Into the Forest With Deep Learning

Foxy Digitalis depends on our awesome readers to keep things rolling. Pledge your support today via our Patreon or subscribe to The Jewel Garden.


Richard Park’s Deep Learning project returned after a four-year hiatus last month with the excellent Evergreen on Oxtail Recordings. Inspired by walks in the Epping Forest with his son during the height of lockdown, the album is a beautiful surprise. Using a unique approach and palette of CD glitch textures gives Pike a repetitious, pointillist foundation to explore themes of regeneration and rebirth. The results are gorgeous and unexpected. 

Additionally, I am thrilled to premiere Dan Tombs’ (2MS Studios) incredible video for “Standard Measures.” It brings new perspectives to the song, heightening its impact and deep into the digital forest. Tombs can be found on Twitter and Instagram, while Pike can be reached via his websiteEvergreen is out now on Oxtail


To start, I’d love to hear about some of your earliest memories of music and sound. What were some formative experiences when you were younger that have stuck with you through the years?

I was put into Suzuki Method violin lessons when I was 5, and I really appreciate my parents doing this. I’m not sure if it’s my first memory, but I was made to play a ‘violin’ from an empty tissue box and wooden ruler, just to learn how to hold it. My parents were not musically trained, but my dad had a ‘best of classical cassette collection, Mozart, Bach, Brahms, etc. When he had a few drinks of an evening, he’d bring out a violin and try and play along in the tiled hallway that had a nice reverb. Mum would yell at him to stop because it kept the kids up. Later around 11 years old, I switched to guitar because I saw a Van Halen video.

What got you interested in electronic and experimental music?

When I left high school, I met a lot of different musicians. I was very interested in jazz, which made sense after the Heavy Metal to Faith No More to Mr. Bungle gateway. As a kid, I was hesitant about electronic music because it wasn’t ‘real music by real musicians, as Prince would say. But that was simply because all I’d heard was quite simplistic 90s house or pop, which sounded quite superficial to me. It wasn’t until a friend played me Autechre’s LP5 that I realized the huge possibilities with computer music. I got an Akai S2000 and began self-producing my first album with my band, Pivot. It was an incredibly exciting change for me.

Did you always want to be a musician?

The short answer is yes! I never wanted to be anything else, and nothing else has really made sense to me. I would fall asleep dreaming continuous music, and in some ways, my whole life has been trying to pull those dream states of sound out of my own head. It is a daily goal that is harder than you think, especially when real life gets in the way.

How’d you get started writing your own songs, anyway?

I was either writing chords and ideas or words and melodies while my brother played drums. We had our roles chosen and had a kind of musical ESP. We never argued about the direction of the music we were writing. It was quite natural as we shared the same listening library and had the same touchstones. Later on, when I went to the Conservatorium in Sydney, I learned more about orchestrating and writing for others. Plus, I discovered technology and taught myself about audio and mixing. That was a significant time for me as well, especially in terms of choosing a path.

So, you’ve done a ton of stuff for quite some time, but I want to focus on Deep Learning specifically. When did you start the project, and what made you want to have a separate outlet for the music you’re doing under this moniker?

I often wonder if all the different projects can be confusing for people. The internet expects you to be quite branded, and often that branded perception of who you are artistically is beyond your control. Even now, with Evergreen, people might expect it to be quite ambient, but it’s not really. It’s not music for sleep or meditation. I started Deep Learning in 2018 to get out of the album cycle I had formed with my band. The writing period was always exciting, but the long tail of the editing-recording-mixing process started to drain me. I wanted to make immediate music that wasn’t labored. 

This is why experimental music is exciting – it is often one step away from improvisation, and I needed to press reset on the artistic process and do something that felt natural. I was connecting more with it because you stop the editing and arranging just before that moment where your ears get tired of it. Like a traditional Japanese calligraphy artist will pick a word or phrase, finish a piece within a few strokes and leave it where it is. That was the idea. Quiet music seemed like a good place to start. I’m slowly ramping things up again now, but focussing on one piece of the music puzzle at a time and re-learning things. This time it was rhythm, micro-textures, and plucked softsynths. 

It’s mentioned how the new record was inspired by walks with your son in Epping Forest, which is an idea near and dear to my heart (my then 8-year-old daughter and I made a collaborative album last fall as part of her homeschooling due to COVID, and it is one of my favorite things ever). Anyway, what was it about those walks and those experiences that brought out the ideas and sounds in this album?

Ha, incredible you made a collaborative record with your daughter. I’d like to hear that. I’ve been wanting to do the same with my son, as he is constantly singing and making up songs. I’m always blown away by the purity of it. His phrasing is sort of like Iggy Pop, I don’t know how. I’d like to produce his songs, but actually making a living from music has prevented me, and now he’s almost 8 years old. Maybe now’s the time. You can hear one home tape recording of him on the track “A Journal Of,” subtly talking in the background. He says, “I love you, and I will marry you soon,” which was a message from one of his school friends during Covid. 

The experience of those walks was quite layered for me. The pandemic gave me an opportunity to work on this glitch idea I had for a long time. Oval’s 94diskont was a hugely influential album for me. It surprised me that his approach was not picked up more in the last 25 years since and taken other places. An idea was to pick up from that point in the 90s where it left off and do my own thing with it. Bjork’s Vespertine was another influential glitch record to me. All those ideas stayed with me.

On top of that, I felt that huge weight from homeschooling as I have never been a natural teacher, and my son is quite hyperactive and possibly has ADHD. We lived in a 2 bedroom London flat with no garden. So it was a bad combo and came to a real melting point for me. Taking my son to a forest daily, sometimes for up to 5 hours of walking, was the greatest solution. It was suddenly full of so much joy, but a surreal joy because we were communing with nature and no one else. I was walking in places I had never been and in ancient land outside of London that had never been built upon. It was a new discovery every day and such a contrast to 2019.

And reading about the processes of making some of these sounds is fascinating. I’m curious, how does an idea like taking a screwdriver to a CDR start and eventually becomes a reality? And what can you get from that approach that won’t work any other way?

Obviously, there is a randomness to the CD glitch – and I usually hate randomness from a machine because it stinks of the ‘death of the author.’ It’s partly why the Modular craze hasn’t appealed to me so far – it seems many people are getting joy out of these modules playing by themselves. I need to control the sound somehow. Otherwise, I don’t have a connection to the sound in my head. It needs to resonate with my brain waves.

So, to get technical, the first step was recording some of my own sounds. Usually, pads or long chords work, some harmony I like, and I’d burn that to CD-R. Then I would lightly scratch them and find moments that would skip in the player to create that CD-laser click, which sounds like nothing else. Also trying some fast forwarding and rewinding too. Then I would pick nice loops and manipulate them further in the DAW, either locking in a groove or adding some filters or delays. That became the bed where I would layer softsynths, voice, and guitar.

I think processes like that are brilliant, and what I love even more is how I can picture these very visceral processes in my head, but then I listen to Evergreen, and it’s not harsh in that way at all. So that adds a different layer to me. What is most challenging about using these approaches to make music that is, for lack of a better word, ‘pretty’?

Yes, I am glad you noticed this! It was actually an aim to make it beautiful, to take what’s essentially a mistake, something that is ugly and what is quite annoying sound to some people, and turn it into beauty.

What surprised you the most when making Evergreen?

I think as a personal experience, I felt outside myself for a lot of it. Which obviously had a lot to do with the circumstances. But it also felt like an exciting new place I hadn’t been. I sort of floated through it.

Obviously, Evergreen is the main focus, but what else are you working on these days, and what might we see or hear in the not-too-distant future?

I am working the score for a new TV drama show in Australia, which will take up all my time until the end of the year. Also been working on some pretty mental new material with my bandmate/labelmate JQ. The other pandemic album I made with him is our band Forgiveness, with our friend Jack Wyllie from Portico Quartet, and I think you’d like that one too. I’d like to write some new material there. There’s my label Salmon Universe too, and a few possible releases are coming up in 2023. There will be loads coming up.


Foxy Digitalis depends on our awesome readers to keep things rolling. Pledge your support today via our Patreon or subscribe to The Jewel Garden.


One thought on “Into the Forest With Deep Learning

Leave a Reply