I've always been under the impression that [sqrt~] was more precise than [q8_sqrt~]The help file seems to reflect this. (saying sqrt~ is 20 bits accuracy, but q8_sqrt~ is only 8) However, the source in d_math.c seems to indicate that they are actually the same class/perform method:
sigsqrt_class = class_new(gensym("sqrt~"), (t_newmethod)sigsqrt_new, 0, sizeof(t_sigsqrt), 0, 0);class_addcreator(sigsqrt_new, gensym("q8_sqrt~"), 0); am I missing something, or is the documentation incorrect about them being different?
Em qua., 14 de dez. de 2022 às 05:50, Sebastian Shader via Pd-dev < pd-dev@lists.iem.at> escreveu:
the source in d_math.c seems to indicate that they are actually the same class/perform (,,,) is the documentation incorrect about them being different?
yup, I can confirm they're the same thing. Code must have changed at some point. I wonder when/how that happened.
cheers
this goes back to version 0,33! changed somewhere before but miller's site has a gap before 0.33 in 'older version'
Em qua., 14 de dez. de 2022 às 08:47, Alexandre Torres Porres < porres@gmail.com> escreveu:
Em qua., 14 de dez. de 2022 às 05:50, Sebastian Shader via Pd-dev < pd-dev@lists.iem.at> escreveu:
the source in d_math.c seems to indicate that they are actually the same class/perform (,,,) is the documentation incorrect about them being different?
yup, I can confirm they're the same thing. Code must have changed at some point. I wonder when/how that happened.
cheers
anyway, I am changing the documentation to say these used to be cheaper objects still here for backwards compatibility, but now they're the same.
I am also changing audio examples that use the older version for no good reason.
Hopefully this makes it into the final 53-1 release about to come out anytime soon
Em qua., 14 de dez. de 2022 às 09:06, Alexandre Torres Porres < porres@gmail.com> escreveu:
this goes back to version 0,33! changed somewhere before but miller's site has a gap before 0.33 in 'older version'
Em qua., 14 de dez. de 2022 às 08:47, Alexandre Torres Porres < porres@gmail.com> escreveu:
Em qua., 14 de dez. de 2022 às 05:50, Sebastian Shader via Pd-dev < pd-dev@lists.iem.at> escreveu:
the source in d_math.c seems to indicate that they are actually the same class/perform (,,,) is the documentation incorrect about them being different?
yup, I can confirm they're the same thing. Code must have changed at some point. I wonder when/how that happened.
cheers
0.53-1 is out already :) I thought the test phase was still on, hahaha
well, I opened an issue and tried to propose a change, but I'm now confused as Miller is saying we don't have a proper [sqrt~], see https://github.com/pure-data/pddp/issues/125 but as I see it [sqrt~] works well as expected.
I should revert my changes anyway and wait to see what happens.
cheers
Em qua., 14 de dez. de 2022 às 14:12, Alexandre Torres Porres < porres@gmail.com> escreveu:
anyway, I am changing the documentation to say these used to be cheaper objects still here for backwards compatibility, but now they're the same.
I am also changing audio examples that use the older version for no good reason.
Hopefully this makes it into the final 53-1 release about to come out anytime soon
Em qua., 14 de dez. de 2022 às 09:06, Alexandre Torres Porres < porres@gmail.com> escreveu:
this goes back to version 0,33! changed somewhere before but miller's site has a gap before 0.33 in 'older version'
Em qua., 14 de dez. de 2022 às 08:47, Alexandre Torres Porres < porres@gmail.com> escreveu:
Em qua., 14 de dez. de 2022 às 05:50, Sebastian Shader via Pd-dev < pd-dev@lists.iem.at> escreveu:
the source in d_math.c seems to indicate that they are actually the same class/perform (,,,) is the documentation incorrect about them being different?
yup, I can confirm they're the same thing. Code must have changed at some point. I wonder when/how that happened.
cheers
-
Alexandre Torres Porres -
Sebastian Shader