Modiirq Listed as deprecated but no replacement exists

iir should replace in most cases.  It doesn’t have audio rate modulation of the cutoff, being the main difference.  Modulating biquad’s quickly can result in instability of the filters per https://music-dsp.music.columbia.narkive.com/r5TJ3GcJ/help-for-biquad-design-needed#post2, so the loss of audio rate modulation may not be a bad thing.

I’m curious what you are using iir for.  Generally the instability of iir filters at low frequencies due to truncation errors makes them a challenge, though admittedly this should be better with the floating point of the H9000, than the fixed point of the H8000, depending on how iir has been implemented, and how closely it matches the H8000.

Yeah, if you are using the audio mod input, for your cutoff modulation, then iir wouldn’t work. You’d need to re-engineer to use the frequency controller input, or switch across to another filter.  I hear you in re-engineering big algorithms. It’s a big investment of time and brain power.

The SVF filters (h3000svf, modfilter or modfiltp) can be cascaded to provide multiple pole options, but it’s nowhere near as simple as the modiirq, as you’ll need to wire in your own audio path controls, which get complicated quickly, and you’ll need to multiply your lfo output by the frequency range you’ll need as the mod input is -20000 to 20000 Hz rather than 0-1 from 0 to nyquist.  I’ve had to do a similar thing for selecting the number of bands in a vocoder recently, and happy to send that bit of the graph if you ever get the energy to reengineer.  You’d need to swap the multiply blocks for switch blocks to cascade filters rather than just turn mixer channels on and off, but will probably still save a bit of time.

Maybe get in touch with Eventide to see if they can provide a replacement in future updates, though for audio applications, the wind is definitely starting to point more in the direction of trapezoidal state variable filters, so not sue how motivated they will be.  Maybe at the very least they can keep the modiirq on the deprecated list indefinitely.

Regarding bringing the processor to it’s knees, I’d love to see a CPU meter for the H9000, so we know how far we are pushing things with our algorithms, though 64 x 6 pole filters seems a decent benchmark.  My 32 channel vocoder that I’m working on must be chewing into available CPU pretty heavily as well.  It has 66 filters, though they are svf and only 2 pole

I’ve responded to Patchen directly, but just so depreciated modules system is clear for other users:

We have depreciated the modiirq module and have not yet provided a replacement for it. The module will show up as crossed out in VSig but it is still available to use.

The module will continue to work up to and beyond the point where we provide a replacement, so you don’t need to worry about reworking your sigfiles when the modules are depreciated.