|
 Kyma Forum
 Kyma Support
 Kyma Filters (Page 1)
|
| This topic is 2 pages long: 1 2 |
next newest topic | next oldest topic |
| Author | Topic: Kyma Filters | |
|
David McClain Member |
posted 24 October 2000 17:54
 
Hi, After many hours of laboring a point, I have finally come to realize that what the Kyma manual says is true... I am posting this here because I may not be the only one to misunderstand these simple filters (LPF, HPF). Coming from a conventional synth background, I expected that the feedback parameter controls the amount of resonance AT THE FILTER CUTOFF FREQUENCY. In fact, as the manual clearly states, this parameter is the attenuation applied to a portion of the filter output fed back to its input. It always helps to set up a high-level example in Kyma to gain an understanding of its internals... Now that I understand this point, I am at a loss to understand the harmonic relationship between the frequency of the resonance and the stated filter cutoff frequency. The resonance seems to vary with filter type (HPF vs LPF) and order. Perhaps it is a matter of art, as maybe it should be. I want to reiterate that these filters are among the best I have ever seen! I just have to learn how to work with them now... Many thanks for a terrific instrument! Cheers, - DM IP: Logged | |
|
John Dunn Member |
posted 25 October 2000 16:50
Perhaps the best you have SEEN, but surely you are not saying they are the best you've HEARD? Look, I love Kyma. And I am in awe of the brilliance of its design. But I have ears, and those ears have listened to a lot of synths, and they report to me that the Kyma filters are perhaps its weakest feature. They are thin sounding, they have an unnatural mid-high resonance in normal use, and they make Kyma sounds instantly recognizable most of the time - not in a good way. They are like hard rubber tires on a motorcycle - superior on paper perhaps, but they kill you on the road, in the curves. Years ago when Emu was a tiny company that built large modular synths, I got into an argument with Dave Rossum about the Emu filters. He took great exception my opinion and (this is true!) took out paper and pencil and proved to me mathematically how the Emu design was far superior to the Moog filters. He was my boss at the time and his math chops are worlds ahead of mine, so the argument ended there. But here we are 20 years later, and people still have fond memory of Moog filters, but no one even remembers how Emu modular filters sound. Except me: they sounded a lot like Kyma filters My guess is that Kyma filters are very close to perfect in design, and Karl, no less brilliant an engineer than Rossum, can easily prove how they are far superior to whatever you want to compare them to. But try listening to an old Moog modular filter - go dig through your record archives and get Carlos' SOB or something else that uses a lot of Moog filters. I guarantee you will, even today, be impressed by the sweet dirtiness of those filters. For a modern digital comparison, go download VAZ2 and play around with those filters. You will be reminded that - while the British are not particularly known for their engineering (at least in the positive sense), or for the clarity of their sound - they are justifiably famous for the warmth of their sound, of a definite character to their sound that can best be described simply as - good. Please prove me wrong here. I would love to discover I have been misusing Kyma filters (the Q issue notwithstanding), and if only I did thus-and-so, they would have all the warmth and mellowness and just generally sweet sound that Bob Moog came up with 20 years ago or Martin Fay gets in VAZ today. But my ears keep telling me what my heart and brain have never wanted to believe: Kyma filters are the weakest part of an otherwise astounding synthesizer. John Dunn IP: Logged | |
|
dennis Member |
posted 26 October 2000 14:50
Interesting discussion! Seems like this is a accuracy vs. aesthetics type of issue. John, I think you're saying that the Moog filters colored the sound in a particularly appealing way but the Kyma filters are more accurate? What's a spectral analysis of the Moog filters show? Could we duplicate the coloring via Kyma? IP: Logged | |
|
David McClain Member |
posted 26 October 2000 17:08
Yes, I have heard many people state that the Moog filters were really something. I have VAZ/2 and I listen to them. But perhaps, with my untrained ear, I can't tell what I should be hearing. I know that the VAZ uses double-precision arithmetic internally. I don't know from what design their filters originate, and I haven't inspected them on the spectrum analyzer yet... but now I will! The Kyma filters fit a near perfect spectral description of filtering, at least in terms of amplitude response. I need to check their phase characteristics. But I get the impression from the old dogs that "dirty" filtering, presumably a filter with lots of out of band ripple and peculiar phase characteristics "sound" better... I am quite interested in this topic. Generally, if I can see an equation I feel comfortable. But when it comes to art, I am struggling. My wife says that I am too left brained! It occurs to me that there is a lot of interaction with psychoacoustics here. What is that quality that makes a filter sound good? - DM IP: Logged | |
|
David McClain Member |
posted 26 October 2000 18:21
John, I'm puzzled by your statement: [They are thin sounding, they have an unnatural mid-high resonance in normal use, and they make Kyma sounds instantly recognizable most of the time... (snip!) ...I would love to discover I have been misusing Kyma filters (the Q issue notwithstanding), and if only I did thus-and-so, they would have all the warmth and mellowness and ...] What is the Q issue? And I wonder if you are running into the same quandry that prompted my original posting. These filters are extremely flat across their passbands. There is a mid-band resonance in the LPF if you set the FeedBack parameter to anything other than zero. You can see the effect directly for yourself if you use a LPF with zero FeedBack, and then implement a 1-sample delay with adjustable attenuation (actually a gain with negative values smaller than 1) at its output, feed this back to a mixer at the input of the filter. This causes a mid-band resonance just exactly like the one that appears when you use non-zero FeedBack in the filter settings. Puzzled... - DM IP: Logged | |
|
John Dunn Member |
posted 26 October 2000 23:41
Good! I was hoping some discussion would ensue here. I would really like for Carla & Kurt to revisit their filters - I have mentioned it to them in the past, but I suspect they "know" the filters are near-perfect, and so discount the subject. Perhaps it is an accuracy vs. aesthetics type of issue - but Kyma is a musical instrument, or why are we bothering with it? In this context accuracy should be a tool to achieve a goal, the goal being aesthetics. I don't know the details of how Bob Moog came up with his filter design, but I do know he had lots of feedback from musicians, and my guess is practically none from his fellow engineers. Feedback would not have been, "filters fit a near perfect spectral description of filtering." It would have been, "Top end is a little harsh, can you mellow it out a bit? And perhaps fatten up the bottom?" David - the "Q issue" I mentioned is simply a reference to what you described at the beginning of this thread, of the analog synth's filters' resonance at the filter cutoff frequency. Most synths labeled this control as "Q." I would like to see an emulation of this in the LPF and HPF modes, but this is a red herring - the issue is the character of the sound, or lack of it. IP: Logged | |
|
David McClain Member |
posted 27 October 2000 01:36
I put together a simple little ResonantLPF yesterday. I have attached it here, along with the synth for which it was intended. It's basically a simple Kyma LPF with no feedback, in parallel with only one channel of a dual formant filter. There is a Q control and a simple crossfade between the two filters to vary the amount of resonance. The Q parameter varies the bandwidth of the resonance. This faithfully emulates the conventional synth filters. Why wouldn't custom Sound modifications like this satisfy the needs for fattening, or brightening, or whatever, of the existing filters? It seems like the Kyma system offers a bunch of sonic erector-set components that can be assembled in a huge variety of ways to achieve virtually anything one wants. - DM IP: Logged | |
|
David McClain Member |
posted 27 October 2000 01:42
BTW, I should mention that John's Web page inspired me to look at this whole issue, because there you mentioned, (or was it a posting from you?) that you sold your synths and now use only a Kyma... I wanted to see just how well this would work because I am initially hesitant to give up my other synths. As it turns out, Kyma does a very respectable job, but I'm still not ready to give up my Kurzweils, E-mu's, and Virus, nor any of the software synths that I use -- especially Tassman which is a physical modeling synth. Physical modeling, in particular is still beyond the reach of Kyma. (...or is it?) - DM IP: Logged | |
|
David McClain Member |
posted 27 October 2000 02:13
Also along the lines of "the sound of filters" folklore... I wonder how much of the difference between synths isn't so much in their filters, but in the nature of the sound they are pushing through their filters. This topic has come up before on the Nord Modular thread, with many of the seasoned old-timers harking back to the same issues of filter sound. As a result, I spent one evening pushing white noise through both the Nord Modular filters and those of the Virus/B. The filters can be made to respond nearly identically, but I do agree that a Nord sounds like a Nord (rich, metallic, bright), while a Virus sounds like it has a mouth full of chocolate balls. I like each of their respective sounds, but the filters don't seem to be the source of their difference... So how much of this is simply folklore? And how much is really a difference in filters? - DM IP: Logged | |
|
David McClain Member |
posted 27 October 2000 11:22
FWIW, I just measured the VAZ filters. Without resonance they appear the same as the Kyma 4th order filters with zero feedback. So this gets me back to my earlier hunch that the difference in sound has more to do with what is being pushed through the filters. While theoretically, a 4th order filter should give a rolloff of 24 dB/octave, my silly spectrum analyzer (based on FFT's) shows a rolloff of 18 dB/octave for both filters. However, after some work I have been doing for the past several days, attempting to produce a correct logarithmic readout for a spectrum analyzer, I believe that the discrepancy lies with the presentation of the FFT analyzer. When it is modified to show a proper logarithmic energy per interval then white noise shows properly with a positive slope with frequency (6 dB/octave in fact), and pink noise appears flat. That would mean that the rolloff of the filter will become steeper - probably closer to 24 dB/octave. ...come to think of it, I can use my incorrect FFT analyzer and just push pink noise through the filters to get the correct result... in fact this works, I just tried it! - DM - DM - DM IP: Logged | |
|
David McClain Member |
posted 27 October 2000 13:28
I wrote a piece of NML code to grab Wave files saved from a number of Kyma runs where I recorded 5 seconds of white noise being pushed through LPF's with orders running from 1 through 8. This code performs the FFT's and computes an averate periodogram with 4096 cells and then computes the mean rolloff for a variety of intervals between 1 KHz and 2 KHz, where the filters had a cutoff of 1047 Hz (6 c). Interestingly, all of the measured rolloffs show a deficit with respect to (6 * Order). Thinking about the problem of implementing a dynamically tunable digital filter leads me to believe that this has to be a very simple IIR filter. Remember we have considerably less than 20 microseconds to compute each filtered sample. Secondly, every measurement I make shows an extremely flat passband, and essentially, no measurable ripple outside of it. Now what kind of a filter behaves this way, is easily computable, and can readily adapt to a dynamically changing cutoff frequency? Well... I think the answer is a Butterworth filter! (Horror of horrors! - since I often hear many purists denouncing the lowly Butterworth). Say what you will, all the facts are leading up to this conclusion, and I, for one think almost all synths out there are using them! So once again, if I am correct, the difference in the sound is the way you use them - what you push through them, how much cutoff resonance you add, how you modulate them, etc. - DM IP: Logged | |
|
kelvin Member |
posted 29 October 2000 13:06
So once again, if I am correct, the difference in the sound is the way you use them - what you push through them, how much cutoff resonance you add, how you modulate them, etc. - DM[/B][/QUOTE] Dave I do tend to agree with you in many way as to the input to the filters have alot to do with the sound. I looked at many vintage synths signal paths with a scope and find that even there sawtooth waves, as an example, well look very different. Also the type of components used, transistors, or opamps, type of caps, etc. But there is also very much a difference in the filter types. I have a 2600 with the old filter, and new more stable osc..It sounds very different from 2600s with same osc. and newer filter. I have also found res. or Q or what ever it's called by each manufactor, to really sound different, and most of all in the digital versions. I too love Kyma, but don't like the filters in high res. For that matter I don't like many digital synths filters in high res. I do like the Nord lead, Nord Modular, and SuperNova, though. The filters are much smoother and do not have the distortion that can be heard when sweeping through the harmonics with high res. Kelvin IP: Logged | |
|
John Dunn Member |
posted 30 October 2000 00:41
David, it is great that you have taken the time to do a clinical examination of both Kyma and VAZ filters. More information is always better than less - unless it is only part of the picture, and you go away thinking it is the whole story. While I can't offer clinical facts or a scope readout, I can offer some 20+ years of working with modular synthesizers. I have owned an Arp 2600, Arp 2500, Emu, and Serge modular, and I have had studio time on a Moog modular. From this I can say, without the least doubt, that each synth does indeed have an overall sound and each synth's sound color is mostly its filter quality. The other components do add to the overall character, definitely. But it is the filter that really counts, and it is the filter that you "play" more than anything else. Kyma is much more than a subtractive synth emulator, of course. So weak filters on Kyma is not the killer it might be otherwise. Still, it would be very nice to be able to start with a rich filter when doing the classical subtractive stuff. And, yes, I agree that the idea of Kyma being building blocks is what it is about, but the basic components still should be rich and musical. It is not enough to emulate the functionality of classical synths, as you did; there needs to be the ability to emulate the rich and subtle sound of them - which was my original point. IP: Logged | |
|
David McClain Member |
posted 30 October 2000 01:53
Hi, Yes, I now agree with much of what both of you have said... I posted a query on the Nord Modular list and got the feedback from a number of experts there too. The general consensus is that a plain LPF can only remove harmonics, and not color the sound in any particular way. It is the "Q" applied that colors the sound. However, I feel that there is still a misconception at large regarding the Kyma FeedBack parameter. They chose the unfortunate parameter name "!Resonance" as its default value. Hence many of us immediately thought that was the same as the "Q" control. It is not! The Kyma FeedBack parameter is just that, and nothing more. It does impart some resonance to the filter, but not where you might expect it to be. It never occurs at the filter cutoff frequency, and large amounts of it send the system into oscillation at very strange frequencies. Positive feedback applied to a HPF gives the sound a resonance well above the cutoff frequency. Oscillation breaks out at around 8-12 KHz regardless of your chosen cutoff frequency. Negative feedback applied to a LPF puts resonance well below the cutoff (typically half). Large amounts of it also send the system into catastrophic oscillation. These oscillations are not the nice controlled ones you get when you narrow down a bandpass filter. These are catastrophic oscillations caused by the feedback moving the poles of the system beyond the unit circle in the Z-plane. Any attempt to use the filter feedback as the analogy to conventional synth's "Q" or "Resonance" will always be fraught with disappointment. But, given what I have just said, is it fair to criticize the Kyma filters for acting this way? I don't really think so. If you want to emulate classical synths then just parallel a narrow bandpass filter and leave the LPF feedback setting at zero. It isn't and wasn't intended for that purpose. I hope this helps to dispel some (widespread, and understandable) misunderstandings... - DM IP: Logged | |
|
gelauffc Member |
posted 01 December 2000 05:56
Hi all, I noticed your nice discussion about the Kyma filters. I also find the Kyma filter very weak sounding. Instead of discussing about this I wrote(in C++) an filter myself! I sended this to Kurt with additional design documentation. So far, no implementation is made for the moto. If you are interested in the sound of it? Send me an E-mail, I can send you some PC executables. If you like the sound of it, we can again ask Kurt to write the moto code for it. IP: Logged | |
|
David McClain Member |
posted 01 December 2000 08:46
gelauffc -- what is your E-mail address? - DM (dmcclain@azstarnet.com) IP: Logged | |
|
gelauffc Member |
posted 02 December 2000 16:18
christiaan.gelauff@philips.com (Work) or c.agelauff@freeler.nl (Home, please do NOT send big attachments to this address I have a slow modem, I also check this box just once in a while) I looked at the kyma sounds posted above and I think none of you found the sound (using the Kyma LPF) like I did. Already a year ago I found a way of controlling the (LP)filter in a (musically) usefull manner. It is not THE WAY, but a way, without the filter going crazy. I took quit some time to prepair this filter demo, you find attached. It is a .zip file + .wav file. Place them both in one directory. Open "filterex.kym" and of you go. 1) KeepFilterFromOverFlowing General remark: In these examples I do not use nice names for every object. Sorry to lazy. Also I use "!cc00 !cc01 etc" as controlling fader names. This is because I use a Peavey PC 1600 Midi Fader thing to control the parameters with real faders. I always have one program on my Peavey with midicontroller 0 until 15 on the faders. I did not create any nice names with a MIDIMapper for all the controls. Sorry for that, again I was to lazy. 1)+2) Have whitenoise as input. The sounds just demonstrate my way of controlling the Cutoff and FeedBack. 1) and 2) have some slidely different way of controlling the feedback part of the filter. 3) Is a more advanced sound, which soul purpose is to let you listen to what the Kyma LPF can do for you. I hope I have included all the files needed. Please let me know if something is missing, or what you think of the filter sound controlled in this way. CAG IP: Logged | |
|
Pedersen Member |
posted 10 December 2000 12:54
gelauffc: what kind of filter did you implement ? (name and type) IP: Logged | |
|
chuck welti Member |
posted 11 December 2000 15:18
Coloration is one of the major benefits of using high end audio gear and the assocaited musical qualities are what make this gear so popular. Look at the high-end mic-pres and compressors, tube-based, and they add coloration on critical tracks. Look at the popularity of old neve mixer modules for the same reason - they add a grit and fatness. If you look at the plug-in market for TDM or VST you will see a lot of plug-ins for these types of effects and for synths like Waldorf, etc. Is there a discussion thread anywhere in Kyma land for making it compatible with one of the major plug-in architectures. There are many types of plug-in EQs, compressors, all the Waves stuff, synthesizers, etc. It would seem to me that one of the biggest values of Kyma is its overall architecture and object oriented design, and many of the outstanding modules, but, if the architecture was extended to include plug-ins, it would not have to be the best of breed for every type of module for all time. Anyone else see this as high on their wish-list? IP: Logged | |
|
Frank Kruse Member |
posted 11 December 2000 17:00
if i wanted ANOTHER machine that "plays back" VST stuff i wouldn´t have bought kyma... IP: Logged | |
|
armand Member |
posted 11 December 2000 20:49
I'm sorry to say, i think it's not interesting the idea of plug-ins because most of them are imitations of vintage gear, and i prefer the original. And filtering i think there's a lot the be learned from this topic, because most of the time making sounds you're looking to get some kind of acceptable distortion rather then a clear and clean signal. The better the filter design, the harder it is to get an intresting sound. It's sound so clean you end up to apply some colour. Well i'm just a complete sound nut, i've heard a lot digital filters (analogue emulation) but i haven't been convinced by any of them. To get a really musical filter you have to go analogue. So i still use my analogue synth's a lot. Why this comment, because most people always refer to Moog, Arp and EMS. regards Armand. IP: Logged | |
|
gelauffc Member |
posted 13 December 2000 06:39
To Petersen: I implemented a 2nd order (butterworth) lowpass filter! It was designed to have a musical purpose. To Armand: Hoi mede bewoner van NL! (=Dutch) Also for others who are interested in hearing my filter please send me a note: christiaan.gelauff@philips.com IP: Logged | |
|
John Dunn Member |
posted 15 December 2000 06:06
Christiaan: I listened to the audio files of your filter and, yes, they sound OK, with nicely controlled feedback into oscillation. It's hard to really make a judgment on just a few sound files, and a lot of what we have been discussing here concerns not only the sound but also the ability to control the sound, to get there reliably - which in most analog type filters is a combination of the "Q" and cutoff parameters. But I liked what I heard. Armand: While I don't care whether I can emulate a particular manufacturer's sound, the ability to do so is a pretty good benchmark for a professional quality digital synthesizer to meet. It is one that, for example, Reaktor and Scope/Pulsor do quite well with, and Kyma fails utterly. Chuck: I agree with you about coloration being mostly what it's about, but I think you are looking at the wrong end of the telescope regarding plug-ins. When it gets its ASIO driver support, Kyma will become a wonderful processor - the mother of all plug-ins. But I wouldn't want to be forced to use filter (or any other) plug ins to prop up Kyma because it puts the load on my desktop computer instead of the offline processor (Kyma). For only a few channels of sound, this wouldn't matter. In fact for only a few channels of sound, VAZ2 or Reaktor do quite well, and you could save the other $3000. The point of having an offline DSP box like Kyma or Scope is to have the CPU power to run lots of channels of sound and/or do very complex things with it. Carla & Kurt: Here is a wish list item from me, and I suspect - given the traffic on this thread - others might want it also: How about putting this whole issue to rest by developing a module for us retro nuts: A RetroFilter. Something that would emulate those old analog filters, with a controllable resonance peak at the cutoff frequency. Perhaps you could even beta test it with your users before finalizing the design, for its retro-musical usefulness. What do you say? John Dunn IP: Logged | |
|
SSC Administrator |
posted 15 December 2000 09:24
quote: John- Yes, we agree that there are interesting aspects to emulating the old filter designs. We love sound and we are fascinated with all aspects of sound! But right now, we are immersed in a project to develop a set of new synthesis/processing algorithms which are going to open uncharted areas for timbre exploration in Kyma. The retro filter is on our list and we do intend to implement it. We will enlist all of you as beta testers when it is ready. IP: Logged | |
|
John Dunn Member |
posted 15 December 2000 11:12
quote: Cool! Enough said. John Dunn IP: Logged | |
|
armand Member |
posted 15 December 2000 18:30
To Christiaan, Hoi mede bewoner van NL! (=Dutch) To John and others, Of cource my opinion is very personal. And about the digital alternatives have been too precise and unmusical, is not entirely truth. Well I don't compare digital and analogue filters. So here is my positive view about digital: for example the Nordlead and the Virus. It sounds like analogue at first, a bit like synthesizers of the mid-1990's. But they themselves now sound like synth's of the mid1990's, and that's their charm. It's also what made them sound exciting at the time, although we thought it was analogue. The ear is always fooled into believing that some new modelling technique has somehow brought us closer to the original, the particular manufacturer's sound. Maybe that's the manufacturer's job: to exite the musician by creating proximity to the synth's. And for me Kyma is just one of the best in digital life. And also I'm looking forwards for a Retrofilter. Armand. [This message has been edited by armand (edited 15 December 2000).] IP: Logged | |
|
Marcus Satellite Member |
posted 24 January 2001 14:54
RETROFILTER: i third that. VERY important for me as a musician to get resonance at the cutoff frequency. i make really cool melodies this way. i have stopped using kyma as a pitched synth for this reason alone and instead rely on the mc505, 303, et al for this class of sounds. this is a shame because the programmability of kyma is such that i have created really nice performance expressions that i prefer over most patches on consumer synths. if a retrofilter were to behave 1/2 as having said that, i continue to use kyma in my mind, the spectral stuff is what distinguishes kyma from the rest of the world. with some of the kyma5 presets i CARLA/KURT: i'm very excited to hear IP: Logged | |
|
Mott Jordan Member |
posted 31 January 2001 12:37
Ahh, the filters, I came up with a useful tweak or two that have added some real "balls-iness" to the filtered synth sounds that I've been working up. The aforementioned idea of paralleling a bandpass filter does go a long way towards making the sound more authentic and analogue-ish but what really took me in the direction I wanted to go was when I was trying to create a virtual guitar amp in Kyma. Obviously the, cleanliness/dirtiness of the sound depends on managing the gain stages and signal clipping. But try this: create your own unique waveshaper curves, (lots of them) and insert them in many spots along your signal path; before, after, between different filters. Sine is great, but not interesting or "exotic" enough, others that come with Kyma are way too extreme (e.g. arctan) and will destroy any pleasing nuances of the signal. Get complex. Complexity is something that happens as part and parcel of control voltages, and the modular nature of Kyma allows us to insert it at every turn. IP: Logged | |
|
John Dunn Member |
posted 30 January 2003 09:29
quote: So, it's been three years. Will this "Retro Filter" - or any filters that have a more musical sound, that is, a constant Q with respect to cutoff rather than the wild and unmusical "resonance" parameter of the current filters - will something along these lines appear in the new software due out Real Soon Now? IP: Logged | |
|
Mark Member |
posted 30 January 2003 12:34
I would just like to add my encouragement to the fine folks at SSC to revisit Kyma's filters. I would tend to agree Kyma's filter's behavior and (for lack of a better word) "tone" have been one of my few mild disappointments with the system. There are years of anecdotal discussions regarding filter behavior from a subjective vs. scientific standpoint. Common points that rise are saturation/overload characteristics of the input stage of the filter, resonance behavior at filter frequency (one the most revered filters in analog synth land, the Minimoog, lowers resonance with filter frequency somewhat, as do many diode filter designs), and most importantly, the behavior of the filter as breaks into oscillation at high resonance. I'd be happy to use more dsp resource's for a more complex Retrofilter that encompassed many of these ideas, and was less likely to "break" (set the frequency of a Kyma highpass filter down very low, and vary input amplitude to see what I mean). I'd also like to ask Carla and Kurt to share/tease us with just a little more info on the new sound shaping they so briefly mentioned! Best- Mark IP: Logged | |
|
SSC Administrator |
posted 30 January 2003 13:00
"I'd also like to ask Carla and Kurt to share/tease us with just a little more info on the new sound shaping they so briefly mentioned!" Sure, that was Aggregate Synthesis/Processing! IP: Logged | |
|
Mark Member |
posted 30 January 2003 14:21
Whoops- I suspect I need to pay closer attention to the post dates on related posts! Nevertheless, consider this still an only slightly embarrased vote for updated filter modules.. Mark IP: Logged | |
|
John Dunn Member |
posted 05 February 2003 10:30
So, by your silence, is it safe to conclude the new release will not have updated "musical" or "retro" filters with a "Q" parameter for resonance at the cutoff frequency? IP: Logged | |
|
JohnCowan Member |
posted 05 February 2003 11:03
John, I am sorry I don't get your complaint. I was able to put a lowpass filter and a bandpass filter in parallel to create a resonant filter with "Q". Using a noise source and a spectral analyser I could adjust the parameters to achieve what I thought to be proper filtering. This took about an hour. Are you wanting a certain type of harmonic distortion or frequency phase distortion? Wouldn't it be more interesting to come up with some ideas collectively? If a result is overly complex, perhaps SSC could write low-level code. John Cowan IP: Logged | |
|
SSC Administrator |
posted 05 February 2003 12:47
"Wouldn't it be more interesting to come up with some ideas collectively?" This sounds like it could be fun. (BTW, generally speaking, the lack of response to a post here shouldn't necessarily be interpreted either negatively or positively. Thanks!) IP: Logged | |
|
John Dunn Member |
posted 06 February 2003 01:58
"I am sorry I don't get your complaint." If you read the other messages on this thread, I think you will get the gist of it, but here is a summery: Kyma filters are, in a word, unmusical. As has been described here and elsewhere, this is largely due to the fact that there is no "Q" that tracks the cutoff, and the "Resonance" parameter exhibits rather severe artifacts. I find it peculiar that while Kyma is so strong on other things, it should be so weak on this fundamental sonic building block. Other digital synths - virtually ALL other digital synths - have solved this problem and provide good, musical filters. Why not Kyma? I have been asking for this for years; in fact this thread was started two years ago, and ended at the time with a promise that there would be a "retro" filter. I'm still waiting to see it. As for making one ourselves, well sure, perhaps some interesting assembles could be made, and I am all for it. But that is no substitute for a set of good, musical, basic filters. Pulsar/Scope has them, VAZ has them, Reaktor has them, Nord has them, hell everybody has them, everybody but Kyma. As a consequence, Kyma generated sounds that use filters are generally instantly recognizable by a thinness of texture and a signature "whining" quality in the high register, possibly a result of the HPF artifact at 8-12 KHz regardless of cutoff that David McClain described above. What I haven't figured out, and I guess I need to, is whether Kyma is an audio engineering tool or a musical instrument. As an engineering tool, it has no peer; as a musical instrument it is still lacking in the fundamentals. I was hoping to get word that this has been fixed, and frankly I didn't think it would be such a problem for SSC to let us know if it was finally forthcoming. IP: Logged | |
|
JohnCowan Member |
posted 06 February 2003 10:45
John, It seems that there are a number of good ideas in the previous posts. It also seems that some "adequate" musical filters can be made without too much trouble from the Prototypes (that is with a resonant peak at the cutoff with adjustable Q). Everybody seems to agree that using the Kyma lp filter "Resonance" parameter is not really useful. Furthermore, there seems to be no one musical filter that is going to satisfy everyone. Yes, it would be wonderful if SSC supplied a dozen musical filters for us to play with. If they don't, does that mean Kyma is worthless as a musical instrument because it is difficult or impossible to emulate some vintage analog synth? It thought Kyma was about sound design not precise emulation of something else. Surely, lots of other companies have taken up that nitch in hardware and software. Does anyone have semi-technical info on some of these analog filters? It seems to me this would be a more interesting direction for this thread. Maybe we could uncover what characteristics we like about them? John Cowan IP: Logged | |
|
photonal Member |
posted 06 February 2003 17:09
I think John Dunn does have a point - The squealing of the filters sometimes is not something to experience loud! I don't think Kyma should necessarily emulate another synth sound, but it should have some filters which are on par with great sounding synths - I mean in terms of being musically desirable and of a recognisable flavour. Of course this all comes down to taste and fashion. That's one reason why I hardly use my Yamaha DX7 anymore, a lot of it's sounds have a recognisable sound which I find undesirable. The great thing about Kyma though, is that the filters could evolve to suit current or required tastes. There was a sound developed by David McClain (Jarre Pad?) which for me was a great sounding analog emulation. Maybe some filters could be based on that? If SSC don't supply 'a dozen musical filters for us to play with' it won't lessen Kyma's worth but it does lessen the usefulness of Kyma to some members of the Kyma community who aren't involved in Electroacoustic music.
[This message has been edited by photonal (edited 07 February 2003).] IP: Logged | |
|
EckardVossas Member |
posted 11 March 2003 08:52
I would be interested in hearing and examining jarrepad2. But unfortunately the file has disappeared. Is it possible to make it downloadable again? Thanks. Eckard [This message has been edited by EckardVossas (edited 11 March 2003).] IP: Logged | |
|
John Dunn Member |
posted 02 October 2003 05:55
Does Kyma-X, whenever it is finally released for real, fix the unmusical filter syndrome of previous Kymas? IP: Logged |
This topic is 2 pages long: 1 2 All times are CT (US) | next newest topic | next oldest topic |
  |
|
Contact Us | Symbolic Sound Home
This forum is provided solely for the support and edification of the customers of Symbolic Sound Corporation.
Ultimate Bulletin Board 5.45c