Hi Hans,
it means that Pd can be compiled -DPD_FLOATSIZE=64, so it will use 64-bit doubles instead of 32-bit floats everywhere, and all built-in objects should still work as expected.
This has advantages (better precision/stability for certain filters, large precise table indicies, etc.) and possible disadvantages (needs more memory, some code paths might be a bit slower).
Currently, we only ship the good old single precision builds, but people can compile double precision builds from source for themselves to experiment with.
Christof
On 29.05.2020 08:58, hans w. koch wrote:
excited about that release, thanks to all involved!
- maybe mention that Pd is now double precision ready?
what does that exactly mean? is it in, but not enabled or is it a subset? ( certainly its not a marketing scam as these tv sets labeled “HD ready” :-)
thanks for clarification,
hans