Charles Henry wrote:
Most Pd objects (externals) use t_sample to define what gets passed to input and output. At compilation time, the externals code includes m_pd.h, which defines t_sample as a float. Which makes sense on 32-bit processors--Pd for 64-bit processors could potentially redefine t_sample as a double, with no loss in performance (with nearly twice as much memory usage).
that's the theory. in practice, even Pd-vanilla was using a wild mix of t_sample, t_float & float as sample-type until 0.41 (when i submitted a number of patches to clean that up)
i have a basically running version of Pd with 64bit double precision, basically only some oscillators have to be ported (osc~, phasor~,... use lowest level bitmanipulation in their core-routines) playing soundfiles works fine :-)
fgmadsr IOhannes