dannyflint

hi, i have typed the following into an expr object;

my problem is that epr doesnt seem to accept it? it keeps rounding the number up to 0.577623?

my problem is that epr doesnt seem to accept it?    it keeps rounding the number up to 0.577623?

expr-is-rounding-up

Search the forum for the word "float" and click on the first link provided by the search engine.

Thanks for replying Christopher, but what i don't understand is;

The inverse of that equation (Which is provided in the mtof maxhelp) allows;

[expr (440. *exp(.057762265 *($f1-69.)))]

it is because of the space behind it. if you want to represent a 64 bit value (or any other symbol which exceeds 9 digits) in an expr because you think it looks nice, you need to keep it in brackets.

(69+(log($f1/440) / (0.57762265)) should work. but of course, only the 32 bit value is parsend inside.

the ".057762265" with the missing leading 0 seems to be a special case. :)

thanks roman, its not cos i think it looks nice .... it's purely because i am not getting the results i expected from the equation, so i presumed that it was because expr was rounding it up.

Its for frequency to midi number conversion.

i know i can do it with sigmund etc. but i want to understand the maths behind the conversion, and learn to use the expr object better

Its for  frequency to midi number conversion.

i know i can do it with sigmund etc. but i want to understand the maths behind the conversion, and learn to use the expr object better


The number at the bottom, should be the same as the int straight after the kslider?

well, if you reverse an expression you will always find that the result does not match the input 100%, simply because of the limited precision of computer programs.

with things like log() the difference can look quite extreme.

dont forget that in a real life situation you would not do it like in your last patch, you would take the linear note number value right from the input instead of converting it forth and back.

the bigger problem in practice is something else ... there are at least 3 different ways how to calculate midi to frequency (using mtof is one of them) ... and so you must be careful not to mix them within one synth or app.

well, if you reverse an expression you will always find that the result does not match the input 100%, simply because of the limited precision of computer programs. 

in the given example, you could precalculate 

 in a pocket calc program or mathlab, and then do 

in the given example, you could precalculate 1./0.57762265 and 1./440. in a pocket calc program or mathlab, and then do *x in max expr.


expr is rounding up?