Back to blog

How to Prepare Your Mix for Mastering: 10-Point Checklist

Peak around −6 to −3 dBFS, take the limiter off and export 24-bit or 32-bit float WAV. A 10-point checklist for mixes that master cleanly.

Nov 12, 2026Track2MasterTrack2Master
How to Prepare Your Mix for Mastering: 10-Point Checklist

To prepare a mix for mastering, export the full song as a stereo WAV at 24-bit or 32-bit float, at your session's sample rate, with peaks around −6 to −3 dBFS and no limiter or loudness maximizer on the master bus. Before you export, check the low end, compare the mix with a reference at matched loudness, keep the reverb tail intact and give the file a clear name. Mastering can polish a good mix, but it can't rebalance one, so these ten checks decide most of how the final master will sound.

The 10-point checklist at a glance

Here's the whole checklist in one table. Each point gets its own section below.

#CheckWhat to aim for
1HeadroomPeaks around −6 to −3 dBFS, export normalization off
2Master busNo limiter, clipper, maximizer or room-correction plugin
3File formatWAV at 24-bit or 32-bit float, never MP3
4Sample rateThe same rate as your session
5Start and endClean start, complete tail, fades only on purpose
6Low endKick and bass separated, sub-bass centered, no rumble
7ReferenceCompared with a released song at matched loudness
8File nameArtist, title, version, sample rate and bit depth
9Stereo or stemsA stereo bounce, unless stems were asked for
10Already loud?Export an unlimited version too, or master gently

1. Leave headroom: peak between −6 and −3 dBFS

The loudest moment of the song should peak somewhere between −6 and −3 dBFS on your master meter. That's peak level, not average loudness. A finished mix that peaks at −4 dBFS is completely normal.

Headroom gives the mastering chain room to EQ and compress without anything hitting the ceiling. It also shows at a glance that the mix hasn't already been squeezed. If you export at 32-bit float, the file itself won't clip above 0 dBFS, but plenty of plugins, converters and fixed-point formats further down the chain will, so a few dB of space is still the safe habit.

The easiest fix is to pull down the master fader. In most DAWs the fader sits after your master-bus plugins, so lowering it doesn't change how your bus compressor or EQ behaves. Also make sure normalize on export is off. That option pushes the peaks right back up to 0 dBFS and wipes out the headroom you just made.

Don't chase a loudness number at this stage. An unmastered mix usually measures well below release loudness, and that's expected. Bringing it up to level is the mastering stage's job.

2. Take the limiter off the master bus

Remove any limiter, clipper or loudness maximizer before you export. A limiter on the mix bus flattens the transients once. The mastering limiter then has to work on drums that are already flat, and the result usually sounds smaller and more tired than a single, well-set stage would.

What can stay: a bus compressor or EQ that you mixed into and that's part of the sound. If the glue on your mix bus is doing something you love, keep it. Just mention it to whoever masters the track, or put a note in the file name.

While you're on the master bus, bypass anything meant for monitoring rather than the music. Speaker or headphone correction plugins and reference-switching utilities should never be printed into the export.

3. Export a WAV at 24-bit or 32-bit float

Send an uncompressed WAV (or AIFF) at 24-bit or 32-bit float. 24-bit is the standard and works with every mastering service. 32-bit float has the extra advantage that anything above 0 dBFS isn't clipped in the file. Use it if the service you're working with accepts it.

Avoid MP3, AAC and audio ripped from video. Lossy encoding throws away detail and adds small artifacts, and mastering makes those artifacts louder along with everything else. A 16-bit WAV is fine if it's all you have, but don't convert a 24-bit mix down to 16-bit just to send it.

On dither, skip it when you export 32-bit float. Dither matters when bit depth is reduced, and that should happen once, at the very end. Track2Master's AI mastering, for example, gives you a 24-bit WAV plus a 16-bit dithered WAV of each version, so the dither is applied after mastering, not before.

4. Keep your session's sample rate

Export at the sample rate you mixed at, usually 44.1 kHz or 48 kHz. Upsampling a finished mix to 96 kHz doesn't add anything that wasn't recorded. Downsampling is a conversion step, and conversions are best done once, as part of the final master.

If you're preparing an EP or album, keep every song at the same sample rate and bit depth. It makes consistent mastering easier and rules out one source of surprises when the songs are sequenced.

5. Keep the start clean and the tail complete

Leave a short gap of silence before the first note and let the reverb and delay tails fully decay before the export ends. Exports that start exactly on the first downbeat can click, and tails that get cut off sound like a mistake once the song is mastered and louder.

A quick pre-export pass catches most problems:

  • Set the export range a few seconds past the last note.
  • Mute click tracks, guide vocals and count-ins.
  • Clear every solo and check that no track was left muted by accident.
  • Play the last 20 seconds and listen for automation that snaps or a plugin that stops early.

Fade-outs depend on who's mastering. A mastering engineer may prefer to make the fade after mastering, so ask first. With an automated service, put the fade in your mix so you get exactly the curve you want.

6. Check the low end in mono and at low volume

Make sure the kick and bass each have their own space, the sub-bass is centered and there's no rumble below the instruments' useful range. Low-end problems are the most common reason a mix won't master well, because mastering works on the whole file and can't pull the kick and bass apart once they're printed together.

A few checks help:

  • Listen in mono. The bass and kick should stay solid and the vocal shouldn't disappear.
  • Listen quietly. At low volume, balance problems in the low end are easier to hear.
  • High-pass tracks that don't need deep bass, such as pads, guitars and vocals, so they don't stack up mud.
  • Listen on earbuds or a phone speaker as well as your monitors.

To check width and phase objectively, run the bounce through our LUFS meter. It shows stereo correlation, and values near or below zero are a warning sign for mono playback. For reference, our AI mastering folds everything below 120 Hz to mono and keeps its tone EQ gentle, never more than ±4 dB per band (±6 dB when matching a reference). That's plenty to polish a mix, but it's not meant to rescue a bass line that's 8 dB too loud.

7. Compare against a reference at matched loudness

Pick one or two released songs in the same genre and compare them with your mix at the same loudness. Commercial masters are much louder than an unmastered mix, and louder always sounds better at first. If you don't match levels, you'll judge the volume instead of the mix.

Turn the reference down until it sits at about the same level as your mix, then compare:

  • Bass versus vocal, the balance that matters most
  • Overall brightness and how harsh the top end is
  • How loud the vocal sits in the chorus
  • Stereo width and how much reverb is on the main elements

Measuring both files makes matching easy. Note the integrated loudness of each and turn the reference down by the difference. If you want the final master itself to follow a reference, Track2Master's reference mastering lets you add your mix and a reference song, matches the overall tone and loudness, and lets you A/B the two at matched level.

8. Name the file so versions can't get mixed up

Use a clear, consistent name like Artist - Song Title - Mix v3 - 48k 24bit.wav. Version numbers stop the wrong mix from being mastered, and the format details save a question later.

A few rules keep files portable. Skip slashes, colons, question marks and quotation marks, which some systems reject. Don't call anything "final", because there's always a final-final. Keep a short notes file next to the mix covering anything intentional that might look like a mistake, like lo-fi noise, a deliberately distorted vocal or a sudden stop.

Distributors and DJs will often ask for the song's tempo and key, so note them now. If you don't know them, our key finder detects key and BPM in your browser.

9. Send a stereo bounce unless stems were asked for

Most mastering, including AI mastering, works on one stereo file. That's the standard deliverable, and it's what you should send unless someone asks for stems.

Stem mastering is a separate service. You export several group stems, such as drums, bass, music, vocals and effects, and the engineer adjusts their balance while mastering. If you're asked for stems, export them all from the same start point, at the same length and settings, so they line up exactly. Either print your master-bus processing on all of them or leave it off all of them.

Sometimes you only have a finished stereo file, like an old bounce or an AI-generated song, and you need to rebalance it. A stem splitter can separate vocals, drums, bass and other parts so you can rebalance them before mastering. Separation isn't perfect, so make small moves. Our stem splitter runs on a cloud GPU, which means the file is uploaded for processing. Only use it with music you own or have permission to use.

10. If your mix is already loud, export both versions or master gently

The best fix is to bypass the limiter and export a second, unlimited version. Then send both, with the loud one as a reference for the sound you're after. Many producers mix into a limiter because it makes the mix feel finished, and that's fine as long as an unlimited version exists.

If the loud file is all you have:

  • Check for clipping first. Drop it into the LUFS meter and look at true peak and clipped samples. If the file is clipped, re-export it if you can, because mastering can't undo clipping.
  • Choose a target close to where the file already is. If the mix already measures around −10 LUFS, a Loud (−11) or Club (−9) target lets the mastering limiter do very little, instead of squeezing an already dense mix even harder.
  • Read the report. Track2Master's mastering report shows how much limiting it applied and warns when a mix is peaking near 0 dBFS.

Hear your prepared mix mastered

Once the checklist is done, the quickest way to learn whether the mix is ready is to hear it mastered. In Track2Master's AI mastering, the mix stays in your browser and isn't uploaded for mastering. You get Warm, Balanced and Bright versions to compare with your original in a synced player, with a Match loudness switch so tone isn't confused with volume.

The full-length preview is free and unlimited, with no sign-up, which is useful for testing whether a mix change actually helped. If you'd like to start without paying anything, the free mastering page explains what the free preview includes. Downloading the 24-bit and 16-bit WAVs of all versions costs 10 credits per track or is included in a monthly plan; the options are listed on the pricing page.

Quick answers

What level should my mix be before mastering?

Peaks between −6 and −3 dBFS, with no limiter on the master bus. The integrated loudness of the mix doesn't need to hit any particular number.

Should I leave compression on the mix bus?

Yes, if you mixed into it and it's part of the sound. Take off limiters, clippers and maximizers.

Is 24-bit or 32-bit float better for mastering?

Both work. 32-bit float can't clip above 0 dBFS inside the file, so it's the safer choice when the mastering service accepts it.

Should I dither my mix export?

Not for 32-bit float. Dither belongs at the final bit-depth reduction, which happens after mastering.

Can mastering fix a bad mix?

Only a little. Mastering can tilt the overall tone, add glue and set loudness, but a buried vocal or a kick fighting the bass has to be fixed in the mix.