total onscreen polys?
Blitz3D Forums/Blitz3D Programming/total onscreen polys?
| ||
| i know this question has probably been asked to death, but i can't find any results on the super search the forum has :-\ so for a mid range computer with say a midrange card (gf2?) how many polys would anyone recommend being onscreen at the most. i believe that includes particles and stuff, and taking into account the textures, cause they slow up things. |
| ||
| simple rule of the thumb, take your target FPS and your target system, fil up the screen with polies and make sure you have about the ammount of surfaces you wanna have ingame active, then raise the onscreen polycount till the fps drops under your target rate, divide that ammount of poly's by 3 and if your a somewhat decent coder your game should run fine with that polycount onscreen. |
| ||
| probably a stupid question but why by 3? oh and thanks for the tip:) |
| ||
| Polies are also often not the bottleneck. Almost all of the rendering power in Juno goes into the same 120 polygons. |
| ||
| Yeah, there's a lot more to graphical complexity than just number of polygons. Textures, effects (eg. shadow casting,) AI and physics calculation, etc. |
| ||
| oh right, i see. i thought they took the longest to draw. thanks for that :) |
| ||
| My advice: *check your video card memory and try to keep it under 8000 kilobytes (availvidmem() - totalvidmem() will help you) *try keeping the NUMBER OF SURFACES as low as possible (5000 surfaces with 2 poly each will definetely kill your performance, but two 5000 poly surfaces won't) *I would estimate something between 5000-20000 polys onscreen (and total of 50-200+ surfaces in-game) so that Pentium 300Mhz with GF2 would run it smoothly (those are my estimations - can't be for certain - but hope this gives you some rough guideline) |
| ||
| so that Pentium 300Mhz with GF2 would run it smoothly Depending of course your target PC... this is good for "everybody can run, even my grannie", but if your aiming bit higher then you can upscale the poly-amounts etc too. Personally my game will "need" something around 1G CPU and GF3 level GPU (not for shaders, but for general speed & memory) to run it smoothly. But I'm certainly not like most ppl here; I couldn't care less about ppl with Celerons and Voodoos... I want to make a good looking game without any artificial restrictions other than what my skills, time and Blitz sets me. |
| ||
| @mustang I agree. i always thought it was a bit silly restrictinging your talents and ideas so they fit round old computers. i've settled on 15000 on screen polys total. and see what computer runs it. thanks for your replys. very helpful :D |
| ||
| @joker: I must remind you on this: if you are planning to sell the game, then you have to consider things like this. If you are doing it just for fun/free then I guess it doesn't matter so much (at least it runs on your computer ;) But in terms of business: It's important. |
| ||
| IMO it's bit funny to say that "if you want to sell games make sure it runs with low-end comps" when EVERYBODY in-the-real-world are doing just the opposite... every new game released needs more and more power. Of course that's bit too much, but still I think that you can sell just as many copies, if not more, when you aim little higher... I don't see why ppl who buy "shareware" couldn't buy games that need more ooomph from the comp than what Celeron can give? Are ppl who buy "shareware" games really using only antique hardware? I don't think so. BTW, I'm not trying to say that you are wrong or anything, I just wonder why everybody are so stuck to thinking that you can not do a game with Blitz3D that needs a high-end PC. |
| ||
| Perhaps it's because most blitz games already have a small market. I mean, no offense, but many blitz games aren't going to make it to well as far as # of sales. Lowering this even more by making a game that only runs on an extremly modern machine is fairly stupid in this situation. Of course, if someone did however create a good game and advertised it well, this problem wouldn't exist. Commercial game developers don't have to worry about this to much, as they often times sell to customers that game alot and thus most likely have a good computer. Satisfying and maintaining these customers would be more bennificial to the game company than satisfying the people who don't game much, and have a lower chance of buying more of the companies games. Note: I know nothing of the gaming business. I just think that for a Blitzer who wants to actually sell a game, supporting low-end machines is a good strategy. |
| ||
| well, for this game, we're going to be using as many tricks as possible to keep the poly count low, but the trouble with low end computers, like my parents, some of them don't have gfx cards and rely on on-board graphics. i dunno about other ppls, but that don't really cut it in most games and any demos or particle system i've tried run pretty slowly. that's without any models with texturing and alpha and such. so keeping the polycount low isn't the only thing you've got to do, and i don't wanna take any main features outta this. oh and it should be a sellable game, its only a two man show, so it'll take a while. :D thanks for the advice |
| ||
| @Mustang: krhm... If you follow the Dexterity forums at www.dexterity.com/forums you can get several reason to cut down the download size: *First simple reason that bigger download size means bigger bandwith ratings -> bigger service provider costs (it's a huge impact to have 10 meg demo instead of 3 meg demo) => for example modem users are really critical when downloading demo - especially if you have to pay for each minute *With DirectX7 (which Blitz uses) you simply can't compete with DirectX9 if you think of graphical elements (this doesn't say you can't have cool looking game - it's just fact that others will have ever cooler looking games) *It's quite different to compete in shareware market than competing in retail business (or similar - as I can see you have quite much experience in the cutting edge tech demos) *Not all people have AMD XP2000+ with GeForce Ti4 - there are simply so many computers with GF2 etc. that trying to aim higher specs will results in lesser sales. But as said: you can do this just for fun, or you can do this for the money. But you can't beat million dollar budgeted retail game visually using just Blitz3D. But hey, I don't want to argue about this. Those were my points (and same as many in the dexterity forums seem to approve) and I can compromise with you by saying: "yes it's possible - but damn, damn, damn ultradamn hard work to compete with retail games using Blitz" :) |
| ||
| @Morphecy Ummm... you're right, if you do games like you described, and for the target ppl / platform / distribution you described. My personal goals are different, what I'm doing will be so big in end (DL size etc) that "downloadable shareware" is probably not possible in my case... on the other hand ppl happily DL game demos that are even 100-200megs in size? But I'm really aiming for boxed game product, sold on retail. you can do this just for fun, or you can do this for the money. Hopefully I can combine those :) But, yes, this is a hobby-project for me more than anything because I already have a fulltime job in the industry. you can't beat million dollar budgeted retail game visually using just Blitz3D. I sure will try to do just that - I may fail miserably, but IMO it's always good to set your goals too high than too low! "yes it's possible - but damn, damn, damn ultradamn hard work to compete with retail games using Blitz" SO totally agree! :) |
| ||
| @Mustang: heh :) You really must show your project to me when you have finished it, and good luck with it! =) |
| ||
| I disagree entirely with your statment that you can't beat million dollar retail games. I think blitz is perfectly capable of handling such. Yes, however it will take months and possibly years of coding, but you could do it. |
   |