Repro-1 (out now)

[aaron aardvark]

Urs,
Is there any temptation to pursue what people prefer, as opposed to what is "best"?

[KnickersDown]

I'm so glad I have great ears and great taste. I voted for filter 3. 69 other people voted with me. We all are cool. If you didn't vote for filter 3, you are very uncool. And you probably stink.

[EvilDragon]

Urs,
Is there any temptation to pursue what people prefer, as opposed to what is "best"?

In the PDF Urs posted, he said they'll go with 3.


Makes perfect sense to me.

[do_androids_dream]

Urs,
Is there any temptation to pursue what people prefer, as opposed to what is "best"?

In the PDF Urs posted, he said they'll go with 3.


Makes perfect sense to me.

That's a shame. Really lacks the thumpy bottom end as the filter is closing with quick envelope times.

[fmr]

I'm so glad I have great ears and great taste. I voted for filter 3. 69 other people voted with me.

Yeah... Some of them, after knowing it was the one chosen by the developer

Before that, filter #1 had more votes then #3. Anyway, this wasn't a competition, and I din't think that anybody's pick was wrong. It's simply down to what kind of sound one prefers, and what kind of things one values most. For me, anything that reveals any kind of artifact, even if slightly, is a no go.

Anyway, the filter is still under works, and what we were presented with was a raw version of several ways of realizing the same model. Urs chose filter #3 as an all-round, more versatile solution (at least, that's what I got from the PDF reading). Totally understandable. And I'm sure the final version will sound much better

[EvilDragon]

That's a shame. Really lacks the thumpy bottom end as the filter is closing with quick envelope times.

Don't discount the possibility of additional tweaks to the filter code. This is an alpha, after all.

Besides, if that's how u-he's Pro-One sounds, and they want to be as close to it as possible, without the "thumpy bottom end at fast envelopes", then that's how it's gonna be. Although, if they add component tolerance tweaks, you might get it your way too. However the main reason they are going with 3 is that it's the most stable and well-behaved method of calculating the filter model under extreme conditions, not requiring 32x oversampling like other methods.


Also remember that ALL filters here are limited to maximum 25kHz, whereas the hardware Pro-One goes more than twice that high (with modulation). I'm pretty sure this limit is going to go away with method 3, since it behaves so well and doesn't blow up in extremes like poor method 1 does

[do_androids_dream]

Besides, if that's how u-he's Pro-One sounds, and they want to be as close to it as possible, without the "thumpy bottom end at fast envelopes", then that's how it's gonna be. Although, if they add component tolerance tweaks, you might get it your way too. However the main reason they are going with 3 is that it's the most stable and well-behaved method of calculating the filter model under extreme conditions, not requiring 32x oversampling like other methods.


Also remember that ALL filters here are limited to maximum 25kHz, whereas the hardware Pro-One goes more than twice that high (with modulation). I'm pretty sure this limit is going to go away with method 3, since it behaves so well and doesn't blow up in extremes like poor method 1 does

What I personally associate with the Pro One is 'hammer' bass lines - really solid, thumpy, 'wet' sounding bass that's pretty unique even in analog. 1, although pretty much broken sounding at higher resonance settings, has this moreso than any of the other filters. It's still going to sound good whatever filter's chosen.

[Urs]

That's a shame. Really lacks the thumpy bottom end as the filter is closing with quick envelope times.

Well, that's the compression-like effect I pointed out in the PDF. It sounds great in one case, but then has disadvantages in other cases. A compressor will do the same trick with the other modes.

[EnGee]

Update May 2nd 2016: Our solution & conclusion available in a geeky little PDF:

http://www.u-he.com/downloads/UrsBlog/R ... veiled.pdf

(sorry for the maths stuff, I just can't explain things without that...)

Great read Thank you
Looking eagerly for the beta release

[Nokenoku]

+1

Thank you Urs for all the information!
Really interesting stuff.

[urosh]

... delta-method described by Andy Simper ...

?

BTW, how big part of overall computational burden (for iterative solver) is calculation of nonlinear functions?

[Urs]

... delta-method described by Andy Simper ...

?

BTW, how big part of overall computational burden (for iterative solver) is calculation of nonlinear functions?

The non-linear function isn't that big of a deal. The bigger deal is solving the linear system imposed by the Jacobian matrix (green area in the PDF)

[Urs]

That said, I have not yet done any optimised builds with the current set of algorithms to actually compare CPU between the various methods.

[mcbpete]

My god, an actual readable & understandable research paper - High five Urs

[Vesius]

Thanks for the great read - I'm pretty happy with how things turned out, and the edification is much appreciated