teeshirtguy5

has anyone ever had any success modelling a cello in max?

im interested really in something that it extremely pitch bendable (3 octaves) and so takes continuous frequencies rather than midi notes for wild bendy string noises (which is why i cant use something sample based)

+ something with some dynamic range, i.e. bow pressure, position etc etc

i've got the percolate externals and they seem to be perfect functionally but i just cant get them sounding anywhere near a real cello

i've just been using the bowed~ external by itself, maybe in conjunction with something else it could be useful?

really i'm just trying to see if anyone else has had a bash at this and had any luck / revelations

modelling-a-cello

then add two further things: the mechanism for the bow incl. modulations and such, and, behind the string model, a bank of resonators which are modeling the corpus´ formants.

If you take the fractional value of the frequency to midi conversion and scale it to whatever [bendout] value corresponds to a semitone, you can match the pitches in-between midi notes. That said, I have yet to hear a synthetic cello that doesn't sound like shit... which leads me to believe that modeling one is not an easy thing to do. Check out: 

https://cycling74.com/tutorials/physical-modeling-synthesis-for-max-users-a-primer/

If you take the fractional value of the frequency to midi conversion and scale it to whatever [bendout] value corresponds to a semitone, you can match the pitches in-between midi notes.   That said, I have yet to hear a synthetic cello that doesn't sound like shit...  which leads me to believe that modeling one is not an easy thing to do.   Check out:  https://cycling74.com/tutorials/physical-modeling-synthesis-for-max-users-a-primer/


yup i have been using bowed~, getting quite good textures from it but still not quite right

@Roman i wonder what would be the equivalent of bow mechanism modulations? i'm driving the glissando pitch bends with an lfo but it might be cool to use that to simulate some bowing mechanics too

@Metamax i had a go with this too but had some trouble with the big pitch ranges, perhaps instead of bending one note i can do a kindof portamento up the notes somehow, i might look into what techniques ROLI use, maybe there's something in the max toolkit for that that could be useful

when i think about it, this is probably already present inside the bowed object. (i am not familiar with it)

in a karplus/strong type of oscillator it would be something like a noiseburst (pink*pink*pink*pink) with envelopes and modulation inputs for vertical position and pressure.

yes it is very true what you say about synthetic cellos sounding like shit

i mean mainly i need this just to write this piece i'm doing that involves lots of harmonics and playing on the bridge etc of multiple cellos. it will ultimately be played by 7 real cellos i just wanted something semi accurate to write it with

i wonder if someone (clever) could build a sample based one using a multidimensional wavetable synth

sample a string at different velocities and bowing pressures and bow positions and match the phase up exactly and then just be able to move in various dimensions

i.e. stay on same velocity but increase bow pressures

i imagine that is quite far fetched though and i guess also a cello is never really static so you'd have to bring that in somehow

i mean mainly i need this just to write this piece i'm doing that involves lots of harmonics and playing on the bridge etc of multiple cellos. it will ultimately be played by 7 real cellos i just wanted something semi accurate to write it with 

i.e.  stay on same velocity but increase bow pressures

i imagine that is quite far fetched though and i guess also a cello is never really static so you'd have to bring that in somehow


i guess if i model some more resonances / put a body after it and some reverb it could sound quite good with some EQ and compression

yes it is very true what you say about synthetic cellos

I should clarify... the synthetic cellos I have heard don't sound convincing as cellos - that's all. I do find it interesting that the current state-of-the-art allows us to visually model a far more convincing cello even though the fundamental elements of sound are much easier to represent. A hundred years ago, I could refer to the note D and someone across the world would be able reproduce it without much trouble. Doing the same with hue was a lot more difficult due to the perceptual nature of color. Yet for some reason, as everything gets more complex, texture easier than the timbre. We can build these amazing 3d worlds but we're stuck with Bach on kazoo playing in the background. Well... not really.. but kinda.

I should clarify...  the synthetic cellos I have heard don't sound convincing as cellos  - that's all.    I do find it interesting that the current state-of-the-art allows us to visually model a far more convincing cello even though  the fundamental elements of sound are much easier to represent.   A hundred years ago, I could refer to the note D and someone across the world would be able reproduce it without much trouble.  Doing the same with hue was a lot more difficult  due to the perceptual nature of color.   Yet for some reason, as everything gets more complex, texture easier than the timbre.    We can build these amazing 3d worlds but we're stuck with Bach on kazoo playing in the background.     Well... not really..  but kinda.


yeah quite crazy, i guess it could be down to lack of effort for development ( aka lack of commercial incentive for audio vs visual stuff)

having said that the best model of a cello i've come across from the physical modelling article so far is actually the cello preset in ableton's tension

play with the force and position on the lower notes and you get a really nice scratchy tone, turn the velocity down and it sounds like the bow is bouncing on the strings. similar to the percolate ones just way more, maybe they just filled out the same model of it with a body and some reverb

tassman has a few good physical models of imaginary percussive things that sounds great and 'real' even if they don't sound like what they are supposed to

i think i was barking down the wrong tree in trying to make my own physical model of one, the ableton one definitely is good enough. i'm going to add some modulation to it to try simulate the bowing, maybe something like a little peak on the bowing velocity on the local minima / maxima of a triangle lfo and a subtle difference (~1%) in bow position moving up and lfo vs moving down

if you haven't tried that one i do recommend having a bash, i was very surprised, miles ahead of what i got (crudely) out of percolate + tassman

im going to just do the fractional midi trick and use this

tassman has a few good physical models of imaginary percussive things that sounds great and 'real' even if they don't sound like what they are supposed to 