Exactly right. RS in crop mode is the same for 1080 as 4K - in fact there’s virtually no distinguishing the images. The FF is a different story, that typical mush you describe. A found very good test charts online here that illustrate the effects of Canon’s two traditional sampling methods in the R - this is how it’s always been for their stills bodies. Have they developed a new method for the R Pro bodies? Or will it be the same old story? Here are those charts:
Just to add a more easily digestible, more immediately obvious way to show what's happening, here's the chart shot from the video in high-res.
Top to bottom: 1080p full sensor width, Video crop mode 1080p, then 4k on the bottom. 4k is scaled down to 1080p and exported to DNxHR HQX before being put next to the other footage to make it fair. The comparison is then rendered to a 4k timeline to remove the bottleneck there. Staggering how much of a difference there is between 1080p FF and 1080p crop. The jump from 1080p crop to 4k is nominal, comparatively.
Sony A7S III wish list and the not-NAB 2020 no-show today
In: Cameras
Posted
Just to add a more easily digestible, more immediately obvious way to show what's happening, here's the chart shot from the video in high-res.
Top to bottom: 1080p full sensor width, Video crop mode 1080p, then 4k on the bottom. 4k is scaled down to 1080p and exported to DNxHR HQX before being put next to the other footage to make it fair. The comparison is then rendered to a 4k timeline to remove the bottleneck there. Staggering how much of a difference there is between 1080p FF and 1080p crop. The jump from 1080p crop to 4k is nominal, comparatively.