sounds good. I will work on the immediate project, when I have a good function solution I'll isolate that and send it to you and we can pick up the discussion then.
good stuff.
Jorge

Thanks both for all your input on this!
I am always looking for simplify the flow for setting up multi-tiered decision systems, and sort the results. In Pd I have done this by sending up one [random x] to tell me how many (x) times to run the next [random y].
Often I also want [random x] to give some result z that alters the thresholds of the subsequent [random y] while (separately) specifying the number of decisions made at this stage.
The mystery object I used once before streamlined this..but it can be just written out as well (in Pd or Supercollider for example,, from existing objects).

What would be quite interesting is to be able to make x random decisions over some array that could be strings or symbols rather than numbers. Of course numbers could be generated then mapped onto symbols AFTER being generated but even so having an array (while even numeric) that isn't a 'range' (ex. 5, 82, 48, 5.3, 2.9, 9 rather than [random 8] being a range = 0 -> ( 8-1 )
this sort of functionality is what some of the older 'shuffle' objects do.

I am already a big fan of Else. I have another look into these object meanwhile! Thanks so much. It has been interesting to revisit these in search of this elusive vision.

@whale-av said:

list-wrandom

excellent. That isn't the one I used before but is also very interesting for the sort of stuff I am doing!
I also found some others...but still not what I was using before. I need to find the old patches...somewhere.....

also of course (just in case anyone is searching this in the future)

[array random]

[cyclone/urn] (also in nilwind library)

perhaps I dreampt it, I 've done this before.
but I recall the existence of randomisers with sorting functions

you could specify "how many random numbers to generate?"
and things like "which outlet they come out of" (like a random route object)
or do what the [else/rand.list] does above.

I am sure I can make something from that, but it would be interesting to find if these actually exist. If so, I will post them here for the next person who looks for them.

oh wow, yes...how did I miss that one?
Thank you @Porres!

Weirdly, I still think there were some others somewhere...that had different sorting functions. But this will work well for what I am doing.

Cheers!
Jorge

should have tagged @Porres who would know about Cyclone and whose collection of random stuff in Else is amazing (but not for this task I don't think) ...maybe you have some idea on this?

I could swear that I once used several objects which I thought were part of Cyclone that did this and related things:

given several numeric arguments
chose a random selection of these
based on:
the number of chosen values being a random number based on the first argument (functioning as our usual [random] object functions)

example:
[mysteryObject argN valA val B ... valN]

returns: a quantity [random argN-1] of random selection from the list valA valB ... valN

I realise that this is basically a random list operation and I have a version of something similar in Vanilla but I could have sworn I used some really great objects that did this a few years ago (and I could have sworn they were from Cyclone) but perhaps I am totally mistaken?!

anyone know about this?

Hi @Lacuna thanks so much for these suggestions and the links.
Certainly, I am reading the data sheets and documentation.
However, I had missed the forum topic you posted and I will check this github.

I have been also using an old laptop with comport for a lot of projects in fact, comport is an old friend. I guess what I am trying to do, is get the most from these bits of technology rather than multiply them, and especially in a single project. I find myself using both raspberry pi AND arduino because I can manage the audio well on the pi (or well enough) and other stuff on arduino, but really I know the pins on the Pi can do that stuff too, so I just thought I would start to look for perspectives on how other people are doing these things.

I am going to explore your suggestion of doing the PWM in Pd and then sending this across the pins on the Bela to control things, I think this is the sort of very direct approach I was finding myself hazy about.
All the best,
Jorge

hello! this question intersects with Pd even if it is not about Pd specifically...
I currently have several (installation) pieces in which the sound is run from Pd and then I have motors or sensors being run off other things.
This means that in one case I am using a raspberry pi for the Pd part and an Arduino mega for the motor part
In another case I have Pd running on a Bela/BBB and then an Arduino doing the motor stuff.

The reason for this inefficiency is that my knowledge is fairly limited on these platforms. However, it is annoying to have to use the extra gear when I don't have to. Especially because I know that both the raspberry and the Bela have pins that could be used in place of the arduino for this purpose. Since the Bela is a simpler case and the pins can be addressed directly from Pure Data, I will ask about that and leave the Raspberry behind for now (as I think I would end up running Pd AND Arduino on that board anyway, whereas with Bela, one can easily address the pins from inside Pd without any other programs or libraries).

My question is this: does anyone know how the standard mappings of pins Analog/Digital/and PWM pins on the Arduino lines up on the Bela? Does the Bela have have PWM pins like, for example pins 8,9, 10,11 on Arduino Mega or Dueminanove? And if so, what is the best way to address them from Pure Data? Basically, I'd like to convert my already very simple Arduino sketches over to Pd, so that both motors and Audio can run from the same patch.

I know this isn't specifically a Pd issue but I have asked a related question on the Bela forum but didn't get an answer so I hoped someone might know or be able to suggest a resource over here...
I have traversed the Bela documentation already of course....and will continue with that meanwhile.

Jorge

In one of the classic pd books (but not Kreidler, Pucketter, nor Farnell I don't think) there is a great example in which three different approaches to making granular synthesis are given using three different ways of overlapping windowing functions.

Does anyone know where or what this book is?

I was trying to demonstrate something related to this the other day and I remembered these examples but can't for the life of me recall where I saw them.

yes, the OSC option is definitely a good one (though I wonder what a good program for this is on android...I have to check if 'control' does this...there is osculator on mac stuff but it is paid...in fact the libpd thread started just before this one (https://forum.pdpatchrepo.info/topic/13969/questions-about-using-lib-pd) is also helpful...

I am interested not so much in grabbing sensor data and sending it to pd on a computer but rather somehow using Pd directly on a mobile platform, and getting sensor data and particular gps data directly in there. There was also a post today on the Pd Forum about webpd being rebuilt, so that could help with that (you will be able to run vanilla Pd in a browser if that happens)

MobMuPlat grabs data from several of these sensors very well already...but you have to deal with the pd patch +java interface with that, and also the limitation of the 32-bit floating point numbers...so I am looking for other options. Especially because of the 32 bit floating point number issue as the GPS values that are interesting to me 6-7 digits long. Hopefully webpd will be 64 bit.