Circuit modeled filter, how to?

[Kyran]

I just skimmed through, but did any of you guys tried expanding the analog filter?

You can get really good approximations of a filters output using field simulators, but the problem is that it's too much data to really be usefull in realtime.
Now if you could expand this data into some sort of series (like a fourier series, but using a much more appropriate basis for the dimensions you're expanding in) you could possibly reduce the amount of data by a huge amount and make realtime processing possible.

[whyterabbyt]

do you have nothing better to do but troll, wr?

exactly how is providing a counter-instance to your analogy 'trolling' ? the point of modelling something is often to replicate it to a particualr level, not copy its 'style'. Try Roy Lichtenstein for an art-world example of that kind of replication.

"in the medium of glass working"
uh, unless i'm mistaken, i do not believe he ever worked in glass. ever.

No-one said he did. Your analogy was based on recreating his work, in glass, though, and that's all that I responded to.

And by the way I believe you are mistaken. Pablo Picasso worked in several other media, and his works include designs implemented in glass.

EDIT : ah, I see you've done your usual post-editing.

let's imagine you're painting onto your glass with your special glass paintbrush. everything is fine except for when the paint starts dripping and you get all kinds of streaks and droplets, bubbles and so on forming. then you're going to have to say "one day my glass will be powerful enough to.."

would you like to try that in english?

duh, you cant paint on glass the same way you paint on canvas, moron.

you always have to resort to the childishness of ad hominem insults, dont you.

for the record, whilst you cannot paint on glass 'the same way', you can paint on glass such that the end result exactly replicates the look of other painting mediums. and depending on how you do it, you can replicate the surface texture as well.
have a look at some books on enamelled jewellery some time, for example; you'll learn a few eye-opening things about replicating one media in another. although they may force you to reconsider your blindly dogmatism, and I guess we couldnt have that.

[mdsp]

Correct me if I'm wrong but isn't Euler the worst form of integration to use (although I did use it because it's so easy to code)!

true, that's why I wanted to talk about exact simulation using exponentials (when analytical solutions exists) instead of approximations.

but as you said, it's the easiest to understand and implement and also the most efficient. So people usually live with its limitations and use oversampling.

the same is true about linear interpolation vs more complex interpolation schemes.

[neotec]

IMHO it's very simple to make a digital filter 'sound analogue', when modelling or designing a filter without resonance. A nice and 'analogue sounding' resonant filter is much harder to do. So I'm currently focusing on 'studying' various design techniques ... but so far no sucess with implementing a good resonant low pass filter^^

[mdsp]

IMHO it's very simple to make a digital filter 'sound analogue', when modelling or designing a filter without resonance. A nice and 'analogue sounding' resonant filter is much harder to do. So I'm currently focusing on 'studying' various design techniques ... but so far no sucess with implementing a good resonant low pass filter^^

Well, I don't know what you mean by "make a digital filter sound analogue". All we can talk about, is how accuratey we can emulate a particular analog system in the digital world.

[aciddose]

the most important things when getting your dsp code to sound "analogue" will be sample rate and quality of control inputs.

no matter how much time you spend fiddling with useless electronic simulation stuff, you'll still have the three most important issues hitting you hard:

1) aliasing!
2) phase error due to group delay
3) choppy control inputs, stepped responses and so on

increasing sampling rate will solve all three of these without any additional effort. as for choppy control inputs, a majority of coders do not seem to realize how important the inputs are. for example, a tb-303 doesnt have a unique sounding filter, it's generic shit. what makes it sound unique is the way it's inputs are connected and the interactions with all other systems in the unit. if you do a '303 filter' and yet never bother to do '303 envelope', '303 amplifier', '303 mixer', '303 oscillator', and most importantly '303 power supply', '303 signal routing / cmos switches' and '303 sequencer', your synthesizer will sound nothing like a tb-303.

it is possible to solve the control input issues without increasing the sampling rate, so long as you put enough effort (aka cpu time) toward the problem. xhip's control inputs for example are perfect - you could output the xhip signals to 'analogue' vco/filter/vca and the results would actually sound better than the original control signals.

the main issue with a lot of 'emulations' is in the fact they're not really emulating anything at all. they figure all they need to do is flop some electronic nonsense over some integrators and call it done. they have absolutely no idea that there are 100s of factors involved in tweaking feedback path, gain into and out of various stages, offsets and all sorts of other stuff.

all the software synthesizers which claim to be 'modeled' suffer from the problems i've listed above. many of them suffer so much as to make them unusable.

[keldon85]

My greatest trick will be when I create a family of synths and effects plugins that don't operate on sample rates

[aciddose]

hm, i wonder if such a thing has ever been invented yet

these issues are not too difficult to overcome, so long as you're not pointlessly trying to exactly emulate something on a completely unique medium. i'd just rather be spending time solving the real issues than fiddling with simulations which will go nowhere without first solving the problems.

[keldon85]

Yes, it has been invented (in my head). Store / process a succession of paths with varying lengths. For example a simple [line] path can state that it goes from -34dBFS to -33dBFS in xx nanoseconds. You can go as far to say that a path is either positive, negative, or static and cannot be a complete wave with both positive and negative amplitudes.

I will refer to this path as a half-wave. You can model half-waves in any way you see fit - splines are an obvious option.

You can then compute its output in PCM for any [appropriate] sample rate. Not sure if I am illustrating my [random] thought well, but by internally representing audio as [sample independent] half-waves would be a great feat worthy of applause (even if it's impossible to process without resorting to converting each half-wave into PCM)

[keldon85]

Wait, that made no sense :S I just have this picture in my head of it, but I just can't [quite] make it out!

[haydxn]

aciddose, while you're clearly capable of great things (though you blow your own xhip trumpet far too frequently), you really are a tool.

your stupid glass analogy is all very whimsical and lovely and whatnot, but ultimately stupid. the discussion is on digital recreation of analogue systems. if you're going to make an analogy to painting, at least have the sense to use an appropriate one.

glass? no. how about a computer? there are already plenty of digital painting systems developed and in-development, using various modelling techniques to simulate paint on canvas. do you think that maybe you might be a bit of an idiot to ignore that while you were making your fairytale story of the-scientist-who-couldn't-get-it-right-because-some-nonsense-about-glass?

You're very special, clearly, and your opinion (that modelling electronic stuff is 'useless' and such a waste of time) is obviously very noble and to be respected. except for the fact that you're a massive tool, and anyone who thinks you're to be ignored is wise to continue with their research. The world is flat, there is no point in bothering to consider the universe as 'it may be' because when dealing with the medium we have, it's effectively flat.

i see the point you're making, but you're tooly-mc-tool for having them cemented to your face and trying to shackle others to it. it's odd to see someone be such a luddite against potential advancements in the field of digital audio technology; why not just admit that you don't know everything for a change, and accept the possibility that one day, someone far cleverer than you may prove beyond any doubt that you are wrong?

[andy_FX]

the most important things when getting your dsp code to sound "analogue" will be sample rate and quality of control inputs.

no matter how much time you spend fiddling with useless electronic simulation stuff, you'll still have the three most important issues hitting you hard:

1) aliasing!
2) phase error due to group delay
3) choppy control inputs, stepped responses and so on

increasing sampling rate will solve all three of these without any additional effort.

This is exacly why I have a real time oversample amount, and a render oversample amount on my synths. I'm pushing in the direction of 3D modeling here, where you have a simple shaded preview, and then render the full on model. In the end it's the result that everyone gets to listen to, so that's what's most important.

I always smooth all control inputs per sample internally in my block based code. The block size is adjustable so the user can set the tradeoff they want between cpu and accuracy.

as for choppy control inputs, a majority of coders do not seem to realize how important the inputs are. for example, a tb-303 doesnt have a unique sounding filter, it's generic shit. what makes it sound unique is the way it's inputs are connected and the interactions with all other systems in the unit. if you do a '303 filter' and yet never bother to do '303 envelope', '303 amplifier', '303 mixer', '303 oscillator', and most importantly '303 power supply', '303 signal routing / cmos switches' and '303 sequencer', your synthesizer will sound nothing like a tb-303.

Completely agree here. Circuits are complex and sound good because of it. Everything slightly influences other things, sometimes this is planned, sometimes not. Picking which of these types of things sounds good and modeling it is critical to a good model. But how do you pick this stuff out? Typically if you don't already know you will have to spend lots of time modeling and probing circuits, and listening and looking to find out.

the main issue with a lot of 'emulations' is in the fact they're not really emulating anything at all. they figure all they need to do is flop some electronic nonsense over some integrators and call it done.

Actually that is if you are lucky. Much of the time people simply look at the output self oscillation of a filter and match the harmonics with some random waveshapers they dream up. I'm not going to mention any names, but there are well known and companies doing this shit. Ewwww.

There are a couple of more things that matter when doing an analog model that you left out. One of them is that although on a schematic you have a value for a component, this is not enough. Sometimes people making the synths deliberately mismatch components to get a better sound. Matched components are typically indicated on the schematic, but mismatched ones aren't. I think there is a bunch of this stuff that people like keeping a little bit secret to protect their designs, no matter if it's a circuit or dsp.

Andrew

[aciddose]

hm, the "intentionally mismatching" idea sounds a little bit like conspiracy-nut kind of stuff to me. i'm sure a couple companies may have done this (moog perhaps? sounds like his kind of thing) however most of the engineers out there are worried about one thing and one thing only: cost!

you can buy matched components and have always been able to for far cheaper than matching by hand. i wouldn't think people would go to the expense to intentionally mismatch components like transistors as that would cost a hell of a lot of money for something that you could let nature take care of. transistors for example of the same batch almost always have similar vbe/hfe, but similar means "could potentially match 10 out of 100", not "all of them are perfect". the only thing you need to do to get mismatched components is pick any two out of a batch at random.

if you really wanted mismatching, you could pick two from two seperate batches. although, thinking of distribution/probability here, most likely you wouldn't increase the parameter diffusion by any meaningful amount.

the vast majority of designs, quality designs that is, take into account the fact that parameters are going to be fairly diffuse. two examples of many circuits should sound exactly the same after adjusting inputs as the minor changes in levels tend to average out to a steady value - that is to say that usually the diffusion should be 'pink noise', or 'normal'.

if you're modeling little amature style stomp boxes though, often any two can sound completely different. they tend to use 20% components and the designs are very sensitive to component parameters. other designs though always use 1% components in areas where parameters matter, or have adjustments. (can be tuned, like trim pots for levels or offsets for example.)

[djsubject]

Analogue or digital?

http://www.skirize.adbe.org/UnknownVirusArp.mp3

EDIT:: reverb has been added the the audio file,

[aciddose]

well, technically the virus isnt analogue, nor are the juno series. they're hybrids. so it all really depends upon the particular definition of 'analogue' being used. 100% analogue? i can tell you no, it isnt.

how?

well it is sequenced isnt it?