Music Composition and Production Case Studies
Creative5 Case StudiesUpdated Sep 2026
These five case studies come from working studios, scoring rooms, and independent production teams that deliver music for film, games, advertising, and recordings. Each begins with a brief that pulls in several directions at once, then follows the decisions that moved the work forward. The cases assume you already know how to track a session and manage a mix. Their real subject is judgment: how a team turned competing demands into a concept, a schedule, and a set of checks it could defend. Read each situation before the actions, and test the proposed steps against your own habits.
Case Study 1
Reconciling a retail anthem with a voiceover-first mandate
Harbor Lane Sound, a nine-person audio house that scores branded video for regional retailers and streaming platforms.
The situation
Harbor Lane Sound received a brief from Brightwater Grocers, a regional chain of forty-one stores, for a spring campaign spot. Gil, the owner, wanted a bright, sing-along chorus with handclaps and a fiddle line, the kind of energy that fills a parking lot. Priya, the marketing director, drew on research with older suburban shoppers and asked for something close to invisible under a voiceover, with nothing competing in the speech range. The delivery schedule added a 30-second hero cut, a 15-second cutdown, and mono and stereo mixes for in-store screens and phones. The brief listed eleven adjectives, several of which pointed in opposite directions.
The challenge
The brief could not be satisfied literally. A chorus loud enough to carry a parking lot would bury a voiceover, and a voiceover-friendly bed would lose the energy the owner had asked for. The team also had a mid-week deadline for the first review, with two rounds of notes already scheduled, and the client's legal reviewer wanted the loudness targets documented before any music reached air.
Analysis
Teams grounded in Music Composition and Production tend to start by sorting a brief into fixed constraints and negotiable preferences, because the two are rarely labeled. Here, the speech range and the 15-second cutdown were fixed. The adjectives were not. Rather than argue over which adjective won, the composer wrote a one-page translation: the chorus would be built from a rhythmic handclap pattern and a fiddle motif that lived mostly above the speech band, while the bed beneath the voiceover would drop to a sustained pad. That turned an argument about taste into a question of frequency and arrangement, which the client could evaluate.
Evaluation criteria matter as much as the concept. The team wrote four checks that any reviewer could apply without a musical vocabulary: whether the voiceover was intelligible on a phone speaker, whether the chorus could be recognized within the first four seconds, whether the cutdown kept the hook intact, and whether both mixes met the loudness target without limiting. Those checks also gave the client a way to disagree productively. When the owner pushed for a louder ending in the second round, the team could show which check the change would break, and the discussion moved to tradeoffs instead of preferences.
Actions taken
- Wrote a one-page translation that mapped each adjective to a musical decision and a measurable check, then sent it before any sketch existed so the client could correct the mapping.
- Built two sketches from the same chorus, one with the fiddle motif above the speech band and one with a sustained pad under the voiceover, and rendered both at target loudness before presenting.
- Created a cutdown plan at the same time as the hero cut, marking which phrases survived at fifteen seconds so the hook was written to carry both lengths rather than trimmed later.
- Logged every revision request with its source, the reviewer's role, and the check it affected, then shared the log at each review so changes traced back to the brief rather than to mood.
Results
The team presented the translated brief and two sketches in the second week. The client chose the pad-under-voiceover arrangement for the hero cut, and the owner accepted a fiddle-led chorus that entered after the voiceover had finished its first sentence. The final mixes passed the loudness check on the first pass, and the cutdown kept the hook intact. Priya reported that the spot tested clearly on phones. Gil still wanted a bigger ending, which the team delivered as an alternate tag that the client used only on in-store screens.
Questions for discussion
- Which of the eleven adjectives in a brief like this would you ask the client to rank before any sketching begins, and why?
- How would you write evaluation criteria that a reviewer with no musical training could apply without deferring to taste?
- When a client asks for a change that breaks one of your agreed checks, what is the most useful way to show the tradeoff?
Case Study 2
Tracing a stalled stem delivery through four handoffs
Tallow Creek Audio, a fourteen-person post-production house that mixes and masters music and dialogue for documentary and short-film clients.
The situation
Tallow Creek Audio received the score for a feature documentary of about ninety minutes, with a delivery date set for the end of the quarter. The work moved through four hands before it reached the mixing suite: a composer working from a home studio, an orchestrator who arranged the sketches for a session ensemble, a scoring editor who cut the recordings to picture, and the mix team. Each handoff used a different folder structure. Tempo changes made during the third round of picture edits were written into a text note rather than the session file. By the time the stems arrived, Owen, the mix engineer, was working from a file labeled final, which was the third revision.
The challenge
The delivery was four days out when the mix team discovered that two cues were in the wrong key relative to the picture edit, that a sample rate mismatch had pitched one ensemble stem, and that the orchestrator had no record of which sketch the session players had actually recorded. Each problem alone was small. Together they meant the team could not trust any file without checking it against the session notes, and the checking was not on anyone's calendar.
Analysis
Work of this kind often looks like a creative problem when it is really a logistics problem with creative consequences. The brief described the emotional arc of the film, but it said nothing about how the score would be handed over, and the mix engineer was an audience the brief never named. That engineer needed unambiguous stem names, one sample rate, and a tempo map matched to the locked picture. None of those needs appeared in writing, so each contributor made reasonable local choices that failed at the next step. Treating handoff requirements as part of the brief from the start would have cost an hour and saved the week.
Review feedback in this chain arrived as a list of problems instead of a list of decisions. The mix team's notes said the ensemble sounded thin and the cue felt late, which pointed at several possible causes. Only after the team compared session logs with the picture edit did the cause become clear for each one. A versioning rule would have prevented most of the confusion: one master tempo map stored with the session, one naming convention for stems, and a single approved revision number written into every file name and the delivery sheet. Reviewers could then confirm a specific version rather than an impression of the latest one.
- 1Composer sketchesHome studio files used local names and mixed sample rates.
- 2Orchestrator arrangementSketch changes shared by email, never written into session.
- 3Ensemble sessionPlayers recorded an earlier sketch than the one on file.
- 4Scoring editTempo changes from round three lived only in a text note.
- 5Mix intakeFiles labeled final were actually the third revision.
Actions taken
- Froze the tempo map and picture lock as a single reference file, then reissued every cue sheet against it so that no contributor was working from a remembered tempo.
- Set one naming convention for stems, with the cue number, instrument group, sample rate, and revision in the file name, and rejected any delivery that did not follow it before anyone opened it.
- Ran a session audit with the orchestrator and ensemble contractor to confirm which sketch each player recorded, then logged the answer beside each take in the session file.
- Scheduled a fifteen-minute check at the start of each review where the mix engineer compared the delivery sheet, the revision number, and the picture lock before any listening began.
Results
The team paused the mix for two days while it audited every stem against the session logs. It found the pitched ensemble stem, corrected both cues that sat in the wrong key, and rebuilt the cue sheet from the locked picture. The delivery moved by four days, which the production company accepted in exchange for a written schedule with a fixed freeze date. Two later revisions went through the new naming rule without a single misnamed file reaching the mix suite. The director noted that the final mix had fewer rounds of notes than earlier work.
Questions for discussion
- Which handoff in a chain like this would you audit first if you inherited the project halfway through, and what would you look for?
- How would you decide whether a version rule belongs in the brief, the session template, or the delivery contract?
- What is the smallest review check that would have caught the wrong-revision file before the mix began?
Case Study 3
Two cue directions split a small audience panel
Lantern Row Studio, a six-person composition studio that writes cues for independent game developers and short documentaries.
The situation
Lantern Row Studio writes cues for independent games and short documentaries. A small game developer, Foxglove Interactive, asked for a theme for the main menu and a looping exploration bed for a hand-drawn adventure set in a flooded valley. The brief asked for music that felt contemplative but not sleepy, since players would spend thirty to forty minutes per session in the valley and return to the menu often. Two directions were drafted. Direction A used solo cello, processed piano, and long sustained tones in a modal scale. Direction B used a pulsing synth bass, a plucked melody, and a steady pulse under the texture.
The challenge
A panel of twenty-four players listened to 30-second clips of each direction on their own devices and rated each on atmosphere, engagement, and a willingness to keep playing. Direction A led on atmosphere clearly. Direction B led on the desire to continue, and several players wrote that A had started to feel heavy by the third time it looped. The developer's producer, Ines, read the split as a sign that the team should choose B. The composer was not sure the test had measured the right thing.
Analysis
A split panel is usually a sign that the test asked two questions at once. The brief said contemplative but not sleepy, which is a pair of goals that pull in different directions over time. Players move through a session in loops of exploration and rest, so a cue has to hold attention without asking for it. Direction A matched the contemplative half of the brief and scored well on atmosphere. Direction B matched the energy half and kept players in the menu longer. The useful question was which of the brief's two goals each direction served across the whole play session rather than a thirty-second clip.
The test itself had three weaknesses that the team needed to name before reading the scores. Clips were cut from the middle of each cue, so neither direction had a clear opening or resolution. Half the panel listened on laptop speakers, which flattened the low synth in direction B and made the cello in A sound thin. The first exposure also favored the more immediate piece. A revised test played each cue as a full loop of two minutes, randomized listening order, and asked for a fatigue rating at the third repeat. Only then could the scores say something about the music rather than the setup.
- In week oneShort clips, mixed devices, one exposure each.
- Early in week twoDirection A led atmosphere, B led continuation intent.
- Mid week twoLoops, randomized order, and a fatigue question added.
- Start of week threeFresh players, controlled devices, repeat listening measured.
- End of week threeBrief goals used to choose the revised direction.
Actions taken
- Ran the revised test with a fresh panel of twenty-four players, splitting them evenly between laptop and headphone listening and recording the playback device for each participant.
- Replaced the clip with a two-minute loop for each direction and asked participants to note the moment they wanted to stop listening, which gave a behavioral measure beside the ratings.
- Wrote the fatigue question as a single scale item at the third repeat, with a plain definition, so that players who liked a cue could still say it wore thin over time.
- Shared a short note with the game team before results came in, stating which brief goal each direction was meant to serve, so the decision would be judged against the brief.
Results
The second round ran with twenty-four fresh players, split evenly across devices. The low synth in direction B had been muffled for laptop listeners, which explained part of the first panel's preference for A. With full loops, B still drew more interest in continuing, but its fatigue ratings climbed by the third repeat while A held steady. The team revised the direction to keep A's sparse texture and moved B's pulse into a soft bass register under a plucked melody. The game team accepted the revision for the menu and asked for one more exploration test.
Questions for discussion
- What would you need to see in a panel result before you treated a preference for one direction as evidence about the brief's second goal?
- How would you design a listening test so that repeat fatigue and first-impression appeal are measured separately?
- When a client's stakeholders read a split result as a verdict, what would you put in the summary to keep the discussion on the brief?
Case Study 4
Adding a live session to a fixed-fee score
Corvid Field Recording, an eleven-person production house that supplies original music and sound design for regional museums and visitor centers.
The situation
Corvid Field Recording agreed to score an immersive hall at a regional history museum for a fixed fee of 14,000 dollars. The original scope covered eight cues, two rounds of revisions, and stems for the installation vendor. Halfway through the work, Ruth, the museum's director, asked for a live ensemble session and four additional cues for a second hall, and the vendor's schedule did not move. The exhibit opening was set for a date the museum had already announced to donors. Ruth's requests were reasonable from her side, but the brief had not described a second hall or live recording, and the fee had been based on the original count.
The challenge
Corvid could not absorb the extra work without either cutting the live session or missing the opening. Every additional cue would take roughly two working days in the studio, and the live session needed a booked venue, a conductor, six players, and a recording engineer, none of which were in the fixed fee. The team also had to protect the eight cues already approved, since the museum had signed off on them and the installation vendor had built timing around them.
Analysis
The brief described an audience, not a scope. Families and older visitors in a dim gallery needed music that stayed under the narration, and that need set a ceiling on how dense the new material could be. The second hall was a different room with different acoustics and a different visitor flow, so the four cues could not simply be copies of the first set. Each one had to be checked against its room and against the existing cues before it could be sent to the vendor. A cue count alone missed most of the real work, and the fee had been built on the count.
Feedback from the director arrived as a series of wishes, each reasonable in isolation. The useful move was to convert each wish into a line item with a cost in days and a consequence for the opening. A live session could be offered as a separate option with its own fee, while the four cues could be sequenced so that the first two were delivered on the original timeline and the last two waited for the second hall's measurements. Negotiation works better when the client sees which priority each addition displaces. The item it displaced was the second revision round on the original eight cues.
Actions taken
- Wrote a change order with separate line items for the four cues and the live session, each showing its fee, day count, and effect on the opening date.
- Proposed the live session as an optional add-on with a separate deposit, and made clear that the ensemble booking could not be held without one, since the venue required it.
- Sequenced the four new cues into two groups, delivering the first pair on the original timeline and holding the second pair until measurements from the second hall arrived, with a fixed date for each.
- Kept a running log of every revision round against the original eight cues, so that any further change was visibly priced against the fee rather than absorbed by the team quietly.
Results
Ruth approved the optional live session within the week, and the museum paid the deposit before the ensemble booking lapsed. The four additional cues went out in two groups. The first pair arrived on the original timeline, and the second pair followed once the second hall's measurements were in. The exhibit opened on the agreed date with the original eight cues unchanged. Because every revision round had been logged against the fee, the final invoice raised no disputes, and the museum asked for the same change-order format on its next exhibit.
Questions for discussion
- What would you have needed to hear in the first client meeting to price the second hall before the director asked for it?
- How would you explain a change order to a client who sees the additions as small edits to work already paid for?
- When a deadline is fixed by an announced event, which part of the scope would you protect first, and how would you tell the client?
Case Study 5
Scoring a community oral history with consent gaps
Juniper Hall Collective, a cooperative of four composers and two sound designers who write for community theater and public art installations.
The situation
Juniper Hall Collective received a commission from a neighborhood archive to build a sound piece from recorded interviews with long-term residents. The interviews were in a heritage language spoken by a small community, and the archive wanted the speech melodies, not the words, to carry the piece. The installation would run in a public square for three weeks and online afterward, with an audio description track for visitors who are blind or have low vision. The original recordings had been made by a visiting researcher whose consent forms covered academic use. The archive's director, Teodora, believed the families would agree once they heard the result.
The challenge
The collective's concern was not whether the piece was beautiful. The consent forms did not cover public performance, several speakers had died or could not be reached, and the audio description track risked describing private stories in English without the speakers' framing. A piece built on the material could expose people who never agreed to be heard in a square. The deadline was set, and the archive had already announced the installation to funders.
Analysis
The creative brief asked for a sound piece about belonging, but the audience for the piece included people who would never hear the original language and people who would experience it only through description. Accessibility here is not a separate track added at the end. The description had to convey the emotional shape of the speech melodies without translating private content, which changed what the piece could honestly say. In community work, the most striking material belongs to the speakers. The team needed a brief that named who the piece was for, and who it was for in a sense that included the people it represented.
Feedback came from two directions. The archive's director wanted the piece to feel welcoming, and a community advisor, Adaeze, who had translated interviews asked that no excerpt be heard without its context. The two comments pulled the edit in opposite directions. The team treated consent as an ongoing condition, not a one-time signature. They wrote a consent plan that separated the families who had agreed to public use from those who had not, and used only the first group's material until the others could be reached. Attribution followed the same logic: every source was named by the speaker's preferred form, and no clip ran without a credit line.
- Check consent status Sort each speaker before any audio enters a sketch.
- Confirm context Advisor writes framing notes in the heritage language.
- Describe with restraint Convey shape and emotion, skip unapproved personal detail.
- Audit and withdraw Log credits and honor withdrawals at each public showing.
Actions taken
- Drafted a consent plan that sorted each speaker into public, restricted, or unreachable, and shared it with the archive before any audio was edited into a sketch.
- Asked the community advisor to write the context notes for each excerpt in the heritage language first, so the English description followed the speaker's framing rather than the team's guess.
- Wrote an audio description script that conveyed the shape and emotion of each passage, including rising and falling melody, without repeating or translating personal details the speakers had not approved.
- Gave the archive a written attribution standard and a takedown route, so that any speaker who later withdrew could have their material removed from the square and the online version.
Results
The square installation went ahead with eleven of the nineteen speakers' recordings. Eight speakers were not reached before the opening, so their material was held back and replaced with field sound from the neighborhood, recorded with the owners' permission. The audio description track was reviewed by two listeners who are blind or have low vision, and they said the description conveyed structure without exposing private stories. A family that had agreed to public use asked for a credit in their own dialect, which the archive added to the online version. The archive now uses the consent plan as its template for future community commissions.
Questions for discussion
- Where in a project like this would you place the first consent conversation, and who should be present when it happens?
- How would you write an audio description for material whose meaning depends on a language most of the audience cannot understand?
- What would make you withdraw from a commission when a funder's deadline conflicts with the consent status of part of the material?
Learning Resources
Creative CASE STUDIES
China, India, United States, Indonesia, Pakistan, Nigeria, Brazil, Bangladesh, Russia, Japan, Mexico, Philippines, Vietnam, Egypt, Turkey, Germany, United Kingdom, France, Italy, Canada, South Korea, Australia, Spain.