America: The Godfather of Game Development

Chapter 687: Same number of faces, double the effect!

The cellular stage, which is the first stage at the beginning of the entire game, is as expected.

After all, it was personally designed by Lin Lixin, and it's a 2D game.

The entire development process went much more smoothly than expected.

But things started to change when the second phase began.

As planned, Spore made use of the exaggerated graphics performance of the NovaFusion fifth-generation machine.

To deliver the most cutting-edge gaming experience, it's natural to make it 3D.

But a vast 3D world?

Leaving aside whether graphics processing is fast enough, even with current video memory, it's simply not feasible.

"In short... that's the current situation. We are considering reducing the number of faces to stabilize performance."

Carmack explained,

"With fog effects and streaming loading, the video memory can be compressed to almost nothing, but the visual effect will definitely be greatly reduced."

Unlike here, Lin Lixin's 2D phase allows for extremely refined visual effects.

Because with current graphics performance, they can already create many very good, detailed pixel effects.

In comparison, the 3D stages will only give players a stronger sense of disillusionment.

This is obviously something that is hard for people to accept.

"I tried several different optimization solutions, but... because the creatures in Spore are created by players, we can't predict the final appearance in advance, so we can't make targeted optimizations."

Carmack pulled up several other optimization schemes on the development machine and demonstrated them to Lin Lixin.

It has to be said that his skills are second only to Lin Lixin in the industry.

Either option performs very well.

But perhaps this is also a limitation of the times.

Due to the changes in the timeline, Carmack hasn't really worked on any UGC projects recently, and he's struggling to find a starting point for dealing with this kind of uncertain final outcome.

He is better at extreme optimization when all conditions are known.

"I see." Lin Lixin nodded in understanding.

No wonder the engine team has been gathering together these past few days.

"Actually, we can try a different approach."

As he spoke, he sat down next to Carmack.

"This technology will also be used in our new FPS projects in the future, and it will be used extensively."

After all, FPS is one of the core genres that made Gamenova famous.

As a first-party title in the fifth generation, it's definitely indispensable.

As for what to do then, Lin Lixin already had a direction.

"Actually, we don't necessarily need to worry about the number of noodles; reducing the number of noodles appropriately is perfectly fine."

Lin Lixin casually picked up a lump of clay, rolled it into a ball, and then gently rolled it a few times on the table.

With just a few simple strokes, this lump of clay was transformed into a regular polyhedron.

"We're making games, not computer graphics. Often, our goal isn't to make it perfect, or to make the computer think it's perfect. We only have one goal—to satisfy the players." That's the difference between the gaming industry and other cutting-edge computer industries.

If it were General Motors, even the slightest error in the modeling would cause a major problem.

The chain reaction can even affect the aerodynamic simulation in the wind tunnel, impacting the entire R&D production line.

This is why General Motors' digital design department has almost demanding requirements for parameters.

This also made them one of Gamenova's largest buyers of high-end industrial graphics cards.

But games are different.

As long as it looks the way the player sees it, then there's absolutely no problem.

"So we need to ask computer graphics to lend a hand."

Carmack was a programming genius, and this talent overflowed, allowing him to achieve considerable success in the field of computer graphics as well.

But he was a self-taught garage prodigy, not a true classically trained professional.

When faced with some classic computer graphics problems, he always managed to pull off miraculous solutions through sheer force and effort.

While this personality trait can sometimes bring surprises, winning a bunch of awards and helping the computer graphics community solve some age-old problems, it can also make him prone to reinventing the wheel and wasting energy on unnecessary things.

"Gouraud shading".

In the world of computer 3D graphics, there are three main shading schemes.

The first option, which is the most reliable, simplest, and most traditional solution given the current graphics performance, is flat shading.

Because three-dimensional objects in the computer world are composed of a large number of triangular pieces, shading schemes can be designed very easily.

When coloring, the color of the entire triangle is calculated only once, and the same color is applied to the entire triangle.

The advantages of this approach are obvious: it is incredibly fast and has extremely low performance overhead.

However, its drawbacks are also obvious, which is why Carmack has yet to find a solution.

stiff.

The colors lack transitions, and the angularity of the entire 3D object is very obvious, making it look nothing like a smooth object.

Admittedly, this can be improved by increasing the number of faces.

Go to Bing Search Flags to see the latest chapter! But this green skin approach of making a big splash obviously made these graphics scientists very unhappy.

So in the 70s, two geniuses stepped forward and each came up with a different solution.

Henry Gollold's Gollold coloring method and Pei Xiangfeng's Phong coloring method.

Gorold's approach is to perform calculations directly using the vertices of the face, and then perform a transition interpolation for each pixel on the face.

This way, the color on the entire surface will definitely be a smooth gradient.

Phong goes a step further, performing a complete shading calculation for each pixel, taking into account factors such as lighting and material properties for a comprehensive calculation.

In comparison, it's easy to see that Phong definitely has the best and most beautiful rendering effect.

But it's not perfect either.

Just like how planar coloring sacrifices visual effects for an absolute speed advantage.

Phong sacrificed performance for absolutely stunning visuals.

It is only suitable for use in applications such as the film and television industry that highly rely on precision but do not require real-time calculation.

As for Gorod... it struck a middle ground between the two.

It achieves a much stronger effect than planar shading despite its lower computational intensity.

The number of faces didn't increase, but the whole model suddenly looked just like the real thing.

that's enough.

It's enough to fool the players' eyes.

Lin Lixin sat down at the development station and quickly modified the shader code.

Recompile and run.

The model that was originally a finite number of faces on the screen suddenly became smooth.

Advanced developers can often cut corners and reduce the workload on the graphics card without anyone noticing.

"I see……"

Carmack glanced at the source code and immediately grasped the brilliance of this strategy.

"Reverse thinking...?" he muttered, as if he had suddenly realized something, and then his eyes lit up.

"Ah! That reminds me. I suddenly have a great idea."

"What's wrong?"

"Sam asked me yesterday if there was any way to design flexible and realistic motion animations for a creature when it's uncertain what it will look like."

Carmack exclaimed excitedly, quickly opening a document to jot down the thoughts that were forming in his head, lest these inspirations vanish in an instant.

"Lin, you know that when we used to make 3D animations, we used existing skeletal data."

"Indeed, but this time our skeletal structure was determined later."

"Yes, so there's no way to pre-design how a creature moves, because we can't even determine where its legs are located, or how many legs it has."

Carmack nodded; this same problem had been bothering him for some time.

But Lin Lixin's words just now gave him a wake-up call.

Sometimes, solving a problem simply requires changing your perspective or thinking in reverse.

He showed the words "[Inverse Kinematics]" to Lin Lixin:

"We don't control how the limbs move, but rather tell the limbs where to place them."

Logically speaking, the establishment of a leg-raising animation should be a sequential chain.

The thigh moves first, and then the knee, lower leg, ankle, and toe are pulled and moved along with the skeletal system.

This is a typical example of forward kinematics.

But IK took a completely different approach.

It first determines the final landing point.

By analyzing the point of impact, we can deduce how the limbs should move.

No matter what the ground looks like, no matter what the creature looks like.

Its legs always fit snugly against the ground.

Of course, the trade-off is that the processor overhead will also be higher.

"This system can even support multiple legs moving together! And it can greatly reduce our resource expenditures." Since everything is calculated in real time by the program, they certainly don't need to store these things on CDs.

In the 3D era, there is a problem that has caused headaches for all manufacturers.

As the number of characters increases and the scenes become more magnificent, the animation resources also continue to expand.

Even CDs, which had a capacity that seemed like alien technology just a few years ago, are now starting to feel inadequate.

Especially Spore, a game that allows players to unleash their boundless imagination.

Theoretically speaking, an infinite number of combinations is equivalent to an infinite number of actions, which is also equivalent to an infinite number of expenditures.

Once programmed, they can do only L2D splitting, just like making a 2D game, which greatly saves space.

Of course, it also saves working hours.

After all, animators no longer need to painstakingly create animations frame by frame. As long as the programmers write the rules, the game can handle everything else on its own.

"Combining IK inverse dynamics with the Gollod shader we just discussed, dynamically generating meshes and drawing skins based on parameterized spine nodes, using IK skeletons to control gait, and adding Will's Jazz AI architecture..."

Carmack excitedly organized the documents.

A whole new world seemed to have opened up to him. A completely programmed, completely free world.

It is also an infinitely large world with infinite possibilities.

If the idea of ​​procedural sandboxes is further extended...

Tap the screen to use advanced tools Tip: You can use left and right keyboard keys to browse between chapters.

You'll Also Like