Optoma 240Hz DLP Projector Released!

[Chief Blur Buster]

There's a New 240Hz DLP Projector!

Has anyone obtained an Optoma UHD50X projector yet? It's a 4K DLP projector with a 240Hz 1080p mode!

• Optoma's Website: UHD50X
• Amazon Page: Optoma UHD50X

_________________

Technologically:
I'd love to compare 1080p 60Hz versus 1080p 120Hz versus 1080p 240Hz with TestUFO Photo Test.

Motion will probably look super-smooth, with a slightly amplified banding effect (compared to 60Hz) caused by reduced bits-per-refresh effects of the DLP temporal dithering. It's kind of a tradeoff effect that occurs.

Basically, the motion blurring will be reduced, but other kinds of artifacts become slightly more visible -- depending on its temporal dithering algorithm (motion compensated or non-motion-compensated), including
-- potential microrainbow artifacts in single-pixelwidth details that is no longer hidden by as much as motion blur
-- potential contouring artifacts (similiar to old plasmas).
Many algorithms by many parties exist to mitigate this to some extent (but not fully).

This is because most single-chip DLPs tend to run at ~1440Hz monochome 1bit temporally generating 24bits (1440/60Hz = 24-bits). 1440/60 = 24-bits per refresh cycles, but 1440/240 = 6-bits per refresh cycles (2 bits per color channel). This varies from DLP implementation to implementation, but the moral of the story is that the pixel pulse rate is not quadrupled at the quadrupled Hz, creating some interesting artifacts worth scientific study. Many people miss them (like many miss 3:2 pulldown detection or miss rainbow artifacts) until they notice, then they notice the artifacts.

There are some necessary compromises in temporal dithering algorithms during fast full-framerate motion (without eye tracking sensors + motion compensated temporal dithering + eye tracking compensated temporal dithering algorithms).

Static imagery will look great (no bitdepth decimation), motion will look better, but never as good as a sample-and-hold 240Hz (full color depth) because of bitdepth decimation per Hz. Blur Busters mainly works on full-color-depth-per-Hz displays nowadays, but I also study the science of display artifacts, so I'm less interested in DLP as an enabler of ultra-Hz displays because of my artifact discoveries.

That said, I'm excited that Optoma has included 240Hz modes in their DLP projector -- this is much needed progress.

As a display researcher highly experienced in display motion artifacts, I'm quite interested to see how well Optoma has done 240Hz temporally with a single-chip (for both fixed-gaze-on-moving-motion situation, as well as trakcing-gaze-on-moving-motion situations).

If you have an Optoma UHD50X, I have a question:
I'm especially interested to see what www.testufo.com/photo looks like with 1080p 60Hz, 1080p 120Hz, and 1080p 240Hz? Such as banding or contouring artifacts in the sky (compared to the reference image: paused photo which should show no banding or contouring). See what advanced algorithms Optoma included to compensate, and how good a job they are doing.

______

<Display Science>
Supplemental info: Long term, it would be neat to use four different 1440Hz DLP chips (with concurrent rainbow wheels) to get 240Hz with the same color-depth-per-Hz (24 bits) as single-chip 60Hz. Even so, the 1-bitness of DLPs generate some difficulties with the Vicious Cycle Effect, given increased resolution, increased DPI, and increased refresh rates -- amplifying visibility of refresh rate limitations. This, IMHO, will collide very hard with the per-pixel-Hz limitations of DLP for ultrafast-motion gaming (at least when compared to a good 1ms GtG 240Hz+ monitor). Though, the big screen size and color quality of a good DLP is impeccable!

Over the years, I correctly predicted some display artifacts -- of a variety of display technologies including DLP -- that many other people disbelieved in (important for businesses to put correctly fair valuations on refresh rate innovations), so this is a artifact-science that I am especially interested in. I might purchase this Optoma projector as an upgrade for my home theater, so I'll possibly get to experience some of these artifact-quirks personally myself.

Although I have not been working on any DLP projects for the last while -- some of you may have seen my recent non-DLP collaborations NVIDIA (TestUFO for the 360Hz monitor) and ViewSonic (the Blur Busters Approved programme). Displays have very interesting interactions (whether it's a Hz-raising, or a BFI-enabling, or other innovation -- for example, OLED and IPS don't degrade noticeably with BFI, while TN and DLP tends to degrade with BFI) that few people other than me understand well. It's like physics versus geometry, and BlurBusters approaches displays in ways that aren't reliably predicted by math or scientific papers or 10% measurement cutoffs -- like the strobed red phosphor situation or the stroboscopic artifacts situation.

For the last nearly ten years, Blur Busters, name sake, I'm always interested in artifact-science as I'm classically a very good predictor of artifacts that even NVIDIA consults me from time to time for my Display Einstein brain matter. So I want to study these DLP artifacts in deeper detail. I frequently surprise manufacturers.
</Display Science>

Nontheless, great seeing the refresh rate race arriving to consumer DLP projectors. Very sorely needed!

[Durante]

Hi, I've been waiting for an affordable >60 Hz DLP for a really long time (and also one with playable input lag at 4k), so I immediately bought this one.

I tested it with basic scrolling text etc. right after I got it ~1.5 weeks ago and was really impressed, it's some of the sharpest text in motion I've ever seen on a display (probably since running a high-end CRT at low resolution and ~ 180 Hz decades back).

That said, your post inspired me to have a much closer look.
I can confirm that in the motion tests, there is contouring. In particular, on scrolling photos and other material I see red contours to the left of black pixels that are moving right, on some color transitions.

Sadly, I don't have the equipment to do a scientific test, but I spent roughly 20 minutes trying to move my shitty cell phone camera roughly in sync with the image on display (at 960 px/s), and I got some images that (other than being much blurrier than real life -- as I said, shitty cell phone camera) capture the artifact pretty well.

Of all the tries this is the most extreme and sharp example of the artifacting which I managed to capture:

(This is running at 240 Hz, 960 pixels/s)

Hope that was interesting for you!

[Chief Blur Buster]

Thank you!

I'm very familiar with this -- have seen this color-lagbehind artifact for many years. That's a temporal side effect of color sequential operation during motion -- also the cause of DLP rainbow artifacts. Higher refresh rates will mean the color wheel can only spin about 1 cycle per Hz, creating easier-to-see color separation (rainbow artifacts) during high-resolution high-refreshrate motion tests on DLP projectors.

The extra temporal resolution is very welcome, but you're still limited by the color wheel, the 1-chip of DLP, (and the effects of 1440Hz 1-bit temporal dithering)

Sadly, I don't have the equipment to do a scientific test, but I spent roughly 20 minutes trying to move my shitty cell phone camera roughly in sync with the image on display (at 960 px/s)

I can help with making it much easier!

For hand-wave pursuit camera, it is much easier with a 3rd party camera app such as DSLRCamera, and using Video instead of photo. Configure camera to 1/30sec exposure per frame for 120Hz, or 1/60sec exposure per frame for 240Hz. Camera exposure more critical with DLP to use exact multiple-per-refresh exposure, to keep the correct number of colorwheel cycles per color channel per camera exposure, while simultaneously stacking more than one refresh cycle to the resulting photograph. The camera shutter doesn't need to be aligned with the refresh cycles, but the camera should be attempted (best-effort) to an exact multiple of refresh cycles where feasibly possible -- to avoid unwanted error-margins (bad colors) in easy pursuit camera operations.

Over the course of a minute or two, video is like burst-shooting 1000 photographs, so you step far away from the screen (not too close, it's too hard to follow) -- preferably standing approximately one screenwidth away from the projector screen -- and rotate using your waist while having both stiff arms holding your video smartphone camera.

Smartphone Hand-Wave Pursuit Camera With Projectors
(Slightly modified version of Easy Smartphone Pursuit Camera, but for projectors)
-- Use a pattern with a pursuit camera sync track www.testufo.com/ghosting
-- Stand approximately one screenwidth away. Not too close!
-- Stiff arms, don't move arms, rotate with your waist instead.
-- Use video mode, not photo mode
-- Please include the sync track in resulting images.
-- Highest quality video mode your phone supports.
-- Use an app to force longer exposures per video frame, 1/30sec for 120Hz, or 1/60sec per 240Hz (built-in camera doesn't always let you)
-- Do a few pursuits per video clip
-- Once done, jog through the video with your finger. Then single-frame-step towards it. Then screengrab the clearest video freezeframe you see in your video.

One alternative is to use continuous burst-shoot instead of video. This maximizes your odds. You'll see many frames with lots of bad fringing and errors, but a few freezeframes that look well tracked.

You'd be amazed how scientific even a hand-wave smartphone can become -- with the right camera settings and enough spray of samples ("1000 photos" of a single video clip) to compensate for a shaky hand. Often 1 in 1000 freezeframes will be scientifically stable enough to be analyzed. Metaphorically, the sync track (the horziontal ladder track at www.testufo.com/chosting ...) is literally almost a cryptographic certificate of scientific accuracy -- so make sure to include the sync track in the resulting photographs.

That's how you get good hand-wave pursuits with a projector. Would you be willing to try again with projector-specific instructions?

No worries if you're all tested-out, but it's great when hobbyists have fun with pursuit camera.

[zalo]

Has anyone reverse engineered Optoma/TI's DLP illumination pattern?

It would be very exciting to possibly remove the color wheel and have an ultra-high refresh rate binary projector.

This technique is fairly naive, but speaks to the interesting sensing applications:
https://www.youtube.com/watch?v=XgrGjJUBF_I

(I would expect a low-level 2880z DLP system (like this one?) could speed up the tracking/acquisition by 48x.
Additionally, it should be possible to replace the photodiodes with passive retroreflectors (and a singular photodiode on the projector) by splitting sensing pulses between retroreflectors on the time domain.)

[Chief Blur Buster]

(I would expect a low-level 2880z DLP system (like this one?) could speed up the tracking/acquisition by 48x.

This DLP chip is 2880Hz?

That's not all that too bad for 240Hz (As long as you're not using wobulation, that destroys bit-depth-per-pixel) -- you'd get 8-bit color depth at 120Hz and 4-bit color depth at 240Hz. One would use multiple-refresh temporal dithering to compensate for lack of per-Hz color depth.

One problem is that since 1/2880sec means 1 pixel offset at 2880 pixels/sec, the temporal splitting of colors will create still nasty banding artifacts during the refresh rate race to retina refresh rates. The Vicious Cycle Effect hurts DLP for ultra-high-Hz, higher resolutions, higher refresh rates, faster motion speeds, closer viewing distances, bigger screens, amplifies DLP artifacts massively. Even 2880Hz is not enough if it's limited to 1-bit.

If a fast-responding eye-tracking sensor is added, eye-tracking-compensated temporal dithering at the full-DLP-Hz granularity, can potentially compensate though to completely eliminate DLP contouring and rainbow artifacts. But that won't happen for a theatre DLP projector with multiple people watching the same screen. Or people who don't want to wear glasses/sensors watching their screen.

Due to these limitations, my opinion is that non-DLP will be easier to reach high-temporal-quality (no motion noise, no contouring, no dither trails, no color misalignments, etc) retina refresh rates for cinema-sized screens, unless you're using many DLP chips concurrently (e.g. 24 different 2880 Hz DLP chips to generate full-color-depth 2880Hz color that has no need for temporal color dithering needed, no banding, no contouring, etc).

DLP is a difficult engineering challenge in the refresh rate race to retina refresh rates. I have been there already, and have already advised multiple people away from 1-chip DLP for high-Hz use due to artifacting reducing the value/worth. Many people did not disbelieve me but most of the parties now grudgingly agreed with me (years later). Some of those people even have a University Ph.D degree but I outsmarten them by analyzing differently (like geometry versus physics). Sure, "240Hz looks good" on some DLPs, but when they look at a proper 240Hz MicroLED / OLED prototype... or even a good 240Hz esports monitor.... then they realize that 240Hz isn't all that exactly good on DLP in noise-free/banding-free/dither-free/artifact-free/fringing-free/colordivergence-free motion artifacts -- it's a very hard whac-a-mole.

High-Hz is still useful for DLP, but high-resolution high-Hz looks better on many other technologies due to the vicious cycle effect.

Improved high-Hz DLP color requires multiple-chip approaches.

I'm VERY glad 240Hz has arrived to DLP and it is the best 240Hz ever seen on DLP. The double-Hz of the chip definitely helps (2880Hz), though it is not enough to completely eliminate artifacts (as you've also seen).

This technique is fairly naive, but speaks to the interesting sensing applications:
https://www.youtube.com/watch?v=XgrGjJUBF_I

Fascinating automated projection mapping!

[CubanLegend]

UHD30 also has 240hz, did anyone take comparitive testufo shots of this one as well? https://www.youtube.com/watch?v=poVxiPTJrW0

It's such a shame in "game mode" you cant do any keystone adjustments..

[Chief Blur Buster]

It's such a shame in "game mode" you cant do any keystone adjustments..

Some things I suspect:
- Keystone adds processing latency, and/or
- Keystone scaler might be unable to perform at 240Hz.