News Gaming industry insiders say cutting-edge graphics cost too much to make for AAA games

Page 3 - Seeking answers? Join the Tom's Hardware community: where nearly two million members share solutions and discuss the latest tech.
I'm pretty sure nobody is working on just TAA, any more. The industry's solution is to use a neural pass, like DLSS. I think TAA has enough tricky corner cases that fusing it with other techniques, via a model that's optimized to pick the best strategy in each situation, is the way they decided to go.
Actually DLSS is TAA but using a neural network to do the temporal math. A game has to support TAA in order for DLSS to work
 
Personally, I'm far more interested in opponent AI, storyline, and escapism than I am in graphics. The gaming industry seems to be allergic to improving things that aren't clearly cinematic (cutscene animations, acting, and similar). The idea that AAA's can't afford to continue pushing graphics isn't much of a concern because graphics have been good enough for the past decade or so.

The problem here is that I think AAA is complaining about graphics cost so they can shift more resources to live services, sequels, and other locked down experiences. I think gaming peaked with the 2000-2013 time period (PS2, XBox, PS3, XBox360). AAA needs to refocus on what that era did right before the Indies eat their lunch. We had such great variety in gaming experiences during that time period with many games encouraging their users to improve the experience for themselves and others. Today it seems like AAA just regurgitates relatively few games designs and IP that worked well in the past -- no risk taking, very few new IP, locked down experiences, and fixation on graphical fidelity.

I hope AAA gaming changes. I don't want to buy a 5090. I can get photorealistic graphics just by walking away from my computer. AAA needs to focus on the non-graphical stuff.

Gaming revenue 2000 - $43 billion
Gaming revenue 2013 - $85 billion
Gaming Revenue 2021 - $190 billion
Gaming Revenue 2023 - $182 billion

2021 is the peak by a considerable amount.
 
I'm not much of a gamer, so maybe my opinion doesn't count for much, but there were a couple recent examples of well-known franchises where I considered picking up an earlier installment of a game made for PS4, but then decided to pass after seeing gameplay videos showing graphics on par almost with what a PS3 could've managed. On the contrary, I've bought a couple PS5-exclusives, just to see what the console can do. Also, when it comes to fighting games, I'm much more a fan of realism (Tekken) than cartoony (Street fighter).

In some sense I wonder if the gaming market hasn't already sorted to a point where you can't be very successful on playstation or XBox without graphics at a certain level, because most of the people who don't care about that stuff are using Nintendo.

Yet PS5 and overall game of the year seems to be Astro. Cartoonish, cute, but well executed.
 
Actually DLSS is TAA but using a neural network to do the temporal math. A game has to support TAA in order for DLSS to work
Yes, I know. Sorry if that wasn't clear, but what I meant is that they're using a neural network to deal with TAA's issues.

The main difference is that DLSS can weight-down or even ignore the temporal sample, in cases where it's not the best strategy. That's why I say you can't rely on DLSS to paper over glitches in quite the same way as TAA.
 
There is plenty of money to be made with games, and having high-quality visuals is pretty important in being able to be rated as a AAA game.

Game developers have altered their target audience to be a small subset of the society, when you are selling to 100% of the society, you can sell tens of millions of copies, and that allows you to spend half a billion dollars making the game.

The problem is that they have started making the games for a very vocal 3% of the population, and instead of having the ability to sell to up to 50 million potential customers, they can sell to at most 1.5 million, but the thing is that this extremely vocal 3% of the population is crazy and thus impossible to actually cater to and convince to buy. When you have a 50 million person target audience, you are still only going to sell about 2-4 million copies. When you have a 1.5 million person target audience you are looking at 60,000 to 120,000 sales volumes if you hit it off with that subgroup.

It is not the problem of having top quality graphics that is causing the losses in the industry, it is the loss of direction in the industry that is losing the audience.
 
Yes, I know. Sorry if that wasn't clear, but what I meant is that they're using a neural network to deal with TAA's issues.

The main difference is that DLSS can weight-down or even ignore the temporal sample, in cases where it's not the best strategy. That's why I say you can't rely on DLSS to paper over glitches in quite the same way as TAA.
Gotcha, thank you for clarifying!

I think there is vital information missing in our discussion which is that: if a console port comes to PC and TAA is a forced on requirement, unless Nvidia forces TAA off, wouldn’t the DLSS neural network see the product of TAA (the glitch fixing, etc.) and determine it must do that as well to produce to successfully emulate the high resolution frames it was taught as its benchmark?
 
Last edited:
This has been going on for along tine... AAA games will focus on graphics, sometimes compromising how it plays even on the game systems, restricting what PCs it runs well enough on, and sometimes they seem to overlook if, you know, the game is actually fun to play. And with the huge costs 1 flop could sink the company (well not Sony or Microsoft. but the standalone game companies.)

To me it's like a special effects laden movie. The special effects are fine, but if the acting is bad, there's no plot, and the writring is bad then to me the movie sucks anyway. and a well directed movie with good writing and director can be great with limited use of inexpensive effects.
 
To me it's like a special effects laden movie. The special effects are fine, but if the acting is bad, there's no plot, and the writring is bad then to me the movie sucks anyway. and a well directed movie with good writing and director can be great with limited use of inexpensive effects.
Well, yes. But, there are certain kinds of movies you just can't make without a big effects budget, like Dune. It would've been a very different movie, if someone tried to do the whole thing with just closeup interior shots (or people standing in a generic desert setting), few extras, no sand worms, etc.
 
Funny, this literal exact same argument was made some 8 years ago and here we are again. Yet ther are several reports from within the industry (SmashJT(youtube)) that these AAA game companies are laying off most of their experienced workers for in many cases unqualified or DEI reasons on top of the checkbox culture going on within right now. So yes it is going to be harder and is brought on by their own doing.

Kids are easily fascinated with the more cartoon style because that's what they see on TV so it makes sense if that is your target audience but if you still want to call yourself a AAA company then you need to put in the work for something that people actually want to play and not play it safe like relying solely on IP.. or shove unhinged DEI directly in our face and say we(the consumer) are the problem.
 
I belive AAA games from the last 5 years or so are already beautiful enough. This push for more graphics is hitting a wall of diminishing returns.

It is.

I'm 100% happy with gaming on my 4K OLED. Same goes for watching movies. No need for 8/16K resolutions. My eyes are 50 years old... and 4K is the peak IMO. There does come a point where your eyes won't notice differences.

Here... let me Google it...

For example, to see 4K, a viewer would need to be sitting 10 feet in front of a 120-inch screen. To see 8K improve from 4K, the viewer would need to be 10 feet from a 280-inch screen.
Most people won't be able to tell the difference between 4K and 8K outside of very large displays. In most living room setups, the step up to 8K will likely be unnoticeable.