4K = Temporary stagnation?

[w35t]

And now you just unfolded that minority, into subgroups that are even smaller and have overlap between them. What's your point?

Since I was arguing for 4K not against it. I don't even get it why do you address this at me.

You're arguing that people interested in higher refresh rates and higher pixel density (within the realm of those interested in panel technology) are a minority, I strongly disagree, that's my point, addressed at you.

 

[Youn]

AA is still useful at high resolutions

 

[M76]

You're arguing that people interested in higher refresh rates and higher pixel density (within the realm of those interested in panel technology) are a minority, I strongly disagree, that's my point, addressed at you.

What is this now about people interested in panel technology? That's new variable you just dropped in. People interested in panel technology are elite enthusiasts, even I don't give a damn about the technology side of displays. Of course if you just look at them they'll want everything, but we are talking about every consumer here. And I'm fairly certain that from that group a very high percentage doesn't even have a clue what refresh rate is.

I stand by my view that anything above 60Hz is just a fancy of a few people. There are a few people who knows about it, just doesn't care, and then there are the rest who doesn't even know that it's a thing.

You act as if the consumer market starts and ends with enthusiasts.

 

[Domingo]

My television is hooked up to everything from my A/V receiver, to my cable, to my consoles, to my PC.
I fully plan to jump on the 4K bandwagon as soon as HDMI 2.0 has fleshed the technology out and prices aren't insane. 'Til then, it's 1080p for me. As long as the details are turned up and my framerates are high, I think it looks "good enough." At least for now. Give me 1080p/60fps over 4K/30-40fps any day.

 

[w35t]

What is this now about people interested in panel technology? That's new variable you just dropped in. People interested in panel technology are elite enthusiasts, even I don't give a damn about the technology side of displays. Of course if you just look at them they'll want everything, but we are talking about every consumer here. And I'm fairly certain that from that group a very high percentage doesn't even have a clue what refresh rate is.

I stand by my view that anything above 60Hz is just a fancy of a few people. There are a few people who knows about it, just doesn't care, and then there are the rest who doesn't even know that it's a thing.

You act as if the consumer market starts and ends with enthusiasts.

I'm sorry I figured we were talking about people who care about monitors. If we're including people that don't even have computers then yes, I suppose you're right. People wanting higher than 60Hz refresh rates and higher pixel densities are a minority.

 

[M76]

I'm sorry I figured we were talking about people who care about monitors. If we're including people that don't even have computers then yes, I suppose you're right. People wanting higher than 60Hz refresh rates and higher pixel densities are a minority.

No problem, imo for a technology to become affordable and mainstream everyone or at least a lot of people have to get on board. We have to sell 4K to everyone to make it cheap. But I don't see how we could sell 120Hz to everyone, as nothing but Highend PCs can benefit from that atm.

 

[pigwalk]

Not really PC gaming related, but all the people I've run into at work that have 4K TV's have way more money than knowledge. Not a single one purchased knowing that there is no 4k Cable or even Satellite in the area, nor even any idea where to get any 4k content.

 

[M76]

Not really PC gaming related, but all the people I've run into at work that have 4K TV's have way more money than knowledge. Not a single one purchased knowing that there is no 4k Cable or even Satellite in the area, nor even any idea where to get any 4k content.

That's to be expected, but regardless of their ignorance they benefit the market. The more 4K sets sold, the sooner cable providers start to realize there is a market for 4K.

In my country HDTVs were used as table decoration early on, everyone wanted one because the guy next door had one, but we still had Analog SD broadcasting which looked awful on early hd ready TVs. It looked immensely better on old CRT tvs.

 

[Domingo]

IMO, the Xbox 360 was the best reason to have an HDTV for a while. There weren't that many stations broadcasting in it and Blu-Ray wasn't really worthwhile yet.

 

[mustang_steve]

Netflix has been offering 4k content for several months now as an opt-in program and now offers it as an add-on service.

The content is trickling in. It's a matter of market potential. Until the hardware is out there, there is no reason to make content in it. The industry has learned from the HD-DVD/Bluray fiasco that it's better to have the change be demand driven. You can't have true demand unless the consumer can actually consume it now however, which is why the hardware has to be out there first.

Doing it the other way around just leads to gambling on a format. We saw how lame 480i was (and likely forgot as it was that much of a failure, being eclipsed by better technologies in mere months).

 

[wabbitseason]

Having to choose between the two I would rather see a focus improvements in real time 3D (such as shaders and lighting) and post process capability (DoF, AA, CC etc.), than an endless increase in resolution alone. I am aware that resolution is straight forward and realistic real time shader technology of decent DoF is not

Completely agreed. MOAR PIXELS does not equal better visual quality. It does mean dramatically higher GPU requirements, though. I'm still delighted to game on a 1080p plasma. Rendering quality still hasn't approached what 1080p is capable of. If anyone needs proof simply look at the CGI in 1080p movies. Do you think anyone would care about 4K if we could play movie CGI real-time? Hell no. 4K is another gimmick to steal our dollars and extend the life of abysmal LCD technology.

 

[wonderfield]

No problem, imo for a technology to become affordable and mainstream everyone or at least a lot of people have to get on board. We have to sell 4K to everyone to make it cheap. But I don't see how we could sell 120Hz to everyone, as nothing but Highend PCs can benefit from that atm.

In a way, they already have: many TVs can display interpolated content at panel refresh rates of 120 Hz or higher. They just don't accept 120 Hz as input.

PC displays are obviously a different animal.

 

[M76]

In a way, they already have: many TVs can display interpolated content at panel refresh rates of 120 Hz or higher. They just don't accept 120 Hz as input.

PC displays are obviously a different animal.

But we were talking about 120hz content, not some interpolation. Anyway for interpolation to work first you'd have to remove motion blur, and since every movie has it, interpolation only makes thing worse imo.

 

[nakedhand]

It's kind of interesting that you chose Monsters Inc. as an example. Pixar's RenderMan was only recently (for 2013's Monsters University) overhauled to support global illumination. We had games with global illumination two years prior. Obviously, RenderMan gets you much closer to the ideal Cornell's box than Enlighten does, but that's not really the point I'm trying to make here. The point is that it's not cut and dry.

Keep in mind that back in 2001, it took Pixar 12 hours to render a single frame with Sully in it. With a render farm. I have no idea how big that render farm was back then, but it was certainly big. Even with the absolute breakneck pace of GPU evolution, that a single chip today (assumably) can't render the same frames in 16 milliseconds shouldn't be too surprising, and certainly shouldn't be upsetting. In what other industry is a 2,700,000% improvement in anything realized within 14 years?

(Just as an aside: for Monsters University, the aggregate frame time was 29 hours)


Rendering isn't really the long pole in the tent as far as those things are concerned.

Let me start by agreeing with you that computational power improvements have been fantastic the last two decades. No one should disagree with that. How ever I think you are missing my point and making a big mistake, when comparing real time and pre render the way you do, to drive home a point about a 2.700.000% improvement.
Your narrative seems to be that we have come a long way in real time graphics and that we should be content with this, and that the current level is good enough for now. My narrative is this: We are in the dawn of real time graphics, the current level is not very evolved, it is visually incomplete even for 1080p (which makes me think 4K is less important right now and tomorrow).

OK, back to your point. Suggestion that there is a correlation between required amount of computational power to process a similar pre render and real time scene as you do in your 2.700.000% example, and trying to impress upon everyone this staggering improvement, is not fair. It is two different worlds entirely. A modern GPU can NOT even process the Monsters Inc compositing tree in Milli seconds, never mind the multitude of Renderman renders. I am sure you know why, but for others who might not, here is an explanation. The whole crux of real time is to get acceptable results with as little computational power necessary as possible (so the methodology and final amount of data is vastly different). That can't be said of pre render. In a pre render environment you are trying to achieve a precise version of the vision the artists had. The bigger the budget, the more likely it is, that this precision is your number one priority, which makes hour long rendertimes possible (acceptable). Things need to look tight and neat "expensive filtering" on the big screen, so tonnes of code is being processed creating those frames and then there is compositing af everything. A real time approximation of a given same scene is built entirely different with much more "cheating" involved. For example baked lights and faked reflections and refractions to mention a tiny fraction. Hence it has to compute way less data. So the comparison is funky and will produce useless numbers like the one you invented.

Regarding rendering not being a big tent pole in a tent etc. I just plainly disagree. And I am someone who will take great game mechanics above all, any day.

My beef is with the prioritization of technologies. Let us do more in terms of real time development for 1080p before adding more freaking pixels. My beef is with the developers and investors, but foremost with the industries (broadcast, consoles). Frankly I can not fathom that people are happy with the current status. Especially, like Wabbitseason pointed out, considering the available CGI in movies today. People have great reference now. Maybe it is lack of education or bad eye sight, sorry for being blunt, but how do people get exited about 4K Crysis 3 after seeing say Avengers in 1080p? Why don't they want more object complexity for instance? Some think they can get everything at once, some sheep are duped into 4K without understanding that they are only getting more of the same pixels, not any other added fidelity. Some are going to love it for good reasons beyond gaming. I for one understand 4K for real footage and as a photographer I am excited about 4K, because the resolution of reality is still pretty good! Real time graphics Anno 2015, not so much. Maybe it is a lost battle hoping for better real time until we are way into the 4K paradigm as you said, then are we going to suffer the 8K hit?

It makes me wonder if the developers drive to improve the difficult real time technologies has faded as a result of the impending 4K future? Have they lost their innovative spirit to the "up res' ing" gig of the industry because they know it will get too taxing for the GPUs to properly run? And who are these people pushing for 4K in the first place anyway? Not the early adopters, the film studios. Is it the broadcast industry?

Did some digging. In 2001 they upgraded their farm to a cluster of 250 sun boxes with eight ultraSPARC III processors each. With a 750mhz clock. So a total of 2000 CPUs. Each one of those were capable of 1.5 GFLOPS, for a combined total of 2000*1.5=3 TFLOPS.

A single Titan Z can push 2.7 TFLOPS in double precision. That's impressive.

Impressive! And keep in mind that those 3 TFLOPS were used to run a software renderer, meaning x86 code, something that the Titan can't execute (correct me if I am wrong, these things change fast hehe).

AA is still useful at high resolutions

Indeed!

As long as the details are turned up and my framerates are high, I think it looks "good enough." At least for now. Give me 1080p/60fps over 4K/30-40fps any day.

I also really enjoy high frame rates, especially crucial in FPS games. One more reason to be hesitant about 4K.

Not really PC gaming related, but all the people I've run into at work that have 4K TV's have way more money than knowledge. Not a single one purchased knowing that there is no 4k Cable or even Satellite in the area, nor even any idea where to get any 4k content.

Classic Many of my colleagues had full HD 3D televisions already a few years ago. Did they know about technology, absolutely not. But they were told they had to have these flat screen TVs

Completely agreed. MOAR PIXELS does not equal better visual quality. It does mean dramatically higher GPU requirements, though. I'm still delighted to game on a 1080p plasma. Rendering quality still hasn't approached what 1080p is capable of. If anyone needs proof simply look at the CGI in 1080p movies. Do you think anyone would care about 4K if we could play movie CGI real-time? Hell no.

Thank you for pointing this out. I could not agree any more. There is so much to be added to a 1080p frame before it looks vaguely like modern CGI. I honestly think most people are unaware of the impact of 4K on real time and the room there still is for improvement.

Netflix has been offering 4k content for several months now as an opt-in program and now offers it as an add-on service.

The content is trickling in. It's a matter of market potential. Until the hardware is out there, there is no reason to make content in it. The industry has learned from the HD-DVD/Bluray fiasco that it's better to have the change be demand driven. You can't have true demand unless the consumer can actually consume it now however, which is why the hardware has to be out there first.

Doing it the other way around just leads to gambling on a format. We saw how lame 480i was (and likely forgot as it was that much of a failure, being eclipsed by better technologies in mere months).

Interesting. So they have learned that the consumers pushes it through demand. But at the same time we are going to get it whether we want it or not? It certainly happening, slowly. I get your point and I think you are right, money talks of course. Nobody wants an expensive failure and the industry must be hoping for great demand since so much is already invested in this. Who drives it to begin with? Broadcast associations?

 

[DLGenesis]

AA is more effective at higher resolutions

 

[DLGenesis]

AA is still useful at high resolutions

AA is more effective at higher resolutions 8)

 

[M76]

Thank you for pointing this out. I could not agree any more. There is so much to be added to a 1080p frame before it looks vaguely like modern CGI. I honestly think most people are unaware of the impact of 4K on real time and the room there still is for improvement.

After you finished patting each other on the back, let me tell you that. 4K is free visual improvement. The developer doesn't have to put in extra work to achieve it. So I can make my games look better, even older ones. Of course there is improvement to be made in other places, but noone says those shouldn't be made because of 4K. I don't understand why would you even have the idea that 4K somehow will kill progress. Or that we should hold off on higher resolution until we can achieve full cgi on smaller ones. By that logic we should still be playing at 320x200. Did progress suddenly stop when we switched from 320x200 to 640x480? NO. did it stop when we went from 640x480 to 1280x720 or 1280x1024? NO, Did it stop when we went from 1280x1024 to 1920x1080? No it still didn't stop. Why would it stop because of 4K then?
Where do you get this idea from?

 

[wabbitseason]

In response to the poorly-worded strawman above:

, let me tell you that. 4K is free visual improvement. The developer doesn't have to put in extra work to achieve it. So I can make my games look better, even older ones.

Let's see:

*4K isn't remotely free, it requires four times the graphics processing power of 1080p. Not to mention buying the 4K panels themselves, thereby funding the despicable businessmen who are desperately trying to extend the life of obsolete LCD technology.
*More pixels does not equal better quality, regardless of your hilarious hang-ups to the contrary. I hope you enjoy many imagined "graphical improvements" running old games with more pixels.
*Nobody said resolution improvements should stop.
*4K is being pitched as the next revolution in display technology, when it fact it does practically nothing to improve gaming quality. As you obviously failed to understand this point, lets say it again just for little old you:

Gaming rendering quality has not even approached the limit of what 1080p is capable of. Look at the CGI in any recent high-budget film. Until rendering quality improves, it makes zero sense to move to 4K for gaming. You can waste your money, in fact I hope you have and continue to. You're the ignorant consumer they are pandering to when they pitch 4K as the next revolution. You're the one who buys into 3D and curved displays, while the marketing department laughs incredulously and says "I can't believe they're falling for this".

Meanwhile rendering quality stagnates. Why? This is a direct result of your existence. It is your fault. Marketing has told research that you will buy 4K and not demand that render quality be prioritized. So everybody goes back to business as usual, just with FOUR TIMES THE PIXELS!!!111

 

[Youn]

I hope they release 8K screens soon... I mean what do you expect them to do, wait for other tech to play catch-up?

 

[wabbitseason]

I hope they release 8K screens soon... I mean what do you expect them to do, wait for other tech to play catch-up?

I'll tell you what I expect them to do: produce truly next-generation display technology at a reasonable price (ex: SED, OLED, crystal LED). All of these technologies are workable. All of them are being held back for a variety of reasons.

SED: The tech is so fundamentally robust that consumers would have no need to upgrade on the desired 3-5 year cycle. That + patent trolling = canned.

OLED: Legitimate concerns over differential aging of the different-colored OLEDs. Plus manufacturers have discovered that by dressing up ordinary LCD technology with "curved screens" and "3D" and "4K", they can continue to sell ignorant consumers inferior technology at higher margins. What happens? Foot dragging and stringing along consumers.

Crystal LED: Cost, too little research done to simplify manufacturing and bring cost down, no economy of scale. Plus we all know Sony is just dumb.

4K LCDs are not a "revolutionary improvement", nor will 8K LCDs be. They've simply packed more pixels into the same area--big deal. The tech is still the same crap they've been selling us for over a decade. It's not impressive, at all. How far will you let them string you along with LCDs? This doesn't even consider the glacial pace of rendering quality improvements of late.

 

[wonderfield]

My narrative is this: We are in the dawn of real time graphics, the current level is not very evolved, it is visually incomplete even for 1080p (which makes me think 4K is less important right now and tomorrow).

Yes, it is visually incomplete. We won't have 'completed' real-time rendering of fully photoreal environments for another hundred years or so, so I'm not sure if incompleteness is really a valid criticism of today's graphics. That's particularly true when you're talking about strict rasterization, in which so much achieved through blunt approximations.

A real time approximation of a given same scene is built entirely different with much more "cheating" involved. For example baked lights and faked reflections and refractions to mention a tiny fraction. Hence it has to compute way less data. So the comparison is funky and will produce useless numbers like the one you invented.

It's not an invented number. In order to achieve Pixar quality, that is approximately the percent improvement we would have required of hardware, to achieve real-time Monsters Inc. at 60 Hz with a single GPU. Don't take the figure out of context — it's there to suggest the magnitude of the problem.

If you want less than Pixar-level quality, the number is not accurate. But I think you again miss the point: we simply don't have the computational power to meet your expectations. Game developers push the current consoles exceedingly hard, and what they're able to squeeze out of them is actually pretty remarkable. Market, economic and business factors simply dictate that more powerful PCs don't usually get much time devoted to exploiting what power they have available, but 4K support is a current zero-cost addition and improves visual quality. Every game engine that enumerates display modes in the standard way can support 4K (or any other arbitrary resolution/refresh rate combo) with no additional effort.

Maybe it is lack of education or bad eye sight, sorry for being blunt, but how do people get exited about 4K Crysis 3 after seeing say Avengers in 1080p? Why don't they want more object complexity for instance?

Who claimed people don't want more complexity? To suggest that current graphics are good does not necessarily mean that they're perfect and that all innovation in that space is wasted. Suggesting that current graphics are good is just that.

Impressive! And keep in mind that those 3 TFLOPS were used to run a software renderer, meaning x86 code, something that the Titan can't execute (correct me if I am wrong, these things change fast hehe).

Kepler chips use a proprietary NVIDIA instruction set.