• Hey Guest. Check out your NeoGAF Wrapped 2025 results here!

Forza Motorsport 5 announced

Sorry but cloud gaming as it offload part of data to clod and send it back to client is simply not possible yet if we want games to be responsive.

Most of the people don't even have less than 70-100 ms ping. And now consider 60 fps game....

(Snip)

Sorry but cloud computing is simply right now a bullcrap. Cloud computing for games will happen in future but that will only happen if most of people will have 20mb/s with 1-5 ms ping.

This is where you find out who has zero imagination.

I'd apply the cloud, in Forza 5 to Weather and the Crowd.

The weather data would simply update F5 on what's happening, in real time, at that location that day. 60fps doesn't matter...neither does ping. The games weather isn't something that needs split second reaction across the track.

With crowds I'd have their physical reaction and AI (like moving around the track offloaded). You could have scripted crowds for offline play, and more dynamic ones for when the system is online. Heck, this could be applied to dynamic wild life too....just for fun.

There are many things that don't have to do with direct gameplay that could be expanded for immersion.

Anyway, don't be so limited in your vision.
 
Nah, he'll save that for the long demo on Gamespot... :-D
There should be a bingo before E3 about Dan's favourite words and sentences !
 
That's one thing I'd like to see : game pulling real-time weather data from the track locations and using the real time of day as well, MSR style. This could open up so many possibilities.
I've been wondering as well, could they use cloud computation to make better replays or at least make better videos out of replays ? I'd love to have the possibility to use cloud computing to basically render a replay with greater AA, effects and maybe even more advanced lighting like ray-tracing. (kinda like what PD and Turn 10 do for there trailers). The output would be videos but even then, they could put an option to stream those directly from the console.
 
I really do wonder what the next generation holds for 'Greenawaltisms'..

Surely he's over the car lovers into gamers mantra by now?
 
I really do wonder what the next generation holds for 'Greenawaltisms'..

Surely he's over the car lovers into gamers mantra by now?

It will be like this:

"When we started Forza our vision was to turn car lovers into gamers and gamers into car lovers. But now through the magic of Xbox One and the power of the cloud, we're able to immerse players into a deeper and richer experience too where you now have a relationship with your car."........
 
Ideas for cloud offloading:
- much more AI cars on the track and better AI - that is obvious one
- changable weather calculation - calculate changing grip on track surface - for example as track is drying calculate grip and drying based on how where cars are passing, where there is no shadow, etc. Much better than prebaked dry lines or same grip on whole width of the track
- calculate damage (collision) consequences on car handling - so if I hit front left make the car handle differently.
- maybe the same for visual damage (how parts of the car bend)
 
Ideas for cloud offloading:
- much more AI cars on the track and better AI - that is obvious one
- changable weather calculation - calculate changing grip on track surface - for example as track is drying calculate grip and drying based on how where cars are passing, where there is no shadow, etc. Much better than prebaked dry lines or same grip on whole width of the track
- calculate damage (collision) consequences on car handling - so if I hit front left make the car handle differently.
- maybe the same for visual damage (how parts of the car bend)

Oh for the love of...
 
Ideas for cloud offloading:
- much more AI cars on the track and better AI - that is obvious one
- changable weather calculation - calculate changing grip on track surface - for example as track is drying calculate grip and drying based on how where cars are passing, where there is no shadow, etc. Much better than prebaked dry lines or same grip on whole width of the track
- calculate damage (collision) consequences on car handling - so if I hit front left make the car handle differently.
- maybe the same for visual damage (how parts of the car bend)

Unfortunately all above can't be done on cloud in game such as forza.
 
Unfortunately all above can't be done on cloud in game such as forza.

why not? all car physics are calculated locally, but server feeds new parameters. This stuff can be late

EDIt I dont say it will be needed, maybe all this doesnt take much of resources to do it locally, but only T10 knows

Personally I am more excited for possible dedicated servers. No more flying cars....
 
I think Dan gave some examples of things that can be done in the cloud when he was talking in the architecture panel.
 
That's one thing I'd like to see : game pulling real-time weather data from the track locations and using the real time of day as well, MSR style. This could open up so many possibilities.
I've been wondering as well, could they use cloud computation to make better replays or at least make better videos out of replays ? I'd love to have the possibility to use cloud computing to basically render a replay with greater AA, effects and maybe even more advanced lighting like ray-tracing. (kinda like what PD and Turn 10 do for there trailers). The output would be videos but even then, they could put an option to stream those directly from the console.
I would be happy just with the videos and bigshots being rendered in the cloud, so you dont need to wait the rendering/upload.
 
Sorry but cloud gaming as it offload part of data to clod and send it back to client is simply not possible yet if we want games to be responsive.

Most of the people don't even have less than 70-100 ms ping. And now consider 60 fps game....

yeah. This is why most of MMO games are just crappy non action games because action games have terrible problems with ping people have. And that is just for simple calculations ! Like positioning or calculating damage ! And now in some way MS created something out of nowhere that will fill the gap.

Sorry but cloud computing is simply right now a bullcrap. Cloud computing for games will happen in future but that will only happen if most of people will have 20mb/s with 1-5 ms ping.

Sorry but you're wrong, they've already made it clear that cloud computing is going to be used on non time sensitive computations. For instance you start doing a burnout and the Xbox One creates the smoke for 1-2 seconds and than the cloud picks up and can keep it going much beyond what a Xbox One could actually do.

Same with spectators, volumetric fog or cloud and so on.

Unfortunately all above can't be done on cloud in game such as forza.

As I agree that most of the things he said wouldn't work but weather would work most definitely as it's something that is out of your control anyway.

It is definitely not bullshit as you would say it.
 
Sorry but you're wrong, they've already made it clear that cloud computing is going to be used on non-sensitive computations. For instance you start doing a burnout and the Xbox One creates the smoke for 1-2 seconds and than the cloud picks up and can keep it going much beyond what a Xbox One could actually do.

Same with spectators, volumetric fog or cloud and so on.

It is definitely not bullshit as you would say it.

Turn 10 has commented on how they will be using cloud compute? Dan was really adamant about it's possibilities during the architecture panel. But he also sounded like he was just towing the company line.
 
Turn 10 has commented on how they will be using cloud compute? Dan was really adamant about it's possibilities during the architecture panel. But he also sounded like he was just towing the company line.

What employee doesn't tow the company line?

"Turning car lovers into gamers and Gamestop employees into McDonald's workers."

My friend used to work at Gamestop it's not much of a step down sadly...
 
Well, he sounded more so than the other guys, but either way, have they said any specifics about their implementations? It sounds like they have based on your post. Just curious.

Dan mentioned some things in the panel but no I can't confirm anything I've said I'm just giving examples based on the information we know.

Any calculations that don't require time sensitivity can be sent to the cloud. I think the biggest difference we'll see when it comes to offline play will be a LoD change and less effects and missing effects entirely(like the fog on the track in the trailer.).

I think I'm going to turn the stream off, if he mentions "turning gamers into car lovers, car lovers into gamers".

LOL, you might as well not even watch it.
 
That's one thing I'd like to see : game pulling real-time weather data from the track locations and using the real time of day as well, MSR style. This could open up so many possibilities.
I've been wondering as well, could they use cloud computation to make better replays or at least make better videos out of replays ? I'd love to have the possibility to use cloud computing to basically render a replay with greater AA, effects and maybe even more advanced lighting like ray-tracing. (kinda like what PD and Turn 10 do for there trailers). The output would be videos but even then, they could put an option to stream those directly from the console.

Yeah, the cloud isn't needed for that sort of weather data.

The high quality replay rendering sounds awesome (I thought about it a lot), but with the "DVR" feature, I'm not hopeful.
 
long time forza fan here guys been playing forza since part 1 cant wait to see beardy on e3 with the maclaren on stage as he walks through the smoke
288142.jpg
 
long time forza fan here guys been playing forza since part 1 cant wait to see beardy on e3 with the maclaren on stage as he walks through the smoke
Yep. It seems the McLaren will take centre stage this year. Fuck man, can't wait too see next gen Forza. I'm dying here!
 
I'm also pretty sceptical of the 'cloud' claims, we all know the extent to the which the cloud has influenced things so far, which is bugger all, so we are going to be sceptical until shown something.

However, I can envisage a few simple use cases..

Take F1 2011 for example (the last one I bought), there is a massive disjoint between the AI times immediately around you, and the rest of the cars in the timing sheets.. Clearly the majority of non visible AI are on some scripted time line, not the same physics based laptimes as the player is doing. This is exactly where the cloud could help, it could be doing the AI calcs for 50+ cars, and ensuring that you get a 'fair' race so to speak.. I could imagine the Nurb24H being perfectly recreated, the cloud is running effectively the same thing as a MP server, telling the local computer where every single AI car is, they'd all react to you as you where near them.. And those near you would be the only ones that the AI is calculated locally (where latency is key), as soon as you overtake and they are out of your 'zone', the cloud AI takes over for them.. you can run extensive tyre/fuel wear and AI could react to being damaged, it'd be very accurate..

Then there is the time of day.. to get really good performance, the tracks have to be use essentially pre-created textures with baked in lighting (all the main shadows are preset), then only the cars shadows are locally calculated.. the cloud could easily be pre-baking the track textures as time changes, the game only needs that slowly downloading in the background so that it's ready for the next lap (or so, time doesn't change too quickly). If the internet goes down for a minute or so, you just get stuck with whatever prebaked track textures you last got..

I fully admit that I expect to see nothing of significance at all at launch, but it seems easy IMO to think of good uses of latency insensitive loads (that's this years incubation by the way)..
 
Ideas for cloud offloading:
- much more AI cars on the track and better AI - that is obvious one
- changable weather calculation - calculate changing grip on track surface - for example as track is drying calculate grip and drying based on how where cars are passing, where there is no shadow, etc. Much better than prebaked dry lines or same grip on whole width of the track
- calculate damage (collision) consequences on car handling - so if I hit front left make the car handle differently.
- maybe the same for visual damage (how parts of the car bend)

Lol you drank too much of that MS PR Koolaid.
 
You should explain to him or her why that stuff can't be done in the cloud.

Ideas for cloud offloading:
- much more AI cars on the track and better AI - that is obvious one
- changable weather calculation - calculate changing grip on track surface - for example as track is drying calculate grip and drying based on how where cars are passing, where there is no shadow, etc. Much better than prebaked dry lines or same grip on whole width of the track
- calculate damage (collision) consequences on car handling - so if I hit front left make the car handle differently.
- maybe the same for visual damage (how parts of the car bend)

1. No, AI cars need to be able to react to the player instantly and cloud would take too long.
2. Yes, this wouldn't be that time sensitive.
3. No, these affects need to be instant.
4. No, these affects need to be instant.
 
You should explain to him or her why that stuff can't be done in the cloud.

It's time sensitive. This is a game and franchise that is based on 60fps.
If you start to offload real time calculations to something that can take potentially quarter of a second, that's 7 to 8 frames missed.

That's a huge time disparity.

Our unpaid MS intern FordGT is trying to sell the idea to other posters by saying that we can get things like Volumetric Fog...

Ooooo sooo gooood right? Because THAT matters in a racing game. Their best bet is course creation with the cloud being able to create a full "24 hour" cycle and populating the track, etc.

EDIT: At least he realizes what's time sensitive and what's not.
EDIT 2: Now I feel bad, because he answered that accurately, but I still think he has too much hope in "the cloud."
 
It's time sensitive. This is a game and franchise that is based on 60fps.
If you start to offload real time calculations to something that can take potentially quarter of a second, that's 7 to 8 frames missed.

That's a huge time disparity.

Our unpaid MS intern FordGT is trying to sell the idea to other posters by saying that we can get things like Volumetric Fog...

Ooooo sooo gooood right? Because THAT matters in a racing game. Their best bet is course creation with the cloud being able to create a full "24 hour" cycle and populating the track, etc.

EDIT: At least he realizes what's time sensitive and what's not.

So you call me out for being a fanboy yet you don't even read my posts...

Not surprising, no wonder I have you ignored, as I've said from the begining that this will be time sensitive.

Volumetric effects can be very important in a racing game, smoke from a burnout for instance. Of course this is only the beginning there are a lot of non-time sensitive effects in a racing game that can be off loaded onto servers.

Better a unpaid intern than a unpaid troll.
 
The only realistic thing suggested here, that wouldn't be a nightmare to implement and that wouldn't ruin more than it would contribute in a game like this - is server-side rendered replays. Microsoft has in fact already done something like this in one of their Windows Mobile games called Galaxy Reign. All the battle scenes (which are mostly non-interactive, except you get to choose the camera angle) are rendered server-side.

Everything else that I can think of (including examples like fog or clouds) would be so unreliable and would require so much overhead and implementation nightmare that I can't see why anyone would bother. Just think about it - you have to have a locally rendered "backup" version of whatever you're planning to calculate from the cloud. Let's say it's some fog clouds done through fluid dynamic, and you can render maybe 500K particles locally for a lower quality effect. Then the cloud computation comes in and gives you maybe three times as many calculated particle positions for you to render. Will you now run into a problem of transparency overdraw, completely killing your performance? You don't even know when those particle positions will come in or what the rest of the scene will look like at that moment so that you should render them or not. It's just one of the few very obvious problems that you'd see when dealing even with the ideal case application of something like this.

I can however see something like this working in games that have no other option but to be always online, in which case you don't deal with balancing between locally rendered effects and effects rendered through this. You deal with absolutes then, and you know that you will have to render whatever data comes back from this cloud computation, and nothing else matters.


Their best bet is course creation with the cloud being able to create a full "24 hour" cycle and populating the track, etc.
Or this, yes, something that keep persistency of the world MMORPG style would be useful. However, Azure cloud is supposed to be computation cloud, not just some persistent storage that anything can do already. It is something I see them selling to research institutions and such, where it's be far more usable and useful than in real time gaming.
 
The only realistic thing suggested here, that wouldn't be a nightmare to implement and that wouldn't ruin more than it would contribute in a game like this - is server-side rendered replays. Microsoft has in fact already done something like this in one of their Windows Mobile games called Galaxy Reign. All the battle scenes (which are mostly non-interactive, except you get to choose the camera angle) are rendered server-side.

Everything else that I can think of (including examples like fog or clouds) would be so unreliable and would require so much overhead and implementation nightmare that I can't see why anyone would bother. Just think about it - you have to have a locally rendered "backup" version of whatever you're planning to calculate from the cloud. Let's say it's some fog clouds done through fluid dynamic, and you can render maybe 500K particles locally for a lower quality effect. Then the cloud computation comes in and gives you maybe three times as many calculated particle positions for you to render. Will you now run into a problem of transparency overdraw, completely killing your performance? You don't even know when those particle positions will come in or what the rest of the scene will look like at that moment so that you should render them or not. It's just one of the few very obvious problems that you'd see when dealing even with the ideal case application of something like this.

I can however see something like this working in games that have no other option but to be always online, in which case you don't deal with balancing between locally rendered effects and effects rendered through this. You deal with absolutes then, and you know that you will have to render whatever data comes back from this cloud computation, and nothing else matters.

Sorry but no...

Non latency sensitive computations are definitely possible no matter how many times you want to say they aren't.

As far as reliability is concerned I'm going to guess that's what the 300,000 servers are for.

Or this, yes, something that keep persistency of the world MMORPG style would be useful. However, Azure cloud is supposed to be computation cloud not just some persistent storage that anything can do already. It something I see them selling to research institutions and such, where it's be far more usable and useful than in real time gaming.

This isn't Azure cloud, in fact the servers are purely Xbox Live.
 
So you call me out for being a fanboy yet you don't even read my posts...

Not surprising, no wonder I have you ignored, as I've said from the begining that this will be time sensitive.

Volumetric effects can be very important in a racing game, smoke from a burnout for instance. Of course this is only the beginning there are a lot of non-time sensitive effects in a racing game that can be off loaded onto servers.

Better a unpaid intern than a unpaid troll.

I apologize. Why do we need volumetric particles? Seems like a waste to me. NFS games have always done their burnout smoke extremely well and they did it on PS360 (granted in 30fps).
 
I apologize. Why do we need volumetric particles? Seems like a waste to me. NFS games have always done their burnout smoke extremely well and they did it on PS360 (granted in 30fps).

This a game, it's a luxury, when you start asking "Why do we need XXXX?" you're already missing the point of gaming.

Why do we need 1200 cars in GT6? Why do we need Nurburgring in Forza?

So what if NFS has done smoke well are you saying we should just stop there and use their technique forever?
 
1. No, AI cars need to be able to react to the player instantly and cloud would take too long.

Here's a post I made on B3D that I think might be viable for racing games:

Another use for cloud computing that might be valuable is more complicated/reflective AI, where the cloud would calculate new sets of AI behavior guidelines while the local cpu would just run down an extensive decision tree provided by the results of that calculation.

For example in Forza after every lap perhaps the behavior of all computer agents and the player is pushed to the cloud. The behavior of every AI and Player can then be calculated into a new overall strategy individually by each AI agent according to what they observed on the road. Say Agent 4 saw the player go wide through left hand turns but were tight through right hand turns, then the AI would know where a players weaknesses are and adapt their strategy to overtake the player or other AIs based on this updated version of their overall strategy rather than just simply attempting to an adhere to a best line method of computerized racing. No idea if this is really all that computationally difficult or being done already on local processing, but really only the moment to moment decisions of the agents need to be on the console, the overall strategy can be continuously recalculated in the cloud to make itself more intelligent.

It would be like having the race team and pit crew telling you what they saw on the track
 
Sorry but no...

Non latency sensitive computations are definitely possible no matter how many times you want to say they aren't.
Please read what I said more carefully. I've given you an exact elaboration of the scenario you suggested and a very obvious problem it could introduce (offloaded compute clogging the rendering pipeline in an unknown moment)
 
I can't wait until Dan comes out and shows what he's been doing on the cloud compute stuff for Forza 5 and people can stop saying it's bullshit...


...oh wait, they won't stop, it will just be another goal post


Edit:
read some other forums and you will find plenty of people with actual hands on and technical knowledge and most are saying of course it's possible it's just how it will e implemented.

so those saying it can't be done are flat out wrong, it's not even an opinion, it's just a fallacy to say this won't work.

will it work system wide and full scale and will devs take advantage? are people ready for it and their connection?

Those are fair questions but as to the technical viability it is without question POSSIBLE
 
I apologize. Why do we need volumetric particles? Seems like a waste to me. NFS games have always done their burnout smoke extremely well and they did it on PS360 (granted in 30fps).

Volumetric particles are quite important for racing games. Stuff like smoke, spray from tires and dirt but you don't need cloud to achieve any of these.
 
This a game, it's a luxury, when you start asking "Why do we need XXXX?" you're already missing the point of gaming.

Why do we need 1200 cars in GT6? Why do we need Nurburgring in Forza?

So what if NFS has done smoke well are you saying we should just stop there and use their technique forever?

Smoke is a passive thing. It isn't a car count or tracks. It can indicate the heat of the tires, but you can already do that with plenty of sprites for much cheaper.
 
Please read what I said more carefully. I've given you an exact elaboration of the scenario you suggested and a very obvious problem it could introduce (offloaded compute clogging the rendering pipeline in an unknown moment)

I'm pretty sure Microsoft has it set up to deal with the rendering and the rendering pipeline is the last "pipeline" we need to be worried about in this scenario.

Smoke is a passive thing. It isn't a car count or tracks. It can indicate the heat of the tires, but you can already do that with plenty of sprites for much cheaper.

What's to say they couldn't do the same with sprites? Also Volumetric effects are a huge step up from simple sprites anyway.

Smoke is not indicative of tire heat, you can not tell how hot tires are based on smoke.

So? It's a luxury nonetheless in the same way a bunch of cars and tracks are luxuries?
 
Sorry but you're wrong, they've already made it clear that cloud computing is going to be used on non time sensitive computations. For instance you start doing a burnout and the Xbox One creates the smoke for 1-2 seconds and than the cloud picks up and can keep it going much beyond what a Xbox One could actually do.

Same with spectators, volumetric fog or cloud and so on.

LOL.
 
While we are bored waiting for the reveal why don't we talk about the new logo? Looks much Much better than before guys.

Dat fixed M in the F'M' logo
 
Top Bottom