What Real Compositional Feedback Looks Like

Most students who want to learn composition have no access to a composition teacher. Not because no such teachers exist, but because private tuition is expensive, rare, and geographically concentrated. A student in a mid-sized town may have nowhere to turn for serious feedback on a score they have written. The critique Gradus provides is not a replacement for a human composition teacher. It is the structured, analytical, bar-specific feedback that a student should be getting between lessons with one — or, if no human teacher is available, the closest practical substitute.

> The short answer: Real compositional feedback is specific to the score, names specific techniques, points to specific measures, and gives the student a specific next action. Vague encouragement ("this is lovely") and vague criticism ("the harmony feels weak") are both useless. The Gradus critique system scores a submitted composition on thirty-two dimensions — voice-leading, counterpoint, melodic contour, harmonic rhythm, and more — and returns three strengths, three growth areas, and one concrete next exercise.

How do I get feedback on my compositions?

Traditionally, through a composition teacher — a human professor who reads the score, points to specific bars, and tells you what is working and what is not. This remains the ideal. It is also, for most students, unavailable at any affordable frequency.

The alternatives are: post your piece to a forum and hope someone knowledgeable responds (quality varies wildly, and most responders are other students in the same position); show it to friends (kind, but usually not technically trained); or work in isolation, which is how most students actually end up and which is the slowest possible path.

Gradus offers a fourth option. Submit your score to Maestro — the composition professor built into the method — and receive structured feedback within seconds, grounded in a thirty-two-dimension analysis of the actual notes you wrote.

What should composition feedback include?

Real feedback addresses the craft at every level it operates on. The Gradus critique framework evaluates thirty-two specific dimensions grouped into eight families:

Intent and style — Does the piece achieve what you set out to do? Is the style fidelity consistent? Does the voice hold together across the work?

Melody — Is the contour shaped or aimless? Is there a thematic clarity the listener can grasp and remember? Is the ornamentation serving the line or cluttering it? Is the phrase structure coherent?

Rhythm — Is there rhythmic variety or monotony? Is the density appropriate to the style? Is the meter clear? Are syncopations purposeful?

Harmony and tonality — Is the harmonic rhythm supporting the phrase structure? Are cadences landing where the ear expects? Is the chord progression functional or incoherent? Are modulations logical?

Voice-leading and texture — Are the voices written cleanly? Do they maintain independence? Is the register and spacing appropriate? Is the doubling correct? Does the texture contrast from section to section?

Contrapuntal craft — Is the motive developed or abandoned? Is imitation used well? Is dissonance prepared and resolved? Is the counterpoint quality defensible?

Unity and variety — Does the piece economize its motives? Is the form clear? Is the development genuine, or is it repetition pretending to be development?

Expression and performance — Is the dynamic shape intentional? Are the articulations serving the rhetoric? Is the writing idiomatic to the instruments involved?

A piece of feedback that addresses only one of these families is incomplete. Good critique prioritizes — names the three most important strengths, the three most pressing growth areas, and one concrete next exercise.

Can AI give music composition feedback?

With care, yes — but the architecture matters. A large language model asked to "critique this score" without structure will produce plausible-sounding prose that may or may not be grounded in what the student actually wrote. This is the failure mode of naive AI feedback tools. They sound authoritative and are often wrong.

The Gradus critique system is built differently. The first pass is programmatic — a deterministic analyzer reads the score and computes specific metrics: interval content, rhythmic variability, contour typology, voice-leading distance, motif similarity, feature vectors compared against style-period centroids, and counts of parallel fifths, voice crossings, and range violations. These numbers come from the notes on the page. They are not opinions.

Only after the scoring is complete does Maestro write the prose commentary — and he does not re-score in prose. His job is to interpret the numbers into specific, actionable teaching for the student. This is the correct division of labor: the analyzer measures, the teacher teaches.

What does a Gradus critique actually return?

A scorecard showing all thirty-two dimensions, grouped by family, each rated one to five or marked N/A where the composition lacks the features required to evaluate that dimension.

Three strengths — the top three scoring dimensions, with specific commentary on what the student did well, grounded in specific bars.

Three growth areas — the bottom three scoring dimensions, with targeted advice about what to work on and why.

One next step — a single concrete exercise the student should do in their next sitting. Not a vague suggestion to "work on harmony," but a specific task: rewrite measures 9–16 with a clearer tonic-dominant alternation; pay attention to where your bass line wants to land on beat one.

This format exists because unfocused feedback is ignored feedback. A student who is told thirty things to fix will fix none of them. A student who is told one thing to fix will fix it.

How does Maestro differ from a chatbot?

Maestro is a composition professor. He responds only to a student's submitted score, never generates music for the student, and never replaces their own compositional decisions. His voice is specific — he is the teacher who developed the method, and he speaks as one, referring to specific techniques by name, pointing to specific bars, holding the student to the standards of the craft.

Feedback is always on request, never automatic. The student decides when a passage is ready for critique. This is how composition teachers work in the conservatory setting, and the principle carries over: the student writes until they have something to show; then and only then do they show it.

Does this replace a human teacher?

No, and it is not meant to. A human composition teacher brings things no system can replicate: they have a personal style preference that shapes their advice, they hear nuances an analyzer cannot measure, and they build a long-term relationship with the student's developing voice over years of work. If you have access to a good human teacher, take it.

What Maestro provides is the critique between lessons — the specific, analytical feedback a student needs the other six days of the week, when the teacher is not in the room. For students with no human teacher available, it is the closest available substitute to structured tuition, at any hour of day, grounded in the actual score.

The craft is earned through practice, and practice requires feedback. Begin Your Journey.