modulate filter with 20Khz lfo ..

[gentleclockdivider]

Exactly ..
Either the developer has been living under a rock or he just has a different understanding of the word "good"' .
Cytomic the drop is just on par with hardware stuff ...and it's creator Andrew Simper is a math genius
His comment is just ridiculous

[gentleclockdivider]

Saying "be carefull with the resonance control proves that these filters are not designed for audio rate modulation" is the same like saying "you cannot say be careful when driving a car, because that would mean the cars are not designed to be driven"

Are MSF filters biquads ?
I think they are , and reason why they become unstable

[MeldaProduction]

Either the developer has been living under a rock or he just has a different understanding of the word "good"' .

Haha and you are surprised someone gets defensive??

Anyways yes these are biquads (pretty much all IIR filters are based on biquad sections), the Filter doesn't have the "analog" feature from TurboFilter and AutoDynamicEQ yet, I could add that I suppose, makes sense. That should do the trick with stability (that's kind of the analog emulation commonly uses), but to be honest, when you modulate something too quickly, you just get distortion (which is what you originally talked about, not instabilities).

[MeldaProduction]

Ok, so I wasn't able to add it to the Filter module, but I added the Analog feature to the Bandpass module and it actually does quite some magic when modulated quickly. I'll also add TurboFilter, but that one is in the development phase, so...

[Gone soft]

Ok, so I wasn't able to add it to the Filter module, but I added the Analog feature to the Bandpass module and it actually does quite some magic when modulated quickly. I'll also add TurboFilter, but that one is in the development phase, so...

Oh, I'm happy just to have an improvement in the BPF module, that's the one I'm using the most for now anyway. And looking forward to that TurboFilter whenever it'll be ready.

[gentleclockdivider]

There is just something verry unpleasant about the filters , the way they are scaled
Just a simple 4 pole filter , cutoff set to default and a 4 seconds linear decaying filter envelope with max filter amount ( no resonance)
The drop in volume is staggering and reminds me of the verry first vst filters .
It's the only thing that's holding me back from purchasing it

[gentleclockdivider]

Either the developer has been living under a rock or he just has a different understanding of the word "good"' .

Haha and you are surprised someone gets defensive??

With that I meant that I was verry surprised you never heard of u-he , cytomic , reaktor etc..
If you would have , you should have known that these are verry highly regarded analogue emulations capable of convincing filter FM , yet you stated you haven't heard convincing filter fm in software .

[bmanic]

Have you actually heard about a software that would allow doing that and sounds good? Because I don't. One of the problems of the digital audio is that it is "extremely fast", so it actually does what you want (while HW doesn't it basically smoothens things out, because the components are just not that "fast").

Absolute 100% bollocks.

Saying "be carefull with the resonance control proves that these filters are not designed for audio rate modulation" is the same like saying "you cannot say be careful when driving a car, because that would mean the cars are not designed to be driven"

Having to be careful so that a parameter doesn't "explode" is never the users fault unless the software in question is a true programming environment. MSF is nowhere near something like that.

Also, It's very possible to make awesome digital filters that are extremely robust and never "explode", even under extreme audio rate modulation.

[zerocrossing]

Have you actually heard about a software that would allow doing that and sounds good? Because I don't. One of the problems of the digital audio is that it is "extremely fast", so it actually does what you want (while HW doesn't it basically smoothens things out, because the components are just not that "fast"). So here you can also smoothen things out, but that's not what you want is it . You know, with power comes responsibility. Saying "be carefull with the resonance control proves that these filters are not designed for audio rate modulation" is the same like saying "you cannot say be careful when driving a car, because that would mean the cars are not designed to be driven"

Well, I don't think anyone sits around and modulates a filter with a 20khz sine wave (well, I'm sure they do... ) but I do regularly modulate a filter with something like a sawtooth wave that has harmonics reaching up pretty high. Anyway, I can get it working in Msoundfactory if I'm careful and set oversampling at the higher rates, but it doesn't sound that great. People mentioned Zebra's XMF filter, but you you could look at Legend's filter, Dune's (must set modulation quality to high) and any number of filters in Reaktor's Blocks, especially if you run Reaktor at twice your project's sample rate. I'm not even sure if I'm asking for the right thing, but if you take a look at Reaktor and use some of its Blocks, like the Monark filter Block, that support overdrive and filterFM, you'll hear it done well. There's a "Monark Mikro" .ens file... not sure if it's a factory .ens or in the user lib, but it's got the filter FM and the source is OSC1. You can crank that puppy up to 2' with a square wave and FM all the way up on the filter and it still sounds good at 44.1khz and better at 88.2. (Reaktor lets you run it at a different sample rate than your host is running)

Also, I think the drive and gain on the filter sound pretty good at lower ranges, but I'd like to have the option to oversample them as well so higher notes ring out with less aliasing.

[gentleclockdivider]

Try turning up the resonance when you do that ...be carefull .
THe 20 kHz was just an example , because that's the maximum range of the lfo's , and it was used as a testcase , they even explode when the lfo runs at 1Khz , or using a noise source as modulation .

Keep in mind that the above given examples ( monark , legend , cytomic , u-he ) are done by people who have been coding and analyzing and refining their efforts to replicate analogue filter behaviour for almost a decade .
It's wishfull thinking all of a sudden to expect that level of detail from melda ( not meant to be condenscending )
I say it's time to leave the biquads filter behind
https://www.earlevel.com/main/2016/02/2 ... rting-out/
https://cytomic.com/index.php?q=technical-papers

[zerocrossing]

Try turning up the resonance when you do that ...be carefull .
THe 20 kHz was just an example , because that's the maximum range of the lfo's , and it was used as a testcase , they even explode when the lfo runs at 1Khz , or using a noise source as modulation .

Keep in mind that the above given examples ( monark , legend , cytomic , u-he ) are done by people who have been coding and analyzing and refining their efforts to replicate analogue filter behaviour for almost a decade .
It's wishfull thinking all of a sudden to expect that level of detail from melda ( not meant to be condenscending )
I say it's time to leave the biquads filter behind
https://www.earlevel.com/main/2016/02/2 ... rting-out/
https://cytomic.com/index.php?q=technical-papers

Oh, I did all that, and yeah, I could break it, but if I stayed in moderate ranges, it was ok. I’m not saying that you’re wrong, but I also can’t imagine that someone who made these synths and effects couldn’t read a few technical papers and with some focus and effort come up with something that worked as we’re describing. But maybe better, in a sense. I’m not looking for emulations. I’ve got plenty of those. Just performance. The filters sound fine as they are, I’d just like them to behave more like Dune’s. Maybe even license someone’s existing tech, like Spectrasonics uses UVI filters. (Which don’t do audio rate modulation, FYI)

[Whywhy]

Cytomic:
"Below are some technical papers that will hopefully contribute some knowledge and algorithms to the DSP community in general."

For sure, it could be useful to implement them to MSF!
Great information, thanks @gentleclockdivider

[MeldaProduction]

Try turning up the resonance when you do that ...be carefull .
THe 20 kHz was just an example , because that's the maximum range of the lfo's , and it was used as a testcase , they even explode when the lfo runs at 1Khz , or using a noise source as modulation .

Got it! From the beginning I thought you were talking about distortion and I quite believe that if you change some relevant parameter of pretty much anything as quickly as many thousand times per second (even a small gain), you will just get some distortion. But of course it may sound good in some context.

Anyways please wait for the next version, the Analog feature, which is now present in the Bandpass seems to do wonders in this aspect. And thanks for the links, will check them out .

[zerocrossing]

Try turning up the resonance when you do that ...be carefull .
THe 20 kHz was just an example , because that's the maximum range of the lfo's , and it was used as a testcase , they even explode when the lfo runs at 1Khz , or using a noise source as modulation .

Got it! From the beginning I thought you were talking about distortion and I quite believe that if you change some relevant parameter of pretty much anything as quickly as many thousand times per second (even a small gain), you will just get some distortion. But of course it may sound good in some context.

Anyways please wait for the next version, the Analog feature, which is now present in the Bandpass seems to do wonders in this aspect. And thanks for the links, will check them out .

...and we all lived, happily ever after!

[gentleclockdivider]

Try turning up the resonance when you do that ...be carefull .
THe 20 kHz was just an example , because that's the maximum range of the lfo's , and it was used as a testcase , they even explode when the lfo runs at 1Khz , or using a noise source as modulation .

Got it! From the beginning I thought you were talking about distortion and I quite believe that if you change some relevant parameter of pretty much anything as quickly as many thousand times per second (even a small gain), you will just get some distortion. But of course it may sound good in some context.

Anyways please wait for the next version, the Analog feature, which is now present in the Bandpass seems to do wonders in this aspect. And thanks for the links, will check them out .

I don't know if this is going to solve the underlying issue and the limitations of the used filters ( biquads )