SetMaster FAQ

My set sounded right when I recorded it. Why master it?

Because it sounded right there. A club system, a pair of headphones, monitors on a desk: each has a sound of its own, and you mixed for it. A room with a heavy low end makes the bass-heavy tracks feel too big, so you pull them down, and in that room it is the right call. Headphones do the opposite to the same tracks.

Then the set leaves, for a phone speaker, a car, somebody's laptop. The room does not travel with the file. Your corrections do. So this is not a mistake being fixed. It is a translation: decisions that were right in one place, made right everywhere else.

Is this the right tool for a single track?

Probably not. It is built for the problem of dozens of already-mastered tracks sitting at different levels. One unfinished track still has a mix to finish, and that is a job for a mastering engineer.

What does SetMaster do to my set?

It finds the drops, measures how loud each one is, and writes a gain envelope that brings them to a common level. That part is automatic, and your tracks keep their own tone, transients and dynamics. Crop, fade, EQ and pan are in the editor too, yours to set by hand if you want them. None of them touches the audio until you move it.

What does it cost?

Free gets you the first 60 minutes of your set. Paid is full length.

Do I need an account?

No sign-up and no password. Your email address and your delivery link are how we know it is you when you come back for an edit, so keep both.

What can I upload?

WAV, MP3, FLAC, AAC or M4A, up to 4 hours and 8 GB. The 4 hours is the real limit; the 8 GB is there to stop abuse. On the free tier a set longer than 60 minutes still uploads, but only its first 60 minutes are edited and rendered.

Why pay for this when I could do it in a DAW?

You could, and it would be correct. Cut at the section boundaries, measure the integrated loudness of each drop, draw the envelope: a few careful hours for a two-hour set. SetMaster does that job and shows you every number it picked. If you would rather do it by hand, nothing here argues against it.

Is my set compressed or limited?

No. The levelling is a gain envelope: a fader, not a processor. Multiplying every sample of a section by the same number leaves the spectrum, the transients and the phase exactly where they were. There is a true-peak safety catch at the end of the chain, and on real material it does not engage. The EQ is flat and the pan is centred until you move them yourself.

Is machine learning involved, and do I have to accept what it decides?

Only to find the drops. The gain values are not machine learning: they come from measuring each section and simple arithmetic. Nothing is generated, replaced or resynthesised, and every boundary and every value is visible in the editor, where you can change it before you render. Your audio is never used to train models, and that is a term of the contract, not a policy we can quietly revise.

My master came back quieter than my mix. What went wrong?

Nothing. A hot recording, with peaks at or above full scale, is turned down to the streaming target of -1 dBTP. Your peak moved down and your loudness moved with it. Your crest, which is the part you can actually hear, stayed where it was.

What loudness do I get, and is it enough for SoundCloud?

There is no loudness dial. The master is peak-normalised to -1.0 dBTP and the loudness lands where your crest puts it. SoundCloud recommends mastering to -14 LUFS, and it applies loudness normalisation as tracks are played, so your master will not always play at the level it was mastered at.

Why -1 dBTP?

It is a safety margin for overshoot. A lossy encode can push peaks higher than they sat in the file, so a master right at full scale can clip after the encode even though it did not before. Measured on a real set: at -0.3 dBTP a master survives our own 320 kbps encode and then clips when a platform re-encodes it at 96 kbps. SoundCloud uses lossy formats, Ogg Vorbis and AAC, and asks for true peak below -1 dBTP so that transcoding introduces no extra distortion. Playback normalisation adjusts the level; clipping from the encode is in the file for good.

How long does it take?

Your set is analysed once it is uploaded and rendered once you finish editing, and the master arrives by email. Under heavy load there can be waiting times; your place in the queue is on screen.

Can I use it on my phone?

You can download a finished master on a phone. The editor is a waveform you drag boundaries on, and it is optimised for desktop use.

Can I change something after the master is delivered?

Yes, for as long as the link is alive. You get the first master plus ten edits, with at least five minutes between renders. Each edit is a full re-render and it pushes the link's expiry back.

What happens to my audio?

It is processed on a server in Germany and it does not leave it.

No tracking, no analytics, no cookies, no third-party scripts. Deletion is on a schedule and we cannot undo it, so download your master and keep your original.

Complaints and refunds

Mail us at mastering@krateo.me.