A Moogle  🏳️‍🌈 💻 🎮 🎶 🇸🇪

I wrote a tweet yesterday about large file sizes for games where I made assumptions about their reasons and had it explained to me that I got it wrong. This was particularly kneejerky of me and I wasn’t, as they say, in my lane on that one.

Honestly didn’t think you were wrong. Optimisation doesn’t appear to matter much these days.

Define "optimisation". Perhaps this is just a failure of my imagination, but every "optimisation" that I can think of comes at the cost of quality - let's talk about that.

Animations can certainly be compressed - when dealing with rigs that are animated with forward kinematics, those optimisations come at the cost of jitter which is increasibly noticeable as render resolutions go up.

Textures can be optimised by running them through compression schemes like DXT1, DXT5, BC6 or BC7 - but you'd better hope that those textures don't need hard edges or otherwise require fine details, because the block compression utilised by all of the above will kill quality.

Audio tracks can be compressed, at the cost of noticeable levels of hiss or atrocious levels of frequency response, which is particularly unforgivable given the number of gamers who use headsets these days.

The simple fact is that the demand for quality assets has increased, while gamers' ability to visualize these assets at the intended quality has increased commensurately as well, leading to double the motivation to avoid lossy compression, and increase asset size.