Increasing performance

Feature Request
1

I think I found a method to dramatically increase performance on low-end smartphones (like mine…): separate sampling from processing.

I think that currently a part of the processing is performed in realtime while capturing data; but on low-end smartphones this leads to data loss and frames loss! Instead, if there was an option to separate first-stage processing from scanning, full processing power could be dedicated to record frames, accelerometer data and gyroscope data, preventing any data loss.

Once the data are collected, than first-stage processing can start, to build the point cloud.

Then, the second-stage processing for calculating splats can be started.

(I stil have to understand if third-stage processing is VR optimization or something else)

This would also allow selecting splat/mesh processing type after the scanning, rather than in advance.

1 Like
2

this sounds like a good idea. the processing part still leads to resolution loss in longer scans

1 Like
3

Hello,

Thanks for your feedback. I’ll pass it along to our team but please do note that we cannot guaranty if/when a change will be implemented.

2 Likes
4

Yes. The preview processing in real-time is useful on iPhones but they are badly implemented on android to the point that it worsens visualization for longer scans.

3 Likes