6 ms·

This winter I went down this rabbit hole and back up, trying to approximate some acoustic instrument sounds. I went through a succession of different workflows,

This winter I went down this rabbit hole and back up, trying to approximate some acoustic instrument sounds. I went through a succession of different workflows, spending a couple weeks on each. Workflow 1: feed a sawtooth oscillator through filters controllable with knobs. Eventually I realized that it will always sound mechanical, no matter how many filters you stack. That led to Workflow 2: feed a sawtooth oscillator through a convolution reverb that uses a custom impulse response. For impulse responses, use random sounds downloaded from the internet (like wood strikes), or mixtures of existing instrument sounds. But that felt limiting, so I moved on to Workflow 3: generate an impulse response wav file with Python, and use that in a convolution reverb to filter the sawtooth. This gave me some more interesting and configurable echoes, but then why start with an oscillator at all? So Workflow 4: write code to generate the sound on the fly, as a sequence of samples. This way I can mimic some nice properties of physical sounds, like "the nth harmonic has initial amplitude 1/n and decay time 1/n". Also I can get inharmonicity, smearing of harmonics (like in PADsynth algorithm), and other nice things that are out of reach if you start with periodic oscillators. If I could go back and give advice to my months-younger self, I'd tell me to skip oscillators and filters, and jump straight into generating the sound with code. You need to learn some math about how sound works, but then you'll be unstoppable. For example, here's a short formula I came up with a month ago for generating a pluck-like sound: https://jsfiddle.net/yd4nv5Ls/ https://jsfiddle.net/yd4nv5Ls/ It's much simpler than doing the same with prebuilt bricks. The whole experience made me suspect that there's an alternative approach to building modular synths, based on physical facts about sound (as opposed to either starting with oscillators, or going all the way to digital modeling of strings and bows). It would be similar to physically based rendering in graphics: for example, it would enforce a physically correct relationship between how high a harmonic is and how long it rings, and maybe some other relationship about what happens to harmonics at the start of the sound, etc. But I'm not an expert and can't figure out fully how such a synth would work.

Synaesthesia 8y ago
FM synthesis can generate a very broad array of sounds. A lot of the time you get bad results but if you’re looking for something that can make all kinds of different sounds that’s one option. There are actually so many synthesis techniques, you should download some software synths and take a listen. Really fun stuff
- zambal 8y ago
  I think FM synthesis is really underappreciated for modeling acoustic sounds. It really shines for mallet instruments, but also brass and woodwind and even plucked sounds like acoustic guitar. By making operators react differently to velocity and other modulation changes, you can create very expressive sounds. However, it requires quite some trial and error indeed.
- posterboy 8y ago
  The idea of a modular synths is to have the potential to create all of that, FM, AM, Sub and more. Isn't it?
  - detaro 8y ago
    Partially/Depends. Yes, they aim for more flexibility, but that doesn't mean they necessarily do a good job of providing the things you want for specific synthesis types.
TheOtherHobbes 8y ago
Subtractive makes a lot of classic synthesizer sounds, which is why it's a thing. It's also easy to understand. It seems to be easy to implement digitally, but it really isn't, because a lot of the nuances and non-linearities that add weight and colour to synthesis with real electronics aren't present in simple digital emulations. For pure DSP the choice is more or less between open additive, modal (which is a kind of constrained additive), AI-constrained additive, which is what Google have been playing with, and physical modelling, which is digital modelling of strings and bows. If you want to "enforce a physically correct relationship" between etc you're going to want AI-constrained additive or physical modelling. The aesthetics of all of this are a different topic altogether.
- cousin_it 7y ago
  > If you want to "enforce a physically correct relationship" between etc you're going to want AI-constrained additive or physical modeling. I was thinking of relationships like these: 1) Decay time of nth partial falls as a certain formula of n. 2) Frequency of nth partial is slightly different from n * fundamental, by a factor which is a formula of n. 3) Spectrum of nth partial isn't a delta function, but a hump whose width is a formula of n. All these ideas come from physical effects, but you can use them to generate sounds directly, without any physical modeling or AI. My hunch is that could be more such ideas and they could play together nicely.
  - bravura 7y ago
    Do you mind emailing me?
zambal 8y ago
> The whole experience made me suspect that there's an alternative approach to building modular synths, based on physical facts about sound Yamaha experimented with this in the 90s with their VL series physical modeling synths, but it never caught on, mostly I think because if you want to have convincing results, you really need alternative midi controllers like a breath controller[1] for woodwind and brass intruments. An alternative take on why physical modeling synths never really caught on is GigaSampler[2]. It was the first sampler (as far as I can remember) that could playback samples from hard disk, by only keeping the first second or so of the samples in memory. This made it possible to have sampled instruments where for example each key of a piano was sampled at various velocity/loudness values. Resulting in a sampled piano that could span multiple gigabytes. At a time where 128MB of RAM was still quite a lot, this was quite revolutionary. While physical modeling can produce convincing sounds with a potential expressiveness that no sample based instrument will ever match, it's base sound still doesn't sound as 'real' as a properly sampled instrument, recorded in a nice room with good microphones. [0] simple overview, including some soft synth alternatives: https://www.musicradar.com/news/tech/blast-from-the-past-yamaha-vl1-634920 https://www.musicradar.com/news/tech/blast-from-the-past-yam... [1] example breath controller: https://www.akaipro.com/ewi5000 https://www.akaipro.com/ewi5000 [2] Review of Gigasampler's successor: https://www.soundonsound.com/reviews/tascam-gigastudio-4 https://www.soundonsound.com/reviews/tascam-gigastudio-4
- byproxy 8y ago
  I feel if you're going to such lengths to approximate physical instruments..maybe just record physical instruments? Synths can do things that are unable to be done physically, so why not use them to that end? Although, I understand sometimes it's an aesthetic choice, having a poorly-approximated physical sound.
  - posterboy 8y ago
    Because playing the violin is hard and emulating the thing is a different challenge. Of course, sequencing a violin track with just note on/off and velocity (i.e. with a keyboard) is a poor-approximation. Better controls and using automation can be transfered back to other sounds for example. It's serious fun at least. We say to play music, after all.
johneffort 8y ago
There are physically modelled instruments like the one from Pianoteq. No samples, pure rendered. Very nice.
byproxy 8y ago
Slightly related: your pluck sounds reminded me of the beginning of "Tonto" by Battles (https://vimeo.com/117914655 https://vimeo.com/117914655)
jacquesm 8y ago
Hello again :) There is this phenomenon called 'stick-slip' that is an essential component of any string based instrument that is bowed. It might be worth modelling that to see what it sounds like, you might end up with some very realistic bowed string sounds as well as a glass organ. https://en.wikipedia.org/wiki/Stick-slip_phenomenon https://en.wikipedia.org/wiki/Stick-slip_phenomenon
campground 8y ago
Plucked string instruments are one of the simpler classes of instruments for physical modelling. Waveguide synthesis uses delay lines and filters to model a wave propagating along a string and reflecting off the bridge and nut. The famous Karplus–Strong algorithm used a burst of white noise as the excitation, but I've had more success using an asymetric triangle-shaped impulse that resembles the shape of the drawn string.
ravar 7y ago
Check out the sub synth in zynaddsubfx, its really easy to get started with it and makes a great sound.
mochomocha 7y ago
In grad school, I had a tiny start-up with a classmate selling a virtual instrument doing just what you describe (for guitar sound synthesis). We had PDE models for the string, finger/string interaction, fret/string interactions etc. We would then discretize these PDEs to generate in real-time a synthetic "force" applied by the string on soundboard, and then convolve this in real-time with measured impulse responses of real acoustic guitar soundboards. Our website is still up (spicyguitar.com) and we give it for free now, but I don't know if it still works in recent DAWs.
- Severian 7y ago
  Loaded it up and I can confirm it still works in FL Studio 32b and 64b versions. I really like the music box preset, very melodious with certain parameters.
  - mochomocha 7y ago
    Thank you!

Synaesthesia 8y ago

FM synthesis can generate a very broad array of sounds. A lot of the time you get bad results but if you’re looking for something that can make all kinds of different sounds that’s one option. There are actually so many synthesis techniques, you should download some software synths and take a listen. Really fun stuff

zambal 8y ago
I think FM synthesis is really underappreciated for modeling acoustic sounds. It really shines for mallet instruments, but also brass and woodwind and even plucked sounds like acoustic guitar. By making operators react differently to velocity and other modulation changes, you can create very expressive sounds. However, it requires quite some trial and error indeed.
posterboy 8y ago
The idea of a modular synths is to have the potential to create all of that, FM, AM, Sub and more. Isn't it?
- detaro 8y ago
  Partially/Depends. Yes, they aim for more flexibility, but that doesn't mean they necessarily do a good job of providing the things you want for specific synthesis types.

TheOtherHobbes 8y ago

Subtractive makes a lot of classic synthesizer sounds, which is why it's a thing. It's also easy to understand. It seems to be easy to implement digitally, but it really isn't, because a lot of the nuances and non-linearities that add weight and colour to synthesis with real electronics aren't present in simple digital emulations. For pure DSP the choice is more or less between open additive, modal (which is a kind of constrained additive), AI-constrained additive, which is what Google have been playing with, and physical modelling, which is digital modelling of strings and bows. If you want to "enforce a physically correct relationship" between etc you're going to want AI-constrained additive or physical modelling. The aesthetics of all of this are a different topic altogether.

cousin_it 7y ago
> If you want to "enforce a physically correct relationship" between etc you're going to want AI-constrained additive or physical modeling. I was thinking of relationships like these: 1) Decay time of nth partial falls as a certain formula of n. 2) Frequency of nth partial is slightly different from n * fundamental, by a factor which is a formula of n. 3) Spectrum of nth partial isn't a delta function, but a hump whose width is a formula of n. All these ideas come from physical effects, but you can use them to generate sounds directly, without any physical modeling or AI. My hunch is that could be more such ideas and they could play together nicely.
- bravura 7y ago
  Do you mind emailing me?

zambal 8y ago

> The whole experience made me suspect that there's an alternative approach to building modular synths, based on physical facts about sound Yamaha experimented with this in the 90s with their VL series physical modeling synths, but it never caught on, mostly I think because if you want to have convincing results, you really need alternative midi controllers like a breath controller[1] for woodwind and brass intruments. An alternative take on why physical modeling synths never really caught on is GigaSampler[2]. It was the first sampler (as far as I can remember) that could playback samples from hard disk, by only keeping the first second or so of the samples in memory. This made it possible to have sampled instruments where for example each key of a piano was sampled at various velocity/loudness values. Resulting in a sampled piano that could span multiple gigabytes. At a time where 128MB of RAM was still quite a lot, this was quite revolutionary. While physical modeling can produce convincing sounds with a potential expressiveness that no sample based instrument will ever match, it's base sound still doesn't sound as 'real' as a properly sampled instrument, recorded in a nice room with good microphones. [0] simple overview, including some soft synth alternatives: https://www.musicradar.com/news/tech/blast-from-the-past-yamaha-vl1-634920 https://www.musicradar.com/news/tech/blast-from-the-past-yam... [1] example breath controller: https://www.akaipro.com/ewi5000 https://www.akaipro.com/ewi5000 [2] Review of Gigasampler's successor: https://www.soundonsound.com/reviews/tascam-gigastudio-4 https://www.soundonsound.com/reviews/tascam-gigastudio-4

byproxy 8y ago
I feel if you're going to such lengths to approximate physical instruments..maybe just record physical instruments? Synths can do things that are unable to be done physically, so why not use them to that end? Although, I understand sometimes it's an aesthetic choice, having a poorly-approximated physical sound.
- posterboy 8y ago
  Because playing the violin is hard and emulating the thing is a different challenge. Of course, sequencing a violin track with just note on/off and velocity (i.e. with a keyboard) is a poor-approximation. Better controls and using automation can be transfered back to other sounds for example. It's serious fun at least. We say to play music, after all.

johneffort 8y ago

There are physically modelled instruments like the one from Pianoteq. No samples, pure rendered. Very nice.

byproxy 8y ago

Slightly related: your pluck sounds reminded me of the beginning of "Tonto" by Battles (https://vimeo.com/117914655 https://vimeo.com/117914655)

jacquesm 8y ago

Hello again :) There is this phenomenon called 'stick-slip' that is an essential component of any string based instrument that is bowed. It might be worth modelling that to see what it sounds like, you might end up with some very realistic bowed string sounds as well as a glass organ. https://en.wikipedia.org/wiki/Stick-slip_phenomenon https://en.wikipedia.org/wiki/Stick-slip_phenomenon

campground 8y ago

Plucked string instruments are one of the simpler classes of instruments for physical modelling. Waveguide synthesis uses delay lines and filters to model a wave propagating along a string and reflecting off the bridge and nut. The famous Karplus–Strong algorithm used a burst of white noise as the excitation, but I've had more success using an asymetric triangle-shaped impulse that resembles the shape of the drawn string.

ravar 7y ago

Check out the sub synth in zynaddsubfx, its really easy to get started with it and makes a great sound.

mochomocha 7y ago

In grad school, I had a tiny start-up with a classmate selling a virtual instrument doing just what you describe (for guitar sound synthesis). We had PDE models for the string, finger/string interaction, fret/string interactions etc. We would then discretize these PDEs to generate in real-time a synthetic "force" applied by the string on soundboard, and then convolve this in real-time with measured impulse responses of real acoustic guitar soundboards. Our website is still up (spicyguitar.com) and we give it for free now, but I don't know if it still works in recent DAWs.

Severian 7y ago
Loaded it up and I can confirm it still works in FL Studio 32b and 64b versions. I really like the music box preset, very melodious with certain parameters.
- mochomocha 7y ago
  Thank you!