@Atrix256 @jb55 @OskSta Machine-learning allows content production at larger scales with equal/smaller teams. There are a number of variables/trade-offs to consider: content-fidelity, content volume, team size, content ‘reception ‘ (i.e. fun), etc ...
@sehurlburt Tough one! Monetization of the DLC might matter. Easiest way might be to identify with a ‘boxed product’. Might make sense to cap number of DLC releases.
@Atrix256 Yeah, nvtt is an easy starting point. There’s a bit of a science and art to this, though. Things that might be ‘correct’ may not look ‘better’ than some Manual Fu.
@Atrix256 Used sinc as well and artists like to sharpen lower mips for things like terrain textures (crisper at oblique angles, but more shimmer as well)...
@sehurlburt Maturity is saying that some software is ‘done enough’ ... and if you update then I’d rather have 1-2 times a year to take meaningful updates than superfluous ‘make customers happy updates’.