[Csnd] Realistic Csound Instruments?

Hallo List,
a rather newbie question: are there realistic sounding Csound virtual instruments for the Jazz Combo around, i.e. most importantly walking upright bass, but also alto/tenor sax, minimal jazz drum kit, maybe acoustic piano or jazz guitar, trumpet, trombone?
I know there is the option midi+soundfont, but since I found something like this (CSound Synthetic Traditional Korean Instrument - YouTube), I wonder if somebody created realistic synth instruments for the jazz combo too?
It’s surprisingly difficult to find Csound music examples that imitate natural instruments instead of creating electronic sounds, but maybe that’s simply out of scope for Csound?

Csound mailing list Csound@listserv.heanet.ie https://listserv.heanet.ie/cgi-bin/wa?A0=CSOUND Send bugs reports to https://github.com/csound/csound/issues Discussions of bugs and features can be posted here


The book “Cooking with Csound: Part 1” by Andrew Horner and Lydia Ayers has recipes for the common wind and brass instruments. The saxes, trumpet and trombone sound pretty good to my ears, but your mileage may vary.

David Bowen

There was a body of work done on a Csound DX7 emulator, some time ago.

I’m not sure if it is around and functional any more. Jeff Harrington, Steven Yi or others
may be able to advise.

While the DX7 algorithms are a simple Csound orc construct, of course its the patches
settings of those algorithms that made the DX7 such a good traditional instrument
emulator. Many keyboards, plucked bass violin, even trumpets and saxes the DX7 did well.

This is an excellent question that requires a nuanced answer.

The short answer is yes indeed, Csound can be used to implement convincing models not only of acoustical instruments, but also of analog electronic instruments and effects.

As what I think is compelling evidence of this, take a look at Jeff Livingston’s physically modelled Spanish guitar (https://github.com/gogins/csound-examples/blob/master/archival/Livingston/guitar.csd) and Lee Zakian’s physically modelled Boehm flutes (https://github.com/gogins/csound-examples/tree/master/archival/Zakian).

The more nuanced answer is that it can often take as much complexity in the Csound orchestra code as you would find in the actual physical instrument, to produce a convincing simulation. So, that can be a lot of work and that work usually requires knowledge both of the physical instrument, and of digital signal processing. Usually, the only people who can do this work expect to be paid, and so they end up working for music software companies that sell proprietary products. Such products are usually written in C++ but they could just as easily have been written in Csound code.

As David Bowen correctly noted, the Cooking in Csound book has also got some pretty convincing Csound instruments; but these were done not with physical modelling, rather with additive synthesis informed by mathematical and acoustical understanding.

For a jazz combo, I have produced an arrangement of Oblivion by Astor PIazzola using Csound instruments, which you can find here: https://github.com/gogins/csound-examples/blob/master/csd/oblivion.csd. Some of the instruments are reasonably good simulations, and some are not. But this could be a starting point. This orchestra does not include a double bass or traps, but my examples repository does have some plucked strings and percussive sounds that could be adapted.


Dear all,
thank you very much for the interesting and profound answers!
The examples given are quite impressive, instruments and compositions, I have to dig deeper into them.
So there are some “natural” instruments available, and indeed they look super complicated. And with the hint about the real demand for this kind of work in the industry it becomes clear, why not everything is free and open source in the csound world.

Personal answer:

Nothing “imitating” acoustic will ever sound as good as acoustic, if the intention of the music is to be as close as possible to an acoustic experience. Too many factors, like the subtle touches of a performer and how those interact with the physics of the instrument, can make the results, at times, disappointing.

That said: not every instrument to emulate is equally difficult. Astoundingly good pipe organs can be made, for instance (maybe that’s b/c traditional pipe organs are like acoustic additive synths?). And some keyboard instruments can be decently emulated with good soundfont sets. Much depends on the stiffness or not of the performance, too. FM bells can be rich, and inharmonic percussion timbres, etc. But something like an imitation violin has almost always been disappointing to me, personally. I find it better to be expressive with an electronic instrument capable of rich timbre and expression w/o “trying” to be an existing instrument.

I agree with Michael Gogins that the best acoustic sounds available tend to be commercial, since the work involved in engineering them tends to be intense, and people want to be paid for that.