Author Topic: is HTC capable of the level of detail of IL2FB?  (Read 2519 times)

Offline Citabria

  • Platinum Member
  • ******
  • Posts: 5149
is HTC capable of the level of detail of IL2FB?
« Reply #30 on: November 28, 2003, 10:38:30 AM »
mold offline boxed single player games have a similar problem to ai games in terms of numbers. for each ai drone in the game requires hundreds upon hundreds more calculations the more detailed the ai the more processing required.

with online games the human brain does all the ai drone processing and the game keeps track of damage position azimuth and speed without such a complex need for detailed ai algorithms.
Fester was my in game name until September 2013

Offline Pongo

  • Platinum Member
  • ******
  • Posts: 6701
is HTC capable of the level of detail of IL2FB?
« Reply #31 on: November 28, 2003, 05:03:33 PM »
In answer to your original question.
No, not with the current business model.

Offline AKS\/\/ulfe

  • Platinum Member
  • ******
  • Posts: 4287
is HTC capable of the level of detail of IL2FB?
« Reply #32 on: November 29, 2003, 01:37:19 AM »
Quote
Originally posted by Citabria
mold offline boxed single player games have a similar problem to ai games in terms of numbers. for each ai drone in the game requires hundreds upon hundreds more calculations the more detailed the ai the more processing required.

with online games the human brain does all the ai drone processing and the game keeps track of damage position azimuth and speed without such a complex need for detailed ai algorithms.


Yes, but there is still a need for far more complex communications processes.
-SW

Offline mold

  • Copper Member
  • **
  • Posts: 305
is HTC capable of the level of detail of IL2FB?
« Reply #33 on: November 29, 2003, 09:13:07 AM »
Quote
Originally posted by Citabria
mold offline boxed single player games have a similar problem to ai games in terms of numbers. for each ai drone in the game requires hundreds upon hundreds more calculations the more detailed the ai the more processing required.


Good point!  Yeah, actually the MMP sim should have less computational load per client.  The damage calculations are distributed.


Quote
Originally posted by AKS\/\/ulfe
Yes, but there is still a need for far more complex communications processes.
-SW


I disagree.  There may be need for higher bandwidth, but not much more complexity than the current mechanisms.  Even bandwidth shouldn't be that much more.  Now we presumably have messages like "wingtip got 500 pts damage".  In the new model, we'd have "rib_1 for 200 pts,rib_2 for 100 pts, rib_5 for 200 pts".

Offline wklink

  • Nickel Member
  • ***
  • Posts: 356
      • http://www.simhq.com
is HTC capable of the level of detail of IL2FB?
« Reply #34 on: November 30, 2003, 12:17:03 PM »
That kind of fidelity is probably possible but probably not within the reasonable download footprint.  I also wonder how many people would want to upgrade to a PIV 2.5 to run the game.  

It is something of a fine edge sword with online video games.  An update for IL2FB with new aircraft requires at times a 75-110 meg download, larger than the entire code for AHII right now.  I am not sure how many of us are willing to wait for that kind of update.  I can imagine how long new terrains would take to download.

It's not a bad idea but I think we all would need broadband to both play it and to download it.  I don't want to alienate the non Pentium IV gamers out there.  AHII pushes that barrier now, to have IL2 as a Massive Multiplayer would probably take them out of the picture.
The artist formerly known as Tom 'Wklink' Cofield

Offline mold

  • Copper Member
  • **
  • Posts: 305
is HTC capable of the level of detail of IL2FB?
« Reply #35 on: November 30, 2003, 10:17:40 PM »
Quote
Originally posted by wklink
That kind of fidelity is probably possible but probably not within the reasonable download footprint.  I also wonder how many people would want to upgrade to a PIV 2.5 to run the game.


Within a year a 2.4 P4 will be so stale that you can use it to kill birds in flight.    Even now a 2.4C is only about $160.  I'm sure that soon enough the local PC store will have 2.4 P4's or 2500XP Athlons in their $300 machines.

Time moves on.  Computers get faster and cheaper.  Games should get better too.  There's a ton of room for improvement.   Why stop now, after all these years of forward progress?  The same thing could've been said 10 years ago, and then where would we be...

Quote
Originally posted by wklink
It is something of a fine edge sword with online video games.  An update for IL2FB with new aircraft requires at times a 75-110 meg download, larger than the entire code for AHII right now.  I am not sure how many of us are willing to wait for that kind of update.  I can imagine how long new terrains would take to download.


Even 35 megs is painfully slow on a modem.  100 megs shouldn't be all that much more pain, due to diminishing marginal effects. :D  Besides...is there anyone reading the thread that wouldn't download 600 megs or more for AH2?  Yeah, there are others out there...but my guess is that anyone who is into AH either as a game or as a plane buff will be quite willing to download anything of any size to play the game.

As far as in-game bandwidth requirements...it would take a bit more, but I don't think it would require broadband. Who knows, until it's tried with an optimized protocol.

Net connections aren't as Moore's Law-ish as computers, but fear not--cheap ethernet to the home is coming soon.  I hope.  Then perhaps will come the death of stone age voice-band modems.

Offline SunKing

  • Gold Member
  • *****
  • Posts: 3726
is HTC capable of the level of detail of IL2FB?
« Reply #36 on: November 30, 2003, 11:41:26 PM »
Quote
Originally posted by 214thCavalier
They did in fact use to do that but think about it, it doubles the work load on a small dev team.



but the up side is if they did put the time in for a his res version.. they would win over all those IL-2 flyers that fly that sim due to graphics. I wish they would consider it.

Offline gatt

  • Gold Member
  • *****
  • Posts: 2441
is HTC capable of the level of detail of IL2FB?
« Reply #37 on: December 01, 2003, 10:39:49 AM »
Mold, IMHO probably the P4 2.4 wont be the average CPU in the next season, but I'm worried about the P3-1Ghz + GeForce2 standard chosen by HTC months ago .... and AH2 looks still .... alpha?
"And one of the finest aircraft I ever flew was the Macchi C.205. Oh, beautiful. And here you had the perfect combination of italian styling and german engineering .... it really was a delight to fly ... and we did tests on it and were most impressed." - Captain Eric Brown

Offline AKS\/\/ulfe

  • Platinum Member
  • ******
  • Posts: 4287
is HTC capable of the level of detail of IL2FB?
« Reply #38 on: December 01, 2003, 12:54:10 PM »
Quote
Originally posted by mold
I disagree.  There may be need for higher bandwidth, but not much more complexity than the current mechanisms.  Even bandwidth shouldn't be that much more.  Now we presumably have messages like "wingtip got 500 pts damage".  In the new model, we'd have "rib_1 for 200 pts,rib_2 for 100 pts, rib_5 for 200 pts".


I didn't say bandwidth, I said "processes". The AI takes up quite a few CPU cycles to compute what it is going to do.

The Networking side of things, which does all of the communication interpretation and packaging, will take up more CPU cycles than the AI will.
-SW

Offline mold

  • Copper Member
  • **
  • Posts: 305
is HTC capable of the level of detail of IL2FB?
« Reply #39 on: December 01, 2003, 01:29:34 PM »
Quote
Originally posted by AKS\/\/ulfe
I didn't say bandwidth, I said "processes". The AI takes up quite a few CPU cycles to compute what it is going to do.


Thought we were talking about damage models, not AI...  Yes, the collision detection will take more CPU power with finer granularity, but if IL2 can do it so can AH2.  IL2 has to collide hundreds of planes, while a single AH2 client has to collide only a couple planes.

Quote
Originally posted by AKS\/\/ulfe
The Networking side of things, which does all of the communication interpretation and packaging, will take up more CPU cycles than the AI will.


How do you figure?

Offline AKS\/\/ulfe

  • Platinum Member
  • ******
  • Posts: 4287
is HTC capable of the level of detail of IL2FB?
« Reply #40 on: December 01, 2003, 01:35:03 PM »
Quote
Originally posted by mold
Thought we were talking about damage models, not AI...  Yes, the collision detection will take more CPU power with finer granularity, but if IL2 can do it so can AH2.  IL2 has to collide hundreds of planes, while a single AH2 client has to collide only a couple planes.


You should probably go back and follow what I was replying to, then read my reply, then read your reply. Cit said that AI calculations do not need to be done, and I said thats true - but the far more complex networking side needs to be done.

Quote
Originally posted by mold
How do you figure?


Because the data communications takes up a lot more cycles than the AI does. Theres a lot more to data transmission than just sending out a simple speed and heading packet, or what is damaged packet, or recieving one.

AI goes by a fixed set of simple rules and use an extremely simplified flight model which doesn't require many CPU computations at all.
-SW

Offline mold

  • Copper Member
  • **
  • Posts: 305
is HTC capable of the level of detail of IL2FB?
« Reply #41 on: December 01, 2003, 02:00:48 PM »
Quote
Originally posted by AKS\/\/ulfe
You should probably go back and follow what I was replying to, then read my reply, then read your reply. Cit said that AI calculations do not need to be done, and I said thats true - but the far more complex networking side needs to be done.


Oh ok.  Slight confusion there.  Anyway, even if you leave AI out of the picture, AH2 should be able to do what IL2 is doing, for the reason I stated above.

 
Quote
Originally posted by AKS\/\/ulfe
Because the data communications takes up a lot more cycles than the AI does. Theres a lot more to data transmission than just sending out a simple speed and heading packet, or what is damaged packet, or recieving one.


Well, I guess we really can't tell unless we know for sure what the AI is doing.  Anyway, leaving aside AI, I think the additional communication requirements are very doable from a CPU load perspective.  It is the additional required bandwidth that might be a question, but even that I deem achievable at modem speeds.

Offline AKS\/\/ulfe

  • Platinum Member
  • ******
  • Posts: 4287
is HTC capable of the level of detail of IL2FB?
« Reply #42 on: December 01, 2003, 02:20:52 PM »
What Il2:FB is doing right now with it's damage model is what AH does - just with additional incremental steps between no damage and full damage, with varying degrees of flight performance being effected.

This is doable on any current PC, works quite well on my machine with full graphics - online or offline against AI.

Now what they plan on doing in BoB is up in the air, all we have seen are screenshots of the underlying structure of the plane's being made. This is not the underlying damage model (as to how a plane gets damaged, which is never seen), but instead the visual model - so we can atleast assume that there will be a skin that will be slowly holed up or blown away. Beyond that, its anyone's guess. This would lead to a lot of data communications, rather than rib1 destroyed - we now have 1x1cm hole input into area 56x 22y, and multiply that by the number of times bullets enter the skin. This would lead to a lot of data being transferred on a lowly 32player game, let alone several hundreds.

This is without even figuring in the actual computations to derive the visual model, so not only do complex calculations for the damage model need to done - now a lot more stuff is being put into packets, or atleast being transferred in multiple packets. This now requires more CPU cycles to recieve, decode, place on the plane, and so on for each plane that is being visually modelled. Then a group of planes from somewhere else on the map become visually modelled, (for sake of argument are damaged), all of their damage has to be displayed and thus communicated... and now your system has to go recieve all of this while still maintaining the usual communications processing it typically does (like what the current AH does).

I doubt how much data can be transferred at once would really be much of a factor, so long as everything can be packaged into small packets and transferred at high rates.

The real problem is with all of these graphical features will occupy a great deal of the computers work load, and then adding in the networking aspect of something this detailed for a scale this large (400 players) would require the state of art computer of two years from now to run it well.

In the case of the current FB, offline I can have combat of up to 32 AI planes on ace level with lots of AI ground object scattered around and achieve the same framerates as playing online with only 16 people with some AI ground objects.

Theres just so much being communicated that the CPU spends the majority of the time sending/recieving/updating locations, object damage/status and each player's aircraft state that something has to give - and thats typically graphics (or sound depending on the sound engine).

Keep in mind that BoB is designed to run on the top of the line computers when it is released in 2005.
-SW

Offline GODO

  • Nickel Member
  • ***
  • Posts: 555
      • http://www.terra.es/personal2/matias.s/fw190.htm
is HTC capable of the level of detail of IL2FB?
« Reply #43 on: December 01, 2003, 02:29:27 PM »
With a very complex damage model, bandwith would be affected noticeabily. When you see a plane, you see the visible damage, so, that information is being sent and distributed by the server. Our actual visible damages are:
2 wing roots.
2 wing tips.
2 ailerons.
2 flaps.
2 elevators.
up tp 2 rudders.
oil x n engines (up to 4).
fuel.
radiator x n engines (up to 4).
fire.
up to 3 gears.
tail section.

26 bits, a single 32 bit word seems enough to transmit our actual visible damage for any plane. Adding much more visible damage will imply the transmission of much more data with the corresponding impact on communications.

Offline mold

  • Copper Member
  • **
  • Posts: 305
is HTC capable of the level of detail of IL2FB?
« Reply #44 on: December 01, 2003, 02:54:20 PM »
Quote
Originally posted by AKS\/\/ulfe
Now what they plan on doing in BoB is up in the air, all we have seen are screenshots of the underlying structure of the plane's being made. This is not the underlying damage model (as to how a plane gets damaged, which is never seen), but instead the visual model - so we can atleast assume that there will be a skin that will be slowly holed up or blown away. Beyond that, its anyone's guess. This would lead to a lot of data communications, rather than rib1 destroyed - we now have 1x1cm hole input into area 56x 22y, and multiply that by the number of times bullets enter the skin. This would lead to a lot of data being transferred on a lowly 32player game, let alone several hundreds.


Yes, I agree.  This is too much for modems, and possibly even cable modems until backbones get upgraded. :)  I was speaking rather of the non-visual damage model.  When Fester posted his pic, I immediately thought of how individual bullets and shells might damage that structure--not how the structure would look in the game or how bullet holes would look.  Actually, see the original post. It ain't about the graphics.

Without the eye candy, the communication and CPU requirements are lowered.

Quote
Originally posted by AKS\/\/ulfe
The real problem is with all of these graphical features will occupy a great deal of the computers work load, and then adding in the networking aspect of something this detailed for a scale this large (400 players) would require the state of art computer of two years from now to run it well.


Perhaps.  When the time comes, the graphics can be filled in.  Furthermore, I believe you underestimate the power of current processors.  I think some of this can be done now.  Especially if it is appropriately scoped and filtered, so that you aren't creating in real-time polygon meshes of torn sheetmetal for planes that are 10 miles away.

Quote
Originally posted by AKS\/\/ulfe
Theres just so much being communicated that the CPU spends the majority of the time sending/recieving/updating locations, object damage/status and each player's aircraft state that something has to give.


I simply do not think that is true.  I have done some multiplayer programming before, and the CPU is not spending the majority of its time in these tasks.  Communication is filtered by the server, so that each client is being updated only with the closeby planes with any great frequency.  Furthermore, even full 400-player updates would not be so bad on the CPU--it is the network that would suffer from this.  Hell, that is exactly what a boxed sim does, and boxed sims have been doing this for a long time!  I remember this boxed game, Out of the Sun--tons of AIs, and no problem updating their positions and states.