Audio Expert

What you'll do

You transcribe polyphonic music into structured notation. The ad's own representative task is the clearest description of the bar: separate four simultaneous soprano, alto, tenor and bass lines out of a mono recording, and convert the complete performance into an exact MusicXML score.

That means identifying pitches, onsets, durations, rests and re-articulations, quantising events onto the right rhythmic grid, and applying correct key signatures, accidentals and enharmonic spelling. Python is part of the job, not a bonus: you use it to inspect audio, analyse spectral information and automate parts of the workflow. You then validate your own transcriptions by musical review, structured checks, or resynthesis.

The ad is explicit about what this is not: not general audio production, and not speech transcription. The engagement is contractor work at roughly 15 hours a week, and you pick your own hours and days, weekends included.

What you need

• Strong Python programming ability
• Advanced ear training and polyphonic transcription ability
• A solid grasp of harmony, rhythm, voice leading and standard notation
• Experience with SATB music, choral arranging, score preparation, or a closely related form of polyphonic transcription

librosa, NumPy, SciPy, music21, Essentia, pretty_midi and MuseScore are named as relevant tools, with no specific one mandatory if your experience transfers.

Formal education and years of professional experience are explicitly not strict requirements. Performers, arrangers, composers, choir musicians, researchers and independent transcription specialists are all named as welcome if they can demonstrate the skills; the wording is that relevant expertise matters more than years of experience.

The job card lists a single required skill, MusicXML, but the body of the ad asks for considerably more. Read the requirements above as the actual bar.

What it pays

$50–100/hour, published in both the card and the ad body. The compensation section then qualifies it: pay is output-based, per task that meets the project specification, not per hour at the desk. The time a task takes is stated to vary with the expert's experience and workflow, so the hourly figure describes a fast transcriber's outcome rather than a rate you are owed for time spent. On work this painstaking, that gap matters more than it would on most listings.

The ad also states a minimum weekly submission requirement and then omits the number: it reads "Experts must submit a minimum of tasks per week", with the figure missing from the sentence. Ask what it is before accepting.

How you get hired

The ad names three stages: apply and answer the screening questions, complete an AI interview of about 30 minutes, then pass hiring manager review. There is no separate technical assessment. None of that time is paid.

Roles are stated to fill typically within 48 hours, with first tasks 24 to 48 hours after onboarding.

Worth knowing

Good:

• Explicitly global and fully remote
• No degree and no minimum years, stated in the ad rather than left implied
• You choose your hours and days, weekends included
• $50–100/hour for musicianship that is otherwise hard to monetise remotely
• Ten openings, so this is not a single seat
• A three-stage process with no separate technical assessment

Less good:

• Output-based pay against an advertised hourly band, and detailed polyphonic transcription is slow work; the gap between the two is real
• A weekly submission minimum applies, but the number is missing from the ad
• The skill combination is genuinely rare: advanced polyphonic ear plus working Python. Most strong transcribers lack one of them
• SATB from a mono mix is one of the harder transcription tasks there is, and the rate is not obviously priced for it
• Contractor engagement: no benefits, no notice period, your own tax to manage
• Expected to start 24–48 hours after onboarding
• Neither the customer nor the project length is named

About this listing

Posted by micro1 on its own jobs portal, confirmed open on 7 September 2026. The $50–100/hour band, the representative task, the tool list and every requirement above are its own published text for this listing.

No language requirement is stated, so the language field is left empty rather than assumed, though the work involves standard Western notation conventions. We could not verify the per-task rate or the weekly submission minimum; neither is published. The ad calls the role global; platform-wide eligibility rules still apply and are on the micro1 page.