.PLY File Format – Data encodings

Order of the data

Nothing marks where one element ends and the next begins. A reader has to read, or step over, every property of every item, which the header makes possible.

ascii

Values are written as decimal numbers separated by white space. By convention each item takes one line, so a vertex is one line and a face is one line, but readers should treat any white space alike. Numbers use the C locale: a full stop as the decimal point and no thousands separators. Floating-point values may use an exponent, as in 1.5e-3.

ply
format ascii 1.0
element vertex 3
property float x
property float y
property float z
element face 1
property list uchar int vertex_indices
end_header
0 0 0
1 0 0
0 1 0
3 0 1 2

binary_little_endian and binary_big_endian

Each value is stored in the size of its type, packed one after another with no separators, no padding and no alignment. Integers are two’s complement and floating-point values are IEEE 754. The two binary encodings differ only in byte order: binary_little_endian stores the least significant byte first, as x86 and ARM processors do, and binary_big_endian the most significant byte first.

A list stores its count in the count type, then its items in the item type, so the size of an item of an element with a list property varies from one item to the next.

Worked example

The same triangle as above in binary_little_endian. Its header is 169 bytes long, ending with the line feed after end_header, so the data starts at offset 169 (A9) and the file is 218 bytes.

ply
format binary_little_endian 1.0
element vertex 3
property float x
property float y
property float z
element face 1
property list uchar int vertex_indices
end_header

In binary_big_endian the bytes of each value are reversed: 1.0 becomes 3F 80 00 00 and vertex number 1 becomes 00 00 00 01. The header is three bytes shorter, 166 bytes, since big has three letters fewer than little.