I was looking at a Pioneer BDP-HD1 player in the store a few weeks ago which was connected to Pioneer Elite 50" PRO-FHD1 plasma that accepted 1080p/24 and displayed I believe @ 72Hz. I changed the settings from 1080i to 1080p/60 to 1080p/24 numerous times throughout the opening sequence of the Blu-Ray Disc "Gone in Sixty Seconds" and couldn't see any difference.

I have Sony SXRD XBR1 which is only 1080i anyway so I can't benefit from it, but many people are saying that 1080p/24 is a real good feature. What exactly does judder look like? I sort of understand it from a frames and cadence perspective, but don't know what to look for.

Is it possible that that Pioneer 50" plasma actually automatically took the 1080i or the 1080p/60 and internally realized it was from 24fps source and displayed it automatically at 72hz and I got the same view as when I switched BD player to 1080p/24? If not, then I just didn't see any difference.

What would be good source material to test this on? I thought that slow pans left and right, like what was in the title sequence of "Gone In Sixty Seconds", was what I was looking for for waviness or something like that? Is it more noticeable in vertical scrolling, like credits? I've turned off the 3-2 pulldown in my old DVD player and watched credits and know how bad that looks, but I think this is a different thing, right?

When you watch movies in a movie theater, what you're seeing is flashing in front of your eyes in 24 frames per second. We're talking about 35mm/70mm here, not digital cinema or Imax. So that's the level of quality being strived.

I don't know how many frames DCinema does it but Imax I believe does it at 30 fps.


fuad

When you watch movies in a movie theater, what you're seeing is flashing in front of your eyes in 24 frames per second. We're talking about 35mm/70mm here, not digital cinema or Imax.

"A typical 35mm movie projector has a two blade shutter (some have three), which at 24fps gives 48 flashes per second, with the shutter open for about 60-70% of the time."

http://66.102.7.104/search?q=cache:HqKsqcy-zCIJ:www.blue-room.org.uk/index.php%3Fshowtopic%3D7810%26amp%3Bmode%3Dthreaded+24fps+% 2235mm+movie+projector%22&hl=en&gl=us&ct=clnk&cd=3

After watching a lot of 24fps movies whether SD or HD at home on 60Hz, your eyes gets trained and used to the judder --- it shows up as subtle changes in the speed of people walking at a constant rate, but the mind will compensate and smooth it out after a while.

A person who watches very few movies on TV but lots of cinema usually picks up this judder immediately.

I think after watching the 24fps again a while, if the connection was done correctly, the movement will be smooth again and it will be easier to pick up differences if you go back to the 60Hz display.

I was looking at a Pioneer BDP-HD1 player in the store a few weeks ago which was connected to Pioneer Elite 50" PRO-FHD1 plasma that accepted 1080p/24 and displayed I believe @ 72Hz. I changed the settings from 1080i to 1080p/60 to 1080p/24 numerous times throughout the opening sequence of the Blu-Ray Disc "Gone in Sixty Seconds" and couldn't see any difference.

If PureCinema was turned on on the FHD1 even 60Hz movies will be shown at 72Hz.

If PureCinema was turned on on the FHD1 even 60Hz movies will be shown at 72Hz.

I'm not very familiar with the display and didn't adjust the TV's settings at all, but I suspect this setting was enabled because I watched through a number of sequences that were slow even pans left and right and saw no difference between the 3 settings on Pio BD.

Hmm, next time I'll check out that setting. On the one hand, I'd like to know how much better it can be, but on the other hand, I really don't want to see it get much better as to ruin my current viewing :( You know, like the first time someone shows you how to see rainbows on DLPs and ruins them for you...

If PureCinema was turned on on the FHD1 even 60Hz movies will be shown at 72Hz.
Advanced PureCinema :

http://img477.imageshack.us/img477/1188/advancedpurecinemaoy2.jpg

After watching a lot of 24fps movies whether SD or HD at home on 60Hz, your eyes gets trained and used to the judder --- it shows up as subtle changes in the speed of people walking at a constant rate, but the mind will compensate and smooth it out after a while.

And therein lies my problem. While I knew there was motion judder with the Toshiba HD-A1 due to limiting output to 1080i, and I could tell you that watching on a 92" screen with the enhanced resolution of HD DVD this judder was rather more noticeable than other sources, I have not been too bothered by this artifact until now. The "problem" is 1080p. First the Samsung BD player , and now the PS3, playing in 1080p are so much smoother on motion than the HD-A1 that watching a HD DVD has become downright irritating. Toshiba can't get the HD-XA2 out fast enough for me.

Thanks for the diagram, DanielTS!

I own Pioneer and never really understood exactly what the PureCinema settings did. So, it's related to the pulldown, right?

I've tried Off, Auto 1 and Auto 2 on both the Pioneer player and display and really could never tell any differences. Maybe I don't know what to look for, but it would be a big help to me if someone could explain what the differences are between these settings. I know it's a bit off-topic...

ss9001

And therein lies my problem. While I knew there was motion judder with the Toshiba HD-A1 due to limiting output to 1080i, and I could tell you that watching on a 92" screen with the enhanced resolution of HD DVD this judder was rather more noticeable than other sources, I have not been too bothered by this artifact until now. The "problem" is 1080p. First the Samsung BD player , and now the PS3, playing in 1080p are so much smoother on motion than the HD-A1 that watching a HD DVD has become downright irritating. Toshiba can't get the HD-XA2 out fast enough for me.

Well. you are mixing-up two things: unless you have CRT FP and display actual 1080i, you are seeing 1080p no matter what. Judder comes from the fact, that display converts 24fps material into 30fps, necessitating repeating some frames (as it was said before). The ONLY way to eliminate judder is to reconstruct (or get it from the disk directly) 24fps (progressive), and display it at 24fps (on digital PJs it will NOT flicker), or multiples of 24 (on CRT-based displays, which do flicker at low refresh rates).
I do just that with LumagenHDP VP, which recreates 1080p/24fps from Toshiba's feed, and outputs it at 48fps for my Marquee CRT PJ. Smooth, film-like motion...

Well. you are mixing-up two things: unless you have CRT FP and display actual 1080i, you are seeing 1080p no matter what. Judder comes from the fact, that display converts 24fps material into 30fps, necessitating repeating some frames (as it was said before). The ONLY way to eliminate judder is to reconstruct (or get it from the disk directly) 24fps (progressive), and display it at 24fps (on digital PJs it will NOT flicker), or multiples of 24 (on CRT-based displays, which do flicker at low refresh rates).
I do just that with LumagenHDP VP, which recreates 1080p/24fps from Toshiba's feed, and outputs it at 48fps for my Marquee CRT PJ. Smooth, film-like motion...

So, at 60HZ one can see motion judder, but at 72 HZ on a CRT display, you can't? Just trying to understand this. I know it takes a refresh rate of 72-75 HZ on a PC monitor for the eye not to see flicker. I use a CRT display, so if the frame reconstruction is taking 1080p24, converting to 1080i60 and then displayed at 1080i72, motion judder is not an issue?

ss9001

So, at 60HZ one can see motion judder, but at 72 HZ on a CRT display, you can't? Just trying to understand this. I know it takes a refresh rate of 72-75 HZ on a PC monitor for the eye not to see flicker. I use a CRT display, so if the frame reconstruction is taking 1080p24, converting to 1080i60 and then displayed at 1080i72, motion judder is not an issue?

ss9001

Not real life example: If you are able to display 1080i at 72Hz, (without deinterlacing it), you have CRT FP (or high-end computer monitor). Neither one alters signal, so your statement ..."converting to 1080i60 and then displayed at 1080i72..." would imply external processing. In that case you just take 1080p/24fps, whether directly from disk, or reconstructed, and output it from processor at whatever configuration you need it, and your display can take. The 1080i/60 is redundant in this case.
(If you feed your processor "i" signal, it always gets deinterlaced for processing, even if you ultimately output it as interlaced).
As for outcome, yes, I think if done properly, it would work well, unfortunately not that many displays can do it. Didn't try it on mine, although I think it would be within it's capabilities, so I might someday...
As for flicker, if you have chance, experiment! I for one found to my great surprise, that 48 Hz refresh on CRT FP doesn't bother me at all. You can see very slight one on large white or bright areas only. (And flicker bothers me on CRT computer monitors sometimes even above 72Hz). All in all, I found 1080p at 48Hz on CRT FP excellent choice (for film-based material) even though many would disagree, just look in to FP Forum. :)

And therein lies my problem. While I knew there was motion judder with the Toshiba HD-A1 due to limiting output to 1080i, and I could tell you that watching on a 92" screen with the enhanced resolution of HD DVD this judder was rather more noticeable than other sources, I have not been too bothered by this artifact until now. The "problem" is 1080p. First the Samsung BD player , and now the PS3, playing in 1080p are so much smoother on motion than the HD-A1 that watching a HD DVD has become downright irritating. Toshiba can't get the HD-XA2 out fast enough for me.

Your complaint makes absolutely no sense. The judder is introduced by the mastering of 24 frames-per-second film content to 60 fields-per-second video. It applies equally whether you are using the HD-A1 with 1080i60 output or the Samsung BDP-1000 or PS3 with 1080p60 output. The conversion of 24fps to 60hz is the problem.

The way to reduce the judder (I won't say eliminate, because 24fps film has plenty of its own judder) is to either output directly at 1080p24 (Pioneer or Sony Blu-ray players) to a screen capable of displaying it at some multiple of 24hz, or to take the 1080i60/1080p60 signal and convert it back to a multiple of 24hz using a video processor. In either case, you need a display that will properly sync to the new refresh rate, and very few consumer devices will display at anything other than 60hz anyway.

If you believe that your current 1080p output is "smoother" than 1080i, that's probably your display doing a poor job of deinterlacing 1080i. But that's a completely separate issue.

You are right of course, as posted almost identical response couple of posts earlier .
What's possibly in play here, making the whole issue even murkier, is motion-adaptive deinterlacing... I don't know how well, or even if it's done in Samsung or Pioneer when they output 1080p/60, but I know from playing around with my Lumagen that it CAN do pretty good job at "smoothing-out" some of the judder, at the same time preventing flicker on CRT display. I am purist so I opted for clean 1080p/24 displayed at 48Hz, which I believe brings me as close to real cinematic experience as possible, but I have to admit that 1080p/60 from GOOD VP DOES look pretty smooth...

Kris

If you believe that your current 1080p output is "smoother" than 1080i, that's probably your display doing a poor job of deinterlacing 1080i. But that's a completely separate issue.

Perhaps that is the issue, or perhaps the HD-A1 can't play as smoothly as some of the other players. Whatever the case, every 1080p player I have put on the Sony projector (Pioneer, Samsung, PS3) are very visibly smoother on horizontal pans than the HD-A1.

My oppo outputs native PAL at 50 hz. (25 fields...) to my PJ so no judder on those discs. Definitely smoother than my NTSC discs, same player,PJ...

When you watch movies in a movie theater, what you're seeing is flashing in front of your eyes in 24 frames per second. We're talking about 35mm/70mm here, not digital cinema or Imax. So that's the level of quality being strived.

I don't know how many frames DCinema does it but Imax I believe does it at 30 fps.


fuad

I use to be a projectionist at an IMAX theatre for about 3 years, actually we use 24 fps as well, although there are some projectors capable of doing 48 fps, which are mainly used on simulation type rides, not movies generally.

Perhaps that is the issue, or perhaps the HD-A1 can't play as smoothly as some of the other players. Whatever the case, every 1080p player I have put on the Sony projector (Pioneer, Samsung, PS3) are very visibly smoother on horizontal pans than the HD-A1.

As an experiment, you could try the 1080i output of the Blu-ray players and see if they produce the same judder as the Toshiba at 1080i to determine if this issue is just 1080i in general or the Toshiba player itself.

Quick question, why would it convert to 72Hz? I appreciate that this is no longer 3:2 pulldown, there's no 3:2 pulldown, correct?

Why not just use 24Hz? Why are we adding 2 extra fields?

Why not just use 24Hz? Why are we adding 2 extra fields?With 24p display, you would see mass flickering. Ever tried to run your old CRT monitor at 24Hz?

However, we can still eliminate judder by displaying 24p content at a multiple of 24, be it 48Hz, 72Hz, 96Hz or 120Hz. With those refresh rates, there isn't any flicker.

The judder seen at 60p results because 60 is not a multiple of 24, and as a result, some frames are repeated while others are not.

Bit off topic..going back to that diagram, the left side shows what happens on PAL displays (50Hz) if I'm correct. I don't quite understand what happens there..it shows 2 frames per 1 original frame, but that's only 48, where do the other 2 come from? How's that speed-up come into picture?

Bit off topic..going back to that diagram, the left side shows what happens on PAL displays (50Hz) if I'm correct. I don't quite understand what happens there..it shows 2 frames per 1 original frame, but that's only 48, where do the other 2 come from? How's that speed-up come into picture?

It is sped up by 4%, affecting audio and makes it almost impossible to watch especially if there is a lot of music.

Here's something I posted recently in another thread that mught help explain the difference between 3:2 cadence judder, and 24p framerate judder.

The issue w/ 24p judder is that the "steps" just aren't granular enough (especially at wide viewing angles) to avoid looking jerky in some cases. For example, a 2-second pan of a skyline on a 96" wide screen. That means in that two seconds, the vertical edge of a building will travel 96 inches accross your field of vision in 2 seconds. That's 48 inches a second. That means in each of the frames on the disc, the bldg edge jumps 2 inches every 0.042 seconds. There will be some camera blur, but you can definitely detect a 2-inch jump as being non-fluid.

Hoever, because the motion and timing is consistent from one frame to the next, your brain can integrate it somewhat as "motion".

In the case of a 60hz presentation of that 24p pan, you've introduced 3:2 judder. That means that the bldg edge will jump 2 inches and stay there for 0.05 seconds, and then another 2 inches 0.033 seconds later, and then another 2 inches 0.05 seconds later, and then back to 0.033 seconds, etc.... Rinse. repeat.

So, now your eye has to deal with the motion that no longer is consistent from a timing perspective, in adddition to the inherant jerkiness of the (low) 24p framerate.

For me, the fact that an object seems to not travel consistently in terms of time/speed across the screen is distracting on my large (133" diag) screen. On my much smaller direct view (32") CRT I very rarely notice it.

Actually, with a display technology based on frame buffers (LCD and plasma), I don't know if there is any 'flicker' anymore. As I understand it, there is no vertical blank or shutter movement ---- the raster lines are no longer scanned, and for all intents and purposes, they just 'show up' and 'go away' as fast as their response time allows.

So 1080P24Hz may be ok, though I have not studied this problem in detail

Here's something I posted recently in another thread that mught help explain the difference between 3:2 cadence judder, and 24p framerate judder.

I think a lot of folks, especially those used to 60+fps games on PCs confuse cadence judder with framerate judder. Then they want 24/48/72Hz displays because they get the impression it will get rid of framerate judder when what it really does is get rid of cadence judder.

I think a lot of folks, especially those used to 60+fps games on PCs confuse cadence judder with framerate judder. Then they want 24/48/72Hz displays because they get the impression it will get rid of framerate judder when what it really does is get rid of cadence judder.

I suspect you are right... there is much confusion on this topic.

Thanks scaesare your comments are right on the money. There are two different kind of judder. As you have pointed out one is due to insufficient sampling of the camera while acquisition and another due to cadence.
For a moment if we disregard that 3:2 pulldown is not present then there are still other factors that contribute to the motion judder.
1) Insufficient frame rate (24 fps is slow frame rate no matter what cinematographers say, For things moving fast there is now we can sample a smooth motion at this slow frame rate). Typically humans can do a smooth pursuit what "scaesare" mentioned upto 100 deg/sec under normal viewing conditions. Which is not much given the seating distance that typical home theater fans have in their home theater.

2) Ambient light level. This is one of most neglected part. The average light level in a movie theater is much lower than a typical consumer front projectors and flat panel displays. So at a higher ambient light level we observe more judder.

3) Object edge definition . This is the reason why we do not get razor sharp edges in a movie. The edge outlines are blurred so that the judder is less visible. Also the depth of field in movies are in general very shallow. So only thing that stands out during a pan or zoom is the central object. Where as when things are shot in a CMOS digital camera the depth of field is not shallow due to the camera senson size. Hence everything that is in the background remains sharp and causes siginificant judder.

So it is still possible to observe significant judder if you are watching a 24 fps film material on a PJ or 72 Hz Plasma. Frame repeats on a sample and hold type of display such as plasma or LCDs do not improve the judder because we are not adding any new information.

This is already a long post. I will write the evils of 3:2 pulldown some time later :)

The way I look for judder is at the end of a movie where the credits are displayed. Without judder, you see a smooth even scroll of the credits as they move from the top of the screen to the bottom. With judder, you see some jerkiness in the movement as the letters are constantly slowing down and then speeding up as they move down the screen.

So what is required to get NEITHER kind of judder?

So what is required to get NEITHER kind of judder?You can't...

We can't change the source. The best we can do is eliminate cadence-based judder (display of 24fps content at something other than a multiple of 24), which is what most people on this forum mean when they say they want to eliminate judder.

Actually you can. Philips did it 7-8 years ago when they introduced the Natural Motion Video Engine. If we can upsample the original signal by doing motion compensated interpolation then you can reduce the amount of judder due to insufficient sampling rate as well as the judder due to the cadence.

How difficult would it be to do that? How much would it cost if implemented?

That's also what is called motion-adaptive deinterlacing which is what all higher-end Video Processors are doing (like LUMAGEN, K-Digital VP-30 VP-50, Crystalio etc).
I tested it on Toshiba's A1 (outputting 180i/60) using LumagenHDP (outputting 1080i/60 - :p YES, it sounds funny, but what it does is to first deinterlace signal, then process it and finally in this case to re-interlace it) to Marquee 8000 CRT FP. In case you wanted to ask: I used output 1080i/60 beacuse it is good combination of res. and refresh rate for this PJ. I looked qite good actually, with both kinds of judder significantly reduced.
At the end, being purist, I opted for 1080p/48 output from Lumagen as giving me, in my opinion ,image closest to actual cinematic experience (yes - including that slight flicker, specially on white/very light, large areas) :) PERFECT!

Kris

EDIT I forgot: Prices for scalers capable of doing decent job like that, start about $1500.00 or so. I got mine used (from authorised dealer) for $1000.00. Phenomenal machine, also has very extensive testing and measurement tools, and flexible adjustments to color, gamma and geometry of image. Well worth the money in my opinion.

K.

That's NOT what motion adaptive deinterlacing (or any deinterlacnig) is.

That's NOT what motion adaptive deinterlacing (or any deinterlacnig) is.

Yes, it is. The only difference in this case is that 1080p/60 generated by scaler is internally interlaced again to 1080i/60. For film material with proper settings on the scaler it produces significantly less judder than direct 1060i/60 from A1.

Kris

Framerate and/or 3:2 cadence judder as being discussed here is not related to de-interlacing artifacts.

A 1080i60 interlaced signal de-intlaced by whatever method (bob, weave, mtion adpative, motion compensated, etc...) to 1080p60 will _STILL_ have to contend with the judder introduced by trying to present 24Hz material in a 60Hz container.

There is no motion-adaptive algorithm that is good enough for framerate upconversion, IMO. Not even the slower-than-realtime ones are IMO good enough.

There is no motion-adaptive algorithm that is good enough for framerate upconversion, IMO. Not even the slower-than-realtime ones are IMO good enough.

You are of course right, I never claimed it eliminates the problem entirely, I am actually amazed how good it is, given complexity of the problem. As I mentioned before, I ultimately opted for 48 Hz refresh of 1080p signal (72 Hz is unfortunately out of range of Marquee...). Funny, the flicker doesn't bother me a bit, even though on CRT computer monitors I find 72 Hz barely sufficient... Go figure :)

Kris

I agree on 72Hz CRT. But it's not as much of a problem on some other types of displays. And I agree that framerate upconversion using motion-adaptive interpolation is a good idea, and I'd love to see a good implementation, but it's just something so complex...

I agree on 72Hz CRT. But it's not as much of a problem on some other types of displays. And I agree that framerate upconversion using motion-adaptive interpolation is a good idea, and I'd love to see a good implementation, but it's just something so complex...

Sure... there are some heavily modded 9 inchers (and supposedly 8ths too per Mike Parker), but my lowly M8000, even having MP neck boards is way too soft at the refresh rate...
Let me know if/when you are in NY area, we are on Long Island, about 25 miles east from the City. You are invited! :) I don't know, if it is "good" implementation though, I have had nothing to compare it with...
As for other displays, absolutely, LCD or plasmas don't flicker in CRT's sense, you can actually drive those at 24 Hz refresh (for film of course), and they should look OK.

Kris

So bottom line--all things being equal--can you SEE the difference beween 1080p/72 and 1080p/60 if a set is fed 1080p/24?

Wil there be a whole lot less VISIBLE judder out of the perfect multiple set?

Wouldn't 1080p/120 solve all problems because it is a multiple of 24, 30, and 60?

One of the reasons I ask this is some displays accept 1080p/24 and some do not.

Is accepting 1080p/24 and displaying it at a perfect multiple such as 48 or 72 or 120 worth purchasing?

Actually you can. Philips did it 7-8 years ago when they introduced the Natural Motion Video Engine. If we can upsample the original signal by doing motion compensated interpolation then you can reduce the amount of judder due to insufficient sampling rate as well as the judder due to the cadence.

Yep - similar to the Intervideo WinDVD Trumension system ISTR.

Both make 24/25fps "film" (or 24/25p HD video) look like conventional 50i/60i (or 50p/60p) video - with much more fluid motion.

I remember the first time I saw a Philips Natural Motion picture in a shop and it is a very bizarre sight seeing something you know was shot on film looking like video.

AFAIK this processing is still only available for SD rather than HD sources (may be wrong on this) - because of the high processing overheads - and it still mangled the picture quite badly in other ways (smearing was an obvious disadvantage) It also got very confused by very fast motion - with picture breakup artefacts appearing as it failed to correctly interpolate missing fields/frames.

That's also what is called motion-adaptive deinterlacing which is what all higher-end Video Processors are doing (like LUMAGEN, K-Digital VP-30 VP-50, Crystalio etc).
I tested it on Toshiba's A1 (outputting 180i/60) using LumagenHDP (outputting 1080i/60 - :p YES, it sounds funny, but what it does is to first deinterlace signal, then process it and finally in this case to re-interlace it) to Marquee 8000 CRT FP. In case you wanted to ask: I used output 1080i/60 beacuse it is good combination of res. and refresh rate for this PJ. I looked qite good actually, with both kinds of judder significantly reduced.
At the end, being purist, I opted for 1080p/48 output from Lumagen as giving me, in my opinion ,image closest to actual cinematic experience (yes - including that slight flicker, specially on white/very light, large areas) :) PERFECT!

Kris

EDIT I forgot: Prices for scalers capable of doing decent job like that, start about $1500.00 or so. I got mine used (from authorised dealer) for $1000.00. Phenomenal machine, also has very extensive testing and measurement tools, and flexible adjustments to color, gamma and geometry of image. Well worth the money in my opinion.

K.

Isn't the Natural Motion system more like motion compensative rather than motion adaptive - in that it interpolates extra frames? Effectively it is more akin to a standards converter (or almost identical to the ill-fated Snell and Wilcox Gazelle slowmotion interpolation device - that attempted to interpolate lower field rate slowmotion replays back up to full fluid field rate. It was tested at Wimbledon in the early 90s and got defeated by fast moving tennis balls...)

AIUI Motion Adaptive systems simply alter how they re-create missing fields/frames by switching de-interlacing algorithms - whereas motion compensative stuff actually tries to track moving picture elements and re-create missing frames/fields entirely, rather than just flipping between bob, weave etc. on a pixel-by-pixel, block-by-block or frame-by-frame basis.

AIUI Motion Adaptive devices are great for de-interlacing SD video prior to scaling - whether they contain 24p/25p or 60i/50i originated material - as scaling a de-interlaced progressive signal with high quality de-interlacing is better than scaling a poorly de-interlaced signal (which is a particular requirement for scaling the video frames that contain fields from two different film frames - which is a 3:2 artefact). However the original motion artefacts of the source material remain relatively unchanged by this process. AIUI the motion adaptive element is a de-interlacing - not a frame rate upconversion process.

The Natural Motion system actually in-betweens new frames with intermediate motion estimates. This doesn't just remove the 3:2 cadence judder (present in 60Hz but not 50Hz "film" transfers), it attempts (and often succeeds) in reducing the 24/25fps sampling judder (which is present in both 50 and 60Hz film transfers) by creating a 50 or 60Hz sequence instead.

Seems like there are a couple of issues here :

1. The inherent efficacy of 24fps motion sampling - i.e. whether the sampling rate is high enough.
2. How visible 3:2 field pull-fown or 3:2 frame repetition is when presenting 24fps material on a 60i or 60p native display. (I suspect those of us used to 2:2 50i presentation notice it more readily than those used to it?)
3. Whether presentation of a 24fps picture is better at 1:1 24Hz, 2:2 48Hz, 3:2 60Hz, 3:3 72Hz etc. on a display capable of these refresh rates.
4. Whether 24p native interconnects are a better method of carrying a 24p signal than a 60i or 60p interconnect (where 3:2 pulldown detection could still be employed to derive a clean 24p signal internally)

Strikes me that 24p interconnects mean that the receiver has less to do - and has no chance of getting 3:2 detection wrong. On the other hand you have to flip the interconnect format when you display a 60p or 60i sourced signal from the disc?

AIUI there is no reason why a 24p disc cannot be replayed on a player outputting 60i or 60p with 3:2 pulldown/repetition, which the display then detects, reverses to create a 24p 1:1 sequence which it can then display at 72Hz if required using 3:3 repetition?

sneals2000 - yes, I would agree with that naming, i might have been using it somewhat incorrectly, but semantics aside, we are talking about the same thing really. What you describe is exactly what I was talking about (less eloquently I suppose...). I meant to describe the system, that "sort-of" creates "missing" frames, using some complicated algorithm which takes into account motion information contained in originals. Of course, it is enormously complicated task, that's why I was so surprised how good was the outcome of it on relatively low-end VP, like LumagenHDP.
As for your points:
1. Agreed - but there is nothing we can do about it, I can't see that standard changed in foreseeable future, because of enormous base of installed hardware (the whole movie distribution system, on top of the movie production facilities). To make things worse, even new commercial digital PJs mostly do not accept higher refresh rates...
2. We here in US ARE accustomed to it, no question about it - it's funny, because if you show unsuspecting viewers the 24/48 Hz (film-based of course) material most of them notice that it looks "better", but can't figure out why?
3. Depends on display technology: I generally agree that "sample-and-hold" displays like LCD and Plasma, which do not refresh given pixel unless it's value changes, do not benefit from higher refresh rates of incoming signal. As experiment - if you can, set your computer's LCD display to as low refresh rate as you can: IT WILL NOT FLICKER! That of course is not true with CRTs, like my PJ.
4. 60i has been in my opinion used simply for compatibility sake: ALL US HDTV displays will accept it, hence less confusion. Displays accepting 24Hz are still rare and expensive (and were even more so when those players were on drawing boards), so designers decided - if you have one of those, you most likely have (or can afford to get) scaler/VP to do the job, which is relatively easy one.
Personally, like I stated I am purist, therefore I would love to have direct 24 Hz output... It sure gives you access to the source with minimum processing.

Kris

3. Depends on display technology: I generally agree that "sample-and-hold" displays like LCD and Plasma, which do not refresh given pixel unless it's value changes, do not benefit from higher refresh rates of incoming signal. As experiment - if you can, set your computer's LCD display to as low refresh rate as you can: IT WILL NOT FLICKER! That of course is not true with CRTs, like my PJ.


Plasmas aren't sample and hold devices though are they? They are sub-field refresh devices - and as such can flicker just as badly as CRTs - and in myriad new ways!

Have you ever looked at a plasma whilst moving your jaw (i.e. whilst eating a chocolate bar or similar) - I find I can see the picture break-up into subfields quite clearly on occasion - and I certainly see large area refresh flicker on plasmas (and you can certainly see flicker when you shoot 60Hz plasmas on 50Hz cameras...)

AIUI Plasmas achieve their grey scales by using a combination of two techniques :

1. Because an individual R,G,B plasma cell can only be "on" or "off" - and has no inherent "brightness" control in the on state - a greyscale is achieved by using sub-fields to implement a form of modulation. These sub-fields achieve greyscale by lasting for different periods of time - if you have a bright "pixel" then that cell is illuminated for all the sub-fields. If it is mid-range, then it may only be illuminated for the shorter sub-fields and not the longer ones, and it it is dark it may only be illuminated for the shortest sub-fields. (I think that rather than having each subfield the same duration, they use a binary style system - where the MSBs are represented by a longer sub-field than the LSBs)

2. However because even with sub-fields running at around 300Hz you can only achieve a reasonably coarse greyscale when displaying a 50 or 60Hz signal, additional greyscale detail is simulated by "dithering" - i.e. adding a degree of noise to minimise visual banding caused by quantisation errors. If you look closely at plasmas - especially the earlier models - you can clearly see this noise.

(One of the better performing early plasma panels marketed in the UK - an Electrohome/Delphi model had a BBC designed SDI input and panel drive circuit that increased the number of subfields and improved the quality hugely)

Out of interest - do LCDs actually alter their display refresh rate with the input refresh rate - I had a feeling they used a frame-store technique (that would allow the display refresh rate to be independent of the input?) on PCs - which is also how many cope with non-native resolutions these days (effectively using a basic scaling algorithm in the display) - though I may be wrong.

Given the poor response time of most PC LCD displays I'd expect any change in refresh rate to be masked by this anyway if they do display native input refresh rates?

sneals2000 - yes, I would agree with that naming, i might have been using it somewhat incorrectly, but semantics aside, we are talking about the same thing really. What you describe is exactly what I was talking about (less eloquently I suppose...). I meant to describe the system, that "sort-of" creates "missing" frames, using some complicated algorithm which takes into account motion information contained in originals. Of course, it is enormously complicated task, that's why I was so surprised how good was the outcome of it on relatively low-end VP, like LumagenHDP.


I guess not having seen the LumagenHDP I can't comment on how it looks but does it :

1. Make 24fps material look like fluid 60i / 60p native material (like Sports, Leno, Dancing with the Stars etc.) ? In other words does it suddenly stop looking like a movie and start looking like a soap, sport or news show?

or :

2. Make 24fps look like a less assymmetrically juddery 24/48fps presentation like you see in the cinema - with no 3:2 judder, but still with the characteristic 24p "look" of low motion sampling and thus no motion rendition above the 24fps level? (Or in other words make it look more like what we see in 50Hz land, where we have 24fps material sped up to 25fps and presented as 2:2 50i)

If 1. how does it cope on action sequences with fast movement - especially those that are not "real" but instead generated using CGI, Compositing etc. Most of the implementations of the Natural Motion type systems I've seen totally fall apart on this kind of material. (The composited sequences in the Mini chase in the original Bourne movie is one I remember trying out with Trimension in WinDVD and it not coping well)

Pioneer 1920X1080 HD Plasmas can do 1080p/72. Are there any other displays that can do 1080p/72? Some people might want to know that don't want to spend thousands on a video processor but would like to be judder free.

Plasmas aren't sample and hold devices though are they? They are sub-field refresh devices - and as such can flicker just as badly as CRTs - and in myriad new ways!

Have you ever looked at a plasma whilst moving your jaw (i.e. whilst eating a chocolate bar or similar) - I find I can see the picture break-up into subfields quite clearly on occasion - ...

Well, I don't think that those are this same things... You are talking about pixels flickering (due to partial illumination WITHIN single frame refresh cycle, generating proper brightness level) while I was talking about whole image, all pixels being turned off-on due to refreshing the whole frame according to refresh rate. I always though that they (the plasmas) were using memory buffer for storing frames, and refreshed those whenever information warranted it. I might be wrong though, never been really interested in flat panels in the first place. I agree that what you described may create those "other" ways plasmas flicker, but by definition that effect would be independent from refresh rate of inputted video, wouldn't it?
As for LCDs you are of course correct as far as their slowness goes, some of the cheap ones go into ghosting on fast moving objects (Oh! How much I hate those :mad: ), which of course would mask any kind of flicker... As for better ones, again what I always understood was that they would display (at the extreme) the static image effectively without refreshing it until it changed, therefore with no flicker, making it somewhat redundant to let's say double refresh rate of video being supplied to them. Of course in case of TV displays there is the problem of displaying (in addition to film-based material) native 60i video, and especially mixed material, where display can't really re-sync each time it detects difference, hence idea of displays refreshing constantly at 120 Hz, which can accommodate both 24 and 60 native material. Problem is I guess, you are getting into really high video frequencies, which makes the whole proposition rather expensive, I guess.

Kris

I guess not having seen the LumagenHDP I can't comment on how it looks but does it :

1. Make 24fps material look like fluid 60i / 60p native material...

Well, I would say qualified #1, although I wouldn't call it THAT fluid, it is still much better that direct 3:2 display... Can't answer your question about action scenes, I really didn't look at the whole thing that carefully... In fact, I discovered it by accident, when I forgot to switch VP and PJ from 60i to 48p output/input, and after couple of minutes of watching movie I realized that it does look different, better than displayed directly in 60i... Classic example: scrolling credits at the end, they were remarkably smooth rolling up the screen.. I was really surprised, but not to the point of considering change in my habits and preferences. I found it still interesting and somewhat amusing. Personally like I said, I prefer that "natural" film look with little 48 Hz flicker and very smooth progressive display of each frame (remember: CRTs DO display "i" signal natively). I think we are here all in the pursuit of perfection it the sense of getting as close to real cinematic experience as possible. That was one of my reasons of getting the Lumagen, other (which turned-out just as important) was it's ability to adjust gamma and color output to correct imperfections characteristic to CTR PJs.
Overall I am very satisfied with the end result,

Kris

Pioneer 1920X1080 HD Plasmas can do 1080p/72. Are there any other displays that can do 1080p/72? Some people might want to know that don't want to spend thousands on a video processor but would like to be judder free.

Don't know, but you should add following questions I guess:

Do they accept 1080p/24 ? If they don't, how do they process 1080i/60, specially film-based?

Kris

Do they accept 1080p/24 ? If they don't, how do they process 1080i/60, specially film-based?The 1080p Pioneer plasmas do, yes.

Are there any other displays that can do 1080p/72? Some people might want to know that don't want to spend thousands on a video processor but would like to be judder free.No other consumer displays. However, plenty will be announced / demonstrated at CES in January. Virtually every 2007 1080p display shipping in the 2H 2007 at >$2000 should have this functionality.

sneals2000 - yes, I would agree with that naming, i might have been using it somewhat incorrectly, but semantics aside, we are talking about the same thing really. What you describe is exactly what I was talking about (less eloquently I suppose...). I meant to describe the system, that "sort-of" creates "missing" frames, using some complicated algorithm which takes into account motion information contained in originals. Of course, it is enormously complicated task, that's why I was so surprised how good was the outcome of it on relatively low-end VP, like LumagenHDP.
As for your points:
1. Agreed - but there is nothing we can do about it, I can't see that standard changed in foreseeable future, because of enormous base of installed hardware (the whole movie distribution system, on top of the movie production facilities). To make things worse, even new commercial digital PJs mostly do not accept higher refresh rates...
2. We here in US ARE accustomed to it, no question about it - it's funny, because if you show unsuspecting viewers the 24/48 Hz (film-based of course) material most of them notice that it looks "better", but can't figure out why?
3. Depends on display technology: I generally agree that "sample-and-hold" displays like LCD and Plasma, which do not refresh given pixel unless it's value changes, do not benefit from higher refresh rates of incoming signal. As experiment - if you can, set your computer's LCD display to as low refresh rate as you can: IT WILL NOT FLICKER! That of course is not true with CRTs, like my PJ.
4. 60i has been in my opinion used simply for compatibility sake: ALL US HDTV displays will accept it, hence less confusion. Displays accepting 24Hz are still rare and expensive (and were even more so when those players were on drawing boards), so designers decided - if you have one of those, you most likely have (or can afford to get) scaler/VP to do the job, which is relatively easy one.
Personally, like I stated I am purist, therefore I would love to have direct 24 Hz output... It sure gives you access to the source with minimum processing.

Kris

This was my point Kris: motion adaptive deinterlacing is not the same thing as frame-interpolation.

Deinterlacing is a very specific thing.

Just because a frame-interpolation technique may use some alogrithms also used in deinterlacing, does not make them the same.

I have used Lumagen, Vantage HD and VP 30 extensively to evaluate video quality and have worked on motion compensated frame rate designs. I can safely state from my experience that no Motion adaptive method can remove 3:2 motion judder while displaying film (shot at 24 Hz) at 60 Hz be it interlaced or progressive. This is mathematically impossible if you use fundamentals of Signal processing.

When we display film source at 24/48/72p on a sample hold display such as LCD we cannot remove the motion judder that is present due to insufficient sampling but can definitely remove the cadence judder which some people call "judder removal". Actually after spending most of my childhood and teenage watching 2:2 (PAL) material I also disliked the cadence judder and had no idea why film looked so funny in American TV channels :)

The main problem of motion compensated frame rate upconversion is to produce a frame which is free of motion vector error artifacts. Philips DNM was somewhat effective in removing sampling judder and cadence judder by initially doing an inverse 3:2 pulldown and then upsampling the 24p to 50p/60p using motion vectors. There will be lot of 120 Hz LCD TVs hitting the market in 2008, hopefully we can then see judder free film rendition.

With 24p display, you would see mass flickering. Ever tried to run your old CRT monitor at 24Hz?


That's because CRTs only shine light where the single beam-spot is touching. So increasing the scan-rate increases the frequency of light that's hitting the screen.

With "on" bulb-technology projetors like LCD, LCOS, and DLP, there's no more or less flicker depending on frame-rate (ignoring color-wheel artifacts here). A frame-rate of 2 Hz would have no more/less flicker than a frame-rate of 200 Hz. A still image would just look "on" in both cases... with no flicker because there's no physical shutter or absence of light between frames.

With a digital projector displaying native 24p source material (like film), a native 24p signal would produce the same image on the screen as one doubled to 48Hz or one trippled to 72.

The reasons for increasing frame-rate in the display beyond the native signal's frequency is because it makes it easy to mix various frame-rates into a single stream (like 72 Hz would be a better way to blend 24 native and 60 native since it would preserve all original frames in both signals versus 48 Hz). In our new world of video, we'll be constantly switching between 24p native material and 60 Hz native material... maybe even 50Hz material. So there's an advantage to picking a single rate that the projector can stay locked onto and having a processor convert everything to this one common interval. As long as that rate is a mathematical integer-multiple of the rate you want to give the greatest pirority, then you won't degrade the image where it really counts.




When we display film source at 24/48/72p on a sample hold display such as LCD we cannot remove the motion judder that is present due to insufficient sampling but can definitely remove the cadence judder which some people call "judder removal". Actually after spending most of my childhood and teenage watching 2:2 (PAL) material I also disliked the cadence judder and had no idea why film looked so funny in American TV channels


Very true. "Judder" in technical terms still refers to the motion-strobbing seen in the native 24p image in the theater. What we're really talking about in these threads when we just say "judder" is really *cadence judder* where the temporal pattern has been currupted by 3-2 frame-repitition.

Is the Pioneer and Sony the only one`s that can output 24p ???

So far. The PS3 will be upgradable for it as well.

So far. The PS3 will be upgradable for it as well.


thanks, can the panasonic be upgraded in the future also?

thanks, can the panasonic be upgraded in the future also?Unknown, but Panasonic hasn't said anything about it.

I wouldn't bet on the current Panasonic BD player getting 1080p24 output.

One thing for everyone to consider.

Right now 1080p24 is something new and esoteric. But the 3-2 cadence is just as easy to detect in a 1080p60 signal as a 1080i60 as a 480i60. It won't be long before projectors and processors start to offer "3-2 inverse telecine reversal" for *progressive* signals... not just for "deinterlacing". In other words, a future 1080p projector could just have 72Hz frefresh rate to handle both native 24 and native 60 and then have a mode that says "display 24p film without judder" where it analyzes the video and when it see the 3-2 in a 480p, 720p, or 1080i/p stream it drops the reduntand frames, plays as 24p, and then refreshes at 72Hz or something to keep the PJ from showing a hickup when native 60Hz shows up.

I think that the new iScan will do something like this. As will future versions of the Lumigen (sp). It would be CHEAP to do once someone slaps it on a chip.

The reasons for increasing frame-rate in the display beyond the native signal's frequency is because it makes it easy to mix various frame-rates into a single stream (like 72 Hz would be a better way to blend 24 native and 60 native since it would preserve all original frames in both signals versus 48 Hz).

Can you explain why 72Hz would be a better way to blend 24 native and 60 native. Why doesn't 48Hz preserve all of the original frames?

How can 48 Hz preserve all the frames in a 60Hz native signal? obviously (math), 12 of them have to be dropped.

Naturally 48 can preserve all the original in the 24 fps signal (with an even cadence).

I thought the 1080p/24 signal was 'folded' into 1080i/60, so could you not de-construct the 1080i/60 signal back to 1080p/24 and then multiply each frame by 2 to show on a screen capable of 48hz?

I agree if the signal was native 1080i/60, but could you not de-interlace the 1080i/60 signal if it was based on 24fps film to 1080p/24 and then multiply each frame by 2 to show on a screen capable of 48hz?

You can, and it's being done all the time: I do it with my VP and CRT Projector (read few posts back, I don't want to repeat myself...), and final result is pretty good, very film-like.

Kris

I agree if the signal was native 1080i/60, but could you not de-interlace the 1080i/60 signal if it was based on 24fps film to 1080p/24 and then multiply each frame by 2 to show on a screen capable of 48hz?

Sigh... when we were talking about "native 60 Hz" material were *were* talking about native 1080i60 (or 480i60). 1080i60 created from 24 fps film is not "native 60" at all... it's native 24. yeah... you use inverse telecine to drop that back down to 24 and then go up from there to whatever scan rate you need (typically 60 in most processors but could be anthing... native 24, double 48, 72 etc.)

The problem is in things like documentaries that mix "native" 24 and 60i into a single 60Hz signal... so the deinterlacer has to switch gears back and forth... sometimes doing inverse telecine 3-2 pulldown reversal for the film stuff and then just motion-adaptive deinterlacing for the actual native 60i stuff. A 72 refresh rate would be best for stuff like that... perfectly fits the 24 fps film stuff (what you want most of all) and fits 60i native stuff pretty well too... with a few interpolated frames thrown in.

Making sense now?

dave :)

And thank you for the explanation!

Thanks for clarifying "native" issue, I thought it was pretty obvious for eonim's post, but it's important that everybody understands the difference between native 60i (live video for example) and 60i created fro 24 fps material.

Kris

Sigh... when we were talking about "native 60 Hz" material were *were* talking about native 1080i60 (or 480i60). 1080i60 created from 24 fps film is not "native 60" at all... it's native 24. yeah... you use inverse telecine to drop that back down to 24 and then go up from there to whatever scan rate you need (typically 60 in most processors but could be anthing... native 24, double 48, 72 etc.)

The problem is in things like documentaries that mix "native" 24 and 60i into a single 60Hz signal... so the deinterlacer has to switch gears back and forth... sometimes doing inverse telecine 3-2 pulldown reversal for the film stuff and then just motion-adaptive deinterlacing for the actual native 60i stuff. A 72 refresh rate would be best for stuff like that... perfectly fits the 24 fps film stuff (what you want most of all) and fits 60i native stuff pretty well too... with a few interpolated frames thrown in.

Making sense now?

dave :)

I guess I thought that somehow (although it now seems not as possible) all of the native film stuff and 60i was converted somehow into all the same type of stream that then was deinterlaced using one method. Is this not possible?

Also, if you don't mind indulging me some more, (I am sure others are interested) how does the deinterlacer know what is native 24 and 60i? Is there some code in in the signal that alerts the deinterlacer? Is the deinterlacer software on a dedicated chip or on a chip set that has other functions. Would you know what chip it is (ie the manufacturer and model?)

So then, if you don't mind indulging me some more, how does the deinterlacer know what is native 24 and 60i? Is there some code in in the signal that alerts the deinterlacer? Is the deinterlacer software on a dedicated chip or on a chip set that has other functions. Would you know what chip it is (ie the manufacturer and model?)

Excellent question. And it's what used to cost $20,000 in the late 1990's.

Basically, a computer chip looks at a series of video images and analyzes them. If it sees a "3-2" repition it makes an educated guess that it was 24 fps film and it strips out every duplicated field to get the cadence even, then it "zips" the fields together that it determined belong to a single actual "frame" of film.

Sounds complicated? It is. That's why those Faroudja line-doublers used to cost so much... they invented the process. Now that same logic is on $20 chips inside most progressive-scan DVD players and TV sets.

DVD players can do it a different way too because the actual fields on the DVD are "flagged" to tag them so, in theory, a DVD player could put them together right from the start in 480p. However, very often these flags are wrong and DVD players that work this way can create bad combing artifacts with improperly authored DVDs.

Any progressive-scan/deinterlacing chip that says "film logic" or "3-2 pulldown" or something of the sort is probably doing the proper job of identifying the 3-2 repitition when it sees it and correcting for it during deinterlacing to ensure you get the right fields zipped back together.

-dave

Hi

I have Vantage HD scaler and I just ordered Pioneer's European equivalent to PDP-5070HD plasma (which scales 1080p24 to 768x1366p24 and then converts it to 72Hz if I'm correct). Currently the Vantage cannot output 1080p24, but it might be possible in future firmwares.

I have one question about the two devices. Is it possible by video processing to slow down 50i PAL films 4% to convert the signal to 24 fps? And is the audio somehow pitched in the European 50i material (sd/hd) to match the speedup, so that even if I could slow them down, I would end up with too low tonality etc.?

Thanks

-Johannes

If you believe that your current 1080p output is "smoother" than 1080i, that's probably your display doing a poor job of deinterlacing 1080i. But that's a completely separate issue.

That is interesting. 1080p/60 actually locks in the 3:2 cadence judder. 1080i SHOULD be no worse.

What is happening if people see worse judder with 1080i than with 1080p/60?! A bob-deinterlacing artifact amplifying judder?

Gary

My bet is that most TVs sold today actually upscales the 540 lines in each field (bob) rather than interleaving them (bob).

This is most likely to be true for the 1366x768, where the 540->768 is STILL!! an upscale, while horizontal downscale has some tricks with color fidelity that can achieve wider range of color than 8bit/channel.

As for cadence detection and 24/48/72. That's still very rare for normal people, and unlikely to show up from the taiwanese and chinese HD-LCDs flooding the market today.

Excellent question. And it's what used to cost $20,000 in the late 1990's.

Basically, a computer chip looks at a series of video images and analyzes them. If it sees a "3-2" repition it makes an educated guess that it was 24 fps film and it strips out every duplicated field to get the cadence even, then it "zips" the fields together that it determined belong to a single actual "frame" of film.

Sounds complicated? It is. That's why those Faroudja line-doublers used to cost so much... they invented the process. Now that same logic is on $20 chips inside most progressive-scan DVD players and TV sets.

DVD players can do it a different way too because the actual fields on the DVD are "flagged" to tag them so, in theory, a DVD player could put them together right from the start in 480p. However, very often these flags are wrong and DVD players that work this way can create bad combing artifacts with improperly authored DVDs.

Any progressive-scan/deinterlacing chip that says "film logic" or "3-2 pulldown" or something of the sort is probably doing the proper job of identifying the 3-2 repitition when it sees it and correcting for it during deinterlacing to ensure you get the right fields zipped back together.

-dave

The DVD flags can be wrong, and that is why upscalers have to ignore these flags --- hence also why brcm's chip does so poorly with upscaling true 480i materials from DVDs.

I have one question about the two devices. Is it possible by video processing to slow down 50i PAL films 4% to convert the signal to 24 fps? And is the audio somehow pitched in the European 50i material (sd/hd) to match the speedup, so that even if I could slow them down, I would end up with too low tonality etc.?

Thanks

-Johannes

It is possible to replay 50Hz European DVDs with 24fps film material mastered as 2:2 25fps with speed-up at 24fps - but the only system I've seen that does this is a PC DVD replay application. (I think it was WinDVD) This would really have to be a player function rather than a display function - as if you tried to do it in the display you'd need a potential vast memory to buffer the incoming video to allow you to store the stuff you weren't showing fast enough (1 frame every second effectively?)

Some 50Hz DVDs are pitch-compensated so that although they play 4% faster and the tempo of music is faster, the pitch of the music remains correct. Other DVDs don't have this pitch correction (and many TV series don't) If you have perfect pitch it can be annoying - and when I visit the US I am surprised that the TV theme tunes sound to be in a slightly different tempo and key!

There is no real reason why pitch corrected 50Hz DVDs couldn't be re-corrected when played back more slowly at 24fps - though two harmonisations might not be a great way to deliver good sound quality.

Sigh... when we were talking about "native 60 Hz" material were *were* talking about native 1080i60 (or 480i60). 1080i60 created from 24 fps film is not "native 60" at all... it's native 24. yeah... you use inverse telecine to drop that back down to 24 and then go up from there to whatever scan rate you need (typically 60 in most processors but could be anthing... native 24, double 48, 72 etc.)

The problem is in things like documentaries that mix "native" 24 and 60i into a single 60Hz signal... so the deinterlacer has to switch gears back and forth... sometimes doing inverse telecine 3-2 pulldown reversal for the film stuff and then just motion-adaptive deinterlacing for the actual native 60i stuff. A 72 refresh rate would be best for stuff like that... perfectly fits the 24 fps film stuff (what you want most of all) and fits 60i native stuff pretty well too... with a few interpolated frames thrown in.

Making sense now?

dave :)

24 fps 3:3 frame repition for 72Hz will look pretty good - however 60Hz to 72Hz sounds like quite a nasty conversion to do well, and would surely require similar levels of processing as a 50 to 60Hz conversion (which are still pretty tricky to do at a broadcast price point at HD)

In Europe a lot of HTPC buffs have been using 75Hz refresh rates to display 25fps European movie releases, but 50Hz native stuff doesn't look great (and 50->75 is a nicer multiple than 60->72) We don't have the advantage of 3:2 cadence detection to flag film sourced material as well - so have to use algorithms to detect motion between fields within a frame to flip between de-interlacing modes.

Surely 120Hz is a much better bet - you can 5:5 frame repeat 24fps material and field/frame double 60i/p native stuff. (Over here 50i to 100i processing has been standard - though again far from great - for a good 10 years on CRT sets as a way of reducing 50Hz large area flicker)

I can't remember the set, but I saw an ad recently saying it was capable of 120Hz refresh. So hopefully this will be a feature we see sooner rather than later on most sets!

I can't remember the set, but I saw an ad recently saying it was capable of 120Hz refresh. So hopefully this will be a feature we see sooner rather than later on most sets!

The Brillian 1080p's accept 1080p/24 and display at 1080p/120. They also display 1080p/60 at 120fps. However, they are expensive and don't do a great job with 1080i. Supposedly, their next generation will properly deinterlace 1080i and at that point they should be a good option for someone willing to spend the money.

After watching numerous Blu Ray discs at 1080p/24, I have now noticed how "choppy" pans are on my HD DVD player. On pans with HD DVD the background is so out of focus. With the BDP-S1 it very clear and smooth.

I am very happy with my BDP-S1 and Sony "Pearl" combination.

After watching numerous Blu Ray discs at 1080p/24, I have now noticed how "choppy" pans are on my HD DVD player. On pans with HD DVD the background is so out of focus. With the BDP-S1 it very clear and smooth.

I am very happy with my BDP-S1 and Sony "Pearl" combination.

24p judder has nothing to do with focus.

24p judder has nothing to do with focus.

I know it doesn't have anything to do with focus.

When viewing panning scenes the "judder" produced by 60hz signals seems to blur the back ground to me.

The reduced judder seems to "clean" up the background.

At CES, I talked to a Panny rep for quite a white. He felt that we are a LONG way away from 120Hz Plasma displays and that one of the problems was that the chips required to 'refresh' the screen 120 times a second are not available and if they were, they would be very expensive. Furthermore they did not see market demand for anything over 60Hz right now, and that the only reason some competing LCD manufacturers went to 120Hz was because they had to in order to get minimize motion blur and artifacts.

One thing I took away from my many discussions with Panny reps was that market demand drives what they do, not what guys like us want in these forums (we are like the other 1/10th of 1% of the market).

I 've been following this thread and find it very interesting as I am a beginner. Let me make sure I have this straight, please correct me if I'm wrong.

I have a Pioneer ProFHD1 plasma TV that the HDMI input accepts 1080p (1920x1080) at 24 and 60 hz signals. I also have a DVDO VP 50 video processor that the HDMI input accepts from 480i up to 1080p at 60 hz signals from my DVD player and my direct TV satellite receiver. However, the VP 50 HDMI input does not accept 1080p at 24 hz. The VP 50 can output 1080p to the pioneer at 24, 48, 60 and 72 hz.

I'm assuming for watching SD and HD programming from my satellite receiver I should set the VP output and Pioneer input at 1080p and 60 hz. Is that correct?

For DVD's what would be the best output from the VP 50 video processor and the best input for the pioneer plasma tv with the limits of the combinations from above? If I set them both to 24 hz will I get flicker? And for a future HD-DVD or Blu ray player what should the settings be?

thanks

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 DMN in another thread.

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 DMN in another thread. I believe this is the thread you mean. (http://www.avsforum.com/avs-vb/showthread.php?t=760549&page=1&pp=30) DNM sounds like a disaster, not to mention resource hogging. It's like Digital Noise Reduction for video or 1080p upconversion; it is building something out of nothing. But I'll reserve final judgement until I can use it and test it with various sources.

The only way to reduce judder is to film at 48fps or 60fps, like how Imax run their films, and watch it at that native speed.


fuad

Now that I am used to 24 frames going back to 60 really is bothersome.
However I don't notice it all the time.

Easiest way I've found to see the difference is to switch between 24 and 60 during the credits on a movie where they slowly scroll upward.

When I watch them closely during 24 setting they are smooth.
When I switch to 60 they jerk in uneven steps rapidly, making me feel dizzy, and the effect is kind of like staring at the end of a fluorescent tube and watching it flicker.

I have a projector, so it is probably less visible on small screen tvs.

Don't know if this is common knowledge yet, but Kris Deering's review of the Sony BD player is up at Secrets.
He talks about 1080/24p output as a big advantage in defeating judder.
He says he could no longer imagine watching BD without it! :eek:

I notice 2:3 cadence judder more than 24p framerate motion judder.

But I recall some people said shooting film at higher than 24p (which would reduce motion judder) would make it look less film like.

Yes, but it would still look AWESOME!! The best examples were the few films shot on Todd-AO 65mm at 30 fps. like Oklahoma and the original cut of Around the World in 80 Days.

My God, the depth of field is amazing and the virtual lack of judder is mind boggling. It looks a little weird at first coming from 24 fps movies, but once you get over the initial shock you wish all films were shot this way.

Dan

Has anybody heard/read anything about 120hz plasmas lately? I can find quite a bit of information on 120hz LCDs, but pretty much nothing about 120hz plasma.

I would really like to see a much "smoother" image in terms of motion artifacts on a flat panel, but not if it means downgrading to an LCD.

Has anybody heard/read anything about 120hz plasmas lately? I can find quite a bit of information on 120hz LCDs, but pretty much nothing about 120hz plasma.

I would really like to see a much "smoother" image in terms of motion artifacts on a flat panel, but not if it means downgrading to an LCD.

For LCDs the 120Hz (100Hz in Europe) is more about improving the motion lag issues than reducing motion judder (and large area flicker isn't a major LCD bugbear) - and AIUI some 120Hz LCDs may still display 24p stuff with 3:2 (or 6:4?) cadence rather than 5:5.

Plasmas have traditionally not suffered from the same issues with motion lag - and so have less reason to require a frame rate increase AIUI. Also - running a plasma with a frame rate of 120Hz would require a major increase in the sub-field frequency, and a reduction in the sub-field display times, which could cause a brightness issue?

(Plasma "pixels" aren't capable of native greyscale display - a plasma cell is either on, or off. To achieve greyscale plasmas split each source field or frame into multiple sub-fields, and then selectively switch the pixels on and off in each subfield to deliver an "average" greyscale that the eye/brain's persistence of vision accepts as a grey scale. I believe DLPs do something similar. As it is tricky to generate enough subfields to deliver a high bit-depth greyscale, plasmas and DLPs also often dither (add semi-random noise) to a signal to reduce the banding issues. Sub-field processing can be in issue if you use plasmas in TV studios if the cameras have shutters that are not correctly set-up - just as you can have problems with single chip colour wheel DLPs for the same reason)

Has anybody heard/read anything about 120hz plasmas lately? I can find quite a bit of information on 120hz LCDs, but pretty much nothing about 120hz plasma.

I would really like to see a much "smoother" image in terms of motion artifacts on a flat panel, but not if it means downgrading to an LCD.

I asked Panasonic this question at CES and they confirmed what the last poster just said. Panasonic also said they have no plans to introduce 120Hz and that the circuitry to do this for plasma would be very expensive (although I am not sure that the last statement is true).

I'd be happy just to see the majority of plasmas at 72Hz, something very few do now.