request

[Pr3t3nd3r]

I don't want to be lame but isn't it possible that for the next version to be made some rendering improvements?
I'm talking about the graphical quality of the rendering, more exactly the mipmaps ... more control over them or somethink.

If you enter the demo is very easy to see some strange predatorian effects ... (i think generated by mipmaps ... or somthing else ... ).
You can easly see how the textures get blured with some levels of bluring at short distances.

Meybe an coeficient for the distances for mipmap. If the map from demo will be scalled 3-10 times biger or somethink, probabily will get more blured

[Guest]

then ill may ask for something too:

1. please, give us new texturemapping functions so that we can use bumpmapping on levels with a lot of textures, too!

2. could you add the possibility of scaling several textures of the same node not equal to each other, eg:

node -> scaleTexture(1, 1.0);
node -> scaleTexture(2. 0.1);

where the syntax is: scaleTexture(TextureNumber, TextureSize);

if this is already possible, im sorry for this, then you can delete my post.


ps: maybe this is a real stupid question, but is it possible to create high quality graphics (like for example age of empires 3) with irrlicht? or even like unreal2 ? i have not seen a screenshot of irrlicht which actually looks better than counterstrike 1.6 or anything like that :/ or can irrlicht actually really be "beautiful" ?

dont take this wrong, i dont want to flame irrlicht!!!

[Guest]

I believe it can look very good. To get an example, enabled Antialiasing and use anisotropic filtering (in you gfx card settings) - use a well modelled lightmapped "world" or model/s with sympathetic use of textures and lights - sprinkle on subtle particle use and you have something that looks decent.

Why people can't understand the beauty in games comes from the effort put in by the artists/coders and NOT from the base engine code - no matter how many features it contains, you can not just flick switches to "Make my tech demo look gorgeous".

The mip map thing and the zbuffer problem do look VERY bad but you can work around them, or they should be fixable but nothing show stopping.

Most people who post stuff here done in irrlicht are keen to just show anything (early shots etc) so they usually look rough. When you start to see finished games (smaller projects not pie-in-the-sky FPS's and MMORPGS, which will never allow small teams enough time to get the art up to scratch) you will see they can look identical to any other last gen/previous gen games. It takes time... it is not easy... Irrlicht gives a massive boost to getting you closer to your goals and should not be dismissed on the strength of a few mistakes in a NON FINAL version.

bottom line.. you want a good looking demo for irrlicht then MAKE one, but be prepared to spend the time that is required the same as every other semi-pro and pro developer out there.

[Guest]

well i thought about creating a demo or something like that that looks simply fantastic but there are several problems:

a. i would only have 2 days left, because then im going on holidays for 6 weeks!
b. when i come back i want to make some big progress with my current rts project, because i've been very lazy actually for a few weeks.
c. why should i do a demo for myself? if irrlicht would have a cool demo that does not look like stuff from 6 years ago i guess that would attract more users to use irrlicht because they see what it can _really_ do, not just unfinished and technical screenshots! but since i am not the creator of the engine, i dont see why i should do it all alone. a community-based project would be the way to go i guess, where everyone can submit a little so that a core-group can construct a cool demo (it does not have to be a whole game nor doom 3 quality, but something that shows what irrlicht is capable off (thats the meaning of demo))

d. i dont even know if i am capable of doing such a thing at all, so i couldnt do it without help i guess.

now to you actual post:

I believe it can look very good. To get an example, enabled Antialiasing and use anisotropic filtering (in you gfx card settings) - use a well modelled lightmapped "world" or model/s with sympathetic use of textures and lights - sprinkle on subtle particle use and you have something that looks decent.
)

antialasing and anisotropic filtering is a nice feature, but it does not make worse graphics -> good graphics. the most thing that bothers me is that many screenshots have textures that obvouisly dont fit together or have very strange color settings at all. it seems that the creators throw everything they have together without thinking. and the lightning is often too bright or too dark at certain points which simply does look bad.

but this is no irrlicht related problem but hobby-game-making in general. i've seen this many times.

i know that a few people now try to flame me about this post and expect me to do everything right, but i dont care. i've never said that i could do everything better, nor that i am perfect. i know everyone makes mistakes and so do i (actually i do alot )

i have to admit that there are also many cool and nice projects around here too. (my absolute favorite is the rts project which pretender is doing )

have a nice day!

ps: why do i care about other screenshots anyway :/ i got too much time

[Pr3t3nd3r]

I wasn't saying any think bad about irrlicht ...

I just released the new version of the game ... anyway i'm too tired right now to make an update on news about what is new ... many thinks ...

[Guest]

there is just the problem that i cannot run it since my notebook has linux since a while already and my other computer with windows has just no monitor because its .. well its destroyed damn

[the one that posted above]

antialasing and anisotropic filtering is a nice feature, but it does not make worse graphics -> good graphics. the most thing that bothers me is that many screenshots have textures that obvouisly dont fit together or have very strange color settings at all. it seems that the creators throw everything they have together without thinking. and the lightning is often too bright or too dark at certain points which simply does look bad.

but this is no irrlicht related problem but hobby-game-making in general. i've seen this many times.

I agree with you, that is what I was implying anyway.. that a lot of "bad looking" stuff comes from the core art / textures / lighting and design which is a seperate issue to Irrlicht itself, but by using AA + AF in combination with great art assets you can get a "hint" of how good irrlicht can look (which is to all intents and purposes as good as any other engine using the same level of tech).. after all the engines all use the same APIS underneath.. a DX call in Irrlicht offers the same results as a DX call in ogre. It is only the unsported, or buggy calls in irrlicht that "may" make it look worse in comparative tests. Underneath all the abstraction it should boil down to 1:1 graphical quality assuming the developers art / design is identical.

The mip mapping, filtering, aliasing etc are what stick out as not looking so "hot" when initially testing irrlicht but this is down to the the non-finished state of the engine and most importantly the tech demos etc which do not push it graphically.

I would love to see some user created tech demos that really show off the engine, I don't think niko should bother though as his time is far more importantly spent on the code, feature improvement, bug fixes and optimisations.

Cheers

[me again]

I'd like to add, I agree with the original posters point - that we could really do with more control and better results concerning:

.Zbuffers (32bit without stencil, flexible, robust fallbacks)
.MipMaps (Work around using AF in card settings but not ideal)
.Anisotropic (very important in modern games while not being too "expensive")

All while maintaining the great dev checks/fallbacks/abstracted nature of Irrlicht.

These would be the 3 most important issues for the core rendering side of things for the next release in my opinion (we now have AntiAliasing which is great)

Good luck with 0.12.0.. I look forward to it.

btw.. niko - any thoughts on optimisation for future versions - would love to see a release with performance boosts!

[niko]

.Zbuffers (32bit without stencil, flexible, robust fallbacks)
.MipMaps (Work around using AF in card settings but not ideal)
.Anisotropic (very important in modern games while not being too "expensive")

Agreed. Already on my list.

[Pr3t3nd3r]

Gama control ...
device->setGama( ....);
will be nice ...

[grayman]

and remeber for the next release

fix memory leaks

fix my3d, (zdimitor is back and the patches will be ready soon)

add CCloudSceneNode

[Pr3t3nd3r]

I just tried irrlichtNX and is incredible how fast it is and the good quality of the rendering ....
When they start the new rendering system i was saying: "lock the nice guys are making a new rendering system for irrlicht ".
Is this true ? Can we use some code from nx?

.....
I search to see why nx is faster ... it doesn't seams to be because of the new rendering ...
what I'm interested is if they are using an different (faster) DrawIndexedPrimitiveUP... ...but it seams that i don't get it

[Pr3t3nd3r]

Lool
Cool .... nx seams to be faster because the camera is moving much faster ....
cool ilusion ...

[Electron]

The main reason NX/LF is faster, at least under OGL, is it uses vbo's instead of merely vertex arrays. It also only updates the transformation of objects when its actually needed, so doesnt slow down on scenes witha huge number of small objects (perhaps this is in irrlicht too by this time, I cant remember)