ED-057 wrote:Sumez wrote:I've heard people complain about too much focus on graphics for about twenty years now.
I believe you. Twenty years ago there were bad games with great graphics, and good games that were perceived as not getting their due respect because of unimpressive graphics. But today, we also have GPUs that eat 200W and occasionally unsolder themselves from the PCB, and development budgets that can reach much higher. Although I wonder if these trends will continue much longer. Maybe the market for expensive cutting-edge games will plateau and devs will shift to handhelds.
There has always been a focus on advanced graphics on cutting edge hardware, except for a minority of "retro" or deliberately austere games. The problem is that this constant pressure to look good has grown to larger and larger amounts of effort, often directly or indirectly damaging other sides of game quality.
In the "8-bit era" game graphics production was drawing little sprites of very few colours on graph paper and converting them to raw data, and bleeding edge graphics programming was playing tricks with the video DAC to display more pixels and/or more colours.
Then graphics resolution increased, and drawing sprites became proportionally more demanding and expensive; memory amounts increased, and the number of graphics assets increased too. The same kind of game needed a much higher graphics budget.
The switch to 3D graphics complicated both game code and graphics production; then the relentless development of new techniques added further complexity to 3D engines, while the push towards quality demanded more complex and detailed 3D assets.
Creating game graphics and displaying them requires vastly more skill and effort than it used to; most game developers tend to increase graphics budget at the expense of attention to everything else [1], or simply to make less games (usually sticking to trite "mainstream" designs).
Moreover, the increased complexity of graphics assets makes them too important and too inflexible.
While an old shmup could easily budget a modest amount of sprites and scripting for a later loop or a TLB that the vast majority of players wouldn't ever see, the complex 3D environments of modern games need to be toured according to an iron-fisted script, or they are wasted money.
Consider the difference of cost and turnaround time between arbitrarily modifying a tile map in a game like Commander Keen or Xenon 2000 (two examples for which I've seen the level editors) and adding even an extra house to a Call of Duty level: the design of modern game levels is too inflexible, another reason to focus on scripts and storyboards and graphics rather than on really good design.
[1] Another episode of me vs. modern FPS degeneracy: the main gimmick of Crysis 2 is putting the player in a fairy powder suit that, among other superpowers, can make him temporarily invisible (with a very nice blur effect on the visible hands and weapon).
The novelty fell apart when I turned invisible to approach and shoot cowardly soldier A (this tactic in itself itself a far cry from the thrilling combat of Doom or Unreal) and soldier B, a couple of meters away,
stood idle (turning his back to me and screaming "Where is he! I don't see him!" on the radio to add insult to injury) without joining the firefight, not even looking around. Boom! Headshot!
I can easily imagine the decision that the suit powers have to matter because players have to use them because they make the game cool, even if they ruin basic aspect of the game like fighting, and the decision that cannon fodder AI isn't a priority.