Input lag really isn't lag?

[deftonesmx17]

The blind at the exit of the tunnel is only present in a few tracks and its delay and intensity is a deliberate effect.

As a whole and computational wise is unrivaled and until the new consoles arrive I doubt that you will see another game doing most of the things that GT5 is doing now at once(1080p/60, volumetric smoke, 16 premium cars on track, realtime day change with dynamic shadows, real time weather, photorealistic lighting engine, realtime damage, physically affected rain particles, etc).

No, it is piss poor game HDR, probably bloom. Most devs have found that subtle HDR or deffered rendering looks much more realistic. Lets also not forget that the type of nvidia card (NV47) in the PS3 can't do proper FP16 blending + AA, those of us who are PC gamers know this. The HDR in GT5 looks like Half-Life 2 HDR which was completely fudged to overcome said limit on said GPU's.

BTW, F1 2010 says hello about all of those except for 1080p and lets not forget that GT5 is not true 1080p and never stays at a constant 60fps.
GT5's weather effects are laughable compared to F1 2010's.

 

[FuriousDemon]

I just did a little test. Played S2U on my PC, I mean, using my pc monitor (1680x1050 max res.) instead of my tv (1920x1080). And what do you know, the lag is gone! So Ian, or someone posting under his name, is completely correct. It has everything to do with how fast your system runs. So if you have a SUPER high end PC (mine is somewhere between mid and high) you can play at 1920x1080, however if you're on PS3 or 360, which are piss poor machines by todays standards, even with obvious graphical sacrifices on those platforms the current engine was not able to process your input quickly enough. If what Ian was saying is true and they were able to change the order of execution, then we should be able to get "lag-free" gaming. If this was a lie and there is no patch in sight, I suggest you get top of the line PC, that is if you really want to play this game that bad Again, the faster your PC (or the worse your graphical settings are), the less lag there will be.

 

[Spagetti69]

Physically effected rain particles.

What is that about?

 

[spapadillion]

I just did a little test. Played S2U on my PC, I mean, using my pc monitor (1680x1050 max res.) instead of my tv (1920x1080). And what do you know, the lag is gone! Again, the faster your PC (or the worse your graphical settings are), the less lag there will be.

I'd imagine the results are skewed by your computer monitor having a much, much faster response time than your TV.

 

[Zer0]

-Is having multi-million poly cars on previous generation level tracks superior to having lower level cars and better modelled tracks? Opinion is important here.

At the end of the day, GT5 does not consistently reach the specs it is advertised as (1080p, 60fps). And it makes major sacrifices to other content (tracks, smoke, shadows, effects) to achieve what it does.

Funny thing is that the city tracks in GT5 are still one of the best in its genre, the smoke/dust one of the most reallistic rendered, the shadows move in real time and affect all, and most of the effects are still good(glares at night, sparks, wind, tyremarks, fireworks, out of track particles, dirt tyres, etc).

No, it is piss poor game HDR, probably bloom. Most devs have found that subtle HDR or deffered rendering looks much more realistic. Lets also not forget that the type of nvidia card (NV47) in the PS3 can't do proper FP16 blending + AA, those of us who are PC gamers know this. The HDR in GT5 looks like Half-Life 2 HDR which was completely fudged to overcome said limit on said GPU's.

BTW, F1 2010 says hello about all of those except for 1080p and lets not forget that GT5 is not true 1080p and never stays at a constant 60fps.
GT5's weather effects are laughable compared to F1 2010's.

Sure, seems they used a vastly superior HDR for these tracks...

http://www.youtube.com/watch?v=1Ggz0M8IYo4#t=1m3s
http://www.youtube.com/watch?v=2n6mb70KFSo#t=1m26s
http://www.youtube.com/watch?v=dJf_k0d_A-A&t=0m20s

The PS3 is not a PC and the gpu renders only a part of the final graphics, most effects are coded out of the gpu and computed by the CPU like in the PS2.

Then what is for you the graphical benchmark of the generation?...

Physically effected rain particles.

What is that about?

Rain drops being individually affected by the car acceleration, jumps, turns, speed, wind and sprays.

 

[oshaylinux]

I do read most of the forums but find it best to respond in those, how shall I put it, well moderated ones

The occasional punter takes the opportunity to have a go at us when we pop our heads above the parapet. Nothing wrong with that if it's at a sensible level of course but we aren't immune to a enjoying a bit of credit also. The appreciation thread and the 'little things' threads here were nice to read

Is't it only right to acknowledge on the official S2U forum, that you all are working on these issues? With the loyalty that they have shown and is showing for the brand, isn't that worth something? I understand no one likes to be ridiculed, but they did buy a product that has manufacturer's defects in it. So one could easily understand their frustration. Isn't that insult to injury, for you all to address other sites but not the official one?

 

[Spagetti69]

Silly question but is it really Ian Bell?

 

[tribolik]

Silly question but is it really Ian Bell?

Ive read in another forum that it is in fact Robert Kubica with a lot of free time on his hands... and he is pulling our leg.... Well at least thats what Niki Lauda wrote...

 

[spapadillion]

Hah!

 

[Devedander]

What I'm pointing out is if the PS3 is capable of better graphics than 'the other console',

Are you saying it is or are you asking if it is?

Remember the two consoles are different hardware and comparing them is like comparing significantly different cars... it's possible that both outperform the other in certain circumstances...

then the game should be bumped up to maximize those capabilities. The PC looks better overall for a reason.

A lot of that reason tends to come from video cards which in most modern PCs are better than what is in consoles by a significant amount.

There is no reason the 360 should have better graphics than the PS3.

Sure there is... it's different architecture. While it may not always result in better results it can easily be the reason for any particular instance.

I remember not that long ago people were assuring each other the PS3 would have the better graphics because it was the lead console on this product and it would be ported to 360... ultimately we see what we see very often, given the same task the 360 often outperforms the PS3.

Why is complicated but why anyone is surprised it's the case with Shift 2 is beyond me.

And if the Wii or PS2 or iPhone had versions of S2u, we would see graphics degradation where appropriate.

And so by the same logic the fact the PS3 version is lower quality is empirical evidence that in this situation the PS3 is the lesser console.

What you keep saying amounts to saying "we all know this car can towe bricks faster than that car so why doesn't it?"

Well the obvious answer is that it can't towe this particular load faster than the other car.

Empirical evidence trumps assumption.

GT5 is the benchmark for console racer graphics PERIOD. I would hope S2u can be tweaked to get closer to that benchmark. Optimizing the game at 1080p and 60 FPS would be the starting point.

Remember GT5 is rendered at 1280x1080 and not 1960x1080 and it's far from a solid 60FPS http://www.eurogamer.net/videos/gt5-performance-analysis-compilation?size=med

So don't forget that GT5 sacrifices in some area to accel in others (massive screen tearing, super jumpy shapdows and smoke issues etc) just like other games may sacrifice other things to meet other ojbectives.

 

[deftonesmx17]

Funny thing is that the city tracks in GT5 are still one of the best in its genre, the smoke/dust one of the most reallistic rendered, the shadows move in real time and affect all, and most of the effects are still good(glares at night, sparks, wind, tyremarks, fireworks, out of track particles, dirt tyres, etc).

Sure, seems they used a vastly superior HDR for these tracks...

http://www.youtube.com/watch?v=1Ggz0M8IYo4#t=1m3s
http://www.youtube.com/watch?v=2n6mb70KFSo#t=1m26s
http://www.youtube.com/watch?v=dJf_k0d_A-A&t=0m20s

The PS3 is not a PC and the gpu renders only a part of the final graphics, most effects are coded out of the gpu and computed by the CPU like in the PS2.

Then what is for you the graphical benchmark of the generation?...

Come on now, the smoke is one of GT5's worst effects.
Not all shadows move in real time, many are static. Dirt effects, really? The rally looks like complete garbage. Wind, what wind? Aside from predetermined waving flags, I'd like some proof of this wind you speak of.

It still looks overly bright an sterile. It is not subtle what so ever and is hardly close to realistic.


Seriously? The PS3 is nothing but a specially designed computer. The GPU hardly does only a part of the rendering...

As a whole, F1 2010 definitely looks better.
Sure, no console game can match GT5's premium model cars, but that is only one aspect and when looking at the whole package most other games do look overall better.

I'd imagine the results are skewed by your computer monitor having a much, much faster response time than your TV.

Not at all the case, at least for me. My Computer monitor runs at 60hz and has a response time of 4ms. My LG HDTV on the other hand runs at 120hz and has a response time of 2.4ms. So My TV has a faster refresh rate and response time.....and those with Plasma's are probably running at 600hz and next to zero response time.
When I played shift 2 on the 360, I could see the lag and feel it. However, on my PC I can hardly see any lag nor do I feel it.

 

[pgagoober]

Awesome news about the patch. If only S2U was the way it was supposed to be at launch I may not have had to go and buy this instead :

And I thought S2U would only cost me $60!!!! I presume SMS will compensate me, right Ian ??

Needless to say, my console has been turned off for the last few days as I've been "testing" my new toy 👍

Can't wait to get the patch installed....

 

[Spagetti69]

Have they given an estimated arrival time for the patch?

 

[Imari]

Have they given an estimated arrival time for the patch?

No, and they probably won't because too much of the process lies outside their control. The patches have to be approved by Sony before they go live.

 

[markao]

Seeing you active in this thread Ian, I firstly would like to thank you for the game, have had some issues, but in general enjoying the game and overall you and your colleagues should be proud of it.


Have two questions regarding FFB settings. As you have seen by now, creating something that feels slightly different results in people posting suggestions on settings all over the net. Having a reasonable understanding about the in-game settings, I still wonder what to use in my logitech profiler. So would it be possible for you guys to post some explanation on what is used (spring/damper/centre/etc.), or the settings used by Doug (Eero?).

Secondly, is there a reason why the controller xml file (.....\ControllerDefaults\PC) are locked this time, why they do not accept changes (always back to default). For example , I would love to alter the "cForceSettings Level="Full" " for example, if I'm not mistaken still adds all the "canned shaking", but right now I cannot do that.


TIA,
Mark

 

[paytonlow]

No, and they probably won't because too much of the process lies outside their control. The patches have to be approved by Sony before they go live.

No approval needed for PC, release it ASAP.

 

[Imari]

No approval needed for PC, release it ASAP.

I agree with you there, push it out on PC and get a headstart on feedback for problems for next patch. Unfortunately, I suspect they'll want to synchronise the releases.

 

[Le Mans 88]

Talking of patches it would be usefull if custom liveries could be shared as race replica liveries added a great deal to Forza 3.

 

[Zer0]

Come on now, the smoke is one of GT5's worst effects.
Not all shadows move in real time, many are static. Dirt effects, really? The rally looks like complete garbage. Wind, what wind? Aside from predetermined waving flags, I'd like some proof of this wind you speak of.

Show me a driving game with more advanced smoke/dust effects:
http://www.youtube.com/watch?v=vFqWITVJRiQ#t=0m40s
http://www.youtube.com/watch?v=6h5YEXDsDCY
http://www.youtube.com/watch?v=I0uA3pywwM4

Show me a driving game with better dynamic shadows/dynamic ambient lighting:
http://www.youtube.com/watch?v=w9piJ4adciM

Wind effect:
-Smoke/dust
http://www.youtube.com/watch?v=KFs9VsuYpyc&t=0m28s

-Grass
http://www.youtube.com/watch?v=mtGK_OrfsKY&t=2m17s
http://www.youtube.com/watch?v=mtGK_OrfsKY&t=2m35s
http://www.youtube.com/watch?v=mtGK_OrfsKY&t=3m01s
http://www.youtube.com/watch?v=mtGK_OrfsKY&t=4m02s
http://www.youtube.com/watch?v=rOyj2qjeK_c&hd=1#t=3m01s

Exactly how much you played the game?..

Seriously? The PS3 is nothing but a specially designed computer. The GPU hardly does only a part of the rendering...

As a whole, F1 2010 definitely looks better.
Sure, no console game can match GT5's premium model cars, but that is only one aspect and when looking at the whole package most other games do look overall better.

You can't pretend to know what a PS3 can do or not treating it like if was a PC just because you think that knows the GPU limits. Is a completely different beast, specially with optimized exclusive content.

An example with Killzone 2:
"Cell-powered pre-processing optimised the amount of geometry RSX actually had to process, allowing for richer, more detail-heavy environments, objects and characters. Post-processing effects work such as the camera and object-based motion blur was hived out to the SPU satellite processors - another example of how the CPU was utilised as a graphics co-processor, with the visual effects work distributed between the main processor and the GPU."

Killzone 3:
"the hardware-driven anti-aliasing from Killzone 2 is now gone in favour of an SPU-driven alternative, so it wouldn't be surprising at all if this change, combined with optimisations to the CPU workload, allow Guerrilla to produce more detailed geometry."

Only by the low memory most of its games would be impossible to run on a pc with a similar 256MB graphic memory/256MB ram.

I don't know what you see superior in F1 2010, there are some minor things but as a whole GT5 can do a lot of things better, you talk about lighting quality and in F1 the HDR is masked by filters, just look at the the screen vignette and the dull sun direct lighting. I doubt that you see much people thinking that the F1 game is more photorealistic or a graphical benchmark for future games.

 

[Imari]

Zero, you're cherry picking great examples of those effects. When it all comes together right, the smoke/dust/spray can look great, but often the edges are pixelated or the dodgy transparency ruins the effect.

There's similar things going on with ambient lighting, and everything else you mentioned. In the right conditions, they're spectacular. But on average, there's too many flaws for them to be anything more than OK. It's great when it all comes together and you get one of those breathtaking moments in game, but they're few and far between because there's always one effect that's not quite keeping up with the rest.

F1 2010 and Shift do a far better job of providing a consistently high level of graphical quality in my opinion. GT5 may be better at it's best, but it's much worse at it's worst and on average it falls behind for me.

 

[Zer0]

Zero, you're cherry picking great examples of those effects. When it all comes together right, the smoke/dust/spray can look great, but often the edges are pixelated or the dodgy transparency ruins the effect.

There's similar things going on with ambient lighting, and everything else you mentioned. In the right conditions, they're spectacular. But on average, there's too many flaws for them to be anything more than OK. It's great when it all comes together and you get one of those breathtaking moments in game, but they're few and far between because there's always one effect that's not quite keeping up with the rest.

F1 2010 and Shift do a far better job of providing a consistently high level of graphical quality in my opinion. GT5 may be better at it's best, but it's much worse at it's worst and on average it falls behind for me.

Yes but I'm not speaking of perfection but innovation(more polygons, more advanced effects, better lighting, more resolution, etc).

For example Crysis 2 has plenty of graphical faults but is a graphical benchmark, is pushing things that no game is pushing in consoles and of course at times is showing their limits. In the other side there are plenty of more consistent games doing less things wich are far from being graphical references.

Is easy to be consistent when you are doing less(GT5 Prologue, Forza 3, Shift, etc...) but GT5 is advancing things that you rarely will see again until the next generation, and that's to me a benchmark.

Shift 2 being directly compared to GT5 instead of F1 2010 or any other game is enought proof of what a graphical benchmark means. Forza 4 will follow the same route.

 

[UnionStrike]

So, will Shift 2 be patched in the near future?

 

[Imari]

Yes but I'm not speaking of perfection but innovation(more polygons, more advanced effects, better lighting, more resolution, etc).

For example Crysis 2 has plenty of graphical faults but is a graphical benchmark, is pushing things that no game is pushing in consoles and of course at times is showing their limits. In the other side there are plenty of more consistent games doing less things wich are far from being graphical references.

Is easy to be consistent when you are doing less(GT5 Prologue, Forza 3, Shift, etc...) but GT5 is advancing things that you rarely will see again until the next generation, and that's to me a benchmark.

Shift 2 being directly compared to GT5 instead of F1 2010 or any other game is enought proof of what a graphical benchmark means. Forza 4 will follow the same route.

Pushing numbers up isn't exactly innovation though. More resolution, more polys, more fps, that's real innovative. HOW they do it is innovative, and had they pulled it off cleanly it would have been. But they didn't. It's not in 1080p, it's in a bastardised half-resolution that is somewhere between 720p and 1080p. It's not 60fps, it's somewhere between 40 and 60fps. It's not 500,000 polys, it's 500,000 polys in certain rare situations where they can control the rest of the load on the console.

I see what you're saying, and GT5 in it's ideal state is the goal that developers are and should be reaching for. But the reality is that GT5 itself cannot produce those graphics itself consistently. And so compared to the benchmark (a series of samples of all the best that GT5 has to offer, a theoretically "ideal" GT5 version) even GT5 itself comes up short.

 

[spapadillion]

Not at all the case, at least for me. My Computer monitor runs at 60hz and has a response time of 4ms. My LG HDTV on the other hand runs at 120hz and has a response time of 2.4ms. So My TV has a faster refresh rate and response time.....and those with Plasma's are probably running at 600hz and next to zero response time.
When I played shift 2 on the 360, I could see the lag and feel it. However, on my PC I can hardly see any lag nor do I feel it.

Numbers aren't everything: Despite the fact that my 46" Samsung 6 630 has an advertised response time of 4ms and a refresh rate of 120Hz, and my 19" Syncmaster 906BW is 2ms and 60Hz, the difference in percievable lag between the two, even when viewing the same material, is absolutely staggering. And that's with the TV's game mode enabled. It is my experience that this is fairly common among televisions, as even the lowest post-processing setting still adds a discernable delay over that of true monitors.

While I can't speculate as to the response times of plasma TVs, as no one I know has one (LCD outsells plasma 10/1), or to whether or not the PC version of the game has less input lag than the console versions, as I've yet to try it, I stand firmly by my previous point.

 

[Imari]

Note that response time and the screen lag are different.

Response time measures how quickly the pixels in the screen can change colour, and is important for avoiding the ghosting effect on images with fast moving objects or sudden transitions of colour. It used to be a big problem with LCD screens, but is largely a non-issue with any semi-modern screen.

The lag effect is a byproduct of post-processing and/or the technology that the screen is based on, and is rarely advertised as part of the stat sheet on monitors unless they're gaming monitors. I've never seen it advertised for TVs, because it's simply not important for non-gaming uses which is what most of the population uses TVs for.

What you're seeing is the difference between the panel types. The Syncmaster is probably a TN panel type screen which gives exceptional response times at the cost of more limited viewing angles and more limited colour ranges. They also tend to be cheap and simple. Your Samsung is probably a more advanced type of panel that gives better viewing angles and colour suitable for a TV while sacrificing some of that lag, which is usually unimportant in TV viewing. The game mode will cut out any post-processing the TV does to enhance the image and speed it up some, but it's still limited by the panel type it's using.

P.S. There's absolutely no way you'd be able to tell the difference between a delay of 2ms and a delay of 4ms in a display, if it were to exist. That would the difference between 1 and 2 frames of delay on a 500fps output. There's the often-quoted-rarely-sourced thing about fighter pilots identifying stuff from a single frame flashed at 200fps. If it takes fighter pilot reactions to spot one frame out of two hundred, no way is anyone spotting one in 500. Sorry.

 

[JackC8]

The steering lock adjustment can give you slow, mushy steering or precise, responsive steering. The default setting is rather slow - turn it up a bit and it's like night and day.

 

[spapadillion]

Imari, I'm not sure what you're apologizing for as I agree with everything you've said there. There's no way in hell I could tell the difference between 2 and 4ms; which is why I brought post processing into the conversation. There has to be something going on, possibly related to the type of panel used (as you said), that's adding to the delay beyond the listed figures of 4ms and 120Hz. I appreciate the added clarity that you've brought to the issue.

 

[FuriousDemon]

Note that response time and the screen lag are different.

Response time measures how quickly the pixels in the screen can change colour, and is important for avoiding the ghosting effect on images with fast moving objects or sudden transitions of colour. It used to be a big problem with LCD screens, but is largely a non-issue with any semi-modern screen.

The lag effect is a byproduct of post-processing and/or the technology that the screen is based on, and is rarely advertised as part of the stat sheet on monitors unless they're gaming monitors. I've never seen it advertised for TVs, because it's simply not important for non-gaming uses which is what most of the population uses TVs for.

What you're seeing is the difference between the panel types. The Syncmaster is probably a TN panel type screen which gives exceptional response times at the cost of more limited viewing angles and more limited colour ranges. They also tend to be cheap and simple. Your Samsung is probably a more advanced type of panel that gives better viewing angles and colour suitable for a TV while sacrificing some of that lag, which is usually unimportant in TV viewing. The game mode will cut out any post-processing the TV does to enhance the image and speed it up some, but it's still limited by the panel type it's using.

P.S. There's absolutely no way you'd be able to tell the difference between a delay of 2ms and a delay of 4ms in a display, if it were to exist. That would the difference between 1 and 2 frames of delay on a 500fps output. There's the often-quoted-rarely-sourced thing about fighter pilots identifying stuff from a single frame flashed at 200fps. If it takes fighter pilot reactions to spot one frame out of two hundred, no way is anyone spotting one in 500. Sorry.

Nice post! That clarifies the issue for me. I forgot my TV had the gaming mode in it Will give it a try and see if it can match my monitor's response.

 

[Imari]

Imari, I'm not sure what you're apologizing for as I agree with everything you've said there. There's no way in hell I could tell the difference between 2 and 4ms; which is why I brought post processing into the conversation. There has to be something going on, possibly related to the type of panel used (as you said), that's adding to the delay beyond the listed figures of 4ms and 120Hz. I appreciate the added clarity that you've brought to the issue.

Sorry, that part was mostly for the benefit of deftones, who seemed to be more under the impression that the difference he was seeing was due to the response times. I was fairly sure you understood what was going on.

 

[Zer0]

Pushing numbers up isn't exactly innovation though. More resolution, more polys, more fps, that's real innovative. HOW they do it is innovative, and had they pulled it off cleanly it would have been. But they didn't. It's not in 1080p, it's in a bastardised half-resolution that is somewhere between 720p and 1080p. It's not 60fps, it's somewhere between 40 and 60fps. It's not 500,000 polys, it's 500,000 polys in certain rare situations where they can control the rest of the load on the console.

I see what you're saying, and GT5 in it's ideal state is the goal that developers are and should be reaching for. But the reality is that GT5 itself cannot produce those graphics itself consistently. And so compared to the benchmark (a series of samples of all the best that GT5 has to offer, a theoretically "ideal" GT5 version) even GT5 itself comes up short.

Is not 1920*1080 but is still more resolution than any other driving game in console, is not always 60fps but is still more than 30fps, is not 500k polys ingame but is still more than any other driving game, etc... is not about not being "this" or "that" is "what more is than?" and GT5 running, even not perfect, is without a doubt more than any other game.

The only effects that really are unbalanced are the lowres external shadows and the pixelated car borders in the smoke clouds. Oh and for some the standard cars.. but still most of the time(except closeups) they are looking better ingame than much of the competition(because the resolution, lighting and reflections) anyway they are not a technical limitation or a graphic flaw.

I think that some people overreact too much with the few flaws and tend to focus at them all the time to negate the rest of what GT5 is showcasing. When playing the game is clearly showing a lot more things good than bad.