[PD] More and more

thewade pdman at aproximation.org
Thu Apr 21 05:53:51 CEST 2005

Hi list!

I got [symbol ,] to output a comma with backslashes (that dont behave
like commas in messages do). Then I saved, closed and re-opend the
patch and the backslash comma\'s were converted into real commas. Not
perfect, but it almost works. Thanks for the help BTW!

I have another question about fft. fft should output two sample blocks
each cycle: a real analysis of the signal in frequency bins, and an
imaginary analysis of the signal (I dont really understand real versus
imaginary, but anyway...)
If I multiply some osc~\'s togeather then that sample block should be
basicaly 1\'s and 0\'s, right? 1\'s in the bins representing the
frequencies of the osc~...
Now if I multiply some fft output block resulting from processing a
sound file, and this osc~ generated fft sample block, and pass the
result through ifft then the resultant signal should only contain
the frequencies common to the two input signals, right? how come
this doesnt work for me?

Lastly, I am using an AMD64 and the tables are all messed up.
Guenter explained the problem

> Array don\'t store floats but unions of t_word, which are 64 bit broad
> on 64 bit architectures.
> - One solution is to store doubles instead of float and change all the
> types of the temporary variables in all objects accessing array to
> double. (externals using float arrays have to be adapted too)
> - Another solution is to ignore the t_word type and store all values
> as floats in arrays (and read them as floats like it is now)
> - We could implement a special iterator or access function to float
> arrays (instead of ++ or []) and change all the occurrencers of all
> float array objects according to this. (Breaks all float array externals).

and said he would post a hacky fix. Has this been done? What will work
and what wont with this workaround? What externals will break if this
problem were solved by method one? Could we make some special header
that these externals read when compiled that would allow them do both
union t_word and double, depending on arch?
Its quite crippling not to be able to use a whole array, and Im not
sure the half array fixes I have been doing work quite right either...

Thanks for all the help, one and all. When I become a rock star I will
kick you all down some dough! :)
I am sooo going to be a rock star. I have the spand-x and everything!

More information about the Pd-list mailing list