How to Record, Mix and Master Your Own Songs at Home
I'm a working musician, and for years I released music that sounded almost right — a bit flat, a bit amateur, never quite like a record — with no studio, no engineer, just a laptop and a spare room. This is the whole chain I worked out to fix that: get a clean recording, mix it so it sits, master it for streaming, and — the part nobody teaches — actually finish it.
It's written for any DAW. The full book shows the exact moves in Logic, but every principle here works the same in FL Studio, Ableton, GarageBand, Reaper, Cubase, Studio One, Pro Tools or BandLab — the menus move, the method doesn't. Specifics and the exact settings sheets are in the £19.99 book; this page gets you most of the way for free.
Recording real music over a Suno or AI track (and making it sound real)
This is the one most recording guides pretend isn't happening. If you've made a track in Suno (or Udio) and it sounds almost there but unmistakably "AI" — too smooth, slightly fake — the fix isn't a better prompt. It's putting a human back on top. Three moves do most of the work:
- Re-sing the lead in your own voice over the AI track, recorded properly (see vocals below). A real lead vocal is the single biggest thing that stops a track sounding generated.
- Replace a part. Pull the AI track apart with stem separation (more on that next) and re-record one real element — your guitar, a real bassline — so the front of the track is genuinely you.
- Mix it yourself instead of leaving the AI's flat, loud bounce. Your EQ, your balance, your space is what makes it a record rather than a demo.
That's the whole idea: AI as a fast session band you direct, then put yourself on top. The book has the full workflow (and where AI actually fits vs where it can't).
Stem separation in 2026 — pulling parts out of a finished track
Stem separation tools can now split a finished song into vocals, drums, bass and other — useful for re-recording one part, remixing, or rescuing a track you only have as a stereo file. They're good, not perfect (busy mixes leave artefacts), and which tool to reach for depends on the source. The book covers when stem separation is worth it, when it'll bite you, and how to clean up the artefacts. (The AI tool landscape moves fast — check current tools before you commit to one.)
Why does my mix sound muddy — or bad in the car but fine on headphones?
Two of the most-searched problems in home recording, and both have honest fixes:
- Muddy is almost always too much energy in the low mids (roughly 250–500 Hz) stacking up across every track. The cure is carving — cutting that region on the instruments that don't need it so one thing owns each frequency band, rather than boosting. A clean mix is mostly subtraction.
- Sounds bad in the car but great on headphones = your mix isn't translating, and it's nearly always the low end. Headphones (and an untreated room) lie to you about bass, so you over- or under-cook it, and the car exposes it. The fix is referencing: A/B against a commercial track you know well, on the same systems, and trust that over your room.
The book has a full troubleshooting chapter (the muddy mix, the weak master, "my vocals sit wrong", the car test) — exact-match for the problem you're actually having.
How to record vocals at home that don't sound amateur
Vocals are where home recordings live or die, so this is the heart of it:
- Mic placement fixes ~80% of harshness before any plugin — the angle and distance matter more than the mic. A pop shield, a stand, and closed-back headphones quietly solve most of the rest.
- Comp the take — record a few passes and build one great vocal from the best lines (the "one good take + comp" rule). Done in any DAW; the book walks the comping moves.
- De-ess the harsh "ess" and "sh" sounds, catch plosives (the "p"/"b" pops) you didn't stop on the way in, and add doubles/harmonies that lift the chorus — automated by section, not left static.
None of it needs a treated studio — it needs the technique, and the technique is what the book is.
How do I make my vocals sit in the mix?
The most-searched mixing question there is, and the answer is a repeatable vocal chain, not luck: high-pass the rumble, carve a little space for the vocal in the other instruments (rather than just turning the vocal up), gentle compression so the quiet words stay audible, then a touch of presence and space. Level it to the mix, not over it. The book has the universal vocal chain plus the volume-balance chart (every element relative to the lead vocal) — copy it at the desk.
Mixing by feel is hard — mix by template instead
You can't trust your ears in an untreated room, so don't mix from scratch every time. The book gives you genre mix templates — pop, rock, ballad, dance — plus the EQ frequency map, a BPM-to-reverb-time cheat sheet and delay-timing divisions, so you start from a proven shape and tweak, instead of staring at a blank session.
Mastering for streaming — the 2026 LUFS targets
Mastering is the final polish that makes a track loud, balanced and consistent across systems — and for streaming it's also about loudness normalisation. Spotify and the rest play everything to a reference level (around −14 LUFS), so mastering louder than that doesn't win the loudness war — it just gets turned down, often sounding worse. The book has the full mastering chain (the canonical order), the 2026 LUFS target table per platform, and an honest verdict on the plugins — so you hit the right loudness instead of guessing.
AI mastering — is LANDR or BandLab good enough?
Honest answer: for demos and a lot of casual releases, AI mastering (LANDR, BandLab, eMastered) is genuinely fine and absurdly cheap. For a track that matters, a proper limiter and the chain in the book — where you make the calls — still beats it, and you learn something every time. The book lays out when AI mastering is the right call and when to do it yourself (or pay a human). (The AI-mastering landscape shifts — treat tool names as 2026 snapshots.)
The discipline that actually gets songs finished
Here's the part that separates the people with releases from the people with 200 half-finished projects: finishing is a skill you have to practise on purpose. The book's method is a 2-month, 4-song parallel workflow — record vocals on four backing tracks, then rough-mix and polish them as a set, with cooling-off periods so your ears reset and you A/B across songs. You get good at the last stages (the part most people never reach) by deliberately finishing in batches against a deadline. It's the most valuable chapter in the book and the one nobody else writes.
