Mix feedback changes the mix. Mastering finishes the mix.
Mix feedback is the review stage where you still have access to the session: vocal level, kick and bass balance, EQ masking, compression, panning, effects, automation, and arrangement density. Mastering works on the finished stereo file. That difference sounds obvious until a producer sends a muddy mix to mastering and expects the mastering chain to separate the kick, bass, and vocal.
The clean rule is this: if the problem belongs to one instrument, one section, or one relationship inside the arrangement, solve it in the mix. If the mix already works and the remaining job is loudness, translation, sequencing, codec safety, and final polish, send it to mastering.
| What you hear | Best next step | Reason |
|---|---|---|
| Vocal disappears in the chorus | Mix feedback | The vocal balance, automation, masking, or arrangement still needs work. |
| Low end is huge on headphones but gone on a phone | Mix feedback | Kick, bass, sub, and harmonic support need a mix decision. |
| Mix is balanced but not release-loud | Mastering | The stereo file can be optimized for loudness and peak safety. |
| Only the vocal stem is too bright, but the session is unavailable | Stem mastering | Grouped stems give the engineer more control than a stereo master. |
| You are unsure what is wrong | Mix feedback first | A diagnostic pass is cheaper than mastering the wrong version. |
What mix feedback should catch before mastering
Good feedback does not mean a list of taste opinions. It should identify the decisions that will get worse when a limiter, encoder, or playback system stresses the mix. The most useful notes are specific enough that you can reopen the session and fix one thing.
- Frequency balance: low-mid buildup, missing air, harsh upper mids, or kick and bass masking.
- Dynamics: crushed drums, uncontrolled peaks, weak section lift, or overactive bus compression.
- Stereo field: over-wide lows, phase risk, narrow choruses, or effects that vanish in mono.
- Clarity: vocal masking, cluttered arrangements, reverb wash, or lead elements fighting the same range.
- Translation: a mix that only works on your monitors but fails on earbuds, phone speakers, or the car.
Mix Analyzer is useful here because it gives you a fast second opinion across more than twenty checks. The report should not override your ears. It should tell you where to listen again before the master chain hides the evidence.
What mastering can actually do
Mastering works globally. A mastering engineer can shape broad tone, control final loudness, catch clipping, manage true peak, prepare files, and make a song or EP translate better across playback systems. They cannot turn down only the snare bleed in the vocal mic unless you sent stems. They cannot unmask a bass line from the kick without affecting the whole low end. They cannot restore punch that a brick-wall limiter already removed from the mix bounce.
Streaming loudness guidance also belongs here, but it should not become a mix-stage obsession. Spotify currently describes playback normalization around -14 LUFS and recommends keeping true peak below -1 dBTP, or below -2 dBTP for masters louder than -14 LUFS. Apple Digital Masters asks engineers to avoid clipping and leave headroom for encoding. Treat those numbers as delivery and playback references, not a reason to crush your mix before mastering.
The pre-master decision test
Before you book mastering, run this test. Listen quietly. Listen loud for one pass only. Fold to mono. Compare against one level-matched reference. Then check the bounce in Mix Analyzer. If the same issue shows up in the report and in your listening notes, it belongs in the session.
- Balance pass: can you follow the lead vocal, snare, bass, and hook at low volume?
- Low-end pass: does the bass line remain readable when the sub is not available?
- Dynamics pass: do drums still hit before any mastering limiter is added?
- Mono pass: do kick, bass, vocal, snare, and main hook survive when the sides collapse?
- Export pass: is the file WAV or AIFF, no normalization, no dither unless this is the final delivery, and no accidental clipping?
When stem mastering makes sense
Stem mastering is not a rescue fantasy. It is useful when the mix is close, the full session is unavailable, and one or two grouped areas need light control. Four to eight stems - drums, bass, vocals, music, effects, and maybe lead elements - give the mastering engineer enough separation to solve small balance problems without rebuilding the mix.
If the vocal rides are unfinished, the bass arrangement is wrong, or the chorus does not lift, stem mastering will still be a compromise. Use it when you need controlled correction. Do not use it to avoid a mix revision that you can still make.
How to use Mix Analyzer in this decision
Upload the pre-master bounce and read the report as a triage tool. Frequency and clarity flags usually point back to mix feedback. Loudness, true peak, clipping, and file-quality flags usually point toward mastering preparation. Stereo and dynamics can go either way, so verify them by listening.
The best workflow is short: analyze, listen, revise one decision, export again, and compare. If the second bounce measures better and feels better, move forward. If the score improves but the song feels worse, keep the song and reject the fix.
The session-level problems mastering should not inherit
Most failed masters start with one of three unfinished mix decisions. The first is hierarchy: the listener cannot tell what to follow because the vocal, hook, snare, lead synth, or guitar line changes role from section to section. The second is masking: two important parts occupy the same frequency and rhythm range, so the mastering EQ can only make both darker or brighter together. The third is unstable width: the exciting part of the mix lives in the sides, but the center does not carry enough of the song.
Those problems are not shameful. They are normal mix problems. The expensive mistake is sending them forward because mastering feels like progress. Reopen the session and make the smallest fix that solves the actual relationship. Ride the vocal. Move one synth octave. Narrow the bass. Pull reverb from the hook. De-ess the vocal before the stereo bus. These are mix moves, and they cost less when you still control the tracks.
What to send when the mix is ready
A mastering engineer does better work when the delivery is boring and clear. Send the clean premaster, a loud reference if you mixed through bus limiting, two or three reference tracks, and short notes about intent. Do not send a paragraph defending every choice. Send the information that changes the engineer's decision.
- The clean stereo mix, lossless, clearly labeled with version and sample rate.
- A limited loudness reference only if it represents the sound you were mixing into.
- Two or three reference tracks with a note about what each reference is for.
- Any known creative risks: distorted vocal, intentionally dark master, wide synth bass, noisy sample, or aggressive drum clipping.
- A simple release target: streaming single, club record, album track, sync cue, demo, or vinyl pre-master.
A useful order for the final review
Review the mix in the order a master will stress it. First balance, because mastering makes balance mistakes more obvious. Then low end, because bass controls headroom and translation. Then dynamics, because limiting exposes over-compression. Then stereo, because encoding and mono playback punish phase tricks. Finish with file quality, because a wrong export can waste a whole revision cycle.
If you can explain the biggest remaining issue in one sentence, you are ready for the right next step. "The vocal is still masked in the last chorus" means mix feedback. "The mix translates and needs final loudness, sequencing, and true-peak safety" means mastering.
Do not judge the rough master as the mix
Many producers keep a loud reference chain on the mix bus while working. That is fine if it helps decisions, but print a clean premaster before the final review. Then compare the limited rough master against the clean bounce. If the song only feels alive with the limiter smashing it, the mix may need dynamics work. If the clean bounce already communicates and the rough master only adds level, mastering is the right next step.
FAQ
Is mix feedback the same as mastering?
No. Mix feedback reviews track-level decisions while the session is still editable. Mastering processes the finished stereo mix for translation, loudness, true peak, and final delivery.
Can mastering fix a muddy mix?
Only a little, and the fix affects the whole stereo file. If the mud comes from kick, bass, guitars, piano, or vocal layers masking each other, fix it in the mix.
How much headroom should I leave before mastering?
Leave practical headroom and avoid clipping. A common target is peaks around -6 to -3 dBFS with no heavy limiter on the mix bus, but clean dynamics matter more than hitting a decorative number.
Should I use AI mix feedback before hiring a mastering engineer?
Yes, if you treat it as a listening checklist. Use automated feedback to find likely problems, verify them by ear, and fix the issues that clearly belong in the mix.