I have noticed that some 3D FPS games age well and some do not. Crysis is still like new. Return to Castle Wolfenstein is so old and dated with its Quake 3 Arena engine, but it is still fun to play. But older medal of Honors look goofy today. And that Wolfenstein that came after looks worse somehow. Doom 3 is creepy, but everything looks like it is molded with clay. But Half-Life 2 is still a good realistic game. They both came in the same time. But Doom 3 had much higher system requrements.

Imho it's a good discussion that bring many questions about what might have been possible even with old hardware instead of needing billions transistors chips and 32GB of ram to just increase resolution textures or whatever.. Half Life 2 was the "perfect" example of how a less features packed game engine might deliver a better graphic overall. I know there're many scene that looks good cause illumination is static and designed as it is by factory but still the graphic result was much better than A LOT of pixel shaders heavily supported games which gfx was quite awful.. I can't even count how many games did use pixel shaders for everything making the game less realistic than a Directx6 game.
I'm not a game developer so just opinion but I suppose it was a feature that even in the early version of Pixel Shading need to be used in a lighter and smarter way cause often not necessary. If it was used to "simulate" what really was needed in the decades and I'm talking about the now possible Global Illumination, it should have been a good thing, but to remember of those game walls/floors reflections with such absurd unrealistic shading was awful in most of those old games. Even the good ones like Far Cry (the soldier skin or leaves reflections, not the mention the saturated color palette).

Kind of of the same but also different is my experience with Farming Simulator (I'm so cool I know)
2013 still looked "computer game-ish" but had the balance right between realism and fun IMHO

Next game I got was 2019 and while the models and graphics overall are 100 times better, the shade model is too realistic with the sun hiding behind a mountain majority of the day making everything dark depending on the camera angle but looking on the web seems that games been struggling with shadows for a while now.

Out of old games I never really cared for the early Win9x games that simply increased the pixel count when increasing resolution. C&C vs C&C95 was my first experience of this.
The lower res kind of blended the pixels together but while the higher resolutions had a sharper image you could also see each pixel within the image.

gerry wrote on 2021-07-02, 09:50:

I always feel these games somehow promise more than they deliver

when i first saw them, with their multiple movable objects, vehicles, desks full of paperwork, lamps, plants and more I was very impressed

and then i tried interacting in some way; surely if there are papers on the desk i can read them, surely a few rocks and a car wreck wouldn't really stop me from walking in this direction, surely this grenade would have at least broken that chair!

Interestingly, this has been brought up before in these forums:
Re: For Duke Nukem 3D/classic first person shooter fans

That topic also has a link to an article I found quite interesting and enjoyable to read (and very glad you brought this up so I can re-read it):
https://eev.ee/blog/2016/06/22/graphical-fide … ng-video-games/

MrFlibble wrote on 2021-07-02, 19:49:

That topic also has a link to an article I found quite interesting and enjoyable to read (and very glad you brought this up so I can re-read it):
https://eev.ee/blog/2016/06/22/graphical-fide … ng-video-games/

Interesting article even if I'd not say Doom to be "easy" when played on the period time computers/hardwares..
The logic of this realism obsession before everything else, is a problem when the whole game lacks "art" in storyline or game logics, design and often realistic games had the features to be realistic but didn't use them well to be enough realistic to be remembered as "next-next-gen" games. As said how many Pixel Shaders early games did use them in the worst way wasting such important feature in the 2001->2010? Often there's no point to play many games at all. Many looks like clones of other previous versions of the same game and in facts not that different from ancient ones like Doom or Quake. For ex. 'Thief The Dark Project' game was a well balanced one on graphic, physic, AI, storyline etc.. some prefers the first or the second but the point is that it still can make you jump on the chair for many reason with its old gfx engine. The rules of the game were simple, the input commands too.. no need of manuals or complicated logics/puzzle, UI for objects or whatever.. you were into the game to explore with some hints and objectives or in-game videos/objects. Also the the physic of that game gave the player the ability to take different paths maybe not even designed to be taken.
For example, in that videogame there was a level in a subterrain where the player, had to reach a sort of internal island to reach some object. I wasn't finding the path to reach it but I tried with an impossible luck to jump on a living sphere of fire and then immediately jump on the river of that island reaching it (loosing like 90% of health that's not the point).

So graphic was more important in the 90's, yes and I suppose it might still be but imho videogames are suffering the way cinema itself is suffering. Lack of ideas, art, vision, something that must be personal as writing a book and then building it in an armony that need developers, graphic artists, optimization, debugging, music composer.. as said in another thread I'm still more interested in the tech demos as I was in the 3DMark2000 times that at least gave some emotions in their cinematic design more than games that are boring to even start. I can remember its demo music still nowdays.

I would say that modern games have too much "noise" in them. Before 2000 there were so many hardware limitations and levels consisted of only what was necessary. They built the gameplay and atmosphere with the bare minimum. Those games are actually much more comfortable to play, because when you encounter a character or found a bit more detailed item, you knew it was there for a purpose, there are less distractions.

Now developers fill the levels with so much detail, but often it adds nothing deep to the game, I call it "noise". You enter a very carefully crafted and visually impressive area filled with many characters, on surface it seems that a lot is going on, but often you end up interacting with only one character/item or you might even just go through the area with nothing even happening and it was all there just for a show. It is such a waste.

Literally everything on the PS5

This reminds me of the concept "uncanny valley" which used to be bandied about quite a lot 10-20 years ago regarding robotics, cinema movie special effects, and computer games.

A term often used to describe the mixed experience of hyper realistic technology : it fools the surface of the brain, but underneath, all the "wrong" micro-details of the simulation are constantly cueing the subconscious brain that things aren't quite right. A conflict which can sometimes lead to an uncomfortable or jarring experience.

antrad wrote on 2021-07-03, 18:53:

I would say that modern games have too much "noise" in them. Before 2000 there were so many hardware limitations and levels consisted of only what was necessary. They built the gameplay and atmosphere with the bare minimum. Those games are actually much more comfortable to play, because when you encounter a character or found a bit more detailed item, you knew it was there for a purpose, there are less distractions.

Now developers fill the levels with so much detail, but often it adds nothing deep to the game, I call it "noise". You enter a very carefully crafted and visually impressive area filled with many characters, on surface it seems that a lot is going on, but often you end up interacting with only one character/item or you might even just go through the area with nothing even happening and it was all there just for a show. It is such a waste.

Cyberpunk 2077 was the latest game to really do exactly what you are saying in my opinion. Why put all this detail into a massive city if its going to be somehow noisy and empty at the same time? Id take a map 1/4 of the size and filled with characters and stories with world ambiance. Instead of a million people with sidequests blowing up my cell phone (that cant be hung up on) and distracting me from enjoying the quests or sidemissions Im CURRENTLY DOING.

Shreddoc wrote on 2021-07-03, 22:25:

This reminds me of the concept "uncanny valley" which used to be bandied about quite a lot 10-20 years ago regarding robotics, cinema movie special effects, and computer games.

A term often used to describe the mixed experience of hyper realistic technology : it fools the surface of the brain, but underneath, all the "wrong" micro-details of the simulation are constantly cueing the subconscious brain that things aren't quite right. A conflict which can sometimes lead to an uncomfortable or jarring experience.

It all comes down to design. The uncanny valley largely isn't real, it's just poor concept work. The technology isn't there to fool you yet, so you have to make compromises, and sometimes developers choose the wrong compromises and end up with something terrible. Realistic graphics don't mean anything if they aren't in service to the game somehow. More pixels isn't automatically more better.

I recently saw a showcase for a lego game that nailed its aesthetic perfectly. It used new technology in an appropriate way to enhance the game and it looks great. Truly a next generation experience.

https://youtu.be/dtbqJXb1UDw

Meanwhile cyberpunk 2077, while impressive for today's standards, will inevitably look like garbage in ten years. CD Projekt Red's previous game Witcher 3 on the hand will probably age quite well despite being technologically inferior. It is what you do with the tech that matters. It's art and you have to execute.

It's never about the graphics for me, all about the gameplay. You can only be temporarily wowed by the graphics. I don't keep going back to games like The Game of Moria, The Settlers, Frontier:Elite II etc for the eye candy. The 80s though, they toughened a man up, the concept art and cover art for games was glorious, and you got this pixel sprite beep boop platformer in the box. To even to begin to think the game could be as good as the advertising was a grave error.

mothergoose729 wrote on 2021-07-04, 00:20:

It all comes down to design. The uncanny valley largely isn't real, it's just poor concept work. The technology isn't there to f […]
Show full quote

Shreddoc wrote on 2021-07-03, 22:25:

This reminds me of the concept "uncanny valley" which used to be bandied about quite a lot 10-20 years ago regarding robotics, cinema movie special effects, and computer games.

A term often used to describe the mixed experience of hyper realistic technology : it fools the surface of the brain, but underneath, all the "wrong" micro-details of the simulation are constantly cueing the subconscious brain that things aren't quite right. A conflict which can sometimes lead to an uncomfortable or jarring experience.

It all comes down to design. The uncanny valley largely isn't real, it's just poor concept work. The technology isn't there to fool you yet, so you have to make compromises, and sometimes developers choose the wrong compromises and end up with something terrible. Realistic graphics don't mean anything if they aren't in service to the game somehow. More pixels isn't automatically more better.

I recently saw a showcase for a lego game that nailed its aesthetic perfectly. It used new technology in an appropriate way to enhance the game and it looks great. Truly a next generation experience.

https://youtu.be/dtbqJXb1UDw

Meanwhile cyberpunk 2077, while impressive for today's standards, will inevitably look like garbage in ten years. CD Projekt Red's previous game Witcher 3 on the hand will probably age quite well despite being technologically inferior. It is what you do with the tech that matters. It's art and you have to execute.

To see a perfect example of what we're talking about, I need only look back to one of the first DOS games I ever played at release: Prince of Persia. Mechner's smooth body animation work, vs the fact his character didn't even have the basic facial definition to have a mouth. And we did not care a single whit.

While uncanny valley may be a deprecated term, some of the logic behind it - that the brain is not fooled into thinking [our current era's] good graphics are analogous to RL, and that the harder you try, the more holes you'll find need plugging - is undeniable.

All that said, it's a long term evolving situation. I won't necessarily be saying the above when 8K is the standard resolution and panels are replicating CRT phosphor glow at accuracy and detail undetectable to the naked human eye....

But none of it changes the fact that some of the most fun things in life are also the simplest, and gaming is ultimately not about technicalities (that's what forums like this are for!), but about having fun.

MrFlibble wrote on 2021-07-02, 19:49:

Interestingly, this has been brought up before in these forums: Re: For Duke Nukem 3D/classic first person shooter fans […]
Show full quote

Interestingly, this has been brought up before in these forums:
Re: For Duke Nukem 3D/classic first person shooter fans

That topic also has a link to an article I found quite interesting and enjoyable to read (and very glad you brought this up so I can re-read it):
https://eev.ee/blog/2016/06/22/graphical-fide … ng-video-games/

that's a great article and well worth reading, thanks for linking it!

The Deus Ex v Deus Ex: Human revolution example in that thread is a perfect example. In the original the game world has very little that is 'spare' or wasted. Almost anything that looks movable can be moved and almost anything readable can be read and so on. The world is sparse but most things in it are 'real' for the player. In the later game there is so much detail and most of it is unavailable for the player which, strangely, makes it feel less real than it should based on appearance

Cyberdyne wrote on 2021-07-02, 11:35:

I have noticed that some 3D FPS games age well and some do not. Crysis is still like new. Return to Castle Wolfenstein is so old and dated with its Quake 3 Arena engine, but it is still fun to play. But older medal of Honors look goofy today. And that Wolfenstein that came after looks worse somehow. Doom 3 is creepy, but everything looks like it is molded with clay. But Half-Life 2 is still a good realistic game. They both came in the same time. But Doom 3 had much higher system requrements.

I still quite like that 2000/2001 era 3D as in MOH and RTCW but i know what you mean, it's aged whereas Doom3 at first looks more 'modern' - but you are so right it looks like clay mouldings, a kind of durable singular clay mass that's been carefully shaped and painted but which in fact is utterly damage proof and static

it's all about the art choices as stated elsewhere in the thread, that has a big impact in our expectations what a game should offer

This remembers me reading somewhere some article about humanoid robots when designed by research labs/institutes, it's much more convenient to design the robot not-realistic at all (for example the face) cause the more realistic you try to make it looks like human, the more the brain "understand" it's not and often those robots might looks sort of "creepy" even when almost perfectly built. As already has been said in the thread, those micro-details that differs from a real human being are easily recognized by the brain probably even before we begin to question about the possibility of it as real human or robot. It might be a not-perfect facial movement, some strange eyes movements or whatever... human being's eye-to-eye contact is part of the whole human life to recognize people we know or we don't, recognize our own emotions by facial movements or voice tone, volume... no 3D or real-robot can (still) deceive those natural capability of the brain.

The Bouncer

386SX wrote on 2021-07-05, 10:21:

This remembers me reading somewhere some article about humanoid robots when designed by research labs/institutes, it's much more convenient to design the robot not-realistic at all (for example the face) cause the more realistic you try to make it looks like human, the more the brain "understand" it's not and often those robots might looks sort of "creepy" even when almost perfectly built. As already has been said in the thread, those micro-details that differs from a real human being are easily recognized by the brain probably even before we begin to question about the possibility of it as real human or robot. It might be a not-perfect facial movement, some strange eyes movements or whatever... human being's eye-to-eye contact is part of the whole human life to recognize people we know or we don't, recognize our own emotions by facial movements or voice tone, volume... no 3D or real-robot can (still) deceive those natural capability of the brain.

Absolutely.

Another way of looking at the same thing: people with (for example) Asperger's Syndrome could no doubt confirm, that the only things required for a human to judge another human as unnatural, are very minor physical cues like the ones you described - "strange eye movement" - and other such micro-details: "facial movements not 100% right", "milliseconds of delay in the wrong place", etc etc.

Illustrating how the concept of convincing realism is difficult even for real humans to achieve - or even properly define! - let alone the monumental task of synthesizing it!

I suspect that the complexities of movement and behaviour dynamics are going to be vastly more demanding than the visual recreation itself. Indeed, having visually-realistic digital avatars which are 'inhabited' or controlled by real humans - taking the monumental difficulties (not to mention risks!) of synthesizing behaviour out of the equation - seems like an obvious stepping-stone before we can even dream of actual behavioural simulation.

Shreddoc wrote on 2021-07-22, 20:48:

Absolutely. […]
Show full quote

386SX wrote on 2021-07-05, 10:21:

This remembers me reading somewhere some article about humanoid robots when designed by research labs/institutes, it's much more convenient to design the robot not-realistic at all (for example the face) cause the more realistic you try to make it looks like human, the more the brain "understand" it's not and often those robots might looks sort of "creepy" even when almost perfectly built. As already has been said in the thread, those micro-details that differs from a real human being are easily recognized by the brain probably even before we begin to question about the possibility of it as real human or robot. It might be a not-perfect facial movement, some strange eyes movements or whatever... human being's eye-to-eye contact is part of the whole human life to recognize people we know or we don't, recognize our own emotions by facial movements or voice tone, volume... no 3D or real-robot can (still) deceive those natural capability of the brain.

Absolutely.

Another way of looking at the same thing: people with (for example) Asperger's Syndrome could no doubt confirm, that the only things required for a human to judge another human as unnatural, are very minor physical cues like the ones you described - "strange eye movement" - and other such micro-details: "facial movements not 100% right", "milliseconds of delay in the wrong place", etc etc.

Illustrating how the concept of convincing realism is difficult even for real humans to achieve - or even properly define! - let alone the monumental task of synthesizing it!

I suspect that the complexities of movement and behaviour dynamics are going to be vastly more demanding than the visual recreation itself. Indeed, having visually-realistic digital avatars which are 'inhabited' or controlled by real humans - taking the monumental difficulties (not to mention risks!) of synthesizing behaviour out of the equation - seems like an obvious stepping-stone before we can even dream of actual behavioural simulation.

Yes and I think it's still good the human being as something that can't be replicated nowdays because it make it feel like he has still something that can do that can't be done by machines, robots or whatever.
About the complexity of digital or robotic realism I agree that is much much more complex that just a visual rendering in a 3D enviroment or in a physical robotic humanoid. Sometimes we don't think about how complex for example the amount of things that are needed in a common discussion between two people is. For example when, how, how long we can look at the other person's eye, when, how, how long we can smile or can't smile, the whole logic of the empathy in a common real life discussion and how easily a wrong word or a wrong facial expression, even its timing or duration, can results in complete different results. And all these variables makes the brain to understand something might not be correct and in a futuristic example the person might be a robot. And instead in the rea life we everyday automatically have discussions that are real without even thinking to how complex they are to be real.

Shader this, DX that, and all they give you are different shades of grey.

Fifty Shades of Duty: Modern Graphics.

Sorry, but anything less that vibrant counts as 'under delivered' in my book.