Imagination & logic are bad friends in the graphic process?

[assurdum]

Well, I'm little embarrassed to open a topic, because my knowledge are more 'amateur' than professional, & my english isn't it so good to substain a deep conversation, by the way, let's try...
too frequently the developers leave me the impression to ignore creativity, imagination for the confort of a classic method when, probably, there is a solution more simple & less expensive then we though. Talking of AA, we have to wait the limits of the ps3 band to find a valid alternative method, with low cost & even more efficient result compared the classic MSAA possible on the console; before that, only a few from what I know have tried something similar in others camps. Back to the topic, I don't understand why a lot of different graphics process continues to be used when other solutions could be more effective, without excessive cost. Talking of buffer why on the ps3 not alternate bitmap with low buffer when the low buffer lose to much pixels, cause the excessive long distance became unpratical for that? It's my question, probably I completely wrong or missed the relative implications. Or why a surface reflection isn't it converted in a simple texture, when we talk of long surface with high buffer... simply just grabbed the real reflection in a picture applied on the surface how a normal texture...or why, for example, not to split the draw of the high buffer to halves, using 2 low buffer to draw a single high buffer? Isn't it possible? I don't know,maybe I use too much imagination when it require more tech knowledge, I know the graphic concepts of the most common features using in the current console generations but I have a terrible memory when we talking of formule, rules & raw numbers...

 

[Cyan]

Well, I think developers are always trying to find new solutions and squeeze every bit of power out of current consoles and they have to face similar dilemmas everyday.

In my opinion, developers should make their work and achievements transparent to the general audience, or in trying to give some info, they should use layman's terms to help the average gamer and ordinary people understand the tech.

There are also rendering techniques developed on a particular console that are exclusive to that console, and then staunch sympathizers start this flaming internet-war between god knows what groups of gamers and "isms" are those...

When developers find an insurmountable challenge, that's when they panic and they come up against new dilemmas, similar to those you are presenting here.

Questions like.., what are we facing in actuality? Who's fault is this, ours or or the lack of power of this machine or a single hardware limitation? And how do we potentially solve it?

I would like to help you with those questions, but I think a developer or any other person here who is well into 3D graphics and programming -this is Beyond3D after all- can type up a post and help you more than me. Can't say much else than that.

 

[AlNom]

In my opinion, developers should make their work and achievements transparent to the general audience, or in trying to give some info, they should use layman's terms to help the average gamer and ordinary people understand the tech.

That's what journalists are for , but quite often (in my experience), PR & IP issues can get in the way, especially if you're expecting something more than a really high-level description of things, at which point the tech discussion can become too general and vague to be of any use.

/me curses PR for foiling several Q&A attempts...

 

[ERP]

On the AA thing, I saw demos of screen space solutions to AA before XBox360 and PS3 were even released. This included using the Color and ZBuffers as hints for edge detection, the results just weren't deemed acceptable at the time.
It's not like developers don't look at these things, they do.
The problem with game development is making something acceptable in the general case is often a difficult task, it takes iteration, it takes time.
There is no Eureka like discovery in the new screen space AA methods, just improvements to the point of acceptability.

A lot/most discovery is this way, understanding what you can do is one thing, making it work in a real setting is entirely different, it's interesting to listen to Carmack talk about it taking 5years to get from Megatexture demo to make it robust enough to ship in a game. This isn't an unusual pattern and most devs can't spend 5 years on a single technology.

If you look at invention in general there is rarely a Eureka type moment, it's usually combinations of existing ideas that provide the breakthroughs.

 

[Laa-Yosh]

I agree with ERP (I guess) - in my opinion all these post "AA" solutions are weak replacements or glorified blur filters and not what I would call antialiasing. I'm in fact a little angry that these inferior solutions will probably become the norm...
(although it's encouraging to hear Carmack talk about the realization that raytraced reflections on a bumpy surface won't look good with a single ray - I could add that no post filter can help there either)


Also, a lot of the "creativity" is far less obvious than one would think. On first glance DICE isn't doing anything special with Battlefield 3, they have an HDR linear lighting pipeline, nice shaders and textures and models, deferred rendering, so basically just a laundry list of features that one could implement today.
And yet, the results are utterly amazing and definitely took a lot of extremely complicated architectural and programming work, lots of creative and unorthodox solutions and so on. But it's not a single case that one could promote on youtube to call for attention...