What sets of transforms between normalized integer pixel values and floats [0..1] produces least error w/ different-sized integer endpoints?
-
-
@cmuratori@pkhuong I'd prefer not to get Bloom filtered. cc@rygorous -
@nothings@cmuratori@pkhuong Graphics APIs (at this point) have float as canonical repr and use round on float->int. For image resizing, - Show replies
New conversation -
-
-
@cmuratori@nothings@pkhuong There's this old Inner product column from@Jonathan_Blow: https://web.archive.org/web/20140220224053/http://number-none.com/product/Scalar%20Quantization/index.html … -
@pervognsen@cmuratori@pkhuong@Jonathan_Blow Yeah, here's the PDF version: https://twitter.com/nlguillemot/status/501488494045585408 …This Tweet is unavailable.
End of conversation
New conversation -
Loading seems to be taking a while.
Twitter may be over capacity or experiencing a momentary hiccup. Try again or visit Twitter Status for more information.