Mix Scoring Rubric: Criteria, Weights and a Template

A mix scoring rubric is a short list of criteria, each rated on a fixed scale, that turns "it sounds good" into something you can compare from one mix to the next. The best ones blend measured data (loudness, peaks, phase) with listening judgements (balance, depth, emotion), and use the total to decide what to fix first, not to pass a verdict on the song. Below is a rubric you can copy, the reasoning behind each line, and how to turn the score into a work plan.

TL;DR

• Score 8 to 10 criteria from 0 to 4, and weight them by genre and goal rather than treating them equally.
• Split the rubric into measured items (loudness, true peak, phase, spectrum) and heard items (balance, depth, emotion). Each side checks the other.
• A score is a map of priorities, not a pass or fail. The lowest weighted criterion is your next move.
• Always rate on level-matched playback, and test translation on at least three systems.
• Re-score after each revision pass so you can see progress instead of guessing at it.

Table of Contents

• What a mix scoring rubric actually is
• The criteria to score (and what to listen for)
• What belongs in the measured half
• Weighting the rubric by genre and goal
• How to run the scoring session
• A worked example
• Turning the score into a work plan
• Keeping numbers honest
• Our perspective
• Get your own rubric-style report
• FAQ
• Sources

What a mix scoring rubric actually is

If you search this phrase, most of what you find comes from classrooms: generic rubrics for essays, poems and algebra. The idea carries over well to music, because the goal is the same: name the qualities that matter, define what each level looks like, and rate against those definitions instead of a mood.

A published mixing and mastering project rubric shows the classroom version. It rates work from "does not meet expectations" to "past expectations", covers mixing, mastering, tools used, organisation and overall sound, and totals out of 40. Its top mixing level asks for every track to be balanced and audible. That is a fine starting skeleton for a student, but a working producer needs sharper, more measurable lines. That's what we'll build.

The same logic shows up in competitions. The 2018 AES Nashville Spring Mixer judged student mixes on eight elements: fidelity and sonics, imaging width, image depth, mix balance, lead presence, mix development, dynamic range and preparation for mastering, and documentation. Notice the blend: some of those you can measure, some you can only hear.

The criteria to score (and what to listen for)

A rubric that is too long never gets used. Eight to ten lines is the sweet spot. Here is the set we'd start with, and what "a 4" looks like in each.

That list isn't invented from scratch. Beat Kitchen's referencing guidance tells producers to compare low-end balance, vocal level relative to the instruments, width, brightness, and whether the dynamics breathe or sound squashed. Warren Huart's mix review checklist runs through volume and headroom, balance, low end, frequency spectrum, clarity, stereo image, transients, compression, vocals, automation, effects, phase, mono compatibility and musicality. Our table is that same ground squeezed into something you can score in one sitting.

Defining the levels

Use 0 to 4 and write one sentence per level for each criterion before you start. Five levels are enough to separate "broken", "weak", "acceptable", "good" and "release-ready", and few enough that you won't agonise between a 6 and a 7. The sentence is the whole point: if you can't describe what a 2 sounds like, you'll score by mood.

One tip we repeat constantly: describe the symptom, not the tool. "Snare disappears behind the guitars in the chorus" is useful. "Needs more compression" is a guess.

What belongs in the measured half

The measured criteria are the ones where a meter can settle an argument. Here is what each one reveals, and what it can't tell you.

Integrated loudness (LUFS). This tells you how loud the whole track is perceived to be. It matters because streaming services normalise playback. The AES recommendation for streaming says the target loudness of a stream should not exceed -16 LUFS, to avoid excessive peak limiting and leave room for dynamics. What it can't tell you is whether the track is good: a mix at -14 LUFS can still be harsh.

True peak (dBTP). This catches inter-sample peaks that can distort after lossy encoding. One practical pipeline guide suggests streaming-oriented masters around -16 to -12 LUFS-I with true peak at or below -1.0 dBTP, and notes that club-oriented masters often go beyond -10 LUFS-I at the cost of more distortion and listener fatigue risk. Treat these as starting ranges, not laws.

Loudness range and crest factor. Both describe dynamics. A very narrow range, or peaks sitting barely above the average level, points to a heavily compressed or limited mix. The same pipeline guide frames these as decision inputs: with a very low loudness range, avoid further bus compression and consider transient enhancement, while a very high one for the genre may call for gentle bus compression or automation.

Spectrum balance. Comparing your spectrum to a reference shows where energy piles up or goes missing. Sean Kim suggests checking low-end energy, muddiness around 200 to 500 Hz, and harshness between 2 and 5 kHz, and he also says to keep the analyser secondary and trust your ears for final decisions. We agree completely.

Phase and mono compatibility. A correlation meter and a mono check reveal parts that cancel when the stereo field folds down. This is the single most common reason a mix sounds great in the studio and thin in a club or on a phone speaker. Our guides on finding mono compatibility problems quickly, a five-step phase scope checklist and stereo imaging analysis with three meters walk through the routine.

Noise, clipping and DC offset. These are pass or fail items. Score them 4 if clean, 0 if not, and fix before anything else.

Weighting the rubric by genre and goal

Equal weights are the classic rubric mistake. A ballad and a club record are not trying to do the same thing. Koby Nelson makes the point about excitement: it should match the emotion of the song, so a ballad need not be as loud or punchy as an upbeat anthem and should still make you feel what it intends.

A simple way to weight is to give each criterion a multiplier of 1 to 3. Here are three examples to adapt.

These weights are our suggestion, not a standard. Change them to suit the record. If a song is going to broadcast, delivery weighting shifts again, because the targets differ: the same pipeline guide lists EBU R128 at -23 LUFS-I and ATSC A/85 at -24 LKFS for broadcast.

For real-world numbers by genre, our mix and mastering benchmarks show measured data you can use to sanity-check your own targets. If you work in dance music, our electronic music mixing guide covers the genre specifics in more detail.

How to run the scoring session

A rubric is only as honest as the conditions you score under. Follow this order.

1. Level-match first. Louder always wins. Listening-test research is blunt about it: loudness should not influence the rated criteria, so normalise loudness unless loudness itself is what you are rating.
2. Listen once with your hands off. Don't touch a control. Note the first three problems you hear.
3. Score the heard criteria before looking at meters. This keeps your ears from being biased by the numbers.
4. Then read the measured criteria. Where a meter and your ears disagree, that's the interesting bit. Investigate before you decide who's right.
5. Compare with a reference. Score the reference with the same rubric to calibrate what a 4 means for this genre.
6. Check translation. Play the mix on alternate monitors, headphones, a phone or laptop, and a car. Each system reveals different problems. Our guide on mixing on headphones shows how to make that one system translate.
7. Total it, then rank the weak spots by weight.

Making the reference step count

Order matters when you compare. The Waves 5-step method matches perceived levels, then dynamics, then EQ, then panning, then imaging and transient response, so your ear narrows the problem before you chase detail. We like that sequence: it stops you EQing a mix that is only too quiet. For the full case for referencing, see our producer's guide to reference tracks. If you also prepare versions for live sets, the basics of a performance track help you pick which version to compare.

A worked example

Say you're scoring a pop mix with the weights above. Raw scores out of 4, then weighted:

The total is 66 out of a possible 80, or 82.5 percent. That looks healthy, but the useful number is the gap. The biggest lost points sit in low end (6 of 12 lost) and clarity (4 of 8 lost). That tells you to start in the 100 to 500 Hz region, not to touch the vocal, which is already fine.

The percentage is for tracking. The gap list is for working.

Turning the score into a work plan

A rubric that ends with a number is half finished. Translate it into actions:

• Fix pass or fail items first: clipping, noise, phase cancellation.
• Then fix the largest weighted gap. In the example, that's low end.
• Make one change at a time and recheck. A practical routine is to change one thing, recheck, and take breaks every 30 to 60 minutes.
• Re-score the criterion you touched, and one neighbour. Low end changes affect clarity, so check both.
• Stop when it translates. A mix is done when it serves the song and holds up across systems, not when every track carries processing.

Keeping numbers honest

A score is evidence, not a verdict. Two habits keep it that way.

First, don't lean only on meters. Beat Kitchen warns against relying solely on "right numbers", and we've seen plenty of technically tidy mixes with no feeling. Second, include perceptual lines, not just technical ones. Academic work on listening tests describes rating production value, clarity and excitement separately, which gives a richer picture than a single overall preference.

The broader lesson from Huart's checklist is that a comprehensive rubric has to bridge objective items (headroom, phase) and artistic ones (musicality, emotional impact). If yours has only one kind, it will mislead you.

Our perspective

Every report we build is a mix scoring rubric at heart: a set of areas, measured and judged, with specific steps attached. We evaluate 15+ areas, from compression, EQ and frequency masking to stereo width, phase, loudness and device translation, and every report gives you both subjective guidance and objective measurement.

We built it that way because numbers alone don't tell you what to do, and opinions alone can't be checked. The measurements come from audio analysis, and the guidance comes from the real-world experience of our founder, a multi-award winning mix engineer with 20+ years in the craft. Our evaluation framework is updated regularly to reflect current genre standards.

Our advice: score your mix, but let the gaps, not the total, drive your next session. A rubric you rerun after every revision is the quickest route to hearing yourself improve.

AubioMix

Get your own rubric-style report

Building and running your own rubric is a great exercise, and it's even better when you have a second set of ears and meters on the same file. With AubioMix you upload your song and get a visual and written report with actionable steps across 15+ areas, plus progress tracking so you can see each revision improve. You can see what the output looks like in our example analyses, such as Mirror by Sigrid.

A single Mix Review is £4.99, a Reference Pass is £2.99, and subscriptions start at £19.99 per month for Professional or £34.99 per month for Professional +. See the pricing page for what each includes.

FAQ

How many criteria should a mix scoring rubric have?

Eight to ten is enough for most songs. Fewer than that and you miss real problems; many more and you stop finishing the scoring. Group related items, such as stereo width and mono compatibility, into one line if needed.

Should I score my own mixes, or only other people's?

Both, but score your own after a break and with level-matched references. Fresh ears and a reference calibrate your scale. You'll tend to rate your own work more kindly the day you finish it, so rescoring the next day is worthwhile.

Is a high total score a guarantee the mix is ready?

No. A total can hide a single serious flaw, such as phase collapse. Treat clipping, noise and mono problems as pass or fail gates, and only then look at the weighted total.

Can I use the same weights for every genre?

It's better not to. Low end, loudness and width matter more for club music, while lead presence and dynamics carry more weight in acoustic or vocal-led songs. Set the weights before you score, based on the song's goal.

Sources

• Mixing and Mastering Project Rubric
• AES Nashville Spring Mixer 2018 press release
• Referencing & Assessment, Beat Kitchen
• Mix Review Checklist, Produce Like A Pro
• AES TD1004 Recommendation for Loudness of Audio Streaming
• From Demo to Master: Automation Pipeline
• Mix Self-Assessment Techniques
• What Makes a Great Mix? 9 Things to Listen For
• Listening Tests, Deep Learning for Automatic Mixing
• The Science of Mix Referencing: The 5-Step Method
• How to Listen Critically to a Mix

Recommended

• Evaluation Framework
• Mix Notes Template: Session Notes and a Self-Review Checklist
• Mix & Mastering Benchmarks by Genre — Real Measured Data
• Producers: Upload First Mix Analysis Workflow for Actionable Feedback