Hi

I'm having a problem creating some of the vcf filters from this library. I can create all of the band pass and some of the low pass filters [vcf_lp2], [vcf_lp6] but not [vcf_lp4] or [vcf_lp8]. However, none of the [vcf_hp] objects instantiate.

I tried the svf~ filter aswell, but the arguments to set filter type with messages doesn't work and throw an error in the post window. So I tried iemlib set of filters.

thanks!

Locate them in your filesystem and add that directory to your paths.

Or you can specify a path in the object like [/iemlib/vcf_lp2~]

a.

thanks obiwannabe -

In fact since writing that email i discovered that the vcf_hp2/hp4/hp6 etc. are abstractions of vcf_filter~. But they don't appear to have the high pass set, so i think I'm gonna have to make them.

Is that unusual for the extended version?

cheers

So they are, I never noticed that before either.

I guess it's not unusual to have object sets that are a mix of externals, abstractions or abstractions that depend on other abstractions that depend on externals Franks list-abs is a very interwoven set of objects, elegant in reuse.

I can't find the help files on my system for iem vcfs. Perhaps you can email Thomas (Musil)) if you can't figure it all out.

Andy

I can't find the help files on my system for iem vcfs. Perhaps you can email Thomas (Musil)) if you can't figure it all out.

will do.

I've just created the missing abstractions. FOr the inlets I copied the objects used in the vcf_lp2/4/6/8. HOwever, they seem to use different ways of scaling the inlet values.

1. For example on the cutoff inlet, what does the [iem_cot4] object do?

2. For the Q inlet (most confusingly) it varies from using:

[iem_sqrt4] on the lp4
to [iem_pow4~ 0.333333] on the lp6
and back to [iem_sqrt4] -> [iem_sqrt4~] on the lp8

I hope the above is clear to understand, but I haven't the foggiest what each of the objects does in particular, except to say they're performing maths signal tasks. What is different between a [sqrt~] and [iem_sqrt4~]?

Surely there's a simpler more consistent way to do this? (he says)

cheers, SI

Just out of interest what do you use for your suite of filter duties?

Those are quick approximations for functions. I'm not sure but I think in iem_sqrt4~ the 4 means the number of bits it operates on to approximate a root not as accurate as standard [sqrt~] but good enough for jazz. That doesn't mean the signal is 4 bit, the algorithm, which I only glanced at and don't fully get works by shifting bits of the mantissa and exponent of a float separately. It's faster than a normal iterative Newton method. (hint shifting divides or multiplies by two)