Underworld Ascendant: 25 Years On
Paul Neurath waxes nostalgic — and comments on the quantum shifts that have happened in the gaming and computing industries — in the latest Stygian Sentinel update on the Underworld Ascendant website:
The original Underworlds were designed to run on 20mhz 386 processor class PC’s. The smartphone in your pocket would crush a PC of that era without breaking a sweat.
These PC’s also lacked any sort of graphics card. You had to do all the rendering in software. That was a huge hurdle to doing real-time 3D texture mapping. Even with some super clever code running in optimized assembly language, we could barely achieve a playable framerate.
In some ways having these performance constraints was helpful, as it compelled us to find creative work-arounds. For instance, there was no way to render fast enough an over-the-shoulder view that would show the player’s character in the foreground and world beyond that. Solution was first-person view, which ended up working well for us, and many games to follow.
Today’s PC’s are ludicrously powerful in comparison. Graphics cards and modern game engines now provide all the building blocks to do sophisticated 3D rendering. The focus has moved from simply trying to get 3D to run, to tweaking the higher end bits of the rendering pipeline to achieve refinements on advanced visual effects. Like getting the fur on that creature to look even more natural than it did in a game from a couple of years ago.
…
Ultima Underworld fit on 6MB of floppy disk space. For a modern PC game 6GB is not uncommon, 1000x bigger.
Where nearly all of the extra space has gone is into making dramatically higher fidelity visuals and audio. Supporting 4K visuals takes a LOT of pixels, as do high-polygon 3D creature models that have fluid animations.
These levels of fidelity facilitate creating beautiful and immersive 3D worlds. But it can take massive effort to build the assets. A modern big-budget 3D game can readily allocate 150 or more person years just to craft the visual assets. That’s on par with the effort it took to build a medieval castle!
He also shares some thoughts on what is required for a game to succeed in today’s massive, but also very crowded, gaming market:
The best way to stack the odds that you’ll deliver a great game is to have a great team building it. A great team is not some collection of ‘super stars’. A great team will have:
- A shared, deep passion for making the game, along with a dogged determination to make the game great. If you set the bar at less than great, nearly always you’ll deliver less than great.
- The skills to pull off the particular game they are building. Genius helps, but no team is filled with super stars. It’s more about having the right mix of competence and creativity among the team members.
- A team that can learn and grow. Making a great games means regularly encountering new challenges thrown your way. To tackle the challenges you frequently need to stretch and try new solutions.
It’s certainly well worth spending a few minutes reading the entire post. Included in it is a small, heavily cropped image of…something. If you can guess what that something is (answers can be submitted here), you might even win an Underworld Ascendant t-shirt.