Installer Steam
log på
|
sprog
简体中文 (forenklet kinesisk)
繁體中文 (traditionelt kinesisk)
日本語 (japansk)
한국어 (koreansk)
ไทย (thai)
Български (bulgarsk)
Čeština (tjekkisk)
Deutsch (tysk)
English (engelsk)
Español – España (spansk – Spanien)
Español – Latinoamérica (spansk – Latinamerika)
Ελληνικά (græsk)
Français (fransk)
Italiano (italiensk)
Bahasa indonesia (indonesisk)
Magyar (ungarsk)
Nederlands (hollandsk)
Norsk
Polski (polsk)
Português (portugisisk – Portugal)
Português – Brasil (portugisisk – Brasilien)
Română (rumænsk)
Русский (russisk)
Suomi (finsk)
Svenska (svensk)
Türkçe (tyrkisk)
Tiếng Việt (Vietnamesisk)
Українська (ukrainsk)
Rapporter et oversættelsesproblem
The problem with UE5 is that the games ARE basically running on "tech demos". This is stuff that used to be restricted and reserved for demo purposes, until devs could optimize and streamline it for a more general and mainstream experience. Now, they figure the "raw hardware power" of Nvidia cards and even some AMD cards can be leveraged to deliver manageable performance for that same fidelity that the demos gave.
Problem of course is that the demos were running on closed hardware with certain factors being constant. No game is ever the same nor the systems running them. Even consoles can hardly be said to be there yet. And yet Sony and Microsoft are forced to improve because publishers want to push "bigger" and "badder" graphics for a new generation.
Mind you, raytracing by itself isn't even necessarily that performance intensive these days. But it seems to work better on in-house engines like 4A Games' Metro 4A engine, Frostbite, RE Engine, and Cryengine etc. UE5 is a free engine suddenly given to a bunch of newbies and expected to deliver a AAA experience most of the time with a AA or indie dev team.
In architecture, product design and much more.
Same reason why Nvidia sucks these days, they are now an AI company.
Which explains the "tech demo" feel. But those other things are rarely processing more than 24 or 30 frames. The engines have always had to streamline it for games though to give more acceptable performance. When UE4 tech demos were first showing off the look, people were saying how much would have to be toned to actually give a playable experience. I guess publishers and devs don't care about that anymore.
People used to say that while Crysis looked great, it was NOT an example of a game you wanted to follow or copy in terms of how to make all games going forward. That's why 2 and 3 toned down the physics and size to make them work on consoles but also because gameplay mattered more than simply looking pretty. Now it seems like we're back to glorified tech demos, with the performance that usually comes with one.
Unreal 3 was a tech Demo as well in the past.
But I doubt I would use Unreal 5 in the first place.
2) Optimization is not magic. It takes time and resources from a studio. Time and resources that might be better spent adding in more content. Therefore, once a game is believed to be within most machines' capabilities, the project lead will probably not want to optimize it further.
3) Unreal does have issues of its own, especially if you work with blueprints as opposed to C++. It's an incredibly versatile tool that comes with a lof of extraneous functions, which can significantly weigh down games.
That said, first-time devs have very few alternatives, and "making your own engine" is certainly not one of them.
Other in-house engines just were better at handling certain things. Unreal also has that "look" that just can't be gotten get rid of.
I wish a lot of the AA and AAA studios would go back to creating their own engines. Not just because the performance would be better for those games, but because the hiring talent needed for programmers who could actually design their own engines would elevate everything else in the game.
As for Unreal, Sweeney can kiss my ass.
edit:
https://arstechnica.com/gaming/2025/01/is-2025-finally-the-real-year-of-half-life-3-confirmed/