Think about music production in six stages

What this chapter helps you decide

This chapter combines official technical material with a proposed production exercise. Generation, performance, pricing and listening comparisons have not been run. It is a supplement to the creative production course and can also be read independently.

• Explain the difference between composition, arrangement and performance.
• Choose the stage to revisit for a particular fault.
• Distinguish a finished audio file from editable musical data.

1. 1Purpose and listening context
2. 2Composition: form
3. 3Arrangement: roles

1. 1Performance: expression
2. 2Render: audio
3. 3Mix: balance
4. 4Master: delivery

1. 1Observed failure
2. 2Locate the responsible stage
3. 3Revise and inspect a new export

This is an original conceptual workflow. The arrows show review order, not a product’s hidden architecture; a failure sends the work back to the responsible stage.

Composition defines the structure

Composition concerns melody, harmony, rhythm and form. Start with a one-sentence purpose: “A 30-second explainer needs music that leaves the spoken voice clear.” A listening piece with a chorus and a continuous background bed have different success conditions. Use your own lyrics or lyrics you have permission to use.

Arrangement decides who plays and when

The same melody feels different when a piano starts alone and when drums enter immediately. Separate bass, rhythm, harmony and lead roles; describe their density in the beginning, middle and ending. Before adding another part, remove a sound that competes with the main subject. This is an editorial method for finding the cause of a busy mix.

Performance gives notes expression

Performance adds dynamics, pauses, articulation and timing variation. MIDI stores performance information such as notes; it is not a recording of sound. Changing the instrument can change the timbre. A generated finished recording may not let you edit each note directly as MIDI would. The choice of an editable project therefore belongs in the brief, before generation.

Rendering, mixing and mastering

Rendering writes the result of instruments and effects as audio. Mixing balances several sounds through level, placement, tone and space. Mastering prepares the finished mix for its delivery context, including overall level, sequencing and format. Rendering also happens during revisions: the stages are a map for locating a problem, rather than a single irreversible pipeline.

One AI output still needs several listening questions

A generator can produce composition, performance and sound together. Review the problems separately: an uneventful piece may need a structural change; a buried voice may need an arrangement or mix change. Making the finished file louder cannot repair the melody or form. The producer decides the intention and acceptance reason, the editor checks the export, and the publisher checks permission and destination.

Preserve the finished recording and its materials

Keep the DAW project, source materials, selected recording and release file distinct. A stem is an audio grouping, such as drums, rather than necessarily one original recorded track. Source-separated stems can contain leakage or missing detail; do not assume they are equivalent to the original session. Ableton’s file-management manual documents projects, audio and MIDI files, and rendering selected or individual tracks. Its Audio Fact Sheet explains why some processing changes the signal. These product descriptions support the file distinction; the six-stage workflow is this course’s teaching proposal.

Use and revise the production record

The input is a brief naming the use, duration and destination, plus materials whose usage conditions can be checked. The outputs are the production note, review sheet and the exercise deliverable. Keep the selected original separate from the editing and release copies.

1. Choose one use and listening situation.
2. Write the roles of the beginning, middle and ending.
3. Decide whether a change belongs to composition, arrangement, performance or mixing.
4. Preserve candidate audio in its original form.
5. Plan separate exports for editing and release.

Symptom	Likely cause to investigate	Revision
The piece sounds crowded	Too many parts share a role, rather than only excessive level.	Remove parts competing with the main subject before changing overall level.
Only one sound needs revision, but it cannot be changed	Only a finished mix was kept.	Specify whether an editable project or stems are required before making the piece.

Design a 30-second piece without generating audio

Design an instrumental cue for an imaginary learning-app introduction. Write the role, instruments, density and ending of each ten-second interval. No audio generation is required.

Deliverable: A three-interval structure note and a responsibility table covering the six stages.

Completion criterion: A reader can tell what to compare and where to revise without hearing any audio.

Execution status: not-run. This chapter contains no generated audio, timing log, listening result or completed publication.

Review checklist

• The brief distinguishes background music from music that is the main subject.
• You can explain the difference between note data and audio.
• Each observed problem has a stage to revisit.

A worked structure, before there is an audio file

The following is an illustrative brief, not a measured or generated cue. At 0–10 seconds, soft keys introduce the app while narration establishes its purpose; at 10–20 seconds, a restrained bass adds continuity without a new lead; at 20–30 seconds, parts thin out and a short decay leaves room for the closing sentence. Assign one owner to each of the six stages. If narration is masked in the middle, first revise the arrangement or mix; if the closing section has no sense of arrival, revisit composition.