@seb-harmonik-ar and @jameslo Thank you both very much for the suggestions! @jameslo I cannot thank you enough for taking the time to dig in and debug the project. I was concerned about that [line] element instantaneously jumping to 1 as well, but wasn't sure how to approach it since ramping the attack didn't seem to make a difference for me. Also did not know that it should've been a [line~].

On first listen I'm still having issues, but I have a lot more to chew on now. I'll play around for a bit and recompile with webpd to make sure it's not just my setup. The abstracted version you provided is much easier to experiment with.

Problematic base patch: calm-generative.pd

I received the base patch above from a friendly stranger online and built upon it, adding vline~ envelopes to each tone to emulate bell strikes and routing the output through a stereo feedback delay network. However, I cannot seem to get rid of pops and clicks that make for a sort of zipper sound in the output. I've attached the original base patch, because that seems to be where the issue is resides.

I initially suspected clipping, but I've tried reducing tone amplitudes with *~ and modifying line parameters to no avail. Changing Audio Settings -> Latency to 50 ms, 100 ms, or even 1000 ms doesn't have an effect either.

The patch is super sensitive to graphics changes for some reason, and it glitches out badly in edit mode when I move my cursor. I've always attributed that to a bug in my setup since I'm running Pd on a Debian Linux machine with PipeWire managing the audio. I had issues with native JACK as well. It's not a big deal, because the final version of the patch is a web app. But it might be relevant to debugging.

My biggest issue is that when compile the patch to WebAssembly using WebPd, the pops and clicks persist. It's actually worse in the web version. The feedback delay network can amplify the pops to a point at which they escape containment and won't even disappear with full dampening. This makes the patch unlistenable on web.

Any ideas?

I'm still a noob, so maybe I'm missing something obvious.

Below are some images of the patch. It's basically a bunch of LFO's driving each other to pseudo-randomly trigger sine waves at set frequencies. There are 14 total possible tones. The attached patch only has 8 of them connected to the dac~ output.

Single tone element with one LFO driver:

Full patch: