SoylentNews is people
SoylentNews
New 8K OLED Displays for Tablets and Laptops: 8.3 and 13.3 Inches
Semiconductor Energy Laboratory, a technology developer from Japan, has developed the industry's first 8.3 and 13.3-inch OLED displays featuring an 8K resolution. The monitors use crystalline oxide semiconductor technology and they are likely preliminary designs for future product commercialization. The company also recently showcased a bendable 8.6-inch OLED panel, potentially for a foldable tablet or smartphone.
Both of SEL's OLED panels featuring a 7680×4320 resolution use a color filter that relies on CAAC-IGZO (c-axis aligned crystalline indium gallium zinc oxide) material. The 8.3-inch 8K panel [boasts] a rather high pixel density of 1061 pixels per inch and has a refresh rate of 60 Hz. The larger 13.3-inch 8K panel features a pixel density of 662 PPI, but has a refresh rate of 120 Hz, which is particularly high for an OLED. The 8.3-inch 8Kp60 OLED was demonstrated last month at SEMICON Japan, whereas the 13.3-inch 8Kp120 OLED currently exists only in SEL's labs.
(Score: 0) by Anonymous Coward on Sunday December 23 2018, @09:23AM (4 children)
If these go mainstream, looks like new VR gear is on the horizon. Any info about refresh rates supported by these tiny 8k screens?
(Score: 2) by takyon on Sunday December 23 2018, @02:36PM (3 children)
Smaller one is 60 Hz, larger is 120 Hz. I assume that is to try to differentiate them and they could make a smaller 120 Hz panel.
The ultimate goal for VR is probably 16K over two panels, with a 220° horizontal, 150° vertical FOV [soylentnews.org] (possibly angled like StarVR, or using some kind of flat meta lens to make it more compact), at 240 Hz.
[SIG] 10/28/2017: Soylent Upgrade v14 [soylentnews.org]
(Score: 2) by Immerman on Sunday December 23 2018, @05:31PM (2 children)
240Hz would be nice, but it's going to take a monster video card to actually drive a display that quickly. Especially a pair of 8k displays.
Of course, one of the other benefits of higher refresh rates is more graceful stepdown when you can't hit every frame - a 240Hz display can also be updated at 120, 80, 60, and 48 Hz, while a 120Hz display degrades straight to 60, and 40Hz. I wonder if something like freesync/gsync wouldn't be valuable for VR, to allow the refresh rate to degrade far more smoothly.
(Score: 2) by takyon on Sunday December 23 2018, @08:56PM (1 child)
Whether or not they actually call it Free/G-Sync, there is plenty of effort being put into lowering latency and delivering smooth framerates:
https://www.amd.com/en/technologies/vr [amd.com]
https://www.pcgamer.com/amd-crimson-relive-1741-smooths-out-vr-performance-on-rift-and-vive/ [pcgamer.com]
This should all be pretty transparent for the user, especially in the case of standalone headsets that pack a display and GPU.
I am predicting that even standalone headsets will get monstrous GPUs, due to advancements [soylentnews.org] that will forestall the end of Moore's law scaling and usher in 3D integrated circuits... with low power consumption. However, they might not even be needed. Foveated rendering [theverge.com] in particular will massively reduce the burden, and the human eye doesn't do much in ~4 ms [quora.com] (1 frame at 240 Hz). Foveated rendering is why Google and LG are already talking about using a 1,443 PPI panel [roadtovr.com] in a standalone headset. Other techniques can also be useful, for example, Seurat [google.com].
If latency needs to be cut further, we can again talk about 3D ICs. Simply integrating DRAM into the chip [monolithic3d.com] could produce massive speedups by nearly eliminating latency between the CPU and DRAM. We could see some further improvements if we can use "universal memory", meaning a post-NAND storage/memory replacement that is as fast as DRAM or faster, with indefinite endurance, and much higher density (terabytes instead of gigabytes).
If you can do something like put 100-1,000 teraflops of GPU performance in a standalone headset, along with CPU, TPU, NPU, etc. as needed, then it should also be possible to do real-time ray tracing. Nvidia's RTX 2000-series GPUs, as hyped as they were, are just 2D trash. Once we go 3D, there will be plenty of room to dedicate to ray tracing and other functions.
[SIG] 10/28/2017: Soylent Upgrade v14 [soylentnews.org]
(Score: 2) by Immerman on Monday December 24 2018, @01:49AM
That's some cool stuff, and a great way to compensate for rendering latency as much as possible, but it's completely independent from variable screen refresh technologies like freesync/gsync, which require hardware support for variable refresh rate. Essentially they're a way to eliminate the "tearing" of non-vsynced rendering by reversing how v-sync works. Instead of trying to sync the rendering engine to a fixed hardware framerate, it syncs the display updates to the rendering engine's infinitely variable output framerate.
And in fact one of the justifications for a high-framerate gaming monitor (at least among those for whom money *is* an object) has long been to allow a wider range of target framerates, since a monitor can essentially display at any integer divisor of its native rate. A 60Hz monitor can only display at 60 and 30Hz - 20Hz is solidly below the perceptual threshold of continuous motion, and even 30Hz is obviously sub-par to most people. A game driving a 120Hz monitor though can target 60, 40, or 30Hz, allowing for more options in the tradeoff between visual quality and speed. 240 would be even better, as it can target 120, 80, 60, 48, 40, 34, and 30. And of course it also helps with stuttering, which is the visual disruption created by shifting between those effective rates when the next frame isn't quite ready in time. The smaller the difference in frame duration between steps, the less obvious the stutter will be.
With a "reverse v-sync" display, that's eliminated as much as possible. If a rendering engine can only manage 57Hz this frame, and 54 the next, then the display refresh rate is just reduced for each frame to match.