#244

@ zoundsabar thanks you for bringing this module to my attention. This would be a far more elegant solution to the problem posed by the J.Haible String Filter (link in one of my previous posts).

On other news: once again, the idea of Performance overshadowing Sound Design is verified. I wasn’t very fond of the Bowed model on MI’s Braids, but after testing it out with a Doepfer a-198 I have to recant my original review. It sounds wonderful for violin sounds, and the same goes for a similar application of Mysteron for cello sounds. As soon as you plug the right parameters out of the ribbon controlling the right components of the sound source, the pieces fall in place and realism arises.

I’m missing the body and the exciter elements but I’m still investigating.

1 Like
#245

http://www.ron-berry.co.uk/page14.html
“The two signals are mixed before going into the both VCA control input and its audio input. The amount of control the VCA has over that mid gain bias setting is set by the voltage control knob. It can go from nothing to a ‘huge effect’ as it does in the sound clip below. The little switch enables the VCA to handle positive and negative control signals if required (the 4 is 4 quadrant operation) which is what’s needed here.”

Could someone give me a hand in this, I’m not sure If I’me getting the mechanics of the VCA in this patch.
Am I suposed to mix [audio+cv+delay feed] into both the CV input and Audio input of my VCA, because that is bugging out the sound completely

#246

More convolution options can only be a good thing for physical modeling in hardware!

By the way, if you think about it when it comes to physical modeling, what you need in Convolution is EQ curves of instrument bodies, right? Which means that the Rossum Morpheus is like a stereo convolution EQ, with pitch, morph, and other control parameters, and 289 presets, instead of allowing you to load your own samples. Not so bad!

4 Likes
#247

That thing looks amazing! It addresses three big issues that have kept me from getting a C0, the length of the IR’s, the ability to crossfade between them, and onboard recording. Looks like you can’t stretch the IR’s though which would be a shame if true, since that makes for some really cool sounds.

The Morpheus kind of lost my interest after I found out you can’t actually build your own filter cubes in it. I think I’d get bored with it, being stuck with only a choice of how much of their baked in morph and transform changes come attached to each resonator instead of getting to create my own new forms. For example, I was really excited about using it to emulate this cello body patch, sending my V/8 signal into the morph control to change the amplitude of each formant peak as I play different notes, but that can’t actually be done.

#248

Korg Kronos perhaps?

#249

Perhaps the Xaox Zadar module?
Great for creating custom waveforms, and chaining custom waveforms together.

#250

I don’t have one but I thought the Zadar was only in the time domain and not the frequency one.

1 Like
#251

Apologies. I thought CV Envelopes could be applied in a variety of ways.

I do not yet have the Zadar.

#252

I was also taken aback by the lack of actual in-depth editing of the filters to create custom ones. First glimpse gave me the impression that you could indeed join different types of curves, but it turns out you can’t, and even though the reasoning behind this is valid, it still leaves you wanting. That being said, 289 filter cubes seem like a more than apposite leverage for this type of sound design.

A large chunk of them are based on actual instruments, from brass, to woodwinds, to stringed, to pianos, to percussion. I’m unsure how good these will be, but from a parameter perspective, I don’t think more than three parameters are needed: morph, transform, pitch sounds good enough to me. Maybe it’ll turn out to be limited, but on paper it looks like an amazingly versatile solution. For instance, I’m not sure I get why I’d need to crossfade between IRs, although it does seem like a fun thing to do. The C0 also seems enticing but if I’m being honest I wouldn’t use it as much as the Morpheus because it requires some ground work to use the IRs.

Thank you for that link. If you mean you can’t actually design that EQ curve, you are right. But you can plug CV on the Morph control to change each peak as you play different notes. Since you brought it up, I’m not sure why you’d want to change the body parameters of a string instrument for each different note; I don’t get the physics. I’d assume the pitch parameter ought to be changed via V/8 signals, and that would suffice. In the meantime, the characteristics of the body (re: morph, transform) would remain the same throughout the range of notes. No? :thinking:

Maybe “pressure” would change one of these, to model the strength of the bow on the string.

#253

I’ve read very mixed opinions concerning the Morpheus – what’s your take? I think that, in principle, it’s a very interesting module for resonator type duties, but some comments about there being clicks, noise and other glitches…

#254

No comments yet as I’m grabbing mine on Saturday from a friend. I’m hoping it will solve my “body” problems, but I’m always moderate in my enthusiasm, so let’s see how it goes.

From all the comments I’ve read on MW I’ve only seen people getting disappointed by the lack of morph between cubes. The sequencer goes from cube to neutral and then to another cube. It can’t possibly switch from one to the next with no “non-filter” in between. Maybe that’s what causing the clicks? I’m willing to assume that noise issues are on the fault of the users not having read the manual and happily perusing the filters. Some require attenuation and/or change of the gain levels. This is not a no-brainer digital filter to have and just fool around with. The manual is mandatory reading, at least the part where the cubes are described, and it might take time to find the appropriate cube for the job, customize it (levels etc), and save it for later use.

These are all assumptions of course. :smile:

1 Like
#255

just to add my 2cents to this body filter topic. i found the best (meaning most natural) results when the body was static, not tracking not changing, just like a regular instrument body it doesn’t change shape, the resonator and exciter do however. best would be to have a big graphic equalizer with 40 bands or something. i just yesterday found new violin body IR and they sound pretty good in the c0 without modulating the module whatsoever, of course everything before has then to be adjusted to the frequency range of the specific instrument and the “barriers” while playing really made it logical to me why the body doesn’t change shape, eg lower notes for example aren’t loud, the body just doesn’t excite these frequencies etc, so in a way i learned to use specific IR for specific registers. of course mixing them up is fun too but tracking the body shape i found sounds not like physical modelling but rather like subtractive synthesis.

4 Likes
#256

and that brings up an idea… if someone has the morpheus filter, it would be really cool to make a few IR’s from the different instrument eq shapes in it :blush:

#257

I think the reason why they are modulating the filter peaks based on note number is that they’ve put together a very simplified model. The cello body has hundreds of resonant peaks, and so different notes will resonate differently. An IR can capture this nuance so you wouldn’t need to modulate that to get good results, but with the 14 peaks of Morpheus there will be spots where the notes just don’t ring out right. I’ve found the same thing when using Three Sisters as a body resonator, some notes within a small range will sound right, but outside of that they start to sound dead or too resonant so I usually send my pitch CV to alter both the resonance and the frequency.

And while you can control the morph or transform functions via a pitch CV, you can’t decide what shapes you’re morphing and transforming the filter into from what I can tell in the manual. So I couldn’t set up different points along that pitch scale where the notes sound right and then morph between them as I play. You can only control how much your incoming CV moves the filter cube along its morph and transform paths to its next preset state.

All this said, I’m sure it’d sound a heck of a lot better than my Sisters and probably better than anything other than a convolver. If it were just a choice between the Morpheus and C0, I’d probably choose the Morpheus because I really don’t like the C0 interface or sample restraints. But this ECR-1 looks a lot more appealing to me, especially for the weirdness that it could do beyond PM patches. I love running sounds through a convolver with weird percussive stuff loaded up into it, so the 20sec recording time and ability to record live on this would make for some really awesome spontaneous stuff!

4 Likes
#258

Yes, exactly. You model the bow and the string, and allow these to determine your pitch and then the body comes as a shell over the two and responds accordingly. I don’t think brass instruments would apply to this concept, as the valves actually determine the length of the body, while the exciter remains static. Not sure yet, just making an assumption here.

That would be something like the J Haible String Filter I linked to recently. PCBs for this amazing concept are available from Random*Source but there is no pre-built option. The design looks massive! The idea was that another preset-PCB would be added with different configurations, but the current design allows each of the 40-bands to be fully configurable from the front panel, if you have the space on your rack and time on your work bench for such a beast. :smile:

I actually read the electro-music thread, and had a preliminary look on some of the other pages regarding modeling of string instruments. Again, thank you so much for sharing the link! It contains very useful info and, as always with this here thread, I now have more stuff to digest and more concepts to struggle with.

I see your point. My original assumption would be that controlling the pitch parameter via pitch-CV would suffice to bring the resonant peaks to the appropriate frequency. The way I imagined this works is that Morph and Transform are the equivalent to changing to body in order to change the response of the instrument, in the same way that Chet mentions changing some parameters of the patch to do multiple recordings of “different” instruments.

From what I understand you are making the assumption that the Moprheus has Morph and Transformation set up in a “linear” fashion, or at least in a way that might not be appropriate for modeling of the sound. Otherwise why would you need to delineate a different relationship between these parameters and your pitch in order to scale “where it sounds right”?

One solution to the problem you’re posing would be to utilize something like the MI Frames, which can morph between points, and map each of your pitch values to different control values on Frames, then patch these values to Morph and Transform in order to have specific notes hit specific Morph and Transform combinations. Shouldn’t be too hard to do actually.

I’m unsure who programmed the cubes but from reading the manual on the C241 String Cube (page 70):

C241 String Cube
Controls high frequency response through CV to Transform, but with the added possibility of emphasising the 2nd harmonic (first overtone) with Morph parameter, à la strings.
Frequency: Maintains harmonic filter tracking with pitch CV patched to Full Level input. Increasing the value opens up the filter to a flat response with a bump at a high> frequency range.
Morph: Controls timbre change; picks out first overtone. -5V value gives a low pass filter with a bump at ~500Hz, increasing the value moves the bypass to lower frequency ranges;
Transform: Determines brightness and volume. Higher values give an almost flat response filter with a bump after ~2kHz

my first assumption would be to tie these parameters to playable things like pressure and velocity. :thinking:

1 Like
#259

I’m not sure what you mean when you say I may be assuming the Morph and Transform controls are set up in a linear fashion, but all I meant is that you don’t have control over what the morph and transform functions do for a particular cube. The Morph and Transform behaviors are pre-defined for each one like it shows in the snippet you posted. I get why they did it though, and in the MW thread one of the designers talks about how the filter’s capable of so much gain that it’d be easy to accidentally produce dangerously loud signals if they gave people full control over all the parameters.

#260

Apologies, I have not made my point clear.

By reading your post it seems to me that you’d rather eschew the way Morph/Transform goes from A to B to C, as particular values might be more appropriate for notes other than the ones you’d get if you just patched your Pitch value at the Morph/Transform CV input.

A solution to that would be to redefine how Morph reacts by having Frames intervene (go from A to C to B), so that different notes would have different Morph values, other than what Rossum has considered appropriate.

What I was trying to say though, is that I’d like to believe that Rossum has assigned Morph and Transform to appropriate performance parameters, in a way that enhances the physical modeling experience. Indeed it seems that you can peceive that Morph/Transform control more than one parameter, in a sort of meta-control way, but even then it would be apposite for a physical model as long as you control that meta-control using the correct performance controller. So, Frequency would be tied to pitch, and maybe bow pressure would be tied to Morph, or velocity to Transform, etc.

Personally I don’t consider their behaviour as predefined, but rather as normal transition from one state of the replicated body to another (e.g. Morph goes from an instrument body experiencing low excitement to high excitement). I could be wrong of course, and I don’t mean to sound like I’m defending the module, but that’s what I get from reading the manual. I understand why you might find that limiting, especially if you’d rather repurpose parameters in a non-realistic way. I’m always interested to read your thoughts, so thanks for sharing. :slight_smile:

#261

Okay I think I get what you’re saying. And yeah, I doubt there’s anything wrong with the way Rossum set up their cubes. I’m not saying these things as criticisms of the module, just how its actual functioning differs from what I had thought when I was first considering it.

Very interested to hear what you do with yours once it arrives!

#262

Stumbled upon a way to better mimic the sound of pitch sliding on a string. For bowed strings I’ve been using a drone oscillator synced to another unheard oscillator that receives the V/oct signal. But I just found that if you invert a copy of your V/oct and attenuate it and send it to the synced oscillator it creates some effect that sounds almost exactly like sliding your finger across a string to change the pitch. I don’t know what physical process it’s mimicking from real strings but it sounds dead on to me. Something about how the harmonics change along with the length.

8 Likes
#263

that sounds very interesting! so you have the attenuated v/oct on the oscillator we hear right? could it be that formants play a role in this sound result? i’m not exactly sure about how formants really work but as i understand it they build up the resonant body of a sound regardless of it’s fundamental pitch. so with sync the heard oscillator forms the formant and the unheard the pitch (or is it the other way around?) so in your example you have a stable pitch but a detuned formant which maybe makes it sound so natural? again i have no idea about this but it sounds really good!

1 Like