Hacker News new | past | comments | ask | show | jobs | submit login

>> "The sonic textures creates by acoustic instruments are arguably richer, as they are capable of much more subtle modes of expression as playable instruments."

Synths are slowly catching up. You can pack quite a lot of expression into Serum with two wavetable oscillators, a noise and sub oscillator, modulating anything by note or velocity, alternate tunings, and a wide range of modulatable adjustments.

It's not a viola, but you can make some very unique instruments with it. Most people use it to make yet another dubstep bass or hypersaw, but the potential is there.

I like Serum a lot, but it has the same set of issues that make all of these digital synths limited as performance instruments compared to, say, a piano.

The issue starts with the low resolution of MIDI. Most instruments implement only velocity connected to keys, with max 128 layers of resolution.

A real piano has near infinite resolution on velocity alone.

This exacerbates the issue that digital synths are generally deterministic, meaning for the same input they produce the same output.

Because the resolution of the controller is so low, attempts to introduce variety rely on randomness. In an acoustic instrument the same sound cant ever be produced twice. But this is a result of a complex, chaotic system, not randomness.

A major factor in the inherently chaotic sound is that each note is exciting the same physical object, even a note played with the same velocity (like a disklavier piano system could) will sound different every time based on the state of the whole large object. As notes are added to a chord or arpeggio, they are not just superimposed over each other, instead they each contribute their energy to the whole object.

The use of mod wheels definitely helps with expressiveness in the hands of a good player.

But my sense is that even “players” are forgoing using these instruments as performance tools, and instead loop the sections, and piece through the whole song, listening carefully, modifying the midi data directly based on what they a hearing.

This has definitely resulted in some very imaginative and striking work, but, to my ear anyway, it has a very different feel as a mode of artistic expression. It’s more “cerebral” in some way.

MIDI 2.0 will likely improve the situation. 4,294,967,296 possible values isn't enough to match a real instrument's range, but it's quite a bit better than 127! Serum has two chaos oscillators that help avoid deterministic-y music. Too many preset designers skip them.

Take a look at Sphere from Echo Sound Works. They sampled a bunch of real instruments and used them for wavetables and noises for Serum presets.


> 4,294,967,296 possible values isn't enough to match a real instrument's range

I don't think this is true -- I doubt our ears can recognize a 2^-32 difference in anything, and there's no way physical instruments can be played with that kind of precision either.

Guidelines | FAQ | Support | API | Security | Lists | Bookmarklet | Legal | Apply to YC | Contact