DLP Color Wheel Speed With XPR Pixel Shifting - AVS Forum | Home Theater Discussions And Reviews
Forum Jump: 
 6Likes
Reply
 
Thread Tools
post #1 of 33 Old 01-08-2019, 09:13 AM - Thread Starter
AVS Forum Special Member
 
Dave in Green's Avatar
 
Join Date: Jan 2014
Location: USA
Posts: 8,115
Mentioned: 147 Post(s)
Tagged: 0 Thread(s)
Quoted: 3680 Post(s)
Liked: 2890
DLP Color Wheel Speed With XPR Pixel Shifting

Consolidating a discussion that has been going on in various individual projector model threads, a question is being debated about how DLP XPR 4K pixel shifting affects color wheel speed rating and its effect on producing the DLP rainbow effect (RBE) that some are more sensitive to than others. Currently all projector color wheels are rated at 7,200 rpm, so that's their actual spinning speed. But projector companies that rated native 1080p RGBRGB color wheels as 4X are now rating XPR pixel shifting models with RGBRGB color wheels at 2X.

This is an important issue for consumers because the X speed rating has traditionally only been used to inform consumers that the higher the X number the less chance of seeing RBE (see color wheel history below). Yet the consensus of professional reviewers and owners to date has been that there's no greater likelihood of seeing rainbows on XPR pixel shifting projectors with 7,200 rpm RGBRGB color wheels rated 2X than there has been on native 1080p projectors with 7,200 rpm RGBRGB color wheels rated 4X.

Projectorcentral.com's @ProCentral Rob will be addressing this issue in a briefing he has scheduled with BenQ this month. Forum member @scottyroo has addressed the question to his contact at BenQ and is currently awaiting a technical answer while @sage11x has also been active in getting color wheel performance data from his contact at BenQ. So we're still awaiting the definitive technical answer to why XPR models are only rated at 2X even though they are apparently no more susceptible to creating RBE than similar 1080p models. When that official data is available it will be posted in this thread.

Color wheel history: The first DLP projector color wheels had one set of RGB segments spinning at 3,600 rpm. When the industry doubled the standard RGB color wheel speed to 7,200 rpm to help reduce RBE they referred to that as 2X speed or double the original (1X) speed. Then they came up with the idea of doubling the number of RGB segments to RGBRGB, which effectively doubled the 7,200 rpm RGB color wheel speed to 4X because R, G and B segments were being flashed twice for each actual wheel revolution. In other words they doubled the X speed rating either by doubling the color wheel's rpm or by doubling the number of sets of RGB segments.

The actual speed at which the color wheel is spinning is defined by the revolutions per minute (rpm). The effective color wheel speed in terms of reducing the rainbow effect is defined by rpm combined with the number of sets of RGB segments. So the first instance of doubling color wheel rpm from 3,600 to 7,200 was an actual speed increase while doubling the number of sets of RGB segments to RGBRGB was a virtual speed increase only in terms of the number of times R, G and B segments were flashed per revolution.
Dave in Green is offline  
Sponsored Links
Advertisement
 
post #2 of 33 Old 01-08-2019, 09:22 AM
Member
 
perkele's Avatar
 
Join Date: Jun 2013
Posts: 132
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 61 Post(s)
Liked: 51
Just to confuse a little more, even the current 7200rpm wheels wont spin at that rate constantly. It varies between 24hz, 50hz and 60hz sources on many modern DLP projectors.

I took a quick video with my Benq HT2050a projector. Ive verified this with some entry level optoma models too, like HD27/hd142x. In this video, i change the source from 60hz to 24hz, and you can hear the wheel starting to spin physically faster.
https://drive.google.com/file/d/1Y9L...ew?usp=sharing

But this is not actually relevant to Pixel shifting 4k dlp projectors like Sage mentions below, because they dont do 24p natively. So the physical speed of the wheel is pobably the same for all the sources with them.

Last edited by perkele; 01-08-2019 at 11:53 AM. Reason: forgot some key details
perkele is offline  
post #3 of 33 Old 01-08-2019, 09:33 AM
AVS Forum Special Member
 
JRock3x8's Avatar
 
Join Date: Feb 2006
Location: Northwest Indiana
Posts: 3,887
Mentioned: 16 Post(s)
Tagged: 0 Thread(s)
Quoted: 1805 Post(s)
Liked: 569
Is this why my PJ sounds like a broken refrigerator?


Sent from my iPhone using Tapatalk
JRock3x8 is offline  
Sponsored Links
Advertisement
 
post #4 of 33 Old 01-08-2019, 09:45 AM
AVS Forum Special Member
 
AV_Integrated's Avatar
 
Join Date: Jun 2005
Location: Northern, VA - Washington, DC
Posts: 6,434
Mentioned: 96 Post(s)
Tagged: 2 Thread(s)
Quoted: 1564 Post(s)
Liked: 1158
I'm not even sure how some of these models that perform quad pixel duty are able to claim a 2x color wheel speed.

I'm pretty sure the original color wheels (RGB) were 1x because they rotated at 3,600RPM and only flashed one color per frame at 60hz video speed. When they doubled the spin rate to 7,200RPM you got a 2x wheel, then again when they went to RGB/RGB you got two flashes per frame resulting in a 4x rated speed. Now, I suppose you could say that the 4x wheel speed still could provide 2x or 4x flashes per frame, but that color time is going to be less than 1/2 (double pixel shifter) or 1/4 (quad shifter) depending on the DLP chip used. It calling a 4x pixel shifter a 2x color wheel speed while using the same RGB/RGB color wheel with a 7,200RPM motor, then I would argue that as not being accurate as there is no way you are getting the full color saturation in that time as you used to with non-shifting chips.

But, it would be an interesting thing to get a technical response about.

I'm still somewhat shocked that they haven't jumped to 6x or higher as a standard after all these years. Probably a reason for that as well.

AV Integrated - Theater, whole house audio, and technology installation in the Washington DC metro area.
AV_Integrated is offline  
post #5 of 33 Old 01-08-2019, 10:59 AM
AVS Forum Special Member
 
sage11x's Avatar
 
Join Date: Dec 2006
Location: Howell, MI
Posts: 6,498
Mentioned: 194 Post(s)
Tagged: 0 Thread(s)
Quoted: 2700 Post(s)
Liked: 2664
DLP Color Wheel Speed With XPR Pixel Shifting

Straight from my contact at BenQ:

All of HT2050A, HT2150ST, and HT3050 are with 7200rpm (2x=60*60*2=7200) for the speed of its wheel, and because there are 2 sets of RGB segments, so we call 4x on its “color wheel speed”.
For HT2550, it’s also with the same 7200rpm on its wheel speed, and because it’s with RGBRGB color wheels and also with pixel-shifting (the 2 sets of RGB segments on 1080p ones are used for one single image after spinning once, the 2 sets of RGB segments on HT2550 are used for different images after spinning once), we call 4x on its “’pixel-shifting’ color wheel speed”.

Pay attention to that last part: the RGB segments on a 1080p DLP are used to flash the same image within one rotation (half a frame) while on the 4K model each RGB pass is used for a different image as each is one of the four images that combine to form the full 4K composite frame-- 1 frame sees two full rotations of the color wheel.

The speed is the same. The difference is a single HD image being flashed four times per frame vs four different HD images-- each a part of the whole 4K frame-- being flashed.

I went into some really deep detail on how this works in the W2700 thread as far as where the shift is within each half and full rotation of the wheel. I’m too exhausted to explain that again here. Suffice to say we need to stop referring to the wheel in 2X or 4X terms. All consumer color wheels are 7200rpm (120Hz) and what matters is the number and compliment of the slices.

Also, none of the 4K DLPs are capable of displaying 24fps natively. Due to the requirements of the pixel shift (120Hz for the DLP66 and 240Hz for the DLP47). everything has to be converted to 60Hz. That’s is why the relative ‘speed’ is the same regardless of content. I.E. you’re not going to see a different quoted speed for 24Hz like you might with a 1080p RGBRGB model.


edit: general cleanup and grammar corrections as my original post was from my phone.

What to do if you find yourself stuck with no hope of rescue:
Consider yourself lucky that life has been good to you so far. Alternatively, if life hasn't been good to you so far, which given your present circumstances seems to be more likely, consider yourself lucky that it won't be troubling you much longer...

-- Excerpt from the Hitchhiker's Guide to the Galaxy.

Last edited by sage11x; 01-08-2019 at 11:25 AM.
sage11x is online now  
post #6 of 33 Old 01-08-2019, 12:48 PM
Member
 
perkele's Avatar
 
Join Date: Jun 2013
Posts: 132
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 61 Post(s)
Liked: 51
Quote:
Originally Posted by sage11x View Post
Straight from my contact at BenQ:

All of HT2050A, HT2150ST, and HT3050 are with 7200rpm (2x=60*60*2=7200) for the speed of its wheel, and because there are 2 sets of RGB segments, so we call 4x on its “color wheel speed”.
For HT2550, it’s also with the same 7200rpm on its wheel speed, and because it’s with RGBRGB color wheels and also with pixel-shifting (the 2 sets of RGB segments on 1080p ones are used for one single image after spinning once, the 2 sets of RGB segments on HT2550 are used for different images after spinning once), we call 4x on its “’pixel-shifting’ color wheel speed”.


The speed is the same. The difference is a single HD image being flashed four times per frame vs four different HD images-- each a part of the whole 4K frame-- being flashed.

So this is the key part i was trying to argue with you earlier and how it could affect rainbow effect. You get less flashes of color for each pixel per frame(or less color time if you do more flashes) on Pixel shifting projectors(XPR). Does it affect rainbow effect in any way?

So far im not actually convinced the RBE is the same even though it is your conclusion. I hope thats true because im shopping for 4k dlp projector right now, while being RBE sensitive to the point that i cant watch 7200rpm RGB**W wheels anymore, ie "2x" speed on non-XRP scale. Returns are not as easy to do in my country compared to USA for example, had to fight hard to get ht2550 returned. So yeah, im really hoping we get more information if XPR actually affects RBE in any way. So far not really convinced in neither way. And like ive mentioned before, my short time with ht2550 i saw more rainbows on it compared to my ht2050a and ht3050. I should have spent more time with it, but it didnt occur to me then.

Last edited by perkele; 01-08-2019 at 12:57 PM. Reason: Typos
perkele is offline  
post #7 of 33 Old 01-08-2019, 01:29 PM
AVS Forum Special Member
 
AV_Integrated's Avatar
 
Join Date: Jun 2005
Location: Northern, VA - Washington, DC
Posts: 6,434
Mentioned: 96 Post(s)
Tagged: 2 Thread(s)
Quoted: 1564 Post(s)
Liked: 1158
Quote:
Originally Posted by sage11x View Post
...we call 4x on its “’pixel-shifting’ color wheel speed”.
Interesting way to look at things. Completely deviates from the history of how and why color wheel speeds were used in the first place, and since DLP asked manufacturers to stop using those terms, I guess it makes it meaningless. But, with less than 1/4 the segment time that the 1080p models have (we'll just deal with 1080p XPR 4K models for this discussion), I simply can't imaging that this doesn't significantly impact RBE to those who are more sensitive.

Quote:
Originally Posted by sage11x View Post
...I went into some really deep detail on how this works in the W2700 thread
I will take a look at what you wrote there.

As I said, I'm quite dubious when manufacturers start spinning reality for marketing purposes. But, perhaps a new breed of terminology always needs to be applied. 4x wheel/4x shift. 4x wheel/2x shift. 4x wheel/1x fixed frame. I don't know, but I'm pretty RBE sensitive. I need to spend some face time with one of the 2x models and a 4x model for some greater insight.

AV Integrated - Theater, whole house audio, and technology installation in the Washington DC metro area.
AV_Integrated is offline  
post #8 of 33 Old 01-08-2019, 01:51 PM
AVS Forum Special Member
 
sage11x's Avatar
 
Join Date: Dec 2006
Location: Howell, MI
Posts: 6,498
Mentioned: 194 Post(s)
Tagged: 0 Thread(s)
Quoted: 2700 Post(s)
Liked: 2664
Quote:
Originally Posted by perkele View Post
So this is the key part i was trying to argue with you earlier and how it could affect rainbow effect. You get less flashes of color for each pixel per frame(or less color time if you do more flashes) on Pixel shifting projectors(XPR). Does it affect rainbow effect in any way?



So far im not actually convinced the RBE is the same even though it is your conclusion. I hope thats true because im shopping for 4k dlp projector right now, while being RBE sensitive to the point that i cant watch 7200rpm RGB**W wheels anymore, ie "2x" speed on non-XRP scale. Returns are not as easy to do in my country compared to USA for example, had to fight hard to get ht2550 returned. So yeah, im really hoping we get more information if XPR actually affects RBE in any way. So far not really conviced in neither way. And like ive mentioned before, my short time with ht2550 i saw more rainbows on it compared to my ht2050a and ht3050. I should have spent more time with it, but it didnt occur to me then.


If you’re seeing RBE on the HT2550 I would avoid DLP. Please understand I am NOT discounting your experience as I 100% believe you saw rbe on the HT2550. What I AM saying is seeing rbe on the HT2550 is rare and indicates you could be sensitive to this particular artifact. Could you see RBE on the Ht2550 but not an HT2050A? Sure— ironically, my sister is exactly the opposite as she saw RBE on the RGBRGB HT1070A when I was reviewing it but not the HT2550. So everyone’s eyes are different.

There are other factors that effect RBE. My contact goes on to say to that color ramping plays a big role here. The trick is to strike a balance between good color saturation and lessening noticeable color separation. Most of these 4K DLPs are attempting to replicate DCI-P3 without actually being able to display DCI-P3. So a lot of times you get manufacturers over-saturating the picture to get that HDR ‘look’ which, in turn, can increase the chance of RBE for those sensitive. For what it’s worth I don’t think BenQ is guilty here as they seem to have erred on the side of caution. But that might help to explain your experience.

What to do if you find yourself stuck with no hope of rescue:
Consider yourself lucky that life has been good to you so far. Alternatively, if life hasn't been good to you so far, which given your present circumstances seems to be more likely, consider yourself lucky that it won't be troubling you much longer...

-- Excerpt from the Hitchhiker's Guide to the Galaxy.
sage11x is online now  
post #9 of 33 Old 01-08-2019, 02:17 PM
AVS Forum Special Member
 
sage11x's Avatar
 
Join Date: Dec 2006
Location: Howell, MI
Posts: 6,498
Mentioned: 194 Post(s)
Tagged: 0 Thread(s)
Quoted: 2700 Post(s)
Liked: 2664
Quote:
Originally Posted by AV_Integrated View Post
Interesting way to look at things. Completely deviates from the history of how and why color wheel speeds were used in the first place, and since DLP asked manufacturers to stop using those terms, I guess it makes it meaningless. But, with less than 1/4 the segment time that the 1080p models have (we'll just deal with 1080p XPR 4K models for this discussion), I simply can't imaging that this doesn't significantly impact RBE to those who are more sensitive.





I will take a look at what you wrote there.



As I said, I'm quite dubious when manufacturers start spinning reality for marketing purposes. But, perhaps a new breed of terminology always needs to be applied. 4x wheel/4x shift. 4x wheel/2x shift. 4x wheel/1x fixed frame. I don't know, but I'm pretty RBE sensitive. I need to spend some face time with one of the 2x models and a 4x model for some greater insight.


I don’t think they’re spinning it (haha... puns). This whole discussion happened when BenQ leaked the HT3550 homepage and spec’d the color wheel at 2X. I would have been happy with them sticking with the 4X moniker because, in my mind, the 4X was signifying how many RGB passes you get PER FRAME. That doesn’t change regardless of the addition of pixel shifting— you are still getting four passes of RGB every 1/60th of a second. But a lot of HT2050 owners got concerned that this meant the Ht3550’s wheel was slower. For the record it’s not.

But in another case you are right— instead of getting 4 passes of RGB for the same frame, on the XPR you are getting 4 passes of RGB where each is being used for a different image within a frame. But the actual speed with which the RGB slices are passing is the same.

Quick note here for everyone else catching up with this convo— you cannot imaging how fast this is all happening. The colorwheel is spinning at 7200 rpm which means it makes 120 full rotations every second. Meanwhile, the DLP47 chip needs to create four distinct images for every frame of 60Hz video. That means four images every 60th of a second for 240 distinct images per second.

While I don’t find this worsens RBE... FOR ME... I DO find it makes a difference for motion handling. Simply put, motion doesn’t have the same surreal clarity it does with my HT2050. I blame that on the fact that each frame is being created from four distinct ‘subframes’ instead of simply being repeated as it would on the 1080p model. To be clear, the motion handling is still much better than what you get with the average LCD/3LCD and you likely won’t notice much difference vs a standard 2X DLP like the BenQ MH530FHD or Optoma 143x. But it’s definitely an area where the premium RGBRGB 1080p DLPs like the HT2050A and HT3050 have an edge. While I can see a difference in film/video this is most apparent in video games. But that’s a little off topic...

What to do if you find yourself stuck with no hope of rescue:
Consider yourself lucky that life has been good to you so far. Alternatively, if life hasn't been good to you so far, which given your present circumstances seems to be more likely, consider yourself lucky that it won't be troubling you much longer...

-- Excerpt from the Hitchhiker's Guide to the Galaxy.
sage11x is online now  
post #10 of 33 Old 01-08-2019, 09:09 PM
Senior Member
 
Join Date: Apr 2017
Location: Dongguan, CH
Posts: 480
Mentioned: 11 Post(s)
Tagged: 0 Thread(s)
Quoted: 260 Post(s)
Liked: 86
interesting read, now my head is spinning @4X speed lol

anyways, do all dlp need/ use color wheels? how about those pico dlp's?
sage11x likes this.

XGimi CC, XGimi Z6, UKCSIS, Vmai M200 Pico
Xbox One S, LG SH4
wheelee is online now  
post #11 of 33 Old 01-09-2019, 02:32 AM
Member
 
LeQuack's Avatar
 
Join Date: Jan 2018
Posts: 61
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 28 Post(s)
Liked: 10
Quote:
Originally Posted by wheelee View Post
interesting read, now my head is spinning @4X speed lol

anyways, do all dlp need/ use color wheels? how about those pico dlp's?

Only single chip DLPs need color wheels and even those can do without, for example RGB LED DLPs don't need it, since the LED switching is fast enough. But the effect on screen is pretty similar ie. alternating colors with time multiplexing, or rainbows for short (unless it's a 3 chip DLP).
sage11x likes this.
LeQuack is offline  
post #12 of 33 Old 01-09-2019, 10:23 AM
AVS Forum Special Member
 
sage11x's Avatar
 
Join Date: Dec 2006
Location: Howell, MI
Posts: 6,498
Mentioned: 194 Post(s)
Tagged: 0 Thread(s)
Quoted: 2700 Post(s)
Liked: 2664
Quote:
Originally Posted by wheelee View Post
interesting read, now my head is spinning @4X speed lol



anyways, do all dlp need/ use color wheels? how about those pico dlp's?


All single chip DLP use sequential color flashes to produce their full color images. Whether you are using a color wheel filtering a lamp or laser light source or rapidly flashing RGB LEDs the result is the same. There was someone on the W2700 thread that was not too excited about color wheels and I don’t think he realized that rapidly flashing LEDs can sometimes result in higher RBE. Like anything it depends on the tech and how it’s implemented.

The advantages of single chip DLP are low cost, small size and a sharp image impervious of convergence errors. The disadvantage is, of course, the possibility of RBE.

At one time there were high end consumer three chip DLPs like what you find in the theater (one chip for red green and blue just like you find in 3LCD and LCoS projectors). But they’re all but extinct now.
wheelee likes this.

What to do if you find yourself stuck with no hope of rescue:
Consider yourself lucky that life has been good to you so far. Alternatively, if life hasn't been good to you so far, which given your present circumstances seems to be more likely, consider yourself lucky that it won't be troubling you much longer...

-- Excerpt from the Hitchhiker's Guide to the Galaxy.
sage11x is online now  
post #13 of 33 Old 01-09-2019, 10:45 AM
Member
 
LeQuack's Avatar
 
Join Date: Jan 2018
Posts: 61
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 28 Post(s)
Liked: 10
I'd also add another disadvantage of DLPs, that is uneconomical light usage, in that only one color can be output at a single time, meaning for the same brightness you need a higher wattage lamp as you would with a 3 chip design. Some projectors also add various other color segments to try to battle that and increase the lumens.
LeQuack is offline  
post #14 of 33 Old 01-09-2019, 01:15 PM
AVS Forum Special Member
 
bud16415's Avatar
 
Join Date: Jan 2006
Location: Erie Pa
Posts: 7,701
Mentioned: 110 Post(s)
Tagged: 0 Thread(s)
Quoted: 2066 Post(s)
Liked: 1033
Quote:
Originally Posted by LeQuack View Post
I'd also add another disadvantage of DLPs, that is uneconomical light usage, in that only one color can be output at a single time, meaning for the same brightness you need a higher wattage lamp as you would with a 3 chip design. Some projectors also add various other color segments to try to battle that and increase the lumens.
I owned a RGBCYW projector for 3years and before that a RGBW. The main reason I bought was when TI introduced the RGBCYW they put out a report explaining this had the possibilities of expanding the color gamut and at the very least it would fully fill the Rec 709 color space and do it with greatly increased efficiencies.

I still think it had way more capabilities than any of the projector manufactures took advantage of. They rather used it as a method of advertising very high overall lumens at the expense of quality color production.

I now have a RGBRGB and it’s wonderful but when I notice it struggling it is normally with bright yellows and bright blues. I don’t know but suspect the industry never tried real hard blending 5 or 6 colors. I did notice when comparing lamp wattages and where each could produce colors accurate enough for my eyes there was a slight advantage to RGBCYW. Not enough though to except half the effective color wheel speed. I never saw rainbows on ether projector so I cant judge that.

What was surprising with a RGBCYW or a RGBW in the most accurate mode they shut everything off except the RGB. Didn’t seem logical at all to pulse RGB once each when you could pulse W once and do the same thing.

Bud
bud16415 is offline  
post #15 of 33 Old 01-09-2019, 05:16 PM
Senior Member
 
Join Date: Apr 2017
Location: Dongguan, CH
Posts: 480
Mentioned: 11 Post(s)
Tagged: 0 Thread(s)
Quoted: 260 Post(s)
Liked: 86
Quote:
Originally Posted by sage11x View Post

At one time there were high end consumer three chip DLPs like what you find in the theater. But they’re all but extinct now.

so even though DLP/ DMD chips price would go down(wishful thinking), consumer 3 DLP pj's wont be in the future & that single DLP is the way forward?

XGimi CC, XGimi Z6, UKCSIS, Vmai M200 Pico
Xbox One S, LG SH4
wheelee is online now  
post #16 of 33 Old 01-09-2019, 05:50 PM - Thread Starter
AVS Forum Special Member
 
Dave in Green's Avatar
 
Join Date: Jan 2014
Location: USA
Posts: 8,115
Mentioned: 147 Post(s)
Tagged: 0 Thread(s)
Quoted: 3680 Post(s)
Liked: 2890
The cost of three DMDs is only a small percentage of the total cost of engineering and producing a quality 3-DLP projector.
Dave in Green is offline  
post #17 of 33 Old 01-26-2019, 01:52 AM
 
Dave Harper's Avatar
 
Join Date: Feb 2000
Location: Paradise on Earth
Posts: 6,554
Mentioned: 62 Post(s)
Tagged: 1 Thread(s)
Quoted: 3159 Post(s)
Liked: 1723
Quote:
Originally Posted by perkele View Post
Just to confuse a little more, even the current 7200rpm wheels wont spin at that rate constantly. It varies between 24hz, 50hz and 60hz sources on many modern DLP projectors.



I took a quick video with my Benq HT2050a projector. Ive verified this with some entry level optoma models too, like HD27/hd142x. In this video, i change the source from 60hz to 24hz, and you can hear the wheel starting to spin physically faster.

https://drive.google.com/file/d/1Y9L...ew?usp=sharing



But this is not actually relevant to Pixel shifting 4k dlp projectors like Sage mentions below, because they dont do 24p natively. So the physical speed of the wheel is pobably the same for all the sources with them.

On my LK970 I actually was just thinking I saw less RBE (barely any anyway) when I play 4K24p UHD HDR discs than when I was watching normal HD via DirecTV at 60p.

Quote:
Originally Posted by sage11x View Post
Straight from my contact at BenQ:

All of HT2050A, HT2150ST, and HT3050 are with 7200rpm (2x=60*60*2=7200) for the speed of its wheel, and because there are 2 sets of RGB segments, so we call 4x on its “color wheel speed”.
For HT2550, it’s also with the same 7200rpm on its wheel speed, and because it’s with RGBRGB color wheels and also with pixel-shifting (the 2 sets of RGB segments on 1080p ones are used for one single image after spinning once, the 2 sets of RGB segments on HT2550 are used for different images after spinning once), we call 4x on its “’pixel-shifting’ color wheel speed”.

Pay attention to that last part: the RGB segments on a 1080p DLP are used to flash the same image within one rotation (half a frame) while on the 4K model each RGB pass is used for a different image as each is one of the four images that combine to form the full 4K composite frame-- 1 frame sees two full rotations of the color wheel.

The speed is the same. The difference is a single HD image being flashed four times per frame vs four different HD images-- each a part of the whole 4K frame-- being flashed.

I went into some really deep detail on how this works in the W2700 thread as far as where the shift is within each half and full rotation of the wheel. I’m too exhausted to explain that again here. Suffice to say we need to stop referring to the wheel in 2X or 4X terms. All consumer color wheels are 7200rpm (120Hz) and what matters is the number and compliment of the slices.

Also, none of the 4K DLPs are capable of displaying 24fps natively. Due to the requirements of the pixel shift (120Hz for the DLP66 and 240Hz for the DLP47). everything has to be converted to 60Hz. That’s is why the relative ‘speed’ is the same regardless of content. I.E. you’re not going to see a different quoted speed for 24Hz like you might with a 1080p RGBRGB model.



edit: general cleanup and grammar corrections as my original post was from my phone.

24 (5x) and 60 (2x) both go into 120 evenly, so there’s no need to convert 24 to 60. It appears to me that the 24Hz material shows 5 frames per spin and the 60Hz shows 2 frames per spin, which could be why I clearly see a smoother more RBE free image when I watch 4K24p HDR material.

I watched Crazy Rich Asians tonight and it was insane how smooth, clear, deep and 3D it looked with no RBE that I could see anyway.
Dave Harper is offline  
post #18 of 33 Old 01-26-2019, 06:27 AM
AVS Forum Special Member
 
sage11x's Avatar
 
Join Date: Dec 2006
Location: Howell, MI
Posts: 6,498
Mentioned: 194 Post(s)
Tagged: 0 Thread(s)
Quoted: 2700 Post(s)
Liked: 2664
Quote:
Originally Posted by Dave Harper View Post
On my LK970 I actually was just thinking I saw less RBE (barely any anyway) when I play 4K24p UHD HDR discs than when I was watching normal HD via DirecTV at 60p.




24 (5x) and 60 (2x) both go into 120 evenly, so there’s no need to convert 24 to 60. It appears to me that the 24Hz material shows 5 frames per spin and the 60Hz shows 2 frames per spin, which could be why I clearly see a smoother more RBE free image when I watch 4K24p HDR material.

I watched Crazy Rich Asians tonight and it was insane how smooth, clear, deep and 3D it looked with no RBE that I could see anyway.
Not "per spin" but per FRAME-- yes, with an RGBRGB 7200rpm wheel you get 4 flashes or sub frames of each image for 60Hz content (2 rotations for 4X) and 5 sub frames of each image for 24Hz content (2.5 rotations for 5X-- technically 5 rotations as 24fps content displays at 48Hz).


The 4K XPR DLPs work a little differently as each of those sub frames is not the same image repeated but actually different images that, when combined with pixel shifting, create the final composite 4K image. With the DLP47 (quad shift) the DMD is operating at 240Hz and creating 4 distinct images per 60Hz frame (4x60=240). On the DLP66 the DMD is operating at 120Hz and creating 2 distinct images per 60Hz frame (2x60=120). Since the XPR operates at this fixed 60Hz it is not able to display 24Hz content natively.


I did a quick google search and Sound and Vision might explain this better than I am:
"One thing we can say about the new chip and its supporting processing, dubbed XPR, is that it comes with at least one inherent limitation that doesn’t necessarily affect other 4K displays. The system always requires playback at a 60-hertz frame rate, with the DLP chip operating at 120 Hz to enable the pixel-shifting, so there is no native 24p playback. If the input signal is a multiple of 60 Hz, it gets converted with 2:2 pulldown; otherwise, 3:2 pulldown is used."
https://www.soundandvision.com/conte...s-4k-really-4k

What to do if you find yourself stuck with no hope of rescue:
Consider yourself lucky that life has been good to you so far. Alternatively, if life hasn't been good to you so far, which given your present circumstances seems to be more likely, consider yourself lucky that it won't be troubling you much longer...

-- Excerpt from the Hitchhiker's Guide to the Galaxy.
sage11x is online now  
post #19 of 33 Old 01-26-2019, 09:06 AM
 
Dave Harper's Avatar
 
Join Date: Feb 2000
Location: Paradise on Earth
Posts: 6,554
Mentioned: 62 Post(s)
Tagged: 1 Thread(s)
Quoted: 3159 Post(s)
Liked: 1723
DLP Color Wheel Speed With XPR Pixel Shifting

Quote:
Originally Posted by sage11x View Post
Not "per spin" but per FRAME-- yes, with an RGBRGB 7200rpm wheel you get 4 flashes or sub frames of each image for 60Hz content (2 rotations for 4X) and 5 sub frames of each image for 24Hz content (2.5 rotations for 5X-- technically 5 rotations as 24fps content displays at 48Hz).


The 4K XPR DLPs work a little differently as each of those sub frames is not the same image repeated but actually different images that, when combined with pixel shifting, create the final composite 4K image. With the DLP47 (quad shift) the DMD is operating at 240Hz and creating 4 distinct images per 60Hz frame (4x60=240). On the DLP66 the DMD is operating at 120Hz and creating 2 distinct images per 60Hz frame (2x60=120). Since the XPR operates at this fixed 60Hz it is not able to display 24Hz content natively.


I did a quick google search and Sound and Vision might explain this better than I am:
"One thing we can say about the new chip and its supporting processing, dubbed XPR, is that it comes with at least one inherent limitation that doesn’t necessarily affect other 4K displays. The system always requires playback at a 60-hertz frame rate, with the DLP chip operating at 120 Hz to enable the pixel-shifting, so there is no native 24p playback. If the input signal is a multiple of 60 Hz, it gets converted with 2:2 pulldown; otherwise, 3:2 pulldown is used."
https://www.soundandvision.com/conte...s-4k-really-4k

Thanks for the info. I understand how XPR works. I had a very interesting “discussion” thread here awhile ago talking about it with another former member and others.

I guess what I don’t understand is why if it’s 120Hz in the end, do they not do as I said and use 24Hz x 5 to equal 120Hz? Converting to 60 and doubling that to get 120 seems stupid. Has this been verified by engineers or just deduced because someone was talking about 60Hz HD video or something?

If 24Hz in the LK970 is being 2:2 pulled down to 60, then its operation is butter smooth based on what I saw!
Dave Harper is offline  
post #20 of 33 Old 01-26-2019, 11:59 AM
AVS Forum Special Member
 
sage11x's Avatar
 
Join Date: Dec 2006
Location: Howell, MI
Posts: 6,498
Mentioned: 194 Post(s)
Tagged: 0 Thread(s)
Quoted: 2700 Post(s)
Liked: 2664
DLP Color Wheel Speed With XPR Pixel Shifting

Quote:
Originally Posted by Dave Harper View Post
Thanks for the info. I understand how XPR works. I had a very interesting “discussion” thread here awhile ago talking about it with another former member and others.

I guess what I don’t understand is why if it’s 120Hz in the end, do they not do as I said and use 24Hz x 5 to equal 120Hz? Converting to 60 and doubling that to get 120 seems stupid. Has this been verified by engineers or just deduced because someone was talking about 60Hz HD video or something?

If 24Hz in the LK970 is being 2:2 pulled down to 60, then its operation is butter smooth based on what I saw!


No, it’s because the full 4K frame is composed of 4 separate 1080p sub frames. You’re locked into 60Hz due to the pixel shift. The DMD needs the full frame— and by frame I mean 1/60th of a second— to produce a full 4K image. On a 1080p DLP the image is just repeated within that frame so it’s no big deal to repeat it 5 times (for 48Hz) instead of 4 (for 60Hz). On the 4K XPR projector the DMD needs that fifth ‘pass’ of RGB to start creating the next 4K frame.

What to do if you find yourself stuck with no hope of rescue:
Consider yourself lucky that life has been good to you so far. Alternatively, if life hasn't been good to you so far, which given your present circumstances seems to be more likely, consider yourself lucky that it won't be troubling you much longer...

-- Excerpt from the Hitchhiker's Guide to the Galaxy.

Last edited by sage11x; 01-26-2019 at 12:10 PM.
sage11x is online now  
post #21 of 33 Old 01-26-2019, 09:50 PM
 
Dave Harper's Avatar
 
Join Date: Feb 2000
Location: Paradise on Earth
Posts: 6,554
Mentioned: 62 Post(s)
Tagged: 1 Thread(s)
Quoted: 3159 Post(s)
Liked: 1723
Quote:
Originally Posted by sage11x View Post
No, it’s because the full 4K frame is composed of 4 separate 1080p sub frames. You’re locked into 60Hz due to the pixel shift. The DMD needs the full frame— and by frame I mean 1/60th of a second— to produce a full 4K image. On a 1080p DLP the image is just repeated within that frame so it’s no big deal to repeat it 5 times (for 48Hz) instead of 4 (for 60Hz). On the 4K XPR projector the DMD needs that fifth ‘pass’ of RGB to start creating the next 4K frame.

Not for the .67” XPR DLP chips. They only flash 2 subframes of 2716 x 1528, not like the .47” XPRs that use 1920 x 1080 DLP chips and flash it four times.
Calibratedlumens likes this.
Dave Harper is offline  
post #22 of 33 Old 01-27-2019, 05:03 AM
AVS Forum Special Member
 
Ruined's Avatar
 
Join Date: Jul 2002
Posts: 6,827
Mentioned: 35 Post(s)
Tagged: 0 Thread(s)
Quoted: 2612 Post(s)
Liked: 1250
Quote:
Originally Posted by Dave Harper View Post
Not for the .67” XPR DLP chips. They only flash 2 subframes of 2716 x 1528, not like the .47” XPRs that use 1920 x 1080 DLP chips and flash it four times.
It doesn't matter, since the RGBY LK970 gets only 1 set of colors per spin at the same physical wheel speed that the RGBRGB ht2550 gets 2 sets of colors per spin. With 2 spins the lk970 gets 2 sets of colors, and with 2 spins the ht2550 gets 4 sets of colors - both in same amount of time. Each matching up with their amount of subframes needed to render a full 4k image.

That being said TI claims native 24fps is possible at 48hz on the 4k xpr pjs though I am not sure if any dlp pj has implemented this nor how it would look.

Last edited by Ruined; 01-27-2019 at 05:06 AM.
Ruined is offline  
post #23 of 33 Old 01-27-2019, 05:13 AM
AVS Forum Special Member
 
Ruined's Avatar
 
Join Date: Jul 2002
Posts: 6,827
Mentioned: 35 Post(s)
Tagged: 0 Thread(s)
Quoted: 2612 Post(s)
Liked: 1250
Quote:
Originally Posted by sage11x View Post
I don’t think they’re spinning it (haha... puns). This whole discussion happened when BenQ leaked the HT3550 homepage and spec’d the color wheel at 2X. I would have been happy with them sticking with the 4X moniker because, in my mind, the 4X was signifying how many RGB passes you get PER FRAME. That doesn’t change regardless of the addition of pixel shifting— you are still getting four passes of RGB every 1/60th of a second. But a lot of HT2050 owners got concerned that this meant the Ht3550’s wheel was slower. For the record it’s not.

But in another case you are right— instead of getting 4 passes of RGB for the same frame, on the XPR you are getting 4 passes of RGB where each is being used for a different image within a frame. But the actual speed with which the RGB slices are passing is the same.

Quick note here for everyone else catching up with this convo— you cannot imaging how fast this is all happening. The colorwheel is spinning at 7200 rpm which means it makes 120 full rotations every second. Meanwhile, the DLP47 chip needs to create four distinct images for every frame of 60Hz video. That means four images every 60th of a second for 240 distinct images per second.

While I don’t find this worsens RBE... FOR ME... I DO find it makes a difference for motion handling. Simply put, motion doesn’t have the same surreal clarity it does with my HT2050. I blame that on the fact that each frame is being created from four distinct ‘subframes’ instead of simply being repeated as it would on the 1080p model. To be clear, the motion handling is still much better than what you get with the average LCD/3LCD and you likely won’t notice much difference vs a standard 2X DLP like the BenQ MH530FHD or Optoma 143x. But it’s definitely an area where the premium RGBRGB 1080p DLPs like the HT2050A and HT3050 have an edge. While I can see a difference in film/video this is most apparent in video games. But that’s a little off topic...
Rbe on xpr is a little different than you'd traditionally think about it. Recall most see rbe with slow wheel speeds and motion, observing a bright white object moving across a black background (color breakup in motion across different frames)

Now, even if the RGBRGB wheel of the ht2550 has its duplicate segments split up to render subframes, remember they are all subframes to render detail on the same static composite frame. So while there may be some intraframe RBE "noise" as a result of this subframe rendering process, this is different and should be much more minor than typical RBE which is observed in motion between different complete frames.

As usual to virtually eliminate RBE on single chip DLP best bet is to completely drop the color wheel for a fast cycling RGB LED/Laser design such as the BenQ HT9060. Designs like this are not bound by the slower speeds of mechanical wheels and hence excel at suppressing RBE.

Note I believe BenQ is coming out with a cheaper 0.47" HLD LED design in the near future that should have similar elimination of RBE.

Last edited by Ruined; 01-27-2019 at 05:18 AM.
Ruined is offline  
post #24 of 33 Old 01-27-2019, 02:31 PM
 
Dave Harper's Avatar
 
Join Date: Feb 2000
Location: Paradise on Earth
Posts: 6,554
Mentioned: 62 Post(s)
Tagged: 1 Thread(s)
Quoted: 3159 Post(s)
Liked: 1723
Quote:
Originally Posted by Ruined View Post
It doesn't matter, since the RGBY LK970 gets only 1 set of colors per spin at the same physical wheel speed that the RGBRGB ht2550 gets 2 sets of colors per spin. With 2 spins the lk970 gets 2 sets of colors, and with 2 spins the ht2550 gets 4 sets of colors - both in same amount of time. Each matching up with their amount of subframes needed to render a full 4k image.



That being said TI claims native 24fps is possible at 48hz on the 4k xpr pjs though I am not sure if any dlp pj has implemented this nor how it would look.

True, but you also have to consider the wheels are the same circumference/size (right?) so on a wheel with 6 segments (RGBRGB), each segment is smaller so the “on time” for each color on each pass is less than a wheel with only 4 segments like the RGCbCy (Red, Green, Clear [Blue Laser], Clear [Yellow Phosphor]).

I am sure ALL these variables factor into how each version of DLP Color and Phosphor wheels present RBE.

As I said, actually seeing is believing.
Calibratedlumens likes this.
Dave Harper is offline  
post #25 of 33 Old 01-28-2019, 03:13 AM
Member
 
Nick Laslett's Avatar
 
Join Date: Feb 2003
Posts: 184
Mentioned: 1 Post(s)
Tagged: 0 Thread(s)
Quoted: 60 Post(s)
Liked: 45
I actually posted a question about dual laser DLPs in the 20,000+ forum, about rainbows.

I wondered how a dual laser single chip DLP impacted the rainbow effect. With a single white light source I could get my head around the different colour wheels. But with dual laser, you have two blue light sources and one red light source. There are two wheels, a phosphor wheel to create green and yellow from the blue lasers and a RGBY colour wheel.



https://www.avaustralia.com.au/optom...dlp-projector/

This approach must lead to a different manifestation of the rainbow effect. Much like how Dave Harper describes the RBE with a just blue lasers being different from lamp based RBE.
Dave Harper likes this.
Nick Laslett is offline  
post #26 of 33 Old 01-29-2019, 05:26 AM
AVS Forum Special Member
 
Ruined's Avatar
 
Join Date: Jul 2002
Posts: 6,827
Mentioned: 35 Post(s)
Tagged: 0 Thread(s)
Quoted: 2612 Post(s)
Liked: 1250
Quote:
Originally Posted by Nick Laslett View Post
I actually posted a question about dual laser DLPs in the 20,000+ forum, about rainbows.

I wondered how a dual laser single chip DLP impacted the rainbow effect. With a single white light source I could get my head around the different colour wheels. But with dual laser, you have two blue light sources and one red light source. There are two wheels, a phosphor wheel to create green and yellow from the blue lasers and a RGBY colour wheel.



https://www.avaustralia.com.au/optom...dlp-projector/

This approach must lead to a different manifestation of the rainbow effect. Much like how Dave Harper describes the RBE with a just blue lasers being different from lamp based RBE.
If the final wheel is a 2.5x RGBY equivalent as benq specs and diagrams indicate, then that becomes the RBE bottleneck even with dual lasers. There is still only one DMD which means only one color can be displayed at a time - and the final wheel config determines the speed of sequential color.

You could get by the RBE bottleneck with 3 lasers or 3 Leds and eliminaing final wheel. Or you could improve it with final wheel RGBRGB equivalent.

Also image u posted has 3 lasers not 2. But not setup for final wheel elimination / low rbe. Instead setup for max brightness.

Last edited by Ruined; 01-29-2019 at 05:48 AM.
Ruined is offline  
post #27 of 33 Old 01-30-2019, 04:46 PM
AVS Forum Special Member
 
Ruined's Avatar
 
Join Date: Jul 2002
Posts: 6,827
Mentioned: 35 Post(s)
Tagged: 0 Thread(s)
Quoted: 2612 Post(s)
Liked: 1250
Alright, here is a way I figured out that MAYBE you could get 5x speed out of a 4 segment RGBY wheel with 2 lasers, but it seems cost prohibitive:

1. The final wheel is physically RGBY equivalent, however the final output per rotation is RGBYGRYB - doubling the speed
2. Way you accomplish this is have two lasers, staggered on opposite sides of the color wheel
3. Each laser only illuminates half of its segment, then it switches over to the other laser to illuminate its half of the opposing segment
4. So assuming R segment opposes G segment, the 1st half of R is illuminated by laser 1 while the 2nd half of G is illuminated by laser 2; then the 1st half of B is illuminated by laser 1 while the 2nd half of Y is illuminated by laser 2. This continues in cycle and eventually all segments of the wheel will be fully lit by alternating beams.
5. The two beams are converged via a prism following the color wheel


I have no idea how this would look with the colors being out of order and further it seems like a much costlier solution than simply using a final 6 segment color wheel. Also, this is not how the BenQ LK970/LK990 work per BenQ, their diagram show a standard 2.5x RGBY 1-beam setup.
Ruined is offline  
post #28 of 33 Old 01-30-2019, 08:11 PM
 
Dave Harper's Avatar
 
Join Date: Feb 2000
Location: Paradise on Earth
Posts: 6,554
Mentioned: 62 Post(s)
Tagged: 1 Thread(s)
Quoted: 3159 Post(s)
Liked: 1723
Quote:
Originally Posted by Ruined View Post
..........

I have no idea how this would look with the colors being out of order and further it seems like a much costlier solution than simply using a final 6 segment color wheel. Also, this is not how the BenQ LK970/LK990 work per BenQ, their diagram show a standard 2.5x RGBY 1-beam setup.

...........I don’t care how they make the sausage, but I sure do like the taste of it!!!
Dave Harper is offline  
post #29 of 33 Old 04-23-2019, 01:03 PM
Newbie
 
Join Date: Jul 2015
Posts: 4
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 2 Post(s)
Liked: 1
Hi all,

I'm trying to clarify something that I've only seen discussed in passing, and with conflicting answers. Do these Optoma pixel shifters deliver a true 24p playback, or only with pulldown or interpolation?

I've heard two answers: (1) that they display 24p as intended, and (2) that they use pulldown to shift it into 60hz. Which is correct?

Currently, I have a Benq 1070, and I actually like the native 24p "judderiness" on, say, slow pans. I certainly hate the buttery smoothness of many frame interpolation schemes.

Do I have to give up 24fps by going to a 4K projector?

Also, is there any difference in this between the Optoma 50/51A and the 60, 65 series?

Thanks,

Josh
Josh C. is offline  
post #30 of 33 Old 08-11-2019, 02:19 PM
Newbie
 
Join Date: Aug 2019
Location: Germany
Posts: 6
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 5 Post(s)
Liked: 0
Question

Quote:
Originally Posted by sage11x View Post
Straight from my contact at BenQ:

All of HT2050A, HT2150ST, and HT3050 are with 7200rpm (2x=60*60*2=7200) for the speed of its wheel, and because there are 2 sets of RGB segments, so we call 4x on its “color wheel speed”.
For HT2550, it’s also with the same 7200rpm on its wheel speed, and because it’s with RGBRGB color wheels and also with pixel-shifting (the 2 sets of RGB segments on 1080p ones are used for one single image after spinning once, the 2 sets of RGB segments on HT2550 are used for different images after spinning once), we call 4x on its “’pixel-shifting’ color wheel speed”.

Pay attention to that last part: the RGB segments on a 1080p DLP are used to flash the same image within one rotation (half a frame) while on the 4K model each RGB pass is used for a different image as each is one of the four images that combine to form the full 4K composite frame-- 1 frame sees two full rotations of the color wheel.

The speed is the same. The difference is a single HD image being flashed four times per frame vs four different HD images-- each a part of the whole 4K frame-- being flashed.

I went into some really deep detail on how this works in the W2700 thread as far as where the shift is within each half and full rotation of the wheel. I’m too exhausted to explain that again here. Suffice to say we need to stop referring to the wheel in 2X or 4X terms. All consumer color wheels are 7200rpm (120Hz) and what matters is the number and compliment of the slices.

Also, none of the 4K DLPs are capable of displaying 24fps natively. Due to the requirements of the pixel shift (120Hz for the DLP66 and 240Hz for the DLP47). everything has to be converted to 60Hz. That’s is why the relative ‘speed’ is the same regardless of content. I.E. you’re not going to see a different quoted speed for 24Hz like you might with a 1080p RGBRGB model.


edit: general cleanup and grammar corrections as my original post was from my phone.

Very interesting read! And a WIDE field of Information to cover and understand...


BenQ Support here in Germany wrote me in an Email the following Color Wheel Speeds for the W1700 (HT2550):
W1700 Farbradgeschwindigkeit:
2D: 24HZ 5X; 50Hz 2X; 60Hz, 2X
3D: 100Hz 120Hz


So does this mean in terms of RBE the 2D 24hz is "only" 240hz (24x5 = 120 x 2 (RGBRGB wheel) =240hz??


On the other hand I´ve seen a couple of Videoclips where People made some Captures of Gaming and/or Movies where I could see on an Benq W2000 (HT3050))with 6x RGBRGB Color wheel some Rainbow artifacts and on an TK 800 I could see not one - even the TK 800 has NOT an RGBRGB wheel...
THXfreak is online now  
Sponsored Links
Advertisement
 
Reply Digital Projectors - Under $3,000 USD MSRP

Thread Tools
Show Printable Version Show Printable Version
Email this Page Email this Page


Forum Jump: 

Posting Rules  
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts

BB code is On
Smilies are On
[IMG] code is On
HTML code is Off
Trackbacks are Off
Pingbacks are Off
Refbacks are Off