This issue has been closed as 'Won't Fix' due to an extended period of time without updates. If this issue is important to you please let us know by posting on the AnswerHub or UDN, and Epic will re-open the ticket for further review.
The new SceneCapture2d feature 'SceneDepth in R' is not rendering the scenes depth in both Perspective and Orthographic projections as expected.
The user originally mentioned he was using Orthographic, and I assumed that this would not be able to capture scene depth since Orthographic projections are flat. Upon further investigation I was able to reproduce the results the user was reporting, and was unable to determine what the expected outcome should be.
This also reproduced in //UE4/Release-4.14 - 3160120
1. Download and Open attached project.
2. Open the 'DepthRT' texture render target and isolate the Red channel.
3. Select the SceneCapture2D actor and move it up and down while observing the texture render target.
Outcome There are different parts of the chair that seem to render with a small gradient as if it were capturing scene depth.
Expected The entire captured area renders with a scene depth gradient, much like the buffer visualization.
Note I tested this in Perspective and Orthographic projection modes and both produce unexpected results.
Head over to the existingAnswerHub thread and let us know what's up.