of the current crop of processing chips (Realta/Reon, VXP, and ABT) can, in theory, output the original 24 frames from a 60i signal. However, apart from the DVDO VP50 (which uses ABT), there are implementation issues.

I have been reliably informed that there is currently no implementation of the Realta in a shipping product which achieves 24p from 60i. Similarly, the Crystalio II currently has issues with the same thing (Pixel Magic Systems are working on a fix).

My understanding (please correct with evidence if wrong) is that neither the Mitsubishi HC5000 (Reon VX chip) nor the Sony Pearl do this. i.e. Both projectors deinterlace 60i to 60p and display the reconstructed frames in the 3:2 telecine pattern. IIRC, the HC5000 will accept 24p and 48p signals, and the Pearl will accept 24p.

Interesting, thanks. I think I'll spend some time today taking a 2nd look at the DDE setting to see what it's doing on my Pearl then. I thought I had seen a difference.

Out of curiosity, what is the ABT chipset you mentioned? I don't recall hearing that. It's also generally accepted that Sony's processor is "home grown", or do you know if it uses one of the chipsets you mentioned?

Thanks.

Cheers for the informative post Viruskiller, its been hard to nail down the capabilities of the different 1080p scalers on the market in regards to this. I don't think having a 1080p24 display and source is going to necessarily get rid of judder. At the What HiFi show at the weekend, they had the pioneer blu ray player connected up to the pioneer 5000ex (both of which display at 24fps or multiples thereof, and the people there seemed to think both were running this) but there was still judder on the blu ray demo disc. The only perceivable difference I could see from other HD displays doing variations on 1080p/i, was that the judder was more uniform. When viewing HD DVD on the HT5000 next door, I noticed less judder on panning shots. Obviously, the processing in the HT5000 is going to be top-notch but none the less I found this suprising. All I can say, is that the judder on the 50" Pioneer did look different but was more distracting than what I saw on the 12' HT5000! I got to the sim2 demo early and they let me watch about half an hour of MI3, so not just the action sequences, and went straight to the Pioneer booth after. The only answer this leads me to is that the processing plays a more important role in reducing judder than native display capabilities. Anyone got any explanantions for this, or experienced similar?

Yes, from my POST (http://www.avsforum.com/avs-vb/showthread.php?p=8810075&&#post8810075) in another thread:

Framerate Judder: "Jumpiness" casued by the FRAMERATE of the picture. Typically ascribed to 24p film source material. Natural result of framerate selection of original film footage.

3:2 Cadence Judder: Uneven "Jumpiness" caused by the 3:2 frame cadence introduced by telecive, that is having to pack 24p in a 60i container. Typically ascripbed to film material played back on a video display with a 60hz framerate. NOT the natural result of the original framerate selection, but rather an anomaly caused by film sourece packed in a video delivery mechanism.

Flicker: Visually dectable "strobing" effect wehn the refresh rate of the playback device is below the visible threshold. Often visble on slow-resfresh CRT's (i.e, < 60hz).

ABT = Anchor Bay Technologies, as used by DVDO in their external video processors. I believe that Denon has now adopted ABT chips for their upscaling DVD players.

Sony's stuff is home grown.

Viruskiller - the Pioneer and Sony Blu Ray players are the only ones that do 1080p24 so far.

scaesare - I understand the difference between the types of judder that are present in HD material, I was just suprised that the processed image of the HT5000 was considerably smoother than the Pioneer running at the native resolution and framerate. The reps did assure me it was 1080p24 on both player and screen. I am sure that the judder on the Pioneer was whatever was present on the disc, it still wasn't as smooth as the sim2.

Viruskiller - the Pioneer and Sony Blu Ray players are the only ones that do 1080p24 so far.


There is hope that 1080p24 will be enabled in future players and it is included as an optional video format in HDMI 1.3 (maybe earlier versions of the spec too, but I have read only the 1.3 version). There is also a likelihood that HD-SDI mods will become available for those players like the Tosh HD-A1 which will enable support for 1080p24 - although this means that the rest of the chain needs to properly support the format too.

ABT = Anchor Bay Technologies, as used by DVDO in their external video processors. I believe that Denon has now adopted ABT chips for their upscaling DVD players.

Sony's stuff is home grown.



-sc

Interesting, thanks. I think I'll spend some time today taking a 2nd look at the DDE setting to see what it's doing on my Pearl then. I thought I had seen a difference.

Out of curiosity, what is the ABT chipset you mentioned? I don't recall hearing that. It's also generally accepted that Sony's processor is "home grown", or do you know if it uses one of the chipsets you mentioned?

Thanks.

Well, I have to reverse my previous position. While I originally thought I had seen 3:2 judder on film credits disappear w/ DDE enabled, I must have looked to quickly. I spent some time with this last night, and it appears that the Pearl does NOT extracy 24p from 60i and display at 96Hz.

Too bad. Now I'm back in the market for a 24p HD DVD player. C'mon Toshiba!

Well, I have to reverse my previous position. While I originally thought I had seen 3:2 judder on film credits disappear w/ DDE enabled, I must have looked to quickly. I spent some time with this last night, and it appears that the Pearl does NOT extracy 24p from 60i and display at 96Hz.

Too bad. Now I'm back in the market for a 24p HD DVD player. C'mon Toshiba!

Speaking of Pearl specific issues, here is a link about possible 24p image tearing with the Pearl. If this problem holds up, it would be interesting to find out if this is due to the VP30, unsynchronized panel refresh rates to frame rates, non-source locking (24p output frames are not locked to the input 60i) or some other artifact.

http://www.avsforum.com/avs-vb/showthread.php?p=8847647&&#post8847647

I won't take the main subject too deeply as many other users here, but I would like to sugest you to try Crystal Player (http://www.crystalplayer.com/index.php?page=downloads). It has a function called 'Motion Morphing' that insert new frames between the originals. The result is very interesting, and can run with any kind of video (except DVDs, but for this we have the amazing TrimensionDNM in WinDVD 6 or above). But I think that this option still have a little problem with artefacts. You can try and post here your impressions. :)

I also would like to invite you to visit this thread: http://www.avsforum.com/avs-vb/showthread.php?p=9046177

[]'s!

I have noticed in some slow panning scenes a sort of pixelation in the image when the camera is turning or moving. Is this the so called 'judder' discussed here happening when there is 2:3 cadence? I have seen it on projector and recently i saw the same in a movie theater, in the movie 'Departed'. So i am a bit confused since it is supposed to have been shot on film at 24fps and shown by the same frame rate in the movie theater.

Ramin

--------------------------------------------------------------------------------

Guys ! Motion judder from film sources has nothing to do with a frame rate conversion ! Want me to prove it ??? 24 film images are sped up to 25 (hence the 4.16%) and matted onto 2 identical fields for a PAL 50Hz display. Meaning there are 2 identical images for each frame of PAL video ... essentially what people try and achieve when they play film at 48Hz on a computer monitor. Notice how the judder remains !?

The problem is not the frame rate conversion. The problem is the low intitial sampling rate at the source ... 24 captures per second for film !

Playing film at 1080/24fps will do absolutely nothing to eliminate motion judder. You are still seeing 24 individual images per second. Playing a film at 1080/72fps will not help in the slightest. You are STILL seeing 24 images per second but 3 copies of each.

Imagine film capturing the world at 72 images per second ? Each moving object is sampled 3x as often so the amount it has to "jump" between frames is smaller and thus it's easier for your brain to believe it's a continuous sequence rather than a series of static images.

Are you guys with me ?

All this talk about 1080/24fps ... ! Gees !

There is only ONE method out there that can increase the number of percieved samples per second of the original source even though the source frequency remains the same.

Search my name and see the massive post I wrote about DNM in another thread.

There's a 3:2 test scene in the new HD DVD Digital Video Essentials. The restaurant sequence (Chapter 6) where the camera passes over the food at the start of the sequence. With 1080i/p/60 3:2 judder is very noticeable but at 1080p/24 the motion is much more fluid.

This sequence is designed to show the difference with displays that perform 3:2 conversion and then use a frame rate that is a multiple of 24.

Dazzer

All this talk about 1080/24fps ... ! Gees !

There is only ONE method out there that can increase the number of percieved samples per second of the original source even though the source frequency remains the same.

Search my name and see the massive post I wrote about DNM in another thread.

There is a difference between the temporal limitations of film motion shot at 24 frames and the judder induced by 3:2 playback process at 60hz.

In my system 24 frame HD DVD film material is much more fluid with a 1080p/24 signal than a 1080p/i/60 signal. :)

Dazzer

Dazzer is most certainly correct.

3:2 cadence judder is distinctly different from 24p framerate judder. The 3:2 cadence judder is particularly noticeable because the time interval is not "even" between successive frames, hence the irregular movement catches the eye.

Certainly 24p judder is distracting as well, however while we CAN eliminate 3:2 judder with the right equipment, there's much less we can likely do to influence the framerate the industry has adopted for shooting movies.

--------------------------------------------------------------------------------

Guys ! Motion judder from film sources has nothing to do with a frame rate conversion ! Want me to prove it ??? 24 film images are sped up to 25 (hence the 4.16%) and matted onto 2 identical fields for a PAL 50Hz display. Meaning there are 2 identical images for each frame of PAL video ...


True for PAL sources, not true for NTSC where the framerate standard is 60hz.

This whole thread topic is about NTSC and not PAL however and with NTSC 2:3 pulldown is required to achieve 60hz from 24fps film sources...


Playing film at 1080/24fps will do absolutely nothing to eliminate motion judder. You are still seeing 24 individual images per second. Playing a film at 1080/72fps will not help in the slightest. You are STILL seeing 24 images per second but 3 copies of each.


This is true. 1080p24 won't help with the inherent jerkiness from slow camera speeds, but this is a different artifact than telecine judder from 2:3 pulldown (which is what 1080p24 corrects).

Wow, this thread's been around for a while. I was looking at my post on the first page, and I noticed that I said that the Pearl deinterlaces 1080i and displays it judder free. I was wrong about that. I believe I made that comment shortly after the Pearl's release (and prior to any good reviews).

I will update my post accordingly. I always wondered where all the new guys were getting the idea that the Pearl eliminated judder on 1080i content, and now I'm wondering if maybe I started all this confusion. Sorry!

If a digital display refresh rate is at only 24Hz it will flicker the same as film (unless the pane has a super slow decay rate). So you need to up the refresh rate to at least 48Hz to avoid flicker. I can even tell when my display is in 48Hz mode by looking and focusing just off screen. I can detect flicker that I can't see in 60Hz mode.

This is incorrect. Digital displays do not "blank" between frames. A digital panel running at 24, 48, or 72 hz will look exactly the same for a 24hz source.

Flicker is caused by the blanking if the light source. In a mechanical projector, the shutter blade blanks the gate during pull down. a single blade shutter is not usable due to extreme flicker problems. A two blade shutter at 48hz is acceptable and the industry norm but still produces visible flicker at 48hz. There rae also three blade shutters the produce a 72hz blanking rate. These are used on high perfromance projectors but they have the another problem. They eat up light so thay are not typically used in a commercial theatre.

A CRT forms an image by exciting phosper dots on a galss surface. As the dot is moving, it can only be on one phosper dot at a given instant. The phosper has chemical persistance which keeps it lit until the beam returns on the next frame. Due to the persistance, a 24hz CRT is not practical. So the minimum acceptable frame rate is 48hz.

A digital display panel does not blank. A pixel remains lit until the program material requests it to change. It doesn's matter if the video signal is being blanked between frames. The pixels just stay where they are. On a solid white noise free signal, there would be no pixels changing state

This is incorrect. Digital displays do not "blank" between frames. A digital panel running at 24, 48, or 72 hz will look exactly the same for a 24hz source.

So, you're saying that Motion Estimation tecnologies can't be useful with LCD displays?

Flicker is caused by the blanking if the light source. In a mechanical projector, the shutter blade blanks the gate during pull down. a single blade shutter is not usable due to extreme flicker problems. A two blade shutter at 48hz is acceptable and the industry norm but still produces visible flicker at 48hz. There rae also three blade shutters the produce a 72hz blanking rate. These are used on high perfromance projectors but they have the another problem. They eat up light so thay are not typically used in a commercial theatre.

Hmmm... Why shutters run in this way? We can make a shutter that blanks all the time, and opens only when the frame is in the right postion. It would be better, don't you think?


A CRT forms an image by exciting phosper dots on a galss surface. As the dot is moving, it can only be on one phosper dot at a given instant. The phosper has chemical persistance which keeps it lit until the beam returns on the next frame. Due to the persistance, a 24hz CRT is not practical. So the minimum acceptable frame rate is 48hz.

Yeah, but 48 Hz in a CRT is terrible. Looks like a strobe, hehehe. Even 60 Hz is very bad. I think that 75 Hz is the minimum, and 100 Hz the perfect one.

So, you're saying that Motion Estimation tecnologies can't be useful with LCD displays?
You said the key word estimation. Yes motion interpolation can be done but the technology is not perfect especially on a consumer price point.


Hmmm... Why shutters run in this way? We can make a shutter that blanks all the time, and opens only when the frame is in the right postion. It would be better, don't you think?
I don't get what you are suggesting. Anytime you chop the light source you will get flicker. The faster the chop rate, the less the flicker. You can't have a mechanical film projector work without a shutter. Now you could use a solid sate LCD light valve as a shutter but it would leak more and still have the flicker problem. Oh and keeping the LCD panel cool enough with a 4K lamp behind it will also be a problem we don't need.


Yeah, but 48 Hz in a CRT is terrible. Looks like a strobe, hehehe. Even 60 Hz is very bad. I think that 75 Hz is the minimum, and 100 Hz the perfect one.
48hz is the minimum acceptable shutter or CRT refresh rate and yes it still flickers. The only way theaters get away with it is due to the low light levels - 16fl max.

At this point in time I don't think we will ever see more use of faster film speeds outside of special effects and scientific uses as it is today. Any jump to faster capture rates for entertainment will be digital based. Film is far too expensive to run faster than 24fps in light of the emerging digital technologies. the DCI spec does allow for 48fps at 2K resolution. 4K is limited to 24fps due to forseeable technology limitions.

Flicker is caused by the blanking if the light source.


This thread is indeed getting circular. As was discussed in this thread way back when there are many causes of flicker and it's really dependent upon the technology itself.

A digital display panel does not blank. A pixel remains lit until the program material requests it to change. It doesn's matter if the video signal is being blanked between frames. The pixels just stay where they are. On a solid white noise free signal, there would be no pixels changing state

This is a case in point. With LCD flat panel technology the LCDs are always changing state because the LC molecule needs to be inverted. If you have a PC with a flat panel display you can exagerrate the flicker with test patterns that show the type of inversion being used. Flat panels typically don't invert the entire screen at once but a subset such as RGB or every other pixel (determined by the inversion scheme).

This was discussed earlier in this thread and I think Wm pointed out that he thought the earlier 2nd gen DILA panels refreshed at 48 and 60hz. From what I gather now I think this is probably true because I don't think that the earlier panels performed refresh/inversion so in fact the refresh rate and frame rate are one in the same (although I don't know this for a fact). The 3rd gen panels in the RS1 on the otherhand invert the panel twice for each frame. I tried running the inversion test patterns on the RS1 but I didn't detect any inversion flicker so I think the entire panel is inverted at once.

You said the key word estimation. Yes motion interpolation can be done but the technology is not perfect especially on a consumer price point.

Hmmm... We missed the point. I thought that you said that because you think that TrimensionDNM (or tecnologies like that) doesn't help when we use LCD displays. Did you think that? I'm asking because I really don't know the answer.

I don't get what you are suggesting. Anytime you chop the light source you will get flicker. The faster the chop rate, the less the flicker. You can't have a mechanical film projector work without a shutter.

Maybe I said bullsh... Sorry. Let me try once more. The flicker is caused by the blank (black) 'images', that occurs when the frame is moving (changing) behind the shutter. When the frame stops right in front of the center of the projector lens, then the shutter 'opens' and the light goes through, right? So, how can we decrease the flicker if we can't show the frame when it's moving? What I'm trying to say is that the time between each frame is always the same, and we can't never change this without changing the framerate. That's why in the cinema flicker has something to do with the framerate.

This thread is indeed getting circular. As was discussed in this thread way back when there are many causes of flicker and it's really dependent upon the technology itself.



This is a case in point. With LCD flat panel technology the LCDs are always changing state because the LC molecule needs to be inverted. If you have a PC with a flat panel display you can exagerrate the flicker with test patterns that show the type of inversion being used. Flat panels typically don't invert the entire screen at once but a subset such as RGB or every other pixel (determined by the inversion scheme).

This was discussed earlier in this thread and I think Wm pointed out that he thought the earlier 2nd gen DILA panels refreshed at 48 and 60hz. From what I gather now I think this is probably true because I don't think that the earlier panels performed refresh/inversion so in fact the refresh rate and frame rate are one in the same (although I don't know this for a fact). The 3rd gen panels in the RS1 on the otherhand invert the panel twice for each frame. I tried running the inversion test patterns on the RS1 but I didn't detect any inversion flicker so I think the entire panel is inverted at once.

DLP is not static either. The mirrors are constantly under motion. However this is not the same thing as blanking. A digital panel does not blank out between video frames.

Now the new Sony broadcast grade video manitor does in fact blank. But this is a new trick to get deeper blacks and requires a very high performance panel.

Hmmm... We missed the point. I thought that you said that because you think that TrimensionDNM (or tecnologies like that) doesn't help when we use LCD displays. Did you think that? I'm asking because I really don't know the answer.



Maybe I said bullsh... Sorry. Let me try once more. The flicker is caused by the blank (black) 'images', that occurs when the frame is moving (changing) behind the shutter. When the frame stops right in front of the center of the projector lens, then the shutter 'opens' and the light goes through, right? So, how can we decrease the flicker if we can't show the frame when it's moving? What I'm trying to say is that the time between each frame is always the same, and we can't never change this without changing the framerate. That's why in the cinema flicker has something to do with the framerate.

All I'm saying is that motion estimantion is just that. It will never replace a higher capture rate in the first place. It is no different then making stereo from a mono source. 5.1 from a stereo source, or HDTV from an SD source.

While the time between frames is determined by the frame rate, you can flash each frame multiple times per each frame time slice. A two blade shutter flashes each film frame twice. A three blade shutter three times. This does reduce flicker. A single blade shutter does not work. The only problem with a three blade shutter is loss of light output which is not economical in a average commercial theater. High end screening rooms however most always have three blade shutters. The advantages of a three blade shutter are as old as Hollywood it's self. This is hardly a new concept.

All I'm saying is that motion estimantion is just that. It will never replace a higher capture rate in the first place. It is no different then making stereo from a mono source. 5.1 from a stereo source, or HDTV from an SD source.

Not the best analogy. I really doubt that you can tell the difference between TrimensionDNM and a real 60 FPS movie in a blind test. But nevermind. Let's go to the main point that I'm trying to ask you all this time: do you think that TrimensionDNM (or tecnologies like that) can work with LCD displays? When I say 'work' I mean do the main propose.

While the time between frames is determined by the frame rate, you can flash each frame multiple times per each frame time slice.

Why? And how? If a frame is moving (and it is), how can we blink it more than one time? IMO, at each flash we're going to see the same frame in a different position, and it's bad. Besides, if we can't show the frame while it's moving (i.e. changing to another frame), how can we reduce the flicker without changing the framerate?

Delerue,

Film projectors use an "intermittent" mechanism that advances the film to the next frame in 1/4 of the frame rate...the rest of the time the frame is stationary in the gate, not moving.

While stationary, the frame can be flashed as many times as you want, tho as glimmie states, doing more flashes decreases the amount of light reaching the screen. At 24fps the film advance time is about 10.4ms of the 41.6ms frame rate.

Cheers,

Film projectors use an "intermittent" mechanism that advances the film to the next frame in 1/4 of the frame rate...the rest of the time the frame is stationary in the gate, not moving.

Hmmm... Now things are getting sense. :)


While stationary, the frame can be flashed as many times as you want, tho as glimmie states, doing more flashes decreases the amount of light reaching the screen.

I only can't understand why 'flashing' the light (i.e. rotating two or more sttuters) results in a better experience. I mean, flicker is caused by a long period between the change of a frame to another, right? But how can we decrease the flicker only 'flashing' the light? Keeping the frame lightened during all the time while it's stationary in the gate would a better choice, don't you think?

At 24fps the film advance time is about 10.4ms of the 41.6ms frame rate.

I'm not doubting you, but as I said before I'm writing an article about the myth of 30 FPS human eye limitation, so if you can provide to me the source of these informations would be very nice.

Thanks
;)

Hmmm... Now things are getting sense. :)



I only can't understand why 'flashing' the light (i.e. rotating two or more sttuters) results in a better experience. I mean, flicker is caused by a long period between the change of a frame to another, right? But how can we decrease the flicker only 'flashing' the light? Keeping the frame lightened during all the time while it's stationary in the gate would a better choice, don't you think?



I'm not doubting you, but as I said before I'm writing an article about the myth of 30 FPS human eye limitation, so if you can provide to me the source of these informations would be very nice.

Thanks
;)

Not sure if this is relevant -

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

I'm not doubting you, but as I said before I'm writing an article about the myth of 30 FPS human eye limitation, so if you can provide to me the source of these informations would be very nice.

Thanks
;)

I'd humbly suggest you get your head around flicker threshold before you start taking potshots at standard frame rate "limitations".

If you want to see what higher frame rate capture looks like then just watch some 60i field based material on a progressive display . It will likely run a bob type deinterlace on the fields giving you nominal 60fps display.

Then decide if its better looking than 24fps material.

Motion interpolated material is riddled with artifacts and is soft.

If you want to see what higher frame rate capture looks like then just watch some 60i field based material on a progressive display . It will likely run a bob type deinterlace on the fields giving you nominal 60fps display.

Then decide if its better looking than 24fps material.

Or you could compare 24p material to 60p material, instead of to 60i material. That way, de-interlacing is not a factor.

ABC, FOX, and ESPN show sports at 720/60p, for example. For comparison, you could watch a similar sports scene in a movie that was shot at 24p.

Or you could compare 24p material to 60p material, instead of to 60i material. That way, de-interlacing is not a factor.



My point is the capture interval is the same for 60i vs 60p unless the deinterlace does something very very silly. You don't have to do anything remotely exotic to see the difference.

I like 24p capture the only time you get objectionable judder( strobing is actually a better description) is when the shutter angle is unusually small and the capture interval minimises motionblur.

I hate the way this issue gets over mystified. Its the usual numbers game over-simplification. I file this one with the poor deluded souls clamouring for 4k and 10bit.

Delerue,

Film projectors use an "intermittent" mechanism that advances the film to the next frame in 1/4 of the frame rate...the rest of the time the frame is stationary in the gate, not moving.

While stationary, the frame can be flashed as many times as you want, tho as glimmie states, doing more flashes decreases the amount of light reaching the screen. At 24fps the film advance time is about 10.4ms of the 41.6ms frame rate.

Cheers,

Thank you! Much better explination.