I take it this is bad for lighting and performance?

[AluminumHaste 1084]

I remember reading that going over three lights for one brush is not good. Well I have this and I split the brushes but of course the compiler put them back together lol.

 

 

See what I mean? the floor is one big piece even though in the editor I split them all up. I think this is the t-junction fix from the compiler?

[greebo 86]

Jeez, can you please stop to open a new thread for each and every question you have? You have an entire forum page going.

 

In case you didn't notice, there is a thread for general stuff like this here: Itches & Glitches - TDM

[AluminumHaste 1084]

Jeez, can you please stop to open a new thread for each and every question you have? You have an entire forum page going.

 

In case you didn't notice, there is a thread for general stuff like this here: Itches & Glitches - TDM

 

can I delete these?

[Fidcal 175]

You might not be able to delete. Try a full edit. if you can't see a delete then no don't worry but just put future ones on the main thread. I personnally don't see anything wrong in a separate thread for each topic - helps to find stuff - but best stick to the guidelines.

 

The only way I know to split them up is with trim and such like but hard to see how to do it here. It's maybe better to reduce your light radii. I had much improvement in my mission when I did that. I even rotated some lights so they missed the boundary of the next one.

[AluminumHaste 1084]

You might not be able to delete. Try a full edit. if you can't see a delete then no don't worry but just put future ones on the main thread. I personnally don't see anything wrong in a separate thread for each topic - helps to find stuff - but best stick to the guidelines.

 

The only way I know to split them up is with trim and such like but hard to see how to do it here. It's maybe better to reduce your light radii. I had much improvement in my mission when I did that. I even rotated some lights so they missed the boundary of the next one.

 

 

Yeah funny thing is I did use trim and cut them all up but the compiler stitched them back together.

[Tels 279]

Yeah funny thing is I did use trim and cut them all up but the compiler stitched them back together.

 

No, a "trim" here means a edge or ridge, e.g. anything that makes the texture or surface non-continues.

 

You can either:

 

* add a small dent/rim/trim/raised edge/ridge in the floor to break them up (but this adds more triangles, which can be bad)

* split them up and shift on texture by one pixel (hor/ver) via the surface inspector. D3 will not join faces if the textures don't match 100%

 

Hope this helps.

[Komag 20]

I also don't mind separate topics for each thing, easier than having to read through a whole big "collection" thread, but I guess there have been a lot of them lately.

 

For the brushes, you might try making them different heights underneath, dunno if that will work but maybe

[Fidcal 175]

Hey, that's not a bad idea if it works Komag. But also as Tels says, if the texture is even slightly different it must be a different brush surface. Try some decoration like a stripe of a different texture.

[Baddcog 114]

I hadn't thought of making the bottoms different heights but I don't think it'll help.

 

When compilied I believe it's coplanar surfaces that are optimized, not the entire brush.

 

What I've done to fix that situation and it worked great was to spilt the terrain like you did, but then I nudged the verts along the split up or down a few small grid lines.

 

This way they can't be optimized as they are not coplanar and if done in small units the player can't see the difference.

 

It did help alot in an area of mine that was very large so many lights far away from each other hit the same large polys. afterwards they only hit smaller polys closer to themselves.

 

The engine can handle tons of polys, that's not the problem, the problem is too many lights hitting the same polys. You have to find the balance where splitting terrain actually limits how many lights hit polys.

 

If you just spilt randomly you could make it worse because you could make 6 lights hit 100 polys instead of 6 lights hitting 4 polys. That would be bad.

[sparhawk 17]

I don't see why having a seperate thread for different topics is a bad thing. After all, that's what topics are supposed for.

 

Also I doubt that a thread would be more readable if you bunch up hundreds of different things into it. Especially if somebody has a similar issue later and tries to find a solution. Then he is told use the search function and has to wade through hundreds of unrelated issues? Doesn't sound like a good plan to me.

 

It makes sense to bunch all related stuff into a single thread. For example, put all questions regarding rope arrows into one thread. Apart from that, I don't see what it should help to have unrelated issues put together.

[SneaksieDave 40]

I think the intent is being misunderstood. The "itches and glitches" topics are for things minor or unsure (as the first post states). Things like "hey, if I put this here does it cause a problem? It's not solid anymore..."

 

No one's saying don't make new threads for individual topics/issues. Cramming everything into the "itches" threads would also be inappropriate and not very user-friendly. The point (as with the thread of the same name in the member forums) is just so that there aren't 700 threads for every little question or clarification. As greebo mentioned, there was an entire forum page of new threads, most asking simple questions that could've fallen into such an "itches" thread.

 

Anyway, it's no fast rule, do whatever you like. Maybe ask yourself "does this deserve a full thread?" If the answer is yes, do so without fear. If not, just stick it in the appropriate itches thread. No problems. Can't be more clear than that, because again, it's no fast rule, just a suggestion. The threads are only there for convenience of all. So you can also ignore it. It's based on how it is done behind the scenes, and it's worked well.

[Komag 20]

Yeah that makes sense, have to just judge each case

[jdude 18]

Personally I try to have at most 2 visible dynamic lights at any one given time, anymore creates a significant fps drop which isn't worth it. You won't be able to have all those lights giving off shadows if you want decent fps. I recommend creating a shader to simulate the shadow instead. 99/100 people won't be able to tell the difference The only draw back is that you will not see the AI shadows, but ask yourself, '' are the shadows from AI worth -40 fps?'' Most answer no.

[OrbWeaver 664]

I remember reading that going over three lights for one brush is not good.

 

It's a myth, don't waste your time with the "OMG NO MORE THAN 3 LIGHTS" nonsense.

 

Minimising shadow counts is extremely important, and reducing light counts in general helps with this and is not a bad thing in general, but the 3 lights thing makes no difference in any normal mapping case. Unfortunately because it is written into several of the Doom 3 tutorials there are lots of people who think that Bad Things will happen if you have more than 3 lights, which is untrue.

[AluminumHaste 1084]

It's a myth, don't waste your time with the "OMG NO MORE THAN 3 LIGHTS" nonsense.

 

Minimising shadow counts is extremely important, and reducing light counts in general helps with this and is not a bad thing in general, but the 3 lights thing makes no difference in any normal mapping case. Unfortunately because it is written into several of the Doom 3 tutorials there are lots of people who think that Bad Things will happen if you have more than 3 lights, which is untrue.

 

Not just 3 lights, 3 lights for one brush.

[OrbWeaver 664]

Not just 3 lights, 3 lights for one brush.

 

It's actually one triangle (brushes don't exist at runtime), and yes, this is the myth I am referring to.

 

Unless you are running on some particularly ancient graphics card, having 4 lights on one brush won't make any detectible difference to the framerate, and the effort you invest in carving up brushes would be much better spent minimising shadow counts and correctly portalising your level.