Hacker News Reader

Comment by nicetryguy | Hacker News Reader

> But I'd really like to have some basic idea of how to put together rudimentary sounds with intent.

There are really only 4 main waves: Sine, Square, Triangle, and Sawtooth. The "complexity" comes from using these waves to modulate and transform each other. Apply Volume: labeled as Attack, Sustain, Decay, and Release; and you have yourself a patch! Find any synthesizer plugin, load up some of the patches, and start twiddling knobs: Disable things and break each patch down to it's individual components. You can get a feel for how complex sounds are built with "digital VCO's".

Music isn't an exact science: You just have to play with it until it sounds good to you. Have fun =)

View on HN

There's only one wave, sine ;) https://en.wikipedia.org/wiki/Fourier_transform

In a sense that reality is "nothing but atoms" you're correct =p

These are the main shapes from a user standpoint =)

Yes, you're right of course. I just felt obliged to do the pedantic HN thing =D

Why are there only 4? Are these like Platonic forms?

I see that sine are "no harmonics", sawtooth are "all harmonics", square are "odd harmonics" but a triangle wave is also all odd harmonics. Okay.

Why these 4 forms? There are lots of other forms, like a cycloid.

Last question: is a sawtooth heard as an octave lower than a square wave, because there is only one pulse per cycle, whereas a square has two pulses per cycle?

They were easy to make with uncomplicated circuitry. That's it.

The original big modulars included oscillator sync - which locked a bank of oscillators together - and waveform mixing. So you could create more complex shapes. (With some effort and imagination, and not in a particularly intuitive way.)

When digital started to become a thing in the 80s you could start using any shape, including your own hand-drawn waveshapes or a custom mix of overtones.

But that turned out to be less exciting than expected, so most people just switched to samples instead.

A sawtooth is not heard an octave lower. The fundamental is the same for both shapes, because the Fourier decomposition has the same f0.

There are some unexpected psychoacoustic effects which change the perception of fundamental pitch. E.g there's almost no fundamental in the lowest octave of a piano, but our ears fill it in anyway.

But simple repeating shapes tend to do what it says on the label. (And in the textbook.)

A 200hz saw, despite having twice the overtones than a square, sounds much lower. I still haven't gotten an explanation of that one, ooh well.

Overall, nice stories, thank you. I'd be surprised by the lack of the fundamental in a grand piano. However, I've heard that's true of small headphones, where they simply mechanically can't produce large wavelengths. Yet they are heard!

Besides the engineering reasons others have pointed out, some of these waveforms tend to occur in wind instruments. For example a flute produces a waveform close (for some definition of "close") to a sine wave and clarinets produce a waveform closer to a square wave. Thus these waves sound somewhat familiar to most listeners before you start subtracting harmonics. (I don't know that a sawtooth is similar to a natural instrument but it's popular in subtractive synthesis because it's a rich source of harmonics.)

In the early days of synthesizers the goal was to emulate non-electronic instruments, so familiar waveforms were desirable. Then Bob Moog came along and said "Synthesizers are instrumemts in their own right and should not be considered as imitating something else" and then Wendy Carlos released Switched-On Bach and the rest is history.

Sawtooth is similar to violins and other bowed instruments. (For convincing plucking, you kinda need Karpluss-Strong, though I’m still partial to very short decay times.) FM synthesis with enharmonic tones is bell-like.

I’d argue that Don Buchla pushed synthesizers further than Bob Moog did when it comes to being their own instrument vs imitating something else, but he was also pushing for a radical departure from 500 years of western music in general. (Read: “what is this godawful noise, and why are you subjecting me to it??”)

Why include a piano keyboard at all? Why think of “playing measures in a score” when you can wire up and interact with something that feels like a living system and see what happens? It’s the musical equivalent of “what is THAT thing!? What if I try to pet it, or poke it with a stick?” You might get bitten.

This gets close to the heart of the East Coast (Moog) vs West Coast (Buchla) styles of synthesis. I’m not dissing Bob Moog in the slightest here: I’d rather listen to more conventional western music the vast majority of the time, and he was integral in carving out synths as a legitimate instrument. You’re not wrong.

Sometimes I just gotta let go and let something like Silver Apples of the Moon [x] roll through my head. Not often, maybe not even all the way through, and I almost always skip the first two minutes. I love being reminded that synths have had this wild streak running through them since basically the beginning though.

One of the cool things that’s happened with the rise of Eurorack and the splintering of musical genres is that you can find much more music that ignores the East/West coast divide, taking elements of each and using them wisely to serve the purpose of song. Silver Apples around 23:00 could easily fit in a Nine Inch Nails track, for example. This has been going on for decades; the Eurorack standard just makes it more likely to continue, for which I am thankful.

[x] https://youtu.be/3G1hRNLlYpg

Good point about sawtooth and strings. I've forgotten a lot. And also thanks for mentioning Karplus-Strong--one of the most amazing algorithms ever. With like 8 assembler instructions you can build a very good facsimile of a plucked string instrument [0]. I implemented the heck out of that thing back in the 80s. It would work quite well today on an Arduino. (I'd expect many people have done so.)

As to Silver Apples, I just could never get into that style of music. When I was studying this stuff in college our instructor was very enamored with the Subotnick/Stockhausen "bleeps and bloops" style. We had a giant E-mu modular in the lab and he wanted us to avoid using its keyboard. But whenever I was on my own I of course used the keyboard. Then one day a truck rolled up to the lab and dropped off a crate labeled "Fairlight" and I fell in love hard but that's another story.

[0] https://en.m.wikipedia.org/wiki/Karplus%E2%80%93Strong_strin...

> Why these 4 forms?

It has to do with Voltage Control Oscillators, or VCO's, and actually controlling Voltages with physical components: These are the shapes you can actually make. Modulation is applying a waveform to another, so, if you apply a Triangle wave to a Sine wave you'll get another more complex wave that looks and sounds like neither of the two. You can theoretically make infinite shapes by modulating these forms with each other again and again. This is the bread and butter of analog synthesizers.

I don't think a square, or any wave, is measured in groups of two pulses. "Pitch" is just how many times the pattern repeats per second, the same as how many times a string would vibrate back and forth on a guitar when you pluck it.

Harmonics is a deep rabbit hole that i don't quite understand myself, so, i don't feel comfortable speaking to that.

Noise generators tend to be pretty important as they allow you to add complexity to sound synthesis that you can't extract from very determinative wave forms

Realizing this was exactly what I learned when playing around with the C64 SID chip in 1983. Neat!

Reply on news.ycombinator.com