The team's initial goal for this feature was to display the player's remaining stamina visually. Instead of using numbers, we chose to base the stamina wheel in this DLC on the one from The Legend of Zelda: Breath of the Wild.

Now, this was an interesting challenge because in Zelda, the stamina wheel moves with the character, but always faces the camera.

Initially, I tried modifying the player model to include the stamina wheel and then rotating it so it faces the camera.

However, as you can see, the wheel appears to jitter. This jittering occurs because the variables that determine the wheel's rotation don't update fast enough to keep up with the real-time data. This desync causes the wheel to jitter.

At this point, I realized I had to pivot to a different system.

The glorious performance optimization fix to all of my problems

I knew the main performance issue was the number of bindings that were calculated on the client side to render the correct wheel state. Therefore, my goal was to reduce this number as much as possible.

I made a few cuts and optimizations to the rendering here and there, which did improve performance, but not enough to be considered a shippable product. Think of like borderline unplayable to merely unpleasant.

A bigger fix was yet to come.

"What if there was a way to not use bindings at all?" I thought to myself one day.

We could already send text directly to the UI without needing a ridiculous amount of calculations to display it. I just wished there was a way we could send images like text to gain all these performance benefits.

And that's when the realization hit.

Glyphs are custom text characters that you can modify to look like whatever you want.

In this sprite sheet, each square represents a frame index of the stamina wheel depletion or delta depletion.

Using it, I could send stamina wheel glyphs directly to the client, so they wouldn't need to do so many calculations to render the whole stamina wheel.