func shaper

[hi-hat]

had some experience with funcshaper -really powerful outstanding plugin for 1000 usd

but have suggestion:

1.more gentle cut filter for upsampling with gentle transition and alias allowed to eliminate ringing in highs. instead of rippled very steep resonant
2.delay compensation for upsampled mode by 5 samples for any 'x'upsampling

this improvements will bring better accuracy for plugin

thanks and good luck!

p.s.i hate maths and physics but i like good music and sound)

[hi-hat]

...would like this plugin to be usable. but as i mentioned delay compensation is needed -try to dry/wet with oversampling-have phase issue!
finally ugly bug: when several instances of funcshaper are enabled on tracks dc biasing on any leads to pops and cracks...

[Music Engineer]

...would like this plugin to be usable. but as i mentioned delay compensation is needed -try to dry/wet with oversampling-have phase issue!
finally ugly bug: when several instances of funcshaper are enabled on tracks dc biasing on any leads to pops and cracks...

hmm...OK. i cant't promise anything right now because i'm currently a bit busy with contract work and the next thing on my priority list will be 64 bit support. but i'll keep these things in mind for the next update.

[Aroused by JarJar]

This is a great plugin. Now I don't have to use Csound for these same effects (or for the filters in the Engineer's filter). At this point the waveshaper does seem to have latency and delay issues though. In my host, Samplitude, I get intermittent(?!) loss of sync between tracks when using this plugin, which doesn't make sense at all to me. It is as if I have a CPU overload and the sequencer starting "rambling", but the CPU meter does not show very heavy usage.

Like hi-hat I am using wet/dry mix and oversampling (4X, usually). I know that hi-hat is referring to the phasing of a wet/dry mix without the dry delayed to jibe with the wet (the hack for this is copying the file to a dry track, then delaying that track) and not about the problem I am having. Nevertheless my problem does seem to arise when using oversampling.

[Music Engineer]

At this point the waveshaper does seem to have latency and delay issues though. In my host, Samplitude, I get intermittent(?!) loss of sync between tracks when using this plugin, which doesn't make sense at all to me. It is as if I have a CPU overload and the sequencer starting "rambling", but the CPU meter does not show very heavy usage.

hmm - this is strange. there actually isn't anything in the algorithm which could produce such delays. it's just upsampling->shaper->downsampling where the resampling filters are both elliptics. i tried to get the trial version of samplitude but unfortunately they do not seem to have an offline installer and my workstation is offline. i asked for the possibility of an nfr-license exchange. we'll see what comes out.

the more i think about it, the more i appreciate the the idea of providing alternative filters for the up-/downsampling. as there is distortion in between, i think it could be a good idea to have some kind of waveshape-preserving anti-alias filter (Bessel or something). that would sacrifice steepness, of course - nothing beats elliptics in this regard.

[Aroused by JarJar]

At this point the waveshaper does seem to have latency and delay issues though. In my host, Samplitude, I get intermittent(?!) loss of sync between tracks when using this plugin, which doesn't make sense at all to me. It is as if I have a CPU overload and the sequencer starting "rambling", but the CPU meter does not show very heavy usage.

hmm - this is strange. there actually isn't anything in the algorithm which could produce such delays. it's just upsampling->shaper->downsampling where the resampling filters are both elliptics. i tried to get the trial version of samplitude but unfortunately they do not seem to have an offline installer and my workstation is offline. i asked for the possibility of an nfr-license exchange. we'll see what comes out.

the more i think about it, the more i appreciate the the idea of providing alternative filters for the up-/downsampling. as there is distortion in between, i think it could be a good idea to have some kind of waveshape-preserving anti-alias filter (Bessel or something). that would sacrifice steepness, of course - nothing beats elliptics in this regard.

The version of Samplitude I use on my laptop came in a magazine for 10 Euros- except for tracks limited to 48, only 4 sub busses and 2 plugins per track, and other such unimportant (to me) limitations, it is just as good as the full version I have in the studio. The problem may not be your filter at all: the included Samplitude plugins, which I don't use much, ( sr limited to 44.1k in this version) and may be the cause of my problem. Occam's razor says it is simply my CPU being overtaxed; it is typical to get ragged timing in this case, but as I said Samplitude isn't showing such enormous usage when this happens. Next time this happens, I will check the Windows CPU usage and see what it shows.

And something I just thought of- when I've got Csound slaved to Samplitude, I have been using MME drivers, and I sometimes forget to switch back to ASIO when I'm not using Csound. Don't know if this could be the source of the problem- MME seems sluggish and sloppy sometimes.

Isn't a Chebyshev type one with many poles but very low ripple a standard filter for anti-aliasing, or is my informatin 20 years out of date (or wrong)?
Chebyshevs of both types do color, even strongly, but I like the general color a lot (kind of orange-brown and dusty, reminds me of Buchla synths for some reason).

[Aroused by JarJar]

Found the problem. The most reasonable explanation turns out to be the best, as usual.

As I do not use very many plugins, this has not happened enough for me to catch the problem sooner. What happens is that any sudden large leap in CPU from any plugin will cause the timing of tracks to swim and hang up (no surprise there I guess), but it seems the program may be "hung" or already stumbling before it registers this leap on the meter. So, I will be seeing a 50% CPU usage on the Samplitude meter, which remains frozen as the actual CPU usage leaps far above this. I can reproduce this problem by putting heavy plugins on objects, so that I may see the point where this happens. My laptop is simply underpowered for its job.

[Music Engineer]

i sometimes see the cpu meters of some hosts fail to show the actual load with my signal-analyzer plugin. presumably this is due to the fact that in this plugin, all the heavy fft-processing is done in the gui-thread whereas the cpu-meter apparently only measures the load in the audio thread. but this is just for info and has probably nothing to do with the case here because you say that the load is actually shown but with delay. perhaps the cpu is under such heavy load, that it also fails to update (redraw) the gui on time?

Isn't a Chebyshev type one with many poles but very low ripple a standard filter for anti-aliasing, or is my informatin 20 years out of date (or wrong)?

hmm - i don't know really. i actually thought that in AD-converters without oversampling, elliptics would be the filter of first choice due to their awesome steepness properties. also, ripple in the stopband would not really matter at all as long as it remains below the noise floor of the analog electronics. nowadays, i think (but i'm not sure) that most AD-converters work with oversampling and use quite shallow lowpasses in the analog domain and FIRs for internal resampling. inside pure software applications, i gather from what i read in the dsp-forum here, that most developers use FIRs as well for resampling applications.

Chebyshevs of both types do color, even strongly, but I like the general color a lot (kind of orange-brown and dusty, reminds me of Buchla synths for some reason).

interesting. i do not have any synesthetic experiences with particular types of filters. for me, however, sawtooths are clearly green and square-waves are yellow (a dark, brownish one).

[miedex]

Do you know of any resources for equations that might be intended for this purpose or otherwise worthy of trying anyways?

[Music Engineer]

Do you know of any resources for equations that might be intended for this purpose or otherwise worthy of trying anyways?

well, i have written a little tuorial about waveshaping myself - you find it in the "Tutorials" on my website.

Miller Puckette has some chapter on the topic in his online book:

http://crca.ucsd.edu/~msp/techniques/v0 ... ode74.html

other than that: i don't know any online resources offhand, but maybe it could be worthwhile to explore what types of transfer-functions typically occur inside analog equipment.

btw.: everyone is welcome to post interesting functions, ideas, presets, etc. here. maybe we can build up some kind of resource here.

[hi-hat]

i am pretty sure it already has all dual triode and triple pentode altering behavior included
but here's some more feedback: new version has degrade-it pops on any parameter change-very hard to monitor
is shaper integer processing and all this pops and cracks due to?
enabling in reaper buggy plugin compatibility mode removes pops while dc bias if several shapers are on tracks-may be this could point you on trouble source.
also precise slider control with some 'shift' key needed for all plugins
some kind of denormal exists for shaper and also eng.filter -noise addition prior processing helps to eliminate load.
bigger gui for eq and shaper would be good too,1-order xover appreciated... thoughts aloud
good stuff though-thanks. do you plan some new?

[Music Engineer]

i am pretty sure it already has all dual triode and triple pentode altering behavior included

well, i have some presets where the transfer function was inspired by something i have read about tubes. but i won't claim anything about how close it resembles the actual behavior of a real tube. probably not at all, since tubes cannot really be modeled properly by pure waveshaping. they do dynamic distortion.

but here's some more feedback: new version has degrade-it pops on any parameter change-very hard to monitor. is shaper integer processing and all this pops and cracks due to?

no. all processing is done in 64-bit float. i'l have a look at this. actually, for some parameters, most notably the a/b/c/d constants that are used inside the transfer function, some popping is expected since the plugin uses an internal table for the transfer function which has to be recalculated whenever one of these parameters changes. but if it also pops when you set up - say - the filter's cutoff frequencies or something ...hmm well, the parameters are not smoothed, so there might be some slight zipper, too. but it has always been like that, so if it only occurs in a newer version, there might be something wrong.

enabling in reaper buggy plugin compatibility mode removes pops while dc bias if several shapers are on tracks-may be this could point you on trouble source.

"removes pops while dc bias" - i don't understand that. what does it do with dc-bias?

also precise slider control with some 'shift' key needed for all plugins

try the mouse-wheel.

some kind of denormal exists for shaper and also eng.filter -noise addition prior processing helps to eliminate load.

see other thread.

bigger gui for eq and shaper would be good too,1-order xover appreciated... thoughts aloud

yes. that occurred to me too. 1st order crossovers have one chief adavntage: there outputs sum back to the original signal exactly (no allpass characteristic is introduced). i shall do that.

regarding the GUI: are you talking about scaling everything larger (including the fonts) or just make the window bigger but leave the the font-size as is? the latter would be easy to do, the former not (i'm using a selfmade pixel-font, but i'm actually considering to switch back to vector-fonts at some point).

good stuff though-thanks. do you plan some new?

well, in my head there are always a long(ish) to-do list for plugins and there are some half-baked thingies in the pipeline. but i'm currently busy with framework-ish stuff and contract work, so there's nothing new close to release, at the moment.

[hi-hat]

if you have shaper on several tracks in mixer then changing dc parameter on any instance leads to pops and in reaper this could be solved by enabling buggy plugin compatibility mode.
but i like this plug among tons of other similar because it also has 1-2pole filters built in and more...
and eq i like because it's dual channel or m/s with single instance already
good engineered plugs!
about larger gui: yes it could be -mostly in order to point eq peaks with mouse easier sitting 1,5m from screen
hopefully everything works even i use previous funcshaper version.
thanks and good luck!
p.s.will check for update periodically...