Here we go AGAIN with Millerwill posting what he does.
At this point, I have read it so many times I could make his post for him from memory.
Then people will post that using the full panel is a weird scale and introduces artifacts in test patterns and then others will post, real images do not present the extremes of test patterns and there is no real reason that a scale slightly larger than the closest interger scale can't be done just as well.
And then don't use artificial sharpening.
Any upscale involves adding lines (easy) and sharpening (difficult). The sharpening is built into the scaling algs. The Sony scale does ring unlike the Lumagen scale.
Probably the Sony scale is optimized by using a whole number scale, i.e., using the panel at 1.78 width, the normal function0.
But good results can obviously be obtained using the full 1.89 panel width. And those using it, will say in real life its grrrrrreat.
I really don't think it makes very much difference what you do. Do you need the extra light available using 1.89? Its a small amount. And probably as you can see for yourself, 178 width with 1.78 input frames (regardless of the image aspect within such frames) appears a little sharper for whatever reasons.Edited by mark haflich - 10/8/13 at 9:28am