On Mon, 2014-02-24 at 13:35 +0100, Ingo wrote:
Sorry,
forgot ta add [change -1] after the [i].
I thought this was meant to be used with a MIDI signal - maybe I got that wrong?
Yes, it is. I'm nit-picking here. The patch you posted before also works, even without the [change -1]. But even adding the [change -1] doesn't address the issues I mentioned. On a fast ramp, it still misses some values and it still suffers from jitter. It's only details I'm talking about here, yes, but since you decided to remove the features from my version, I hoped to be able to illustrate them with words.
Roman
-----Ursprüngliche Nachricht----- Von: pd-list-bounces@iem.at [mailto:pd-list-bounces@iem.at] Im Auftrag von Roman Haefeli Gesendet: Montag, 24. Februar 2014 10:34 An: pd-list@iem.at Betreff: Re: [PD] smooth random numbers
On Sun, 2014-02-23 at 04:20 +0100, Ingo wrote:
Starting from Roman's patch I would probably do it like the attached
patch.
Many ways might solve a certain problem and in Pd those many ways can often be divided into a "subtractive" approach - more than necessary is generated and the overhead is filtered out afterwards - and an "additive" approach - exactly the data needed is generated.
I believe you totally missed the point why I chose the latter here. Using a constant time grain for [line] generates too much data for slow ramps, leading to many duplicates. Attach a print to our patch and you'll see. At the same time it misses some integer numbers for fast ramps. Also, by having a fixed time grain the result looks like a resampled ramp (which it basically is), which means it is jittery and doesn't emulate a steady movement of the fader.
Roman
Pd-list@iem.at mailing list UNSUBSCRIBE and account-management -> http://lists.puredata.info/listinfo/pd-list