A short while ago (or rather two weeks), I told you about a thing I was doing with „dub techno“. Whatever that may mean. I believe I know a little better now than I did back then.
DeepMEE
The project is called „DeepMEE“ which is a name some AI came up with based on architectural decisions we made (MPC as master sequencer, Elektron Analog Rytm as drum sequencer and drum sound source, and we’re considering sounds into the 32′ foot register, get it?).
And its goal? Defined as:
The DeepMEE project develops a hardware-only live dub techno jamming setup in F# minor (113–126 BPM), inspired by Yagya – Rigning, Monolake – Hongkong, Basic Channel – Quadrant Dub I/II, DeepChord – Hash-Bar Loops, and Fluxion – Spaces. Core setup includes Akai Professional MPC One, Elektron Analog Rytm MkI, DeepMind 6, Access Virus B, SSL BiG SiX, and effects (VD400, SMMH, Eclipse, M9, UC200, etc.).
So what happend since (June 29th)?
A breakdown in chapters
Chapter 1 was about Project Initiation and Requirements Gathering. A lot of design decisions for a functioning prototype were made back then, and the actual mission statement drafted.
Chapters 2 and 3 dealt with turning into the idea into a setup – and into music.
In chapter 4, we’re taking what we learned before, and make a Detailed design.
That is then run through a Proof of Concept in chapter 5, and chapter 6 ist nothing but saying „we’re done“ (or moving back to chapter 4 for another loop).
But how do we even know what to do? Paralogue 1 is about defining a User Story – which mostly happened (and still happens) through listening to music. There’s two further Paralogues – Paralogue 2, which tries the same thing using only a single component (MC-707), while paralogue 3 does the same, but with the modular bigrig. We might drop paralogue 2, and not even start paralogue 3, though.
It’s requirements engineering
Music or not, a lot is about system architecture, requirements engineering and the boring details.
I also enlisted AI help, with varying degrees of success. Lesson learned: a token window of 128k (that’s like 1984 technology?) is hardly enough for a complex project.
Still, generating formal artifacts works, as well as defining mapping charts for the MPC, or Verfahrensanweisungen.
What about music?
There is some. I have recorded some bits, although not for more than a week. But there will be some, I promise!
Until then, more photos.
