You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Sinc interpolation now uses up 74kB RAM instead of 2.4MB and thus is less hard on the CPU cache. In some cases this means even faster rendering. The way this change works is that the polyphase look-up tables now use 256 phases/branches instead of 8192, coupled with linear interpolation between the phases when in use. This method is equal to having many more pre-computed phases than 8192, which in theory also means better precision.
Cubic spline interpolation now doesn't use a look-up table. Again, this is less hard on the CPU cache.
(Hopefully) fixed "Alt Gr" key not working on some systems
"About" screen rendering has been optimized (big whoop)