Math/exp and /log confusion

[floodcomics]

When using the math/exp and math/log objects I've noticed some strange behaviour:

A dial with a value of 0.00 running through /exp outputs 0.25, and at 64.00 it outputs 63.83. Some sort of rounding error?

Similarly, with /log, a 0.00 input dial outputs a value below 0.00 or some other invalid output (0.00 in red).

If I put a positive inverted, either through "inv_pos" object or math *(-1)+64, into /log it will work as intended, until it reaches an input value of 64.00, when it will snap back to 64.00 (in red) rather than 0.00.

Whats going on here?
I'm awful at maths, please forgive me if I'm just missing something super basic.

Example patch:
math_wtf.axp (4.4 KB)

[jaffasplaffa]

For the log object you can try my version. I made a version that when set to 0 it actually outputs 0:
jaffa/math/log2

When log is set to 0, I think it will actually output the lowest possible value, which is why you get a red number, the number is out of scale of the disp/dial p. You can try connecting the Technobears "tb/disp/print f" object to the output of log, to get a value in the log window, to see what you ar working with.

[floodcomics]

/log2 does indeed fix the sub 0 output! Thanks!
"/print f" gives me -144 rather than 0 when using the original /log.

But the wrapping around/snapping problem (when the dial is first inverted) persists also with your /log2 object.
"/print f" gives me 872 at 64.00 (same for original /log)

Tested this on both my axoloti cores, direct USB and via USB hub.

[jaffasplaffa]

Is it essential to invert the dial before the log function?

You could try inverting after the log function, I think that would give the right result.

[floodcomics]

No, I don't think it's essential. I'll test it when I'm home.
Thx!