What Is Matchering? (Open-Source Mastering)
Published · updated · by Loopin
Matchering is an open-source (Python) reference-based mastering library: you give it your track plus a reference song you love, and it matches your master to that reference’s loudness, tonal balance and dynamics. It’s a clever, free approach — and if you’d rather skip the code and the reference entirely, Loopin’s browser masterer is a no-reference alternative that targets streaming loudness automatically.
The short answer
Matchering is an open-source tool that masters by matching a reference. Instead of applying fixed presets, you hand it two files — your unmastered track and a professionally mastered song you want to sound like — and it analyzes the reference and reshapes your track to match it.
It’s written in Python and released as a free library, so it powers command-line workflows, scripts and web front-ends. The most popular front-end is Songmastr, which puts Matchering behind a simple web page.
How reference matching works
The idea behind reference mastering is intuitive: a great master already encodes good decisions about loudness and tone, so why not measure those and apply them to your song? Matchering does exactly that in three broad moves.
It compares your track to the reference and adjusts to close the gap:
- Loudness — it brings your track’s overall level in line with the reference
- Tonal balance — it matches the frequency curve, so your lows, mids and highs sit like the reference
- Dynamics — it shapes the peak-to-average relationship toward the reference’s feel
- Peaks — it limits the result so it doesn’t clip
What you need to use it
The core Matchering library runs in Python, so using it directly means installing it and pointing it at two audio files from the command line or a script. That’s straightforward for developers and a barrier for everyone else.
You also need a good reference track — and this is the crux of the whole approach. The result is only as good as the reference you pick. Choose a song in the same genre, well mastered, with the tonal balance and loudness you’re aiming for. A mismatched reference produces a mismatched master.
- Your track — the unmastered mix you want to finish
- A reference song — a professionally mastered track to match
- Python (for the core library) or a web front-end like Songmastr
- A good ear for choosing a reference that fits your genre
Songmastr: Matchering without the code
If installing Python isn’t for you, Songmastr is a free web front-end built on Matchering. You upload your track and a reference, and it runs the same reference-matching engine in the browser front-end without any command line.
Songmastr is free up to roughly seven songs a week, which is generous for hobby use. Like the underlying library, it still needs an upload and a reference track to work — so the reference-picking decision is still yours to get right.
When reference matching shines
Reference mastering is genuinely powerful when you have a clear target in mind. If you’re trying to sit alongside a specific artist’s sound — the same warmth, the same brightness, the same loudness — matching to their track is a direct way to get there.
It’s also a great learning tool: comparing your mix to a reference reveals exactly where your low end is heavy or your top end is dull. For producers who love that kind of control, it’s a rewarding, free workflow.
When you might not want a reference at all
The reference requirement cuts both ways. You have to have a good reference and pick it well, and you have to be comfortable uploading both files. For a lot of quick jobs — finishing a single, a beat or a demo for streaming — you just want a clean, loud master without hunting for a reference or opening a terminal.
That’s where a no-reference, no-code option fits. Loopin runs a real chain entirely in your browser and targets streaming loudness on its own — no reference track to choose, no upload, no account. It applies corrective EQ, multiband compression with the bass band kept independent and mono, character EQ, saturation, mid/side stereo with the low end kept mono, two-pass loudness targeting to about −14 LUFS, and a true-peak limiter at −1 dBTP, then exports a 16-bit WAV.
- No reference needed — it targets streaming loudness automatically
- No code — it runs in the browser, no Python
- Nothing uploaded — your unreleased track stays on your device
- Free WAV — no signup, no watermark
Try the no-reference route in your browser
Want a finished master without picking a reference or installing anything? Here’s the fastest path.
- Bounce your mix as a WAV with headroom (peaks near −6 dB, no mix-bus limiter).
- Open Loopin Mastering and drop the file in — no account, no reference track.
- Pick a feel (Warm, Punchy, Crisp, Wide, Club, Lo-fi…) and A/B Original vs Mastered.
- Download the mastered WAV — free, no watermark, nothing uploaded.
Matchering vs a no-reference masterer
Both approaches are valid — they just suit different jobs. If you have a specific track you want to sound like and enjoy the hands-on, open-source workflow, Matchering (or Songmastr) is a smart, free choice. If you want a finished file with the fewest decisions and total privacy, a no-reference browser masterer is faster.
The targets end up in the same place: a clean, streaming-loud master with a safe true-peak ceiling. The difference is whether you steer with a reference song or let an automated chain hit the numbers for you.
Frequently asked questions
What is Matchering?
Matchering is an open-source Python library for reference-based mastering. You give it your track plus a reference song, and it matches your master to that reference’s loudness, tonal balance and dynamics. Songmastr is a free web front-end built on it.
How does reference mastering work?
It analyzes a professionally mastered reference song and reshapes your track to match it — bringing your loudness, frequency balance and dynamics in line with the reference, then limiting peaks so the result doesn’t clip. The quality depends heavily on choosing a good, genre-appropriate reference.
Is Matchering free?
Yes. Matchering is open-source and free to use. The Songmastr web front-end built on it is also free, up to roughly seven songs a week. Both require an upload and a reference track to work.
Do I need a reference track to master?
For Matchering and Songmastr, yes — reference matching is the whole point, so you must supply a reference song. If you’d rather not pick one, a no-reference tool like Loopin targets streaming loudness automatically in your browser, with no reference and nothing uploaded.
What’s the difference between Matchering and Loopin?
Matchering matches your track to a reference song you provide, and runs in Python (or via Songmastr with an upload). Loopin runs a fixed real chain in your browser, needs no reference and no account, uploads nothing, and targets about −14 LUFS with a −1 dBTP ceiling automatically.