And the end results. The numbers in the paper are correct. however: 1) The settings in the paper were lies. This will supposedly be fixed. This was suppose

And the end results. The numbers in the paper are correct. however: 1) The settings in the paper were lies. This will supposedly be fixed. This was supposedly unintentional. 2) The clips are indeed cherry-picked (libvpx loses by ~1db on foreman with the same settings). This may or may be fixed. This was intentional. All other criticisms of the paper are valid.