Something wrong with the HL2 Story
- Thread starter palmerston
- Start date
Testiculus Giganticus
Newcomer
Guys, take care, there may be some hardcore AD&D players here, and trolls are generally hunted and killed by AD&D heroes-it may be because they`re big, ugly and carnivorous? 
No game I look forward more to than HL2 (and TF2 and CS2). I play atleast an hour a day on the Half-Life 1 engine (in a CS clan).
No higher wish than to have everyone upgrade their systems to DX9 and have all game developers target this.
Strongly disappointed with past claims (game engines claiming they targeted DX7 and DX8)
Strongly disappointed with past claims of "scalability" (progressive geometry, fallback shading)
Skeptical of Valve's claims. Company's bread and butter is people with crap systems. Company has put huge effort into milking HL1 with as little effort as possible and funky content distribution schemes, reselling HL1 and not even fixing old bugs.
Selling a game engine designed with DX9+ level content on nice screen shots and movies to their bread and butter audience could lead to buyer shock, as what happened with Enter the Matrix as people looked over screenshots, and then found out that their graphics got scaled down massively (and fugly) on their own systems and consoles.
Let's just say that I am reserving judgement on the hype coming out of Gabe et al, because I've heard it all before from others, promising the world, and coming up short. (Daikatana anyone?) I have a DX9 card (9700 PRO) so it is not a worry to me, and I hope Valve succeeds because if love HL1/TF2/CS and want more of the same, but with awesome graphics and physics.
No higher wish than to have everyone upgrade their systems to DX9 and have all game developers target this.
Strongly disappointed with past claims (game engines claiming they targeted DX7 and DX8)
Strongly disappointed with past claims of "scalability" (progressive geometry, fallback shading)
Skeptical of Valve's claims. Company's bread and butter is people with crap systems. Company has put huge effort into milking HL1 with as little effort as possible and funky content distribution schemes, reselling HL1 and not even fixing old bugs.
Selling a game engine designed with DX9+ level content on nice screen shots and movies to their bread and butter audience could lead to buyer shock, as what happened with Enter the Matrix as people looked over screenshots, and then found out that their graphics got scaled down massively (and fugly) on their own systems and consoles.
Let's just say that I am reserving judgement on the hype coming out of Gabe et al, because I've heard it all before from others, promising the world, and coming up short. (Daikatana anyone?) I have a DX9 card (9700 PRO) so it is not a worry to me, and I hope Valve succeeds because if love HL1/TF2/CS and want more of the same, but with awesome graphics and physics.
I think Sacrifice scales successfully. No card could run it with full geometry and texture settings when it first came out, and it looks damn nice now on my 9700. The characters do look better bringing all that power to bear, although the models aren't "realistic" so the scaling is made easier.
Hyp-X said:
Those use at maximum one pixel outside the sampling area due to 2x2 filtering (AF is another matter all the way) so adding the one pixel border shouldn't be a problem.
In my opinion the fact that they pack multiple textures in one is very understandable from the perspective when the Source engine development was started, only power of two, possibly only square sized textures. In a texture intensive game like HL2 you'd want to have "interesting" shaped textures to be grouped into square, power of two sized ones to minimize amount of wasted memory and state changes. For example have all textures for sides of a crate grouped in one as you usually have to use more than one of them to draw the crate anyway (and as fps seems to be the genre of crates you definitely need lots of them). Even today you can't have mipmapping for non-power of two textures with any architecture afaik (might change with Longhorn but anyway).
Unknown Soldier
Veteran
Some of you may or may not have seen the Bink HL2 Movies. While the E3 HL2 Movies looked impressive with what the Source engine could do, it really did not justice to what the Game really looked like.
D/l the bink Movies below and be prepared to be blown away.
ftp://ftp.sunet.se/pub/pc/games/gamershell/movie/HL2-G-Man.exe
ftp://ftp.sunet.se/pub/pc/games/gamershell//movie/hl2-small_lab.zip
ftp://ftp.sunet.se/pub/pc/games/gamershell/movie/hl2-docks.exe
US
D/l the bink Movies below and be prepared to be blown away.
ftp://ftp.sunet.se/pub/pc/games/gamershell/movie/HL2-G-Man.exe
ftp://ftp.sunet.se/pub/pc/games/gamershell//movie/hl2-small_lab.zip
ftp://ftp.sunet.se/pub/pc/games/gamershell/movie/hl2-docks.exe
US
Don't count on it. Regardless of what one of the whitepapers said FAA is still multisampling so it will have the same problem.hkultala_ said:now this reminds me about Parhelia:
AFAIK it should be immune to this effect and it's FAA is much faster than any SSAA...
so parhelia might be quite a good card to run HL2
DemoCoder said:
Semantics, perhaps, but an important distinction, I think: "target" is not a word I've seen used often in this context; but "supporting" DX7/8 certainly is. I think to some degree you are confusing "targeted" API versions with "supported" API versions. I've known a lot of game engines which support to one degree or another more advanced API versions--few that target them.
Yes, agreed. But I think it's fair to state that while many developers eagerly embraced scalability as a desirable goal, most of them found it much easier said than implemented...
Bread & butter of all developers is the median hardware level (the crap level.) I see little difference between Valve and other companies here. A distinction I do see is that Valve has taken 5 years to bring out HL2. This raises expectations for it far higher than would be the case in even a 2-year development cycle. IE, I think Valve will be harshly criticized if HL2 is not dazzling. From what I've seen of the demo movies being circulated it seems likely that HL2 will satisfy in most respects. But again, I'll reserve judgement until I have the game.
This is a pre-release PR job that hopefully Valve will not skirt in the coming months. They need to dazzle, but they also need to clearly spell out the hardware requirements for the game at the same time.
Heh....
Diakatana was exactly what I thought it would be as the hype never impressed me....it was crap just as I thought....*chuckle* I see many, many differences between HL2's upcoming release and Daikatana's Swiss-cheese release schedule. Not to mention Romero's "suck-it-down" mentality surrounding the game. I can't say for certain, of course, but I think Valve will do a much better job in meeting expectations for HL2.Johnny Rotten
Regular
DemoCoder said:
Thats unfair in my books. I think there is ample evidence that Valve has NOT spent the last 4 and half years 'milking' Half-Life, but rather developing HL2 while still providing a level of support to their customers that is unparalleled.
You're justgoing to create lots of little trolls that way. You want a either a Flamebrand or even better (and for all those megalomaniacal Magic Users out there) a nice 'little' Meteor Swarm.
However back to HL2. Demos got a right to worry, previous track record on supposedly scaleable engines ain't that great. However I am willing to bit that Maybe Valve can get it right this time around. Who knows, Maybe we'll be arguing about the scalability of the HL3 engine in another 5 years.
Whatever happens I reckon it's gonna be a great game.
jpaana said:
This is for the first miplevel only.
The second miplevel consist of down filtered texels.
So bilinear filtering the 2nd miplevel uses a 4x4 area of the original texture.
You need to have 2 pixel border to prevent bleeding in the 2nd miplevel.
As for the 3rd miplevel ... I hope you get the point.
And yes AF is a problem even if you disable mipmapping.
