Reflection Model Used: Gibbs Reflective Cycle
Description
For my Final Major Project, I created an EP called RECHARGED. The EP was focused on the production of original music inspired by horror and atmospheric video game soundtracks, with a focus on Metroid Dread. I wanted to explore the difference between human-made and AI-made music by producing a mixture of tracks with both.
I was initially planning to make four tracks with one of the been an AI generated track, but I ended up making five original ones. I composed, mixed, and mastered all of them on my own using Ableton Live 12. I tried some of the AI programs like AIVA, Magenta Studio, and LANDR to determine if they would assist in music composition and mastering. The AI was not able to provide the musical support that I needed, but it assisted in mastering and generating ideas.
Along with the music, I also made my own artist name (ZebesDweller) and album cover for the release of Recharge. I even made my own website, crmusic.co.uk , to showcase my work and I will use this website going forward in my University Course. This involved uploading the completed tracks, to my website, so I had a space to point people towards for future collaborations, projects and employement opportunities..
My final tracks were all uploaded and distributed through LANDR, and awaiting for the EP to released on Spotify and YouTube Music. This actually turned the Project into an actual industry experience.
Feelings
First, I was happy with the Project since it was what I wanted to do in the future game music production. I believed in my sound design and production abilities and was curious about what AI tools were capable of doing. One of the various things that made me happy was how Landr had a mastering tool available if I fell flat on my progress – which I did at times – and it could point me in the right direction if I made errors in my work. This meant I had tools to fall back on incase I ran out of avenues to explore.
Halfway through the Project, I was really stressed. I fell ill and could not focus. I was also let down since the AI programs did not work out as I expected. They were meant to help with producing complete tracks, but they were not up to the dark vibe that I was going for. That discouraged me a bit, and I was worried about the amount of work that I had left to do. The thing that upset me most with this work was that the AI tools (besides mastering) were all very complicated to import into the Ableton studio themselves, and I was drag-dropping each time I wanted to experiment with the tools, making it a chore anytime I wanted to step out of my comfort zone.
Ultimately, things got better. When I completed the RECHARGED artwork and saw the website and tracks come together, I felt proud and inspired again. Getting the music uploaded to LANDR and onto the streaming sites made everything official and worth it.
Reflecting on this project, I recognise the benefit of seeking the help of other individuals in future projects. Be it for art or MIDI compositions, I believe that collaborating with other individuals will not only remove the pressure of creative burnout, but also enhance the project with the diversity of points of view and skill sets. I look forward to finding and being involved in communities where I could learn from and give back to others.
Evaluation
Planned vs Actual Outcomes
| Week | Planned Outcome | Actual Outcome | Key Differences & Learning |
|---|---|---|---|
| 1 | Set up project and complete introduction. | Successfully wrote introduction but forgot to set up digital structure at first. | Learned how important early organisation is to stay on track. |
| 2 | Complete proposal and presentation. | Completed early due to timetable change. | Gained flexibility and adapted quickly to new priorities. |
| 3 | Begin research on Metroid and horror music. | Completed fully as planned. | Research gave me clearer direction and confidence. |
| 4 | Start studio production. | Faced challenges due to gaps in music theory. | Learned to prepare better before studio work. |
| 5 | Research AI tools and begin track creation. | Purchased AI tools but didn’t start creating yet. | Realised I needed more time to explore new tools. |
| 6 | Begin gameplay recording and promotion. | Was ill; only managed basic planning and progress check. | Learned to factor health into my project schedule. |
| 7 | Test AI tools and tweak tracks. | Still unwell, no AI testing; focused on promo planning. | Discovered how to adapt when energy is low. |
| 8 | Begin mastering with LANDR. | Couldn’t access LANDR at first due to login issues. | Understood importance of tracking tool access and passwords. |
| 9 | Catch up and resume track production. | Prioritised practical work and shifted away from AI. | Learned to focus on what’s achievable when behind. |
| 10 | Create artwork and explore publishing options. | Tried AI track again but abandoned it due to poor results. | Accepted AI wasn’t suitable and focused on core music. |
| 11 | Finalise and master tracks. | Exported and mastered 3 tracks; shifted to cover design and gameplay footage. | Managed creative fatigue by changing task focus. |
| 12 | Publish music and collect feedback. | Designed album art; did not submit yet. | Focused on visuals more than expected; delayed publishing. |
| 13 | Submit and reflect. | Dealt with LANDR genre issue and resubmitted; completed reflection using Gibbs. | Learned the importance of accuracy in publishing and personal resilience. |
Project Timeline
What went well:
- I did five tracks, which is more than I had planned.
- I created my own brand ZebesDweller and my own album cover.
- I built my website (crmusic.co.uk) to present the Project professionally and provide an identity and place to showcase my work and skills.
- I experimented with Ableton innovatively — trying effects, layering sounds, detuning, and playing with dark textures.
- I tried different AI tools and discovered what they cannot do.
- I finished my tracks on LANDR, which is a great method of learning how music is released and distributed.
- I expanded my library of tools and asssistance with Landr’s catalogue, and I know various methods to work to an achievable goal.
- The album artwork I created gave the release a professional feel and helped me establish an identity for me as a producer. It complemented the music and how I present myself visually and promote my work.”
What did not work:
- The computer programs could not assist in completing complete songs — they were either not in the proper style or too straightforward.
- I suffered burnout and sickness, which retarded progress.
- LANDR’s interface was at times infuriating, and re-typing track details took forever.
- My Bluesky promotion was not seen much or noticed; I should have used other platforms like Instagram or TikTok to be noticed by more people.]
- My time management was rough, and trying to come up with evaluations and reflections weekly was difficult. I had to resort to quarterly 4 weekly reflections on my work which were only short in the timeframe I had.
Analysis
This project was a learning experience for me about how to complete a big creative project from beginning to end. I did not only make music — I also designed the visuals, planned releases, wrote about my work, and designed a website. This allowed me to get a taste of what music production is like in the real world.
I employed a wide variety of creative software to assist my workflow during this project. My primary tool was Ableton Live 12, and I utilised it to compose, arrange, and mix all five pieces. I discovered it to be trustworthy for sound layering and trying out effects, particularly when composing ambient and sci-fi soundscapes. For mastering, I tried to use the LANDR plugin but didn’t find it flexible enough for my genre, so I used more manual EQ and compression in Ableton.
I also tried AI music tools like Magenta Studio, AIVA, and Soundraw, but these did not deliver what I needed. With AI software, I might be able to see how they worked into the workflow. They are not capable of creating music themselves, at least not emotional or atmospheric music, but they can be used to help generate ideas, master, or edit MIDIs. I had to change my strategy, and in changing it, I learned to adapt when things are not working out.
Changing from music to art when I was tired enabled me to keep working. It demonstrated that creative work does not always move in a linear way, and it is all right to switch between tasks to make headway.
Getting my website up was a huge leap. It had me thinking about how I represent my work and how I want to be seen as an artist. Having everything compiled together made me feel professional and made me comfortable sharing my work with other individuals, like potential listeners or future producers.
Conclusion
I am really proud of what I did for this Project. It’s the best, most professional work I have ever done. I showed that I can make good music and do all that is required to be a digital artist, like artwork, branding, publishing, and marketing.
Despite the fact that I changed some parts of the original plan, I still got down to the main goals. I tried out AI software, learned from my experiences, released quality tracks, and produced something that is true to my style and interest. I also made something that reflects my skills, not just in music production but in branding, planning, and solving problems as well. I feel as if the music they use to train their models (which is all legally sourced) is mostly pop/rock music, and they haven’t trained it up for my purposes yet. Maybe in the next five years, it could create a piece completely for me based off the genres I want, but currently? Not at all.
Prior to this course, I did not know whether I would have the confidence to release my music. Now I have done it—five tracks, original artwork, an artist name, and actual distribution. This is a large step for me. I feel more ready for university and more confident as a person who wishes to pursue a career in game music production.
I chose this project purely out of inspiration for my university course in September, which allowed me to expand my skills in studying BA (Hons) Creative Music Production at Huddersfield University. Working on a whole project from research through to final release has given me confidence in my working procedure, broken down creative barriers, and prepared me for bigger projects where I must manage and time myself. Studying AI, learning tools, and self-publishing has also opened my eyes to real-world tools that I am excited to apply in my degree.
I have a better insight into what goes into releasing tracks, promoting albums/EPs, and I am quite excited to begin a higher level of education. This has helped me develop my skills ahead of time before going into the four year course, and I have nothing but high hopes for what comes in my future.
Action Plan
- Get additional breaks between marathon creative sessions to prevent burnout.
- Only use AI software for minor tasks, like MIDI editing or mastering, but not full songs.
- Share on social media platforms, including Instagram or TikTok, to get more viewers.
- Pre-schedule my uploads and metadata so as not to rush certain things like LANDR publishing.
- Keep working on my branding and release strategy (art, titles, artist identity).
- Search for game jams or indie developers to begin working on actual games.
- Research dynamic music systems in which the music changes as you play.
- Look at collaboration with other students as I need to be able to work with other students in future.
- Get in to the habit of weekly reflections – I can do this by reflecting on my emotions in a certain week and building a norm out of writing in this format.
- Showing progress/screenshots of what I’m working on. Posting variations and iterations of tracks to show my creative process will be good for evidence and for promotion.
- In future projects, I need to get better at using proper in-text citations when I talk about research I’ve used. I did include a full bibliography and mentioned where some ideas came from. Still, I now understand that putting the source in the text helps make my work look more professional and shows where the information came from. I’ll use Harvard-style referencing properly next time, especially in university where it’s expected.
Final List of What I Did: –
- Five completed tracks, written, recorded, and mastered – Album cover and artist name (Recharge, ZebesDweller)
- Personal music website established – crmusic.co.uk
- Experimented with AI tools (AIVA, Magenta Studio, LANDR)
- Complete final assessment and weekly written reflections. – Project posted on online websites (Bluesky, Discord)
- Tracks published on Spotify/YouTube through LANDR – TBC