Include the following details (edit as applicable):
- Platform category: Scaniverse Web
- Device Model & OS version: Gopro Max 2
- NSDK or Scaniverse version: Unknown
I am experiencing a scan render failure for a 360 scan uploaded to Scaniverse. The video uploaded successfully and passed the initial validation checks, but asset generation fails immediately with a generic “Failed” status and no further error details.
Environment & Context Details:
Hardware, File & Export Settings:
-
Camera: GoPro Max 2 360 camera
-
File Size: 5.81 GB
-
Container / Format: MPEG-4
-
Export Codec: HEVC
-
Resolution: 8K
-
Framerate: 29.97 FPS
-
Telemetry: Checked “Retain GPMF data for GPS telemetry” during export.
Steps to Reproduce:
-
Uploaded the 5.81 GB GS029221.mp4 file to the “Kingston Drive” site.
-
The initial pipeline completed with a “Passed” status.
-
Triggered asset generation.
-
The processing failed, returning an explicit status of “Failed” without any descriptive logs or error codes.
Hi Daniel,
Thanks for reaching out let me check in with our team that handles 360 scans and I’ll get back to you. I see you’ve set some of the settings to what’s required per our documentation but are you also able to confirm that you’ve set the settings below:
- Adaptive tone: Off
- Video encoding: H.265
- Color: Standard
- Sharpness: Low
Hi Daniel,
Could you please provide a link to one of the failed splats so that someone on our team can have a look at it?
Hi Daniel,
Some folks on our Scaniverse team took a look, and it appears that the 360 video was taken from a vehicle. Unfortunately this is not a supported use case, but we are looking to see if we’re able to process it for you on our end. I can’t make any guarantees, but I’ll keep you posted.
Hi Daniel,
Unfortunately we were unable to get it processed on our end either so the scan creation from a moving vehicle is currently unsupported. May I know the use case for scanning a 360 video from a car? Was it mainly due to wanting to scan a large area at a faster pace?
Thanks for trying - I actually was able to get a :30 video to render, you can see the results in one of the Kingston Drive 2 (if you can access)
The use case is generically wanting to be able to create a splat tour over a larger area
Ah got it. So then the vehicle in this case is mainly just for convenience correct? Since the recommended way would be to walk an area to generate a scan but since you’re trying to create a tour of a large area I can definitely see how walking wouldn’t be as feasible.
Another thing that came to mind and thought I’d recommend, although again do note that the vehicle use case is unsupported, is creating the recordings in chunks under the same site and when you’re done generating an asset out of all of them. If you’re having some success with smaller scans then maybe generating an asset built off of multiple small scans as opposed to a single large one could be a potential workaround. I just can’t speak on the quality of the final generated asset or the possibility of this being successful.