Attribute precision in POPs (e.g. 32b float vs 16b vs 8b attributes)

Hi,

Is there any way to manually define the precision as less than 32 bits on a per attribute basis for POPs? e.g. in the attribute? I see the attributePOP and other create attribute pages have options for int(), float() double() etc. As far as I understand, float() is 32bits and double() is 64, and int() is also 32 bits

Is there any way to define a half() type of 16 bits, for even a POP version of ‘8 bit fixed’? Main use case being optimization of attribute fetches

Hello @watershed

It’s not currently possible to manually define attribute precision to less than 32-bits (such as a 16-bit half or 8-bit fixed type) within POPs. Right now, the available attribute types map to standard 32-bit floats/ints and 64-bit doubles when supported.

Finer-grained control over attribute precision could still arrive in a future update as POPs continue to evolve.

thank you for the information! would love to see as a future feature - believe would allow for great optimization in any cases, especially when working with large geometries and point clouds