Why are shells and bullets so detailed?

"Oh hey, let's post in a thread without reading the comments. But man, I'd rather just post a great one-liner."
-PrimePriest

There is literally no performance problems with a projectile that detailed. At most your computer would be rendering 50 of those shells at a single moment. That doesn't even touch the processing power necessary for a fully modeled infantry with bones and animations. Not even a thousand of those simple objects would require the same amount of processing as one infantry. Why? Because GPU processing and CPU processing are two completely separate things.

It's good for screenshots.

I'd suspect this adds to the performance degradation in large battles. Reduce all particle/projectile complexity and see if it doesn't run like trash when 100 people are shooting on screen.

Everyone is saying that, "Oh, not very many tank shells going on at once". Could someone try to get a screen shot of what the bullets and small-arms fire looks like? Then maybe we can have a decent conversation. Because drawing conclusions only based on one type of projectile (ie, tank shells) is kind of pointless in a combined-arms game.

Small arms projectiles are just sprites. Nothing to worry about. This whole thread is pointless.

Detail is pretty average, in my opinion. It's just a simple model with a nice texture.

This is a good example of player ignorance.

Tank round fired by a tank, as every round and bullet and plasma ball, becomes an object controlled by either the server or the rendering engine. The problems in the 96+ player fight isn't the rounds being fired by all those players, but the players being controled by individual clients. All the information has to come in, get processed, and then go back out.

And the graphical image is NOT processed in those calculations. What you see is an image. Does your computer lag when you try to view a 512x512 72dpi bitmap? If not, then the tank round isn't going to either.

Three classes any PC gamer should take at their local community college:

1. Basic HTML course
2. Programming and Logic course
3. PC Maintenance and Repair (also called Basic Hardware in some places)

That way they would not only understand what is going on in the games they play, at least at a Very basic level. But they'd be able to maintain their rigs and not complain when a game buzzes out their integrated video cards, and maybe even keep their stuff updated to run the stuff they want to play.

What... the... *****. I guess you aren't that old, as people who have seen this change consciously marvel at what we can do now (no offense meant, but it is just so damn impressive).
We marveled in 1996 when Quake came out. Then, in 1999, Unreal Tournament came to light. It had detailed bullets (a smiley on the flak cannon's grenade, the redeemer rocket was more detailed than the image above, with "Adios!" written on it), breathing, and a whole lot else. Looking back, it is blocky and not very impressive, but it was at that time. Skip forward 2 years.
GTA 3 came out in 2001. You needed a pretty powerful rig at that time for full graphics. For the 10 year anniversary (2011), they released it again, for the phones. A phone that has a chip that has easily double the computing power than that freakishly big rig. On a side-note, that phone also has a higher video capability, a handheld battery that outlasts any battery of that time by a factor of who knows what, and still has more flash storage available, and possibilities to upgrades.
After maybe 2005, you don't see that much change directly, as the graphics look good for a long time now. Still, grab a game and tell me those engines aren't worse. We can now destroy environment, handle an incredible amount of objects, have better lighting, movement, particle and movement simulators and all that stuff. Every shooter now tries out another level of full destruction in their games.
For a long time, and still ongoing, our computing power doubles with every about 18 months. This doesn't filter though to the consumer levels quickly. Marketing wise you will invest a lot in one level of technology, giving out a few classes, and already invest in the next leap of a few levels and release that at a much later time. Investing in every level is too costly, but after the i3/i5/i7 processors for example, you will see another big leap. The same goes for the GPU, but their battle is much quicker.

So yes, everything has gotten much, much more powerful, and still is getting much more powerful.

I do agree with your point though. The game isn't graphics heavy, but the bottleneck is at the amount of stuff happening at once by autonomous players. The making of a shell like this was already possible in 1999, so I'm actually surprised so little detail. The engine is from what, 2008? It should be capable of much more, and I would have suspected one guy to just go overboard one day and make an art of a lot of bullets/rockets/shells in a bored afternoon.

I wondered this when watching other Skyguards shoot upclose, you can see little physical bullets being fired off into the distance with the trail behind them.

I have always been like "wow is that really necessary? .. it already has the trail"

Well, the advancement and progression computational power has tapered off a bit over the past 5 years due to manufacturing companies focusing more on efficiency than raw power. They've still improved significantly more than what Ripshaft implied but not near as much as normal in recent years.

I wouldnt mind if shells were endered in 2D tbh.

I saw one up close once...but then it hit me in the face.

I don't much of a point. Like someone else stated, programming/texturing tricks makes rendering processing pretty low. You probably wouldn't even notice a fps hit unless you were running on a pentium 3. Besides, PS2 having performance issues? Maybe for you but not me. I understand you're argument and you're right, I never noticed until this thread. But not everyone has performance issues with PS2, so it doesn't make sense to remove it entirely.

In a nutshell, no. The way modern hardware accelerates rendering leans heavily on the concept of batching and pipelining. It's not a simple process, but an explanation can be simplified...

• (1) CPU: collate visible objects (culling) and detail (LOD)
• (2) CPU: batch objects using a common shader
• (3) GPU: set shader
• (4) GPU: draw
• repeat...

Some key points are:

• step (3) is expensive because it can stall the rendering process
• step (4) is a concurrent operation parallel to the CPU; as soon as the drawcall is sent the cpu gets on with steps 1 & 2
• the GPU is a pipeline, with triangles fed in at one end and passing through several steps of transformation and rasterisation
• Maximum performance comes from feeding the pipe with enough data that it doesn't have to stall

So...
Game entities that share a common shader (e.g. 'all projectiles', 'all vapor trails', 'all volumetric particles', 'all sparks', etc.) will be batched together and drawn in one go. Essentially the polycount of a projectile is irrelevant (within reason) because feeding the GPU with several hundred triangles is the same as giving it a single triangle. In the latter case it will finish drawing before the CPU can organise the next batch of objects.

Instead of us speculating and posting "standards" in game design yada yada, I would like to call upon an actual SOE DEV to read this and fill us in on whats going on.

We can speculate all we want, but at the end of the day, we did not design nor code this engine. I want to hear it straight from the Devs mouths on why this is or is not a resource hog.

PS2 Devs pls

Wow, it really does look like a Nerf football. I was joking when I said that's what our tanks fired but there you go.

It has nothing to do with speculation. Its called understanding how computers, processing and programming work.

I want an SOE Dev to comment on this to put our speculations at ease. Did you code this engine? Probably not, so lets hear it from a Devs mouth. This would be the best course of action rather than posting stuff lifted from wikipedia. (This is not an attack to your personal experience or knowledge)

I understand how computers, processing and programming work. I've been doing it damn near 15 years. I don't understand how these servers + clients, these processes and their programming works, specifically with regard to rendering - hence the caveat about speculation.

So why doesn't this GPU processing power go into for example terrain textures? These are ugly and low res as from some kind of 2000's game. Are you trying to tell me that rendering projectile and rendering a rock are two different things?