Advertisement
If you have a new account but are having problems posting or verifying your account, please email us on hello@boards.ie for help. Thanks :)
Hello all! Please ensure that you are posting a new thread or question in the appropriate forum. The Feedback forum is overwhelmed with questions that are having to be moved elsewhere. If you need help to verify your account contact hello@boards.ie
debate about Halo 2 and rendering and other stuff
Options
-
15-05-2011 3:10pm#1Join Date:Posts: 50058
Surely there's a case to say that the framerate issues in most games, that early on in a consoles life, are more likely to be down to poorly optimised code rather than an inability of the hardware to meet demand?
Even though it was a first party game.
A good developer will try to get a balance between looks and a solid framerate so that gameplay is unaffected. It's failure on the developers part. For a team like Bungie that managed to get Halo to run so well and so late in the Xbox's life it's pretty much inexcusable. They should have cut back on the amount of pixel shaders that the game used because it was obvious that the console was having pretty bad. The game shipping with such terrible texture pop up in cutscenes ws really bad as well. Orta on the other hand used every trick in the Xbox's library but used them in moderation and looked and played far better as a result. Halo 2 on the other hand was just as technically advanced but was a really ugly looking game most of the time with technical issues.
It's come out that Halo 2 was a total disaster of game development and the final build was thrown together in 9 months, and it really shows in the gameplay. Perhaps they could have gotten the game to run better with more development time but the game you release I'm afraid is your legacy.
Ninja Gaiden would be another gorgeous Xbox game. It forgoes the use of advanced shaders and other effects to really pump out a huge amount of polygons and a high level of geometry, an effect that the even more gorgeous Metroid Prime used.
0
Comments
-
Retr0gamer wrote: »A good developer will try to get a balance between looks and a solid framerate so that gameplay is unaffected. It's failure on the developers part. For a team like Bungie that managed to get Halo to run so well and so late in the Xbox's life it's pretty much inexcusable. They should have cut back on the amount of pixel shaders that the game used because it was obvious that the console was having pretty bad. The game shipping with such terrible texture pop up in cutscenes ws really bad as well. Orta on the other hand used every trick in the Xbox's library but used them in moderation and looked and played far better as a result. Halo 2 on the other hand was just as technically advanced but was a really ugly looking game most of the time with technical issues.
It's come out that Halo 2 was a total disaster of game development and the final build was thrown together in 9 months, and it really shows in the gameplay. Perhaps they could have gotten the game to run better with more development time but the game you release I'm afraid is your legacy.
Ninja Gaiden would be another gorgeous Xbox game. It forgoes the use of advanced shaders and other effects to really pump out a huge amount of polygons and a high level of geometry, an effect that the even more gorgeous Metroid Prime used.
How do you know anything about how many shaders they used?
0 -
Join Date:Posts: 50058
Just look at the game. Halo 2 covers absolutely everything in Pixel Shaders. The fact that the game takes a few seconds to load the pixel shaders in and the pop up makes it very noticeable that every texture is covered with a bump mapped pixel shader on top of it.
Panzer Dragoon Orta is again a case of looking at it. The enivorments aren't covered with pixel shaders but the NPC and enemy models use them very subtley. Check out the model viewer in pandoras box to see it. It's a far more subtle and frankly better use of pixel shaders. There was a time that pixel shaders where plastered about in every game and it was like coloured lighting and lens flair in the 90's they were over used and made games look ugly despite being technically of a very high standard. It took Half-Life to show people that a more subtle use of lighting and coloured light makes far better looking games and funnily enough it was Half-Life 2 that thought developers to be more subtle with the use of pixel shaders.
0 -
That's usually the case. If you're not a graphics programmer with an authority on the subject, you shouldn't pretend to be and shouldn't spam the forum with misinterpreted drivel on the pretense of being a graphics programmer. Just simply cite your source, instead.SirLemonhead wrote: »Just comes across like someone who's read a bunch of technical terms on Wikipedia or something and is just regurgitating it here. You don't seem to know how they actually work and are applied..
0 -
NeoKubrick wrote: »That's usually the case. If you're not a graphics programmer with an authority on the subject, you shouldn't pretend to be and shouldn't spam the forum with misinterpreted drivel on the pretense of being a graphics programmer. Just simply cite your source, instead.
I am, and have released a game on the xbox (sorta)
It's more the "stating something that's really just my opinion in a way to convey it as being" fact that I disagree with. The very idea of just *looking* at something and being able to discern what the cause of the performance issue is a bit mad0 -
-
Advertisement
-
Join Date:Posts: 50058
NeoKubrick wrote: »That's usually the case. If you're not a graphics programmer with an authority on the subject, you shouldn't pretend to be and shouldn't spam the forum with misinterpreted drivel on the pretense of being a graphics programmer. Just simply cite your source, instead.
Never said or pretended to be a graphics programmer I just read up on this stuff because I'm interested in it. Also if like me you had a PC around the time running up to the release of Doom 3 and Half-Life 2 you knew what pixel shaders were due to the hype around the technology from these games and wanting to own a PC that could carry out these effects properly.
I'm not writing a disertation here I'm posting on a forum. Bungie hasn't commented on specifics on why performance was so bad and why there were technical issues with Halo 2. I'm making an informed assumption on what I know. The GPU in the Xbox wasn't the best for normal and specular maps which is why it's used in moderation in most games. Halo 2 had them plastered on everything. So you've got a texture with a specular map and a normal map, maybe just the one shader or a custom shader but a shader none the less, on top being loaded in for every texture. That's a hell of a lot of memory for the RAM to handle and it shows with the texture pop in in the cutscenes. When there's insufficient ram you get a lot of memory swapping which leads more processor overhead. It's not hard to extrapolate that the overuse of a very intensive process that the hardware wasn't built for is a cause, but not the only cause,of the poor performance. There's a reason the PS2 supported normal maps but were only used in a handful of games and only late in it's life, Path of Neo being the only game I can think of. It's a very processor intensive process.
Perhaps you'd like to apply occams razor and come up with a more plausible theory or just sit there sneering and adding nothing to the discussion.SirLemonhead wrote: »I am, and have released a game on the xbox (sorta)
It's more the "stating something that's really just my opinion in a way to convey it as being" fact that I disagree with. The very idea of just *looking* at something and being able to discern what the cause of the performance issue is a bit mad
I didn't say I just looked at the game and knew what caused the poor performance. Read the posts again. You asked how did I know about how many shaders they used. As I said it's very obvious. Just look at the game and you'll see that practically every texture has at least one shader map applied to it. It's very obvious and it's applied far more extensively than any other Xbox game I've seen.
Also having made an XNA game doesn't make you an expert either just because you know a few post processing shader effects. If you're willing to add some discussion and maybe teach me a thing or two then fine. One of the reasons I know a bit about shaders is because a few years ago someone corrected me about them one this board.
0 -
-
-
-
-
Advertisement
-
-
Join Date:Posts: 50058
It's not going to stop me discussing and potulating about it even if makes me sound like Sheila from Joe Duffy. Anyway you don't need to hold a qualification in a subject to know something about it and I know far too many graduates and even postdoctorates that haven't a clue about their respective fields and still work in them.
There's no point citing sources when there's no one that specifically points out what the performance issues are. I still think it's very obvious what some of the performance issues are and what caused them and there's no harm in speculating. If you want to know where I got some of the information it's from digital foundry articles, articles and videos on sahders by Valve, ATI and Nvidia and also the article on eurogamer about Halo 2's development. Anyway you are one to talk when you cite in an article a PR presentation by bungie on how great they are which is probably worse than citing wikipedia in an article. Hardly independent or peer reviewed.
0 -
Moderators, Category Moderators, Arts Moderators, Computer Games Moderators, Entertainment Moderators Posts: 29,147 CMod ✭✭✭✭
Join Date:Posts: 27859
-
-
Moderators, Category Moderators, Arts Moderators, Computer Games Moderators, Entertainment Moderators Posts: 29,147 CMod ✭✭✭✭ Join Date:Posts: 27859
Retr0gamer wrote: »You're solving nothing. Do you not see the amount of texture applied pixel shaders on the CGI?
We can go more traditional if you like.
0 -
johnny_ultimate wrote: »We can go more traditional if you like.
Ah make it a LITTLE bit more up to date yeah?
FIGHT! FIGHT! FIGHT! FIGHT! etc...0 -
-
-
Join Date:Posts: 50058
I'm not bitter just defending myself from personal attacks from two sources that so far have contributed nothing to discussion. Also one of them doesn't seem toosure of himself what qualifications he has or has backed them up.
I'm not a graphics programmer with experience of coding graphical solutions but I think I've more than enough knowledge to come up with and defend the solution to Halo 2's technical issues. I actually think it's very obvious and doesn't require these qualifications, just a bit of knowledge. Pixel shading on textures is a very ram and processor intensive. Just think about the amount of calculations that have to be run for normal map real time lighting effects. The Xbox GPU is based on a low spec Geforce 3 which has very rudimentary support for pixel and vertex shaders and would be hopeless in a PC at running games that required a lot of these effects like Doom 3. So you are saying that my hypothesis that cutting down on pixel shader effects in Halo 2 wouldn't help Halo 2 run better. In fact just like Halo 1, Half-Life 2 or Panzer Dragoon Orta I think a less is more approach would be better, allowing higher res textures to be used and giving the game less of a 'everything smeared in vaseline look'. The low amount of video ram and only a few shader pipelines on the GPU coupled with the sheer amount of these effects is certainly putting a strain on the system. So cutting them down is my solution. Find some one to tell me it wouldn't work.
Of course there's probably more in Halo 2 causing problems. It could and probably is be bad programming but I'm not qualified to say that. There's probably other ways to fix the game as well but my approach to it I believe is valid.
As for GDC 2002, do you need reminding that recently developers complained about the lack of peer review at GDC and it had become a place for developers to shill their wares. It only recently got turned into what it was originally intended for and it was a few years after 2002. Anyway anyone who reads the presentation will know it's a load of bull. Increasing the difficulty makes gamers think the AI is smarter. I'd rather a paper on how to actually make the AI smarter and how not to select your test group from dumb asses.
0 -
-
Advertisement
-
Join Date:Posts: 50058
Here's something with pictures of what a normal and specular map is in case anyone is interested:
http://ciardhubh.de/node/18
Basically a grey scale or another colour scale texture with the scale of the colour denoting the height. You throw that over a texture and when you hit it with light in the game it will change the colour of the texture underneath and add specular highlights to give the illusion of depth to the texture increasing it's detail and making it seem 3D instead of flat. The calculations for lighting it are predefined so it's much less processor intensive than it should be.
So you've got a texture ovelaid with one or more of these shader textures and all the lighting calculations. So as any person with a small amount of computer knowledge will probably come to the conclusion that yes these are very resource intensive.
0 -
Retr0gamer wrote: »I'm not bitter just defending myself from personal attacks from two sources that so far have contributed nothing to discussion. Also one of them doesn't seem toosure of himself what qualifications he has or has backed them up.
As for GDC 2002, do you need reminding that recently developers complained about the lack of peer review at GDC and it had become a place for developers to shill their wares. It only recently got turned into what it was originally intended for and it was a few years after 2002.
Do not justify your embarrassing personal attacks by falsely claiming I personally attacked you. That only exists within your own noggin'.
Don't weasel out of your utterly incorrect assumption that 'The Illusion of Intelligence' was a PR presentation and wasn't peer-reviewed. It was. You cannot question the definition or quality of 'peer-review' or the peer committee now after you based your pathetic personal attack upon it a post before. The Game Developer Conference is for the game industry: therefore, the only people shilling "their wares" are the software and hardware companies (ATI et cetera), not the "game" companies, because for software companies, the customer is the game developer and for game companies, the customer is the gamer. Clearly, you don't know what the Game Developer Conference is or the dynamics of the game industry.
Anyway, you missed the fact that you contradicted yourself by holding 'peer-review' as irrefutable and synonymous of a standard but simultaneously refuting academic qualifications as not synonymous of a standard. Your comments on Butcher's and Griesmer's seminal work is too embarrassing to respond to.EnterNow wrote:I hate to say it guys, but I'm with Retr0 on this one (again). You don't need a piece of paper to prove you know something, if you've a genuine interest in something, you'll know about it. He may be technically wrong, I've no idea its all over my head, but he's not automatically wrong just because he doesn't have a cert or qualification in the field
Strawman argument: it's entirely irrelevant if he is right or wrong as it is if a monkey randomly scribbles and imitates 2 + 2 = 4 on a blackboard. The monkey may have made a correct calculation, but it's irrelevant if he does not understand the equation or the mathematics. By Retr0's own admission, he is not a graphics programmer, and therefore, he is either talking gibberish or he is interpretating technical expertise and applying it as if he were a technical expert without any code testing the application and comparing the algorithms, which is obviously stupid and creating a baseless argument.
"Code or it didn't happen". If Retr0 were a graphics programmer, then he would obviously understand the import of that statement.0 -
Join Date:Posts: 50058
NeoKubrick wrote: »Do not justify your embarrassing personal attacks by falsely claiming I personally attacked you. That only exists within your own noggin'.
Don't weasel out of your utterly incorrect assumption that 'The Illusion of Intelligence' was a PR presentation and wasn't peer-reviewed. It was. You cannot question the definition or quality of 'peer-review' or the peer committee now after you based your pathetic personal attack upon it a post before. The Game Developer Conference is for the game industry: therefore, the only people shilling "their wares" are the software and hardware companies (ATI et cetera), not the "game" companies, because for software companies, the customer is the game developer and for game companies, the customer is the gamer. Clearly, you don't know what the Game Developer Conference is or the dynamics of the game industry.
Anyway, you missed the fact that you contradicted yourself by holding 'peer-review' as irrefutable and synonymous of a standard but simultaneously refuting academic qualifications as not synonymous of a standard. Your comments on Butcher's and Griesmer's seminal work is too embarrassing to respond to.
Well I was wrong there. I assumed it wasn't peer reviewed because of how bad it was. Anyway GDC definitely shifted more towards actual game developer presentations now due to complaints that it was focusing more on big reveals. It's why platformer holders don't hold keynotes now if I remember correctly. It's hardly great peer review when you let a presentation through on game development that is just product reveals and demos. Also Butcher's and Griesmer's 'seminal work is a load of balls imo. My opinion and I'm sure a lot of other people that read it will agree.NeoKubrick wrote: »Strawman argument: it's entirely irrelevant if he is right or wrong as it is if a monkey randomly scribbles and imitates 2 + 2 = 4 on a blackboard. The monkey may have made a correct calculation, but it's irrelevant if he does not understand the equation or the mathematics. By Retr0's own admission, he is not a graphics programmer, and therefore, he is either talking gibberish or he is interpretating technical expertise and applying it as if he were a technical expert without any code testing the application and comparing the algorithms, which is obviously stupid and creating a baseless argument.
"Code or it didn't happen". If Retr0 were a graphics programmer, then he would obviously understand the import of that statement.
So what you can't teach yourself these things? You're putting too much weight on a piece of paper here. There's plenty of people that have doen great things and become experts in fields that they have thought themselves in. Anyway you seem to have less knowledge in this area than I do so what gives you the right to say I'm wrong? For all you know I could be right. And no I don't need to know how to code to know what a resource intensive process is. What's coding got to do with this? It's efficient use of assets that's got very little to do with code optimisation.
0 -
Join Date:Posts: 50058
Also here's a lesson in academia. There's different levels of peer review and some peer reviewed papers will let any old **** through. There's impact factors of the papers to consider, the higher it is the harder it is to get a paper accepted since your work must be of a very high standard. It's why in what I'm doing you have to be very careful where you reference. Just because it's peer reviewed doesn't mean it's a load of old balls. It's also why China is pumping out tonnes of research papers at the moment, usually to badly peer reviewed journals. They're peer reviewed but of terrible quality but still published and there to make the republic look good.
0 -
Can we please cease and desist the 'cite your sources' bs... it's a bloody internet message board, not an academic debate. And I've been to academic debates, it's the same petty bickering about minor inconsequential details and nobody ever learns anything!
Just pit your wits against each other, mano e mano, in a battle to see who can sound the smartest and get the most thanks
0 -
-
Join Date:Posts: 50058
hooradiation wrote: »This is not to say one can apply shaders endlessly, as there is a cost in terms of reading and writing to the RAM, but they're not this massive drain on resources you make them out to be.
But isn't this the point I'm making, that they've gone too far applying shaders, increasing reading and writing to the ram and draining resources?hooradiation wrote: »The entire point behind normal maps is to reduce those calculations to the lowest possible. That's what a normal map is, it is a map of pre-calculated vertex normals so that cycles aren't wasted working that out on the fly.
Oh I know but it's not exactly a process with no resource overhead either.
I was just speculating. The game was rushed and there's bound to be other technical issues contributing as well. However some people just seem to think I'm stating fact and take a mickey fit over it. It's nice to have an intelligent conversation like this and don't mind being proven wrong.
0 -
I've been either busy today or been ill for the evening (stuck in front of the telly watching movies all night)
I didn't mean to come across the way I did, I apologise for that.
I'm a bit apprehensive of stating what I've done, as I'm not sure how "impressive" it comes across.
It's easy to find what I did if you throw my name into google, so i'll tell you. I took an old game, wrote a new rendering backend for it so good that the company who originally made the game stole my work and re-released it on Steam without crediting me. After emailing them, all I got was a vague "yeah thanks" and a promise of a copy of their new game. As that new game turned out to be Aliens Vs Predator 2010, i'm glad I declined
So yeah, AvP Classic 2010 on Steam brought to you by..me. I've got my proof it was me and I don't really feel like backing it up now, not especially nearly a year and a half later.
I ported it to the xbox also, as well as doing similar work for the game Forsaken (I wrote a basic d3d9 backend for that which got the guys started on writing an OpenGL version)
That's just the stuff I've released or has seen the light of day.0 -
Join Date:Posts: 50058
Oh yeah you mentioned that before. That was pretty nasty what rebellion did to you and pretty much unacceptible but you hear about it happening all the time with companies stealing modders work.
0 -
Advertisement
-
https://www.youtube.com/watch?v=ei5l3r1dV4I
This discussion has been closed.
Advertisement