Audio Exploration Report | MAR-23
Modulations
There have been two main pieces of software I have been exploring and using this month. Massive X, which I touched on last month, has proven to be a great synthesizer for exploring different kinds of modulations. As I am currently working on my second ambient album, where the “rules” of music are a lot more relaxed, it has given rise to some really interesting sound design and rhythmic developments. For example, in Digital Wisp Corridor (below), it is only a single layer of Massive X, just with a lot of modulation and automation on it.
One key aspect that you may have heard in a few of my recent releases is the use of the “Osc” speed LFO in Massive X, as opposed to the “Free” setting. This correlates the LFO speed with the pitch of the oscillator, which means that higher pitched notes will generally have faster LFO modulations on them — emulating the pitch-shifting sound of many sample-based synthesizers, despite the fact it is all synthesized.
Get John Oestmann’s stories in your inbox
Join Medium for free to get updates from this writer.
Another key feature that has really helped to polish the sound is the use of the VR (Voice Random) modulator. By hooking this up to the Panning, for example, each note (voice) you play will come out at a slightly different position in the stereo space. A held note (voice) will retain its position, while the next one will appear at a new stereo position. This is one of the best parts of stereo-based polyphonic synthesis. For minimalist compositions such as my recent ones, this is a great feature for widening the sound to help immersiveness in stereo.
SuperCollider — Designing from the Ground Up
One tool I’ve been wanting to get into for a while now is SuperCollider. It is an audio programming language — text-based, as opposed to the visual-based Max or Pure Data. As it is free, open-source, and easy to reverse-engineer (once you understand the language), it was always one of the perfect options for my ideal compositional process (see my earlier article about my reasons for project-sharing). Suffice to say, many open-source tools are far from intuitive to get into. I had tried a few times with SuperCollider, but fortunately as there is now a few more grassroots learning resources popping up, I found one that was easier for me to digest, and it kicked things off. Given that I have been dreaming of being able to create music using entirely free + open-source software, I am proud to say I finally have. Bacchus’ Empty Comet (below) was created — sound design and composition — entirely in SuperCollider with code. This code is up on the GitHub, and is free to be studied or re-used under the DCC license. While this is far from an entry-level way of making music, it does make me feel like I have complete control over the design of the music, from the ground up. And working with this level of granularity is teaching me a lot about synthesis.
While I have never been into computer-generated compositions, as I have found them always sound a little too random, I was recently put onto the concept of Markov Chains and Finite State Machines. One example of how they work is by laying out a series of possible notes that can be played. Each of these notes then has a list of other possible notes that it can jump to. C5 might be able to jump to D5, E5, F5, or anything else you choose. This can be set for each note in your note list, and so that D5 might only be able to move to E5 or A6. As you set up these possible pathways for each note, and set the probabilities of different pathways being followed, your computer-generated composition moves from random beeps to regular patterns and motifs emerging, still flourished by plenty of computer-generated variety. This doesn’t have to just be done with notes either. In Bacchus’ Empty Comet, I used Finite State Machines both for note selection, and for percussion rhythm selection. When I had the different perc rhythms available for equal random choice, I would end up with cool drum rolls happening too many times in a row, becoming distracting. By using PFSM — the SuperCollider implementation of Finite State Machines — I could ensure that these perc rolls would only happen occasionally — much like a real drummer. You can listen to the results below.
I release as much of my new music under the DCC license as possible, and so all of these explorations can be found in the project files hosted at https://github.com/Soundworlds-JO/Soundworlds-Datapedia
If you want to keep up with what I’m up to, you can also follow me on: https://mastodon.gamedev.place/@soundworlds
