CM article - Graphics

[The Commissar]

Originally posted by Gunny Bunny:

Here is an interesting occurence in the AI I am experiencing:

It is dark with 25 meters of visibility. My choice was made by the AI and it selected three mortar carriers. This is a low point game, so allot of armour points went into these.

The problem is you need at least 100 meters minimum to fire them. So I am really in a jam.

My opponent has all usable armour.

My infantry are fighting like pigs, but it is hard.

I think in future I am only going to play game in which I manually choose my forces, otherwise the game becomes a little strange if the AI is cruel -

And the, uh, point of that was...?

Originally posted by Gunny Bunny:

I have only offered constructive criticism.

Regards,

Gunny Bunny

Ahem, *cough-cough-bull$@*!-cough-cough* Whew, damn cold weather, I knew I should have worn a sweater yesterday!

------------------

"...Every position, every meter of Soviet soil must be defended to the last drop of blood..."

- Segment from Order 227 "Not a step back"

[This message has been edited by The Commissar (edited 01-29-2001).]

[Guest Rollstoy]

Another attempt to get back on the subject ...

tom_w, you wrote

Their demo video of the game play, if it is to be believed, shows FULL Dynamic Lighting effects and renders more polygons than CM in a true 3D environment.

I watched the video and did not see anything in terms of dynamic lighting. Period. The only noticeable feature was the accurate (?) shadows thrown by the soldiers and vehicles. While certainly impressive, that does not make 'FULL Dynamic Lighting effects'. Did you check the explosions? Simple billboard graphics. Did not light the environment at all. Same for rockets. Now effects there, either.

Now for the polygon count. I will try to throw in a screenshot here:

Do you notice the FOG! That is about the oldest trick to lower the render load. These are certainly not more polys than in CM .

Worth mentioning is the dust thrown by the tracks and wheels. A nice effect.

Other than that, the woods are done nicely, at least from a distance. This is something CM could improve, too. However, the technology used is conventional.

All in all, I do not think this game is the quantum leap you want it to be ... which once more prooves that good ideas are needed. Bring em' on ...

Regards, Thomm

By the way: theoretically it should be impossible for any FPS to handle more polygons than CM, because in CM the calculation and the rendering are decoupled, i.e. when showing the movie, the CPU does not have to bother with AI and turn resolution.

[Guest Germanboy]

Originally posted by Rollstoy:

All in all, I do not think this game is the quantum leap you want it to be ...

I would concur from just looking at the screenshot - I think it would be correct to assume that in the screenshot we see a bar for armour, which implicates that armour is whittled down gradually, according to number of hits? So, how many jeep-MGs does it need to take out the Apache? If this assumption is correct, it means that there are no complex calculations about armour penetration going on in the background.

------------------

Andreas

Der Kessel

Home of „Die Sturmgruppe“; Scenario Design Group for Combat Mission.

[Guest Rollstoy]

I watched the video and did not see anything in terms of dynamic lighting.

Hmm. I was wrong. Watched the video of the parked Mi-17 during the course of a day: shadows moving depending on the position of the sun, helicopter shaded accordingly. Impressive!

What to conclude from this? I think they calculate shadows and shading from sun-light, but not from point light sources like explosions (like Quake III, e.g.). So there is a compromise here, too. The latter would make night fighting more impressive, but during day time there would not be much of a difference.

Furthermore, I think that only the terrain is affected by shadows, not other objects (another simplification).

At last, the FAQ says that you can only move into the first 50 meters of the border of the woods ... due to technical problems ... well, here we go again - they all sit in the same boat.

So, after all I have seen, I think that dynamic lighting is most important if you specialize in night fighting. But since night fighting is not the most interesting game mode in CM in my opinion, I think that DL is not much of a issue at the moment.

We got the simple shadows for vehicles! THIS was a must have, anything that goes beyond is certainly nice, but not obligatory ...

I'd say throw in some dust clouds next, more polygons for soldier faces, damage on buildings, but that's about it ...

Regards, Thomm

[Bruno Weiss]

Well I don't know ofcourse, what everyone else is running, but I use most all of the terrain, vehicle, uniform, and building mods out there. And on my FSAA card, my CM outshines whatever that screen shot above is by far.

Of course, it's fairly difficult to judge things from a screen shot, but the woods, buildings, terrain, everything save for possibly the fog on my CM is better looking than that. And I'm not even certain but what the fog isn't also.

I don't know why shadows according to the sun's movement would be an improvement, in a short 1 hour or so battle which is CM's general scale, why would that even matter? Improved faces, maybe with lil whiskers okay sure fine, explosion damage to buildings more realistically detailed, yeah sure I wouldn't turn it down, dust from tires and tracks to give them that Peanuts Pigpen look, maybe even windshield wipers swiping too and fro in the rain, but at what price glory?

If those things meant a drastic slowdown towards the release date, or that their inclusion meant more for the program to deal with at the expense of less of something else, then at that point I'd opt for priority to go to the game, and not the ambiance. I myself don't know the answer to that, but I'm comfortable with the BTS response earlier in the thread.

------------------

"Gentlemen, you may be sure that of the three courses

open to the enemy, he will always choose the fourth."

-Field Marshal Count Helmuth von Moltke, (1848-1916)

[This message has been edited by Bruno Weiss (edited 01-30-2001).]

[Gremlin]

My last comments on the issue (applause), meant as reasonable and hopefully constructive observations on the topic at hand:

* Gameplay and graphics are intertwined, with one affecting the other, at least to an extent. Try listening to Debussy in piano transcriptions only, or take down all the paintings in the Louvre and hang up black and white photocopies of them. The analogies aren't far fetched if you think about them and how games work.

* A game that takes advantage of the latest hardware can still run on older machines. Most all games that are released are scalable via resolution, detail settings, dynamic lighting, and other toggles and settings. It's not an either/or proposition, though obviously there's only so much time in BTS's day and money in their budget, and they'll naturally have to choose priorities.

* Not everyone fully agrees with BTS's take on the game industry, their estimation of their current and potential customer bases, their estimation of what makes CM an enjoyable and successful game, and/or what should be included in future CM games.

------------------

New to Combat Mission?

Visit CM Boot Camp at Combat Missions for tips.

[Quenaelin]

It was very clearly stated in their interview in some web page (what I can't remember) that Operation Flashpoint will be Action oriented shoot'em up and not anything more. If something is worth lookforward in simulation genre it is Steel Beasts and Steel Beasts 2, not so good in graphics but gameplay is marvellous, I tried the demo. Maybe I am still going to buy its first version, even that is lacking air support and 3-D HW, but I saw great desert mod for it and I know it will be just as easy to modify as Combat Mission.

[aka_tom_w]

Originally posted by Rollstoy:

..

Regards, Thomm

By the way: theoretically it should be impossible for any FPS to handle more polygons than CM, because in CM the calculation and the rendering are decoupled, i.e. when showing the movie, the CPU does not have to bother with AI and turn resolution.

Hello Thom

Yes I totally agree that theoretically it should be possible for CM to handle more polygons because in CM the calculation and the rendering are decoupled, i.e. when showing the movie, the CPU does not have to bother with AI and turn resolution.

So as Gremlin points out, why not make the D/L eye candy optional and scalable? If some players would like to wait for a longer crunch time then they could toggle on their optional ray tracing for the render for the one minute turn. Since the ray tracing and the Polygons don't have to be rendered on the fly, it would just take longer in the turn crunch.

Thank you also for being open minded enough to look at that demo moive preview and see the D/L effects in Operation Flashpoint I was refering to. The Dust those AFV's kick up is very impressive, I thought the explosions were very WELL done.

AND and this is the kicker.... Some how those designers have concluded that the "average consumer computer" can and WILL render ALL that eye candy in REAL real time!

So unfortunatly I have to agree with Germanboy when he speculates that there is NO way that that game can do the kind of armour penetration calucalations that CM does AND render all the that 3D eyecandy on the fly.

so it is now just a big fancy version of Warcraft or Age of Empires with hit points for tanks. I must admit that I am disappointed by that conclusion.

in case you want to see that dust those AFV's kick up try this link, but it is a 25 meg download.

http://www.flashpoint1985.com/video/fpinvex.mpeg

check it out

-tom w

[This message has been edited by aka_tom_w (edited 01-30-2001).]

[Chupacabra]

Originally posted by Gunny Bunny:

I find it rather interesting that people believe I think CM is a bad program... I have only offered constructive criticism.

Regards,

Gunny Bunny

Which is why you're still bitching about brainwashing over on Usenet?

STEP 1 - Decapitate victim

STEP 2 - Scrape out brain with large spoon

STEP 3 - Place brain in brain-wash machine until it simmers

STEP 4 - Place brain back in skull and reattach to body

STEP 5 - While in a zombie state force victim to repeat the following:

*** Everything BTS states is the truth

*** I shall never compare apples with oranges

*** Steve is my one and only GOD

*** I shall make an annual contribution of US$ 60 annually to BTS *** I shall remain a faithful BTS party soldier for all time

STEP 6 - Find next victim and repeat steps 1 to 5

There is, of course, room for improvement in CM's graphics. And they will improve. If you think that acting like an scat-flinging monkey is going to induce anyone to see your point of view, you are quite delusional. Wake the hell up, kid.

And tell me, how does it feel to get constantly flamed on a usenet forum you started?

------------------

Soy super bien soy super super bien soy bien bien super bien bien bien super super

[Guest Germanboy]

Originally posted by aka_tom_w:

So *snip*, why not make the D/L eye candy optional and scalable?

Simple really - resource constraints at BTS. What many people who argue for optionality seem to forget is that somebody still has to code the stuff, even if only 10% (or whatever) of the customers use it. This does not matter if you are Microsoft, because you just employ another bunch of summer students on next to no salary, but in BTS's case it matters a lot. I believe they will be hard-pressed with CM2, even without coding in tons of optional stuff that some customers want. Every optional feature (note: these features being optional means they are not crucial by definition) coded means more time to the release of CM2, more time for BTS to generate a new revenue stream, well you get the point I am sure.

What is the proposal you, or Elijah who called for optionality earlier have to get around this problem?

------------------

Andreas

Der Kessel

Home of „Die Sturmgruppe“; Scenario Design Group for Combat Mission.

[This message has been edited by Germanboy (edited 01-30-2001).]

[Gremlin]

Andreas, before lumping me in with others, please read my words and not words attributed to me. I raise a point similar to yours. Thank you.

------------------

New to Combat Mission?

Visit CM Boot Camp at Combat Missions for tips.

[Guest Germanboy]

Originally posted by Gremlin:

Andreas, before lumping me in with others, please read my words and not words attributed to me. I raise a point similar to yours. Thank you.

My apologies. I edit it out.

------------------

Andreas

Der Kessel

Home of „Die Sturmgruppe“; Scenario Design Group for Combat Mission.

[Gremlin]

No problem

------------------

New to Combat Mission?

Visit CM Boot Camp at Combat Missions for tips.

[Guest Rollstoy]

Dear tom_w!

You wrote:

So unfortunatly I have to agree with Germanboy when he speculates that there is NO way that game can do the kind of armour penetration calculations that CM does.

Without wanting to step on anybodys foot here I would presume that the amount of CPU time necessary to evaluate the outcome of shell/target interaction is negligble in the large scheme of things. Once you found the angle of incident of the shell you can possibly rely on precalculated look-up tables to determine the result. The hard part is to find exactly where the shell hit. In CM (if I am not wrong) this is calculated first with a random function, then the flight path is calculated. (this, by the way, would explain how that Stuart was hit through the house by the Panzerschreck:

(1) CM finds clear line-of-fire at the moment of missile LAUNCH;

(2) CM calculates where and when shell hits;

(3) CM does not recheck if the actual line-of-fire, with added lead, intersects with a building;

(4) PS hits Stuart through building ...

as the unfortunate combination of a slow missile and a fast vehicle causes it too. Possible solution: recheck line-of-fire after lead was added).

So, if they solved the collision detection part (and they have to) the actual penetration algorithm should not take any time because it has to be evaluated extremely rarely compared to other things.

That said, I guess that the really time consuming part about proper armour penetration algorithm is the RESEARCH and the IMPLEMENTATION, but not the actual evaluation.

Regards, Thomm

[Quenaelin]

Main point is that these ray-tracing calculations (which somebody are suggesting) have to be still calculated in realtime so you can pan and zoom, otherwise it will be just like 3-D animated 2-D movie (and we don't want that.)

[Guest Rollstoy]

Uh-oh, ray-tracing is a entirely different beast and will not be seen in games for the next 10 years, since it means that you render every individual pixel on-screen seperately and not whole polygons at once.

Raytracing is only necessary for the mathematically exact calculation of multiple reflections. Total luxury in games, where we have Z-Buffers, Cube-Environmental-Mapping and double rendering of the same environment for simple mirrors (i.e. just render the room and the objects twice, one time in front, one time behind the mirror) to do the job.

So do not worry about ray-tracing .

Regards, Thomm

[deanco]

Originally posted by Germanboy:

Simple really - resource constraints at BTS. What many people who argue for optionality seem to forget is that somebody still has to code the stuff, even if only 10% (or whatever) of the customers use it. This does not matter if you are Microsoft, because you just employ another bunch of summer students on next to no salary, but in BTS's case it matters a lot. I believe they will be hard-pressed with CM2, even without coding in tons of optional stuff that some customers want. Every optional feature (note: these features being optional means they are not crucial by definition) coded means more time to the release of CM2, more time for BTS to generate a new revenue stream, well you get the point I am sure.

What is the proposal you, or Elijah who called for optionality earlier have to get around this problem?

Here's an idea :

Hire outside help.

Possible advantages:

1) Will not appreciably slow down (except for added 'administrative' work) research and development of CM2. (That is, Steve and Charles' work.)

2) A chance to hire the best minds in the business to attack this problem of graphics quality vs. computing power. At the very worst, fresh minds can sometimes see the problem from a different angle due to their 'neutraity' to the game. They can also do poor work, run behind schedule, etc. I know. But the idea would be to get the best and brightest minds possible.

3) Most importantly, expanded graphic quality could make the appeal of CM even wider than it is now (more sales) and bring other gamers from other areas (RTS or whatever) into the wargame fold. This is good for BTS and good for wargaming in general.

Possible disadvantages :

1) Costs more money. Yes, I know this is not a 'possible' disadvantage. See below.

2) Security issues : If 2 people know a secret, it's no longer a secret, I read somewhere. Could be minimized by only releasing the 3D portion of the game to work on, thereby insuring that Tac AI, battle algorithms, etc stay in the family.

3) Enlarges the company. The more people you have, the less absolute control over the product and the harder it is to oversee everything. Again this can hopefully be gotten around by hiring good people who are responsible and can work on their own.

---------------------

The bottom line in all this is this : If CM2 is going to cost X dollars more than CM to make, then BTS MUST sell Y more copies of CM2, due to graphic improvements made to the game, to get their investment back and show a profit. Anything less than that, and the pain-in-the-ass factor becomes too great. This question can only be answered by market research, mixed in with a bit of gut feeling. I don't have the figures, but I do have the gut feeling. Sorry, I know that's not rational.

Keep the flames low folks, I DID try to present a balanced argument.

------------------

DeanCo--

CM interface mods: http://mapage.cybercable.fr/deanco/

so many games...so little time

[Guest Germanboy]

Originally posted by deanco:

Here's an idea :

Hire outside help.

I appreciate your even take on it, I know that you like graphics effects a lot. Some things you may have forgotten about:

It changes the risk structure dramatically.

You have commitment issues, and need to manage your staff instead of doing research/code.

I believe that assembling code from different people and then test it is a lot more resource intensive and increases fault probability (if someone does not agree, they can have a copy of any Microsoft product from me...), so again this takes time away from production. Effect is that if you hire one additional person, you only get 1-X net-effect in additional resource.

I believe the management bit is really crucial. I should think that Matt and KD are pretty much self-managed. A sweatshop of coders would not be.

These people's salaries would have to be financed out of CMBO sales until CM2 hits the road. This may be difficult.

Internal communications suddenly become a huge issue.

Increase in company size is not a straightforward thing, it means that a lot of things change, and this change in itself can be time-consuming and costly. I see that in our organisation at the moment. It is a risky strategy - whether BTS wants to pursue it further than hiring Matt and KD I am not qualified to say, but I also have a gut-feeling about this.

------------------

Andreas

Der Kessel

Home of „Die Sturmgruppe“; Scenario Design Group for Combat Mission.

[deanco]

Originally posted by Germanboy:

I appreciate your even take on it, I know that you like graphics effects a lot. Some things you may have forgotten about:

It changes the risk structure dramatically.

You have commitment issues, and need to manage your staff instead of doing research/code.

I believe that assembling code from different people and then test it is a lot more resource intensive and increases fault probability (if someone does not agree, they can have a copy of any Microsoft product from me...), so again this takes time away from production. Effect is that if you hire one additional person, you only get 1-X net-effect in additional resource.

I believe the management bit is really crucial. I should think that Matt and KD are pretty much self-managed. A sweatshop of coders would not be.

That was what I meant by 'admistrative' issues. I probably underestimated the problem. But see below.

These people's salaries would have to be financed out of CMBO sales until CM2 hits the road. This may be difficult.

Internal communications suddenly become a huge issue.

Increase in company size is not a straightforward thing, it means that a lot of things change, and this change in itself can be time-consuming and costly. I see that in our organisation at the moment. It is a risky strategy - whether BTS wants to pursue it further than hiring Matt and KD I am not qualified to say, but I also have a gut-feeling about this.

Can we have some of the experts weigh in here? I have a question.

I agree with you Andreas, if BTS had to hire 50 people ('a sweatshop of coders') it would destroy the nature of the company! I surely wouldn't wish that on them. I was thinking of a smaller, more managable number, but perhaps I have been spoiled seeing the incredible work Charles has done by himself. So here's my question :

How many people will it take to do the job?

Oops, I'd better define my terms. Say, one light source, (the 'sun') on all the time and bathing everything constantly in diffused light. And a bunch of smaller lights, on intermittantly, that only affect a small area (muzzle blasts and explosions). Large map, lots of buildings, and a 1000 point assault worth of units on the map too. Sun reflects off the armor but not the men, the small flashes light up everything. How many monkeys and how long? Educated guesses?

Happy to get this discussion on a constructive track.

------------------

DeanCo--

CM interface mods: http://mapage.cybercable.fr/deanco/

so many games...so little time

[Jeff Heidman]

I think is you look at what Charles damaged to accomplish by himself, the addition of a single additional person to handle some of the "Grunt" code work would add a lot, but only if they find the right person.

That is the hard part. If they could grab an experienced, self-motivated game coder to crank out the graphics portion after Charles and Steve do the design specification, that could add a huge amount of value for a relatively low cost.

But it is risky. If they hire the wrong person, it could end up taking valuable development time away through dealing with management issues.

BTS is in a hard spot. To some degree the amount of work available is greater than the number of resources, but at the same time expanding resources might change the nature of the business that has got them to where they are today.

They have two choices. Either expand anyway and hope they can manage that expansion, or refuse to expand. Both choices have dangers.

On the one hand, expanding (which they have already done to some extent) runs the risk of losing the very vision that made CM what it is.

On the other hand, the market is not static, and neither is the business. CM1 is successful because it is revolutionary, but the cat is out of the bag. Other groups and companies will certainly steal their ideas and try to match that formula.

This is the classic problem of entrepreneurship. The skills that make people successful for the first release of a product are sometimes not the same skills that sustain that success.

Id software went through much of the same problems, albeit on a greater scale.

Jeff Heidman

[Gremlin]

It's imperative, too, that any entrepeneur or game designer, for that matter, avoid resting on his or her laurels, in order to stay innovative and competitive. Every game designer is sure they know what they're doing, but look at all the commercial and critical failures in every entertainment genre, including games.

Doom-->Daikatana should scare anyone in the business

------------------

New to Combat Mission?

Visit CM Boot Camp at Combat Missions for tips.

[Guest Rollstoy]

Levels of Hypothetical Dynamic Lighting:

(A) Directional light (the SUN) [see Flashpoint]

( Point light sources (Explosions, Muzzles, Fires?) [see Quake III]

which ...

(1) Affect Object Shading (polygon shading) [state of the Art]

(2) Affect Object Shading (vertex smooth shading) [Quake III]

(3) Cause Objects to throw shadows on terrain [Flashpoint, Rune]

(4) Cause Objects to throw shadows on other objects [???]

About how many man-hours it would take to implement any of those is hard to tell, because I am not familiar enough with the APIs. I guess (A1) would be a matter of weeks if the architecture is ready for it to be plugged in.

Regards, Thomm

[This message has been edited by Rollstoy (edited 01-30-2001).]

[Jeff Heidman]

I would be most interested in graphical improvements that have a direct effect on the game play.

Currently night scenarios are basically unplayable because of the lack of dynamic lighting, either in graphics or the underlying engine. Night is basically treated like really thick fog, which is a gross over-simplification. Gun flashes, flares, fires, tracers, etc., etc. are all necessary to have any hope of re-creating night time fighting.

Jeff Heidman

[deanco]

Originally posted by Rollstoy:

Levels of Hypothetical Dynamic Lighting:

(A) Directional light (the SUN) [see Flashpoint]

( Point light sources (Explosions, Muzzles, Fires?) [see Quake III]

which ...

(1) Affect Object Shading (polygon shading) [state of the Art]

(2) Affect Object Shading (vertex smooth shading) [Quake III]

(3) Cause Objects to throw shadows on terrain [Flashpoint, Rune]

(4) Cause Objects to throw shadows on other objects [???]

About how many man-hours it would take to implement any of those is hard to tell, because I am not familiar enough with the APIs. I guess (A1) would be a matter of weeks if the architecture is ready for it to be plugged in.

Regards, Thomm

[This message has been edited by Rollstoy (edited 01-30-2001).]

Rollstoy, 2 questions :

Which of those levels of lighting would vehicle reflections come under? I'm guessing Level 2.

I know your time estimate is a ballpark figure of course, but how many coders were you thinking of to do it in that time frame? And I know it's just an educated guess, I take it as such.

------------------

DeanCo--

CM interface mods: http://mapage.cybercable.fr/deanco/

so many games...so little time

[Guest Germanboy]

Originally posted by deanco:

How many monkeys and how long? Educated guesses?

Happy to get this discussion on a constructive track.

Deanco, I agree. Unfortunately I am not qualified to answer that question, since I am about as knowledgeable about coding as I am about heart transplant operations. I know a lot about what can and does go wrong in organisations though

------------------

Andreas

Der Kessel

Home of „Die Sturmgruppe“; Scenario Design Group for Combat Mission.