Sonic Performance
On this page is documentation, context, and draft versions of my Sonic Performance project.
01
Research Stage
My starting point for this project was investigating how contemporary artists approach song creation with electronic devices - specifically, how complex and layered music can incorporate live performance without sacrificing its emotional depth. My two primary references were 'GODSTAINED' by Quadeca, and 'These Chains' by Mid-air Thief.
'Godstained' brought me in with its production density and beautiful chords, and the way he layers harmonic content and builds emotional weight with its texture. This forced me to ask myself how I can create something of similar emotional complexity when performing with a controller?
'These Chains' offered an additional perspective, where the track juggles intense dynamic depth with an equal importance on restraint and space as density. This informed me and helped me remember to manipulate the volume and dynamics of the track as I go.
As for the performance in the end, I struggled with deciding on what equipment I could use and how I could make it interesting. In the end I settled on a simple but deeply customisable Ableton clip sequencer - a Launchpad mini, as it offered me the most interesting options and is very cost effective.
Both tracks share emotional authenticity, which is an important part of my performance and song. Rather than aiming for a flashy performance, I wanted to create something considered and intentional. This research stage helped me understand what kind of performance I wanted to create.
02
Prototyping and Song Creation
The track I chose to create and perform, nicknamed 'Milkteeth', was
constructed in Ableton Live across multiple sessions. From the beginning, the arrangement was built with layering and structural variation in mind. The two demo audio files illustrate how the track developed from an early sketch
into a more rhythmically and instrumentally complex piece. The addition of
a four on the floor drum beat may sound different when the track is a lot
slower and more free flowing, but it was important to create elements of my
song that allowed for further variation in my performance.
The production is organised around six stem groups - drums, percussion (recorded by myself), melodics, chords, bass, and effects, visible in the arrangement window opposite. Each group represents a distinct layer within the mix, and also made using my Launchpad significantly easier.
It was also important to export the stems and required careful attention - all tracks were bounced from bar + beat one, and at 125 BPM to ensure perfect alignment on important. Any mistake here would derail the entire performance, as it would become misaligned when launching the clips.
After bouncing out my stems, I had to go through a process of finding sections in the track (e.g. intro, chorus, bridge), and cutting up the stems into these distinct sections, which can be seen in the arrangement screen attached. This would allow me to have the option of playing whole sections, as well as each layer individually, making it much easier to construct full songs. A significant challenge rised here - how to cut up as many sections as possible, whilst still having room on my controller? I initially wanted to use 8 bars for everything, but in the end I had to use a mixture of 8, 12, and 16 bar sections to make it possible to play each section effectively. This was an important lesson in how the limitations of hardware directly control creative and technical decisions, and something I wasn't familiar with originally.
03
Rehearsal and Drafts
With the stems chopped and in sections, the next step was to paste it into session view and start experimenting. First, I had to map the six track groups to the Launchpad Mini and beginning to rehearse the performance. The top screenshot shows an early version of the session view layout - at this point the clips were misaligned with the Launchpad grid, with scenes extending beyond the 8x8 boundary of the controller. This made certain rows inaccessible during performance, which was a critical problem that needed resolving before any meaningful rehearsal could take place. This was early in the process, and I was still deeply unfamiliar with how session view operated. After returning to the arrange view and reprocessing the cuts at 16-bar intervals, the corrected session view - shown in the middle screenshot - produced 6 clean scenes across 6 columns, sitting comfortably within the controller's grid. I simply just had to close the groups and then move the tracks up slightly in the red box, and it worked perfectly. This was the most satisfying part of this project, as it was a huge hurdle and bugging me a lot.
The two rehearsal videos show the Launchpad in use during early attemps of the performance. The first test was particularly rough - the priority was understanding which pads corresponded to which stems and building muscle memory for the layout. Because of this, it's essentially just me triggering each section of the track (verse, chorus, bridge, etc). The second test shows my more intentional approach, with slightly transitions between scenes and more deliberate decisions about which layers to bring in and out. Comparing the two videos reveals my clear progression in performance confidence, which was the goal of this stage. Rehearsing multiple times before filming the final performance was essential in ensuring the final take looked and felt considered rather than hesitant.
04
Final Project
The final performance was recorded in a one-shot take, with the launchpad mini mapped to the completed session view layout show in the screenshot below. The six column groups, clearly labeled, are visible on the video opposite with the colours helping make each one more visible, easier to play, and more interesting to watch.
Rather than simply playing the track through from start to finish, the performance involved making real-time decisions about which scenes and individual clips to launch. Sections were deliberately re-ordered to try and make the track different to the pre-made arrangement, which makes it more challenging and more fun. This approach reflects the core idea behind session view - that the arrangement is not fixed, but shaped in the moment by instinct.
The final performance was functional but not without its weaknesses. The most significant limitation was the lack of real-time effects processing - the performance relied entirely on launching and stopping clips, which, while effective, is a relatively passive form of live electronic performance. Artists like Four Tet or Actress, who perform in similar session view contexts, typically incorporate live parameter manipulation - filter sweeps, reverb sends, delay feedback - to create a sense of genuine real-time sound shaping. My performance did not achieve this level of interactivity, and as a result the distinction between a performed version and simply playing the track back is perhaps less pronounced than it could be. With more practice the performance would have also felt more fluid and confident. These are all areas I would address in future iterations of this kind of project.
