I think the best suggestion so far was to walk through the level yourself. Put yourself into the shoes of your PC actually walking around in the level. What would you expect to see in such and such corner, up above you, or down that dark alley? Even better sometimes, what would you NOT expect to see? It is those WTF things a player sees that is really going to draw their attention ... which tends to be where the baddies come in [wink]

Quote: Original post by mntlinstituteflr
Sorry, no. It's not textured and lit at the moment (otherwise than the light for renders), I will do these after all the geometry is modeled. Current poly count is 6,450. To answer your question though, the entire level will be played and available to the player.

What I meant was: how will this level be played? First-person shooter-styled? Third person? Top-down view? What will be going on in it? Lots of action (which takes focus away from tiny details, and also brings polygon budget down because enemies and effects need some budget, too)? Or calm, athmospheric (in which case details are more important)? How much polygons can you reasonable use at all? Have you tested this in-game already?

These are important factors. When I designed levels some time ago, I used to test frequently. If something got broken, I would notice it pretty fast. If something would decrease performance too much, I could quickly implement an alternative, without wasting too much time. If the level wasn't fun to play, I could easily change the core layout - because I would test this before adding details and fluff. Every time I detailed levels before playtesting them, I found it extremely hard to tweak the gameplay, because I had invested so much time in the details already, which made breaking up the level for some serious reworking much harder.

As for texturing and lighting, I believe these affect the outcome of a level significantly. You should not wait with them untill all geometry is done. Put a basic texture layer on your geometry, maybe as simple as just applying colors for now. Things can look pretty different, and details that are obvious now can blend together with the environment once textured. Their effect is too large to be ignored early on.

Quote: Original post by Captain P

Quote: Original post by mntlinstituteflr
Sorry, no. It's not textured and lit at the moment (otherwise than the light for renders), I will do these after all the geometry is modeled. Current poly count is 6,450. To answer your question though, the entire level will be played and available to the player.

What I meant was: how will this level be played? First-person shooter-styled? Third person? Top-down view? What will be going on in it? Lots of action (which takes focus away from tiny details, and also brings polygon budget down because enemies and effects need some budget, too)? Or calm, athmospheric (in which case details are more important)? How much polygons can you reasonable use at all? Have you tested this in-game already?

Alright, now we're on the right track here. It's an FPS based in the future (won't go too far in depth, it's irrelevent). Poly limit for the levels is around 50-100k. And now I've added a bit more and I'm at 18k so I'm doing fine. Yes, we have tested those poly counts in our level editor with a decent PC, nothing special. We also did an experiment where we used 3 of the levels (around 54-55k) and we still had very high framerates so poly count so far isn't an issue, at least for this level.

Quote: These are important factors. When I designed levels some time ago, I used to test frequently. If something got broken, I would notice it pretty fast. If something would decrease performance too much, I could quickly implement an alternative, without wasting too much time. If the level wasn't fun to play, I could easily change the core layout - because I would test this before adding details and fluff. Every time I detailed levels before playtesting them, I found it extremely hard to tweak the gameplay, because I had invested so much time in the details already, which made breaking up the level for some serious reworking much harder.

As for texturing and lighting, I believe these affect the outcome of a level significantly. You should not wait with them untill all geometry is done. Put a basic texture layer on your geometry, maybe as simple as just applying colors for now. Things can look pretty different, and details that are obvious now can blend together with the environment once textured. Their effect is too large to be ignored early on.

Thank you, good points. I have nothing to say about that, but I did soak what you said in. It makes sense to play test a lot and I plan to do so, both for poly counts in the level editor and for gameplay in the demos. As for lighting and texturing, I can understand it helping you get a grasp of things better, but I want to finish the majority of the geometry first so I don't waste time texturing something I will soon delete or change drasicly. But I will texture and light some of the more obvious to stay parts of the level.