Re: Rokit 3D phones
toggle quoted messageShow quoted text
The article seems to prefer the depth map pictures. However I find that they all flatten the depth and rather resemble the results from my old Thrill phone which has very close lenses. The anaglyph images conceal a lot of faults. I would want to see actual color images as sBS high resolution copies. The Mickey mouse W3 closeup did not look bad to me. The statue picture had some depth problems which look like poor 3D conversions. The L arm of the statue is too far back, and some of the trees look like paper cutouts. The car photo in the depth map had some strange floating pieces on the hood, also very similar to poor 3D conversions.
Overall I would say that the depth map was a very inferior 3D image. Yes, the cell phone image has better detail, but it is possible to refocus and denoise the W3 images to improve them. It is doubtful that the depth disparities of the depth map images can be easily fixed. Once the depth has been determined by the sensors, only hand editing can fix it.
Ghosting is caused by the viewing system, not the photograph. The author complains of ghosting, but I would submit that viewing an SBS print in a Holmes viewer or in a good VR headset would get rid of the ghosting. The Mickey photo would pose problems if it were displayed on a large screen, so the solution is to have the horizontal alignment adjusted to prevent wide separation of the background. Mickey might then pop out. Ghosting is obvious when you have images with very high contrast between light and dark areas with wide horizontal separation between L&R. That is why the author saw ghosting in the car and Mickey images.
I have seen some fairly good images created by depth maps, but I am sure that there will be a number of random images that have obvious depth artifacts. The truth is that the average viewer would not notice them, but experienced 3D photographers would be disappointed. The only good reason for using the tested 3D phone is convenience.
I found the article discouraging because the author preferred the flattened images with obvious depth artifacts. I suspect if the depth map photos were adjusted to match the W3 depth, the artifacts would be much more obvious. Limiting the depth conceals the problems.
John M. Clement
From: main@Photo-3d.groups.io <main@Photo-3d.groups.io> On Behalf Of mrthreed via groups.io
Sent: Tuesday, March 30, 2021 1:25 AM
Subject: Re: [Photo-3d] Rokit 3D phones
I just found this interesting article comparing 3D photos shot with a Huawei P30 Pro using depth maps and photos shot with a Fuji W1:
-------- Original message --------
From: "robert mcafee via groups.io" <geargod2@...>
Date: 3/29/21 4:37 PM (GMT-09:00)
Subject: [Photo-3d] Rokit 3D phones
I noticed on Rokit's website most models of their phones were sold out. I contacted them to inquire if they were discontinuing the phones or the 3D phones. Below is their response. Perhaps we will see a new model with a stereo camera or at least one that will save images with embedded depth maps