Zero Delay Feedback Filter (How to test if your synth has a )- Xils-Lab White Paper -

[braj]

people can kill other people in a car accident and apology

A bit of a ridiculous comparison, don't you think?

Well, he probably said that impulsively

[Fritze]

how it is easy to hate and now to apology ?

Okay, I guess that admitting mistakes and apologising for them is a bad thing?

It's one of the best things you can do in life.

[olikana]

creases.

A commonly held belief is that non-0df filters are unable to produce sounds with rapid onset transients, like percussion because of the delay introduced. This is incorrect, the 1 delay lag is not perceived by the ear. The 0df filter problem affects sounds at more or less mid resonance settings, when the filter frequency is varying (when is generally the case).

whoever wrote this shouldn't be writing tech papers and not with such contempt.
it's very sketchy and inaccurate.
if anything fast modulations with mid res is where the filter feedback lag (or the consequences of it) are most likely to be "exposed" in real musical uses .( mid resonant fast shaped sounds )....which leads to the typical plasticky sound associated with dps filters from 1996 to 2011 (till the release of Diva)
the way you wrote it makes it sound that a synth without 0df can produce analog like percussions to perfection...but that all depends on how the sound is shaped and at what resonance ( there are plenty of percussive sounds that are made with 30%-70% res)...in those sounds u will not get 100% analog sound without 0df (btw it should be 0dfF !!!!)

the most common belief on KVR has always been that softsynths were not capable of delivering good analog punchy sounds..like PUNCHY BASSES (THIS HAS BEEN ALWAYS THE N.1 COMPLAINT ON KVR SINCE THESE FORUMS EXIST),and that softsynthssounded "plasticky" in comparison.
guess what??? PUNCHY BASSES are made with "rapid onset transients" at about MID RES!

so kvr members which u r trying to ridicule from your non stature were actually right...while developers (including the one who wrote this paper lol) were releasing emulations like the minimoogV which were unable to even deliver convincing typical moog bass cos of filter feedback lag ( a typical moog funk bass uses about MID RES and sounded plasticky on that so called "emulation").
but ofc such developers were flogging their emulations as being accurate....

Final Note: Synths that are not equipped with 0df filters are definitely NOT "bad" synths. They have their own character, filters, etc, and in a lot of situations, you wont be able to tell the difference. The same applies to synths equipped with 0df filters:

not true...any synth without 0df should not be used for someone looking for true analog sound . it will have the same limitation that characterized DSP filters in the past. ok for trance for example but not ok for early 80s pop and electrofunk which are full of ressy punchy sounds .

At Xils-Lab we're always trying to better our algorithms, and the 0df part is no exception. There are several possibilities and approaches to achieve 0df filters in the digital world, and we hope that our own way to do things will continue to evolve
There are several possibilities and approaches to achieve 0df filters in the digital world, and we hope that our own way to do things will continue to evolve

my suggestion is do it the way that diva or saurus did it cos they actualy got it right

[KBSoundSmith]

my suggestion is do it the way that diva or saurus did it cos they actualy got it right

Only 101 pages to go to beat out the Tone2 Saurus thread. Literally eating some popcorn right now...

[Lotuzia]

creases.

A commonly held belief is that non-0df filters are unable to produce sounds with rapid onset transients, like percussion because of the delay introduced. This is incorrect, the 1 delay lag is not perceived by the ear. The 0df filter problem affects sounds at more or less mid resonance settings, when the filter frequency is varying (when is generally the case).

whoever wrote this shouldn't be writing tech papers and not with such contempt.
it's very sketchy and inaccurate.
if anything fast modulations with mid res is where the filter feedback lag (or the consequences of it) are most likely to be "exposed" in real musical uses .( mid resonant fast shaped sounds )....which leads to the typical plasticky sound associated with dps filters from 1996 to 2011 (till the release of Diva)
the way you wrote it makes it sound that a synth without 0df can produce analog like percussions to perfection...but that all depends on how the sound is shaped and at what resonance ( there are plenty of percussive sounds that are made with 30%-70% res)...in those sounds u will not get 100% analog sound without 0df (btw it should be 0dfF !!!!)

the most common belief on KVR has always been that softsynths were not capable of delivering good analog punchy sounds..like PUNCHY BASSES (THIS HAS BEEN ALWAYS THE N.1 COMPLAINT ON KVR SINCE THESE FORUMS EXIST),and that softsynthssounded "plasticky" in comparison.
guess what??? PUNCHY BASSES are made with "rapid onset transients" at about MID RES!

so kvr members which u r trying to ridicule from your non stature were actually right...while developers (including the one who wrote this paper lol) were releasing emulations like the minimoogV which were unable to even deliver convincing typical moog bass cos of filter feedback lag ( a typical moog funk bass uses about MID RES and sounded plasticky on that so called "emulation").
but ofc such developers were flogging their emulations as being accurate....

Final Note: Synths that are not equipped with 0df filters are definitely NOT "bad" synths. They have their own character, filters, etc, and in a lot of situations, you wont be able to tell the difference. The same applies to synths equipped with 0df filters:

not true...any synth without 0df should not be used for someone looking for true analog sound . it will have the same limitation that characterized DSP filters in the past. ok for trance for example but not ok for early 80s pop and electrofunk which are full of ressy punchy sounds .

etc etc ......

Thanks for your POV Olikana and your sweet words,

Transcients : Unfortunately, while you pretend to represent the KVR community and call it to rescue, no one in this thread did agree with you on this particular point. (if you read the thread )

one/Two pages ago a poster was even wondering who could say such things, and thought you were a strawman.

A lot of posters also declared that non 0df filters could lead to good results.

But everybody has the right to express his POV. And now I'm sure that strawmen don't exist only in my imagination.

As for the technical data we provided about transcients, its just .... very simple maths.

LtZ

[hakey]

one/Two pages ago a poster was even wondering who could say such things, and thought you were a strawman.

Stuff and nonsense.

My reading of olikana's post is that he's claiming that non-0df filters can't replicate typical analogue response when processing fast transients.

"analog like percussions... delivering good analog punchy sounds"

That is NOT the same as claiming that non-0df filters can't produce fast transients.

[olikana]

i wrote my post quickly...i have got nothing to sell here...and i don't claim to be writing any papers on this topic.
but i do stand correct that filter feedback lag (or the consequences of it) get exposed in punchy or fast modulated sounds with mid res.
but it's not the transients per se...it's the shaping of the filter.


when i used "transients" i only used it as a quote from your "paper".
what i said in my post is that filter feedback delay is most likely to be heard in musical terms in fast or punchy modulation settings.

and the example of a punchy moog bass (with middle res) is appropriate.
you can hear in a synth without 0dff how different it would sound if you oversample it (which among other things like aliasing reduces also filter feedback lag).
infact that's what dcam and other synths did to reduce filter feedback lag...oversample (it gave u the option to)..although esponentially oversampling it would get u close to 0 it would never reach it and never sound "right".


btw i don't say that a synth without 0dff cannot make a good mid res punchy bass...but it cannot do it 100% like analog....it will sound 99.7% same...for some is good enough...but it's not 100% analog like.

[Lotuzia]

this is what the paper says :

" A commonly held belief is that non-0df filters are unable to produce sounds with rapid onset transients, like percussion because of the delay introduced. This is incorrect, the 1 delay lag is not perceived by the ear. The 0df filter problem affects sounds at more or less mid resonance settings, when the filter frequency is varying (when is generally the case). "

No more, no less

No mention in this of anything else like moog basses, or percussive analog percussions wich would be more or less analog than the analog of my neighbour. So either people find this is correct, according to maths, either they think its not. It's as simple as that.

As for Dcam, how do you know what's exactly under the hood ?. It has visible oversampling parameters, but thats it. They could use other techniques together/in parralel with these ones, and its not like oversampling is not used by a lot of developpers, including Xils.

Its not like if we said that 0df filters will sound exactly similar to non 0df ones. If it was the case why would we have done it ? We say that non 0df filters can sound "good" and can have a nice character of their own. That's all. ( and we dont sell some so see how we're open minded .... )

[AdmiralQuality]

Oh and off-topic note: Playing with the Voyager was healthy (for wallet), as I don't really want one anymore. I found the design rather painful in some ways actually.

Try a Little Phatty. They're surprisingly easy to program and it's a very complete synth voice, despite the multi-function knobs. I'm told they also don't suffer from the zipper noise that the Voyager shows in its pitch bend wheel at wide range.

[mystran]

Oh and off-topic note: Playing with the Voyager was healthy (for wallet), as I don't really want one anymore. I found the design rather painful in some ways actually.

Try a Little Phatty. They're surprisingly easy to program and it's a very complete synth voice, despite the multi-function knobs. I'm told they also don't suffer from the zipper noise that the Voyager shows in its pitch bend wheel at wide range.

I've been thinking about Slim Phatty actually (got enough keyboards, but there's space in the rack ). Occasionally it would nice to have some basic analog with patch storage, and one of those wouldn't break the bank either.

On totally different topic though:

I don't understand why people insist on these zero-delay filters being expensive. You don't need to use iterative feedback solvers unless you want to. Just linearize the filter around it's known state once per sample and you can solve analytically. The overhead is pretty insignificant and it works quite fine in practice, at least until aliasing starts becoming a bigger problem anyway.

[AdmiralQuality]

Oh and off-topic note: Playing with the Voyager was healthy (for wallet), as I don't really want one anymore. I found the design rather painful in some ways actually.

Try a Little Phatty. They're surprisingly easy to program and it's a very complete synth voice, despite the multi-function knobs. I'm told they also don't suffer from the zipper noise that the Voyager shows in its pitch bend wheel at wide range.

I've been thinking about Slim Phatty actually (got enough keyboards, but there's space in the rack ). Occasionally it would nice to have some basic analog with patch storage, and one of those wouldn't break the bank either.

Yeah, I want one. (Probably a Little. I like keyboards on my keyboards.) I was really impressed when I played with one in the music store.

On totally different topic though:

I don't understand why people insist on these zero-delay filters being expensive. You don't need to use iterative feedback solvers unless you want to. Just linearize the filter around it's known state once per sample and you can solve analytically. The overhead is pretty insignificant and it works quite fine in practice, at least until aliasing starts becoming a bigger problem anyway.

Oh? I'm not much of a math guy, I approach most of this stuff intuitively and fill in missing math where needed (which is relatively rarely). I do think I understand what Urs is doing with iterative prediction though. How's this other method work?

If you'll share an algorithm and I like it I'll put it in the next update of Poly-Ana and you can have free AQ software for life! (Mine or yours, whichever comes first. )

[aciddose]

it's all just the application of algorithms that have been applied for decades in other areas.

it has to be more expensive than the trapezoidal integration because of the simple fact that it exists purely as an attempt to solve the short-comings of trapezoidal integration!

yes you can find areas in which you can get away with less computation, the new method being the optimal solution. when you attempt to work your solution in the range of applications trapezoidal integration already works and demonstrate a static improvement however you'll find that you're facing having to pay a huge expense.

many methods are simply trading expense in one area for that in another. they're merely moving issues around.

personally i prefer in most cases the same solutions i've used for nearly a decade now. they're elegant, easy to implement and fast. all the trade-offs are already selected to match my preference. i consider application of different predictions or integrations still limited to the experimental realm outside of the application they see when used in simulation.

[Kriminal]

I just tried the hack method lotty posted, seems Steinbrrg Neon had 0DF

[AdmiralQuality]

I just tried the hack method lotty posted, seems Steinbrrg Neon had 0DF

[Ingonator]

I just tried the hack method lotty posted, seems Steinbrrg Neon had 0DF

Hi,

i just thought this is a joke but i tried with the now free version of Neon and you seem to be right...
Of course this synth is so old that it should not include any 0df filter design.

I already recognized that for a few synths like e.g. Diva, Saurus and Xils Lab synths the result seems to be quite clear but also found it could be very unclear for some others.

I also found that the 12dB and the 24db filter in Saurus behave quite differently. While the 12dB filter seems to work exactly like in Lutuzia's video in the 24dB filter the Resonance peak seems to slowly disappear at high Cutoff frequencies.

So far i have not checked with my Moog Slim Phatty but i'll check soon to get a kind of "reference".


Ingo