Encodes vector using 32 bits of precision per sample in IEEE floating point format.
The number of bytes required to encode a scalar in this format. A vector will nominally require dimension * byteSize bytes of storage.