@jameslo I wonder if that's much different from DX-style fm. I know linear fm is technically possible with analog circuitry if it's designed to implement the equivalent of negative frequencies. I'd imagine it would have different results to lookup table FM. MY tx81z is at the bottom of a big box, so I can't test

The feedback routing I had was based on the poly setup of a 4op patch, I forget where I heard previously that send and receive was the right way to implement feedback, but I'm pretty sure that's the problem. What I don't understand is why s~ and r~ don't seam to be affected by block~, although i'm not sure. Was using an Array object as a buffer something you've seen for the purpose before?

That would be cool. Messing with libpd was one of the many untouch things on my projects list. VCV's code is really well laid-out and readable too. I know I'm too distracted probably get to it, but I bet someone will.

@jameslo I am about to dl and check it out now. One thing though, I think I did have the modulator as the feedback operator, it's a bit hard to see because the feedback goes through inlets and outlets in the operator subpatch (the top one), and then the send and receives are outside that subpatch. Edit, you saw that. Yes having it post-attenuation and post-adsr was a change I just made, but it wasn't better before when I had it coming straigt from cos~. The thing is you are suposed to be able to get a decent saw just with the modulator so I wonder if it's something about how I was using block~ or the way I had the feedback routed outside the patch..

The switch to get the direct signal from the modulator is because as that video was demonstrating you are supposed to be able to get a saw just from the modulator alone with feedback, and a better saw from the carrier. Either way my patch didn't sound right.. I will check your patch out,,

Edit2: Yours definitely creates a saw perfectly. I noticed you defined a feedback buffer specifically instead of just using sends and receives, I think that or maybe something about the subpatch structure I had was meesing up my block size? I will have to look when I have more time later,,

Hi, I have been working for a while on a 4 op FM synth, but eventually I noticed I saw not able to get some sounds I expected to get despite any setting I had tried. I know feedback is essential and I have that implemented with send and receives, as well as a block~ 1 setting. Still I seem to be missing something.

I attached a simplified test patch, with just one operator fed onto another, the modulator is capable of feedback and the carrier is sent to the dac. I looked around and this video:

describes getting saw sounds from a DX7, but replicating those settings still doesn't work.

Anyone have an idea of what I'm missing? I included a 2op test patch.

Thanks
fmtest.pd

Ok, that solved it. thanks : )

I'm using Purr Data and noticed I can't create an expr~ object with a condition although it works with expr.
Has anyone had this problem? The documentation seems to suggest that it would work...
The expression I tried was: [ expr~ if($f1 >= 0.5, 1-$f1, $f1) ]

Hi, I just thought I'd ask about the trouble I'm having getting poly to work as I expected.

I am trying to get polyphonic keyboard input working using poly. I previously had some other solution, but I thought poly and route would be the proper way.. In my patch, I have the note and on/off values packed together which seems correct to me, but the output of poly shows it's cycling through notes on evrey input even when the midi note is the same.

I'm stumped. If anyone could let me know where I went wrong I'd appreciate it.

Thanks
key.pd

Hmm, yes that didn't seem to work,

I noticed that in the download from the pd externals page, there is a separate folder for the mp3 stuff with the source, so perhaps I could try to compile it myself, but I probably won't end up doing that.

I don't think I have the same readme's that you mentioned (I have the dekken installed version 1.7, and the 1.7 zip from the pd site). Shame, because it would be a good way to have a large library of source material for processing sounds, but I suspect it just isn't compiled to work on current windows machines...

I'm wondering if any windows users have had success with any mp3 playback externals?
I tried mp3play~ in iemlib which wouldn't create, and also readanysf~ which also wouldn't create.
I think I followed the installation instructions correctly and it's more likely that it just isn't working.

thanks

It seems like the libraries are loading now, I have cyclone working in 0.47 by adding it's subfolder to the path menu. It still says "cyclone: can't load library" in the console if I have cyclone as a startup entry, but if I remove it the error goes away although the library still loads. The other libraries seem to be loading also, except for zexy which seems to need the startup entry.

I had already tried that.

I recently uninstalled and removed everything, then built pd 0.47-0 from the source and tried again, but I am getting the same result. Zexy works but cyclone it says "can't load library"

Just an update: I had actually installed pd-zexy via the linux terminal earlier, so I removed it from the pd-externals folder that deken was installing to and it still works. I then removed everything from the pd-externals folder, but when I open pd I still have zexy loading fine and cyclone unable to load, although instead of "can't open", it now just says: "cyclone: can't load library"

It is the same for iemlib and ggee. The only libraries that seem to work for me are zexy and bassemu~
Still no luck.

I am setting up pd vanilla to be compatible with some old pd-extended patches.

I got the deken plugin to work and installed zexy and cyclone, I added them to the startup options and zexy loads, but cyclone does not.
I just get the message "cyclone:can't open" in the pd console
I also tried installing with apt-get (I am on linux), I'm not sure what to try next.

Thanks, any help is much appreciated!

hi,

I uninstalled my old version of pd-extended and installed the newest stable release (4.25) but I cannot get it to run. Ive tried uninstalling, rebooting and installing again, etc, but when ever I try to run the program I just get this error message:

Error in startup script

no such variable

(read trace on "::env(AppData)")
invoked from within
"file normalize $::env(AppData)/Pd"
("1" arm line 1)
invoked from within
"switch $pd_nt {
0 {set ::pd_path

}
1 {set ::pd_path ..."
(file "c:\Prograam Files\pd\bin\pd.tk" line 278)

I am running win7 32 bit. I tried installing the latest beta release and that worked but has very limited functionality it seems and didn't load most of the objects in the patch I am working on. I could try an older version, but I was having some bugs with GEM so I'd like to be using the latest version.

Sorry to bring this kind of problem up on this forum and let me know if it is not the type of problem addressed here usually, it does seem pretty incomprehensable at least to me, but maybe someone here would recognize this error or hopefully some developers might know of a fix?

thanks very much!!!

Hello. I put together this guitar patch for live effects and some recording capability.
I quite like it, and the sounds that I can get, but its usability is really limited by latency which by me estimate is close to half a second.

I am running it on a dell inspiron 1420, which was a cheap laptop even when it was released about 4 years ago... So I was wondering, especially since I am planning to buy a new (laptop) computer, if it is even realistic to run this patch at usable latencies (>50ms or so). and if so, how "good" a computer would have to be. I attached the patch...

Thanks in advance!

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

hi,

I put together a guitar effect patch to play live into with an added 3 track recorder.

I anticipated that pd isn't really built to function as a DAW, but I only needed some capability to sketch things out, so I've only been using approx. 15 second length arrays.

even still, when I try to save a patch with recorded material, however short, it takes forever. something like 5 seconds for each second recorded.

the sliders I've got to set the array length can go up to 60 seconds worth, but that would be completely unusable. opening a patch takes a while too, but it's saving that's really slow.

I read on this forum that using tables is faster, and it seemed to work perfectly, but it seems you cant save data in a table, even though the graph inside, does have a "save contents box" when I reopen a patch, the recordings are gone and the box is unchecked with the graphs back to 100 samples.

anyway, is there any way around this? thanks a lot in advance!

is there a variation out there of the toggle object, where the "face" of the toggle turns completely opaque instead of just displaying an X?

I use the toggle just as an indicator alot of the time, so anything that is out there like this would help

Ive heard of people loading they're own dsp code onto Analog Devices chips,kind of like a homemade non-patchable nord micromod kind of thing. but I thought it would be amazing be able to run a realtime puredata or max type environment from hardware dsp chips. with just a few chips, if it could be realtime and computer interfaced, it would be insanely powerful compared to a nord modular.