Playing the Spears & Munsil HD Benchmark 2nd Edition from my UDP-203 to an LG OLEDC9, I'm finding that the UDP-203 appears to be applying some type of bi-cubic upscaling to 1080p source material. Alternating pixels in 1080p result in one bright pixel, followed by 2 gray, 1 black, 2 gray, 1 bright, and then repeating, both horizontally and vertically in 4k. When using Source Direct output and allowing the LG C9 to do the upscaling the alternating pixels are displayed as a block of 2 bright pixels, 2 dark, 2 bright, and then repeating which creates a 2 x 2 block of pixels to represent each 1080p pixel. While this sounds like it would produce smoother results with the UDP-203 performing the upscaling, the LG C9 applies anti-aliasing when upscaling (with its sharpness control set at a setting of 10 which applies no other edge enhancement) which provides considerably smoother diagonals. Diagonals on the S&M 1080p scaling pattern look tremendously cleaner with the LG C9 performing the upscaling. The same applies to the S&M 720p scaling pattern, but the UDP-203 does not appear to apply the bi-cubic upscaling in this case. Each 720p pixel is represented by a 3x3 block of pixels in 4k regardless of which unit is upscaling.
However, the highest frequency chroma bursts appear muted when the LG C9 is performing the upscaling (see the multiburst patterns and the Color Space Evaluation pattern), while they appear to be perfect with the UDP-203 performing the upscaling. However, the with the LG performing the upscaling and applying anti-aliasing, the diagonal chroma bursts are obviously smoother. With the LG performing the upscaling, banding appears in the color ramps which is absent with the UDP-203 performing the upscaling.
So for the score:
Smooth jaggie-free diagonals: LG C9
High-frequency chroma information: UDP-203
Smooth band free color ramps: UDP-203
Smooth jaggie-free chroma burst diagonals: LG C9
This would seem to be a tie, if all the above factors are weighed equally. This leaves the difference between what appears to be bicubic upscaling vs. ratio integer scaling on 1080p sources. I'm really not sure which to consider superior here. Bicubic is certainly considered preferable when applied to up-SAMPLING chroma to be added back to luma, but I'm not sure if the same applies to pure up-scaling.
Ultimately the only difference I really notice on movies and regular viewing content is that video or computer generated titles that contain diagonal lines usually (but not always) look considerably smoother with the LG C9 doing the upscaling. I would like an objective conclusion of which is better.
Last edited by KC-Technerd; 10-21-2019 at 08:22 PM.