Hi,
I'm trying to get an upward moving straight line (a phasor as the subject indicates) from an [osc~].
I've built a patch that does this, but the frequency is rather restricted. Here it is in ASCII

[0\ <--frequency
|\
| \
| \
| \
| [expr 512 / pow(2, ($f1 - 1))]
| |
| [set $1(
| |
| [block~] <--set block size so you get bangs with [bang~] at an appropriate rate
|
[epxr 10.75 * pow(2, ($f1 - 1))]
|
| [bang~] <--bang so you get 1/4 of the sine
| |
| [0.5(
| |
[osc~]
|
[expr~ ((asin($v1) / pi) + 0.5) * 4]

Any other ideas as to be able to get float number frequencies? Cause now the decimal numbers in between all integers get higher that 1.

Thanks,
Alexandros

Actually the numbers are there this way cause I'm sending the result to a table with the size of 512, if this gives any clue. It's quite possible that it's rather messy cause I don't get the results I want. I'm trying to use it as a [phasor~] to send it to a [tabread4~]. For now I get sawtooths from 0-1 for the frequency numbers 1-4 (integers only).
Anyone out there has something to propose?

I would imagine you're likely to run into problems because the block size has to be an integer power-of-two, so non-integer frequencies won't work.

I'm also not really familiar with the method you're using, so I don't know if the math is right. But I do know of an alternate method. You can use the Hilbert-transform to convert a real signal into a complex one. The imaginary part will be pi/2 radians out-of-phase with the real one. Then, you can convert the signal to it's polar form to get the instantaneous amplitude and phase (this assumes the input is a sinusoid, which you are doing).

Pd comes with the [hilbert~] abstraction that does the conversion from real to complex for you, though it's not perfect. IIRC, the Hilbert-transform isn't really ideal for real-time, so this abstraction approximates it by sending the input through IIR all-pass filters with phase responses that are about pi/2 radians apart for most frequencies. It gets a little less accurate in the low and high extremes, but they're fairly close for much of the frequency range.

I've attached a patch that illustrates. Again, it's noticeably curvy in the lower frequencies, but it becomes straighter as you move up.

http://www.pdpatchrepo.info/hurleur/osc2phase.pd

i didn't test this thoroughly, but i think it does what it should

http://www.pdpatchrepo.info/hurleur/osc2phasor.pd

seems to work fine at all frequencies. here's a test patch converting the phasor~ back into an osc~ and then overlaying that onto the original osc~

http://www.pdpatchrepo.info/hurleur/osc2phasor-test2.pd

Thanks to both of you guys. Mod's version seems to work fine for all frequencies indeed. Thanks!