In Stewart Brand’s book The Media Lab: Inventing the Future at M. I. T., Steward Brand describes the phrase “Demo or Die”; A phrase that was coined by the students who attended the Media Lab at its infancy, and is now part of the now-famous Media lab’s culture of innovation. The meaning of the phrase is short but dire; Create something show-able, or become irrelevant.

We love this concept at Codeate. It’s incredibly important, especially during game development, that developers aggressively demo their work to discover and learn about what is fun and what is annoying to the player. We demoed our game at the Creative Technologies Center this Saturday and found that although most of the players had fun, many ran into various annoyances that could’ve been improved.

Take the following for example.

The graphic on the left shows the original implementation of a VR based body rig. As only the headset is available, the game makes a best-guess to the orientation of the player, and create a “Body” where equipment can be stored.

The original method used the Y rotation of the camera to rotate the body, locking the X and Z rotations. (Without additional data, the best guess is to assume the player is upright.) This poses a problem when the player looks down too far, as the game suddenly thinks that the player is facing the opposite direction, which spins the ammo belt to the right side.

The fix for this issue was to prevent the body from rotating when the player is looking down. Although this does not perform well in the case where the player is looking down and rotating, we felt that it was a good usability trade off. In the example on the right, the player’s body rotation is locked when looking greater then 50 degrees below the horizon, which allows the player to look at and reach for items on the belt without the belt rotating. This also fixes the issue with the belt flipping 180 degrees when looking too far down.

Another issue with the body rig was the transition between standing and kneeling. As the game itself encourage the use of cover around the environment, players often find themselves standing, kneeling, and sometimes even crawling.

This poses an interesting problem for the body rig, as the original approximation fails in any pose other then standing. The original implementation uses average data of a human body, establishing the hip as 0.65 meters below the head. However, this model does not work well when the user is crouching or significantly shorter (For example, a teenager). In this case, the virtual hip almost reaches the floor!

The new implementation solves this by smoothly changing the ratio of the hip between the head and floor, which results in the second graphic on the right. This version transitions between a maximum ratio (Standing) and a minimum ratio (Kneeling), based on a reference height (Which can be customized per player using a method similar to the one seen in Thrill of the Fight).

We’re making huge strides based on the feedback received from our demos, and we’re looking forward to getting more input in the coming weeks!