Doom 3@60fps

Rookie

Newcomer
IGN reports Doom3 will be capped to 60 fps.

"At a recent NVIDIA Editors' Day, id Software CEO Todd Hollenshed announced that DOOM 3 will be capped to 60 frames per second in the rendering engine.

We checked with John Carmack himself about why DOOM 3 will be hard-capped at 60fps in the renderer, and he had this to say:

"The game tic simulation, including player movement, runs at 60hz, so if it rendered any faster, it would just be rendering identical frames. A fixed tic rate removes issues like Quake 3 had, where some jumps could only be made at certain framerates. In Doom, the same player inputs will produce the same motions, no matter what the framerate is."

Will this change the prospect of gfx industry?Will this urge IHV to focus on IQ rather than persuit the 100+/200+ meaningless fps?

Ofcoz,I know that means 60fps is the highest fps in doom 3,the ideal status is that avg or lowest fps@ 60fps.
 
hey thats a 71.5% fps increase from old dooms. Maybe carmack couldn't get the fx 5600 to run doom 3 at 1024 x786 @ 90 with textures so he decided to cap it lower.

Being serious I wonder what sort of cpu we will need to handle the physics because they have made their own physics engine which is what all this capping is about.
 
This cap is driven by peculiarities of the physics engine.
Personally, I think this is a crummy design decision.

60Hz tick. Ugh.
It will be impossible to build a game around this engine that is as fast-paced as Quake3. Or rather, you could, but it would suck.
 
Entropy said:
60Hz tick. Ugh.
It will be impossible to build a game around this engine that is as fast-paced as Quake3. Or rather, you could, but it would suck.
Of course you could. D3 will run with 60hz tick, but if you used the engine to make a fast paced game then you could use a different tick for that game, like 120hz, or something else. The user might not be able to change the tick rate, but developers using the engine to make another game will be able to change it for that game.
 
Rookie said:
The game tic simulation, including player movement, runs at 60hz, so if it rendered any faster, it would just be rendering identical frames.
Why ain't it interpolating between two physics frames for rendering??? :oops:
Nearly everyone already does that for ages (that is, run logic at low fixed rate, interpolate before rendering)...
 
Entropy said:
60Hz tick. Ugh.
It will be impossible to build a game around this engine that is as fast-paced as Quake3. Or rather, you could, but it would suck.

Quake 3 only uses 30 ticks by default.

Edit:
30 is the defaulth tick rate on servers in Q3 the tick rate on the client is dependent on the framerate.
 
NeARAZ said:
Why ain't it interpolating between two physics frames for rendering??? :oops:
Nearly everyone already does that for ages (that is, run logic at low fixed rate, interpolate before rendering)...

Because, as explained, this is what leads to the big problems previous games have suffered for mainly online play.

Quake2 resulted in all the "pro" players ramping to damn near wireframe output to get 200fps as the interpolation between engine ticks, renderer refresh and network refreshes led to compromises between them to allow players with higher framerates to pull off incredible feats and thus have an advantage.

Quake3 reduced this through tighter interpolation, but still had problems with jumps, recoil and reversals. The physics again became less accurate/dissimilar once renderer framerate was increased.

I'm sure John Carmack is being forward thinking for Doom3-engine online games down the road. It's always been a huge problem with prior id software engines and left the 3rd party market scrambling for ways to cap client framerates and client packet updates to be as close to 1:1 as possible to keep like-physics for all players. 60hz seems a bit arbitrary for single player, but it is nearly perfect for network/online use to tween to client updates if such uses come out of the engine down the road.
 
Entropy said:
60Hz tick. Ugh.
It will be impossible to build a game around this engine that is as fast-paced as Quake3. Or rather, you could, but it would suck.
Well, the fastest a human being can react to a signal is about 100ms (sprinters who leave their blocks less than 100ms after the gun are given a false start, even though they left AFTER the gun!). 60 updates/second is 6x faster than this. Should be plenty, I'd have thought.

If this was a problem, then playing network FPS with 100ms pings would be pretty much impossible, instead of being the point when the latency just starts intruding a bit.

I'm sure someone will say they notice even the effect of DSL in-country pings at around 50ms, of course.
 
Remember the shutter of a camera remains open for a time, and with a flash the time is a prescribed amount, therfore if you blink after the flash goes off the shutter may still be open and the negative will show you blinking. You see what I mean. I think the flash stays on for 60ms though. BTW usually when you blink you actually blink before the flash goes off, i.e. you see the finger pressing down and involutarily blink.
 
This is very interesting...

The brightness does not decay or rise instantly. It may be that the lag between the 'flash starting' and the 'brightest point of the flash' (at which the greatest part of the film response will occur) is around 100ms, thereby giving you enough time to react. I don't know what the profile of this is, a quick Google search didn't reveal much.

Another theory is that the blink response is, as a reflex response, faster than the brain having to process and say 'do this' - perhaps the response time is only 20-30 ms... again, the Google search didn't reveal much.

Anyone know more on either of these?
 
Sharkfood said:
NeARAZ said:
Why ain't it interpolating between two physics frames for rendering??? :oops:
Nearly everyone already does that for ages (that is, run logic at low fixed rate, interpolate before rendering)...

Because, as explained, this is what leads to the big problems previous games have suffered for mainly online play.
No, the physics/sim ticks are still running at the same rate between all clients, so there is no way for one to be at an advantage - it's merely that what you see rendered on screen is the interpolation of two EXISTING sim ticks; ie the previous and current one.

Honestly this move suprises me greatly, as if I didn't know better, I'd think that John has lost his marbles. As NeARAZ stated, a huge number of games have been doing it this way for a long time, so it's definitely nothing new or confusing that could account for a bad implementation. Other than this fact, the 60Hz tick seems very high also - I've worked on racing simulation games with ticks as low as 8Hz and RTS's with around 12-15Hz, but I've never heard of anything above 30Hz. Seems like there's a lot of unnecessary processing going on there at 60Hz.
 
Dio said:
If this was a problem, then playing network FPS with 100ms pings would be pretty much impossible, instead of being the point when the latency just starts intruding a bit.
Playing network FPS with 100ms ping is pretty much impossible.
With 0ms ping vs 100ms ping I seriously believe I would beat fatality even in 1on1. And he is much MUCH stronger than I was at my strongest.

I haven't played fatality, but I have played vs. Lakerman (twice) and his fellow players in All* (more times). Clans quit Div. 1 Q3CTF because having a ping disadvantage of 30ms was absolutely hopeless. It takes a huge advantage in skill to compensate for a 70ms vs 40ms ping disadvantage.

I never play with triple digit ping. Ever. Not even slack play. It's just too damn frustrating.

Tim said:
Quake 3 only uses 30 ticks by default.
However, you do get interpolated frames, which helps smoothness immensely. Try playing with com_maxfps=30..... The claim in the original article was that if the frame rate was higher than 60 fps, the game wouldn't interpolate between ticks. Those 60 frames would be repeated if the rendered framerate went higher. Ideally, the effect would be as if you set com_maxfps=60 in Q3. Which is horrible.

If the claim in the original article was wrong, and the engine does produce interpolated frames, then there's not much of a problem. Even if the article was correct (which I believe) the issue is probably negligeable for DOOM3. But I stand by my claim that lack of interpolation will make the engine troublingly flickery, and thus ill suited for a game the speed of Quake3.
 
Dio, since you asked for theories, here’s mine.

Actually I was being only somewhat fictitious with the “cheeseâ€￾ remark. It is human nature to have heightened reflexes in anticipation of a major event ( to the brain a flash going off is a major event). Before having our picture taken or level of anxiety is raised and we know that a flash is going to go off, thus we are prepared to blink at the slightest event. This could be saying the word “cheeseâ€￾, the movement of the shutter finger, etc. Compare photo’s taken candidly to those that are posed. I am sure looking at the candid ones (where the subject was unaware that a photo was going to be taken) you will find very few were the subject blinks.
 
Well, I knew the quake-heads would argue the call about ping latencies.

All I can say is that for the first time recently I've gone online with FPS (Jedi Academy finally suckered me into doing it). I would say I haven't noticed any problems with around 60ms ping and (referring back to that earlier thread) the LCD response time. And I came top on team quite a bit more than I expected to. But there you go. I'll never be a real FPS player.

Maybe we've all missed the fundamental point: a 60Hz tick is fine as long as the rules are the same for everyone :)
 
Back
Top