Independent career guide

Voice Recording for AI Training: Scripts, Cars and Studios

What micro1 audio recording listings describe: scripted sentences, in-car sounds, studio-grade samples — and what to settle before you record.

On this page

For native and near-native speakers who can produce clean recordings on request and want a route through, in order, from “which listing is mine” to the interview. The steps report what the full listing descriptions say, and mark the blog’s own practice material where it starts.

Step 1 — Find out which kind of recording a listing wants

On the date in Sources, the audio listings on micro1’s portal were written per locale — English (US), English (UK), English (AUS), Spanish (Mexico), French (France), French (Canada) and others — and their descriptions asked for three different things.

Reading sentences, with staged background sound. The voice recorder listings (a language plus English) describe recording yourself reading provided sentences, in a stationary car or a quiet office, and creating specific background sounds on request — seatbelt clicks, a door, an indicator. They say the car must be stationary. Two of them add listening to clips and transcribing or annotating them. They ask for native-level proficiency in the language and a stated minimum level of English.

Studio-grade voice samples. The audio recording expert listings for several languages ask for professional work: years of professional recording and editing experience, professional equipment and software, edited and cleaned files, and a sample recorded on your best equipment as part of the selection. The English ones ask for experience in audio recording, engineering or voice-over and for natural delivery with post-recording cleanup.

Recording plus transcription and tagging. The French and Mexican Spanish listings combine the two: record audio in the language (and, for French, in English), transcribe it, and annotate the files with metadata.

One title misleads. The listing titled Spanish Audio Recording Expert (Spain) has no recording task: its description is correcting machine-generated transcriptions and adding metadata. That is transcription work, and it is outside this page.

Locale is the job. A French (France) listing and a French (Canada) listing are different listings, and the same goes for the English and Spanish ones. Apply to the locale you actually speak, and write it down precisely.

Browse all expert opportunities

Opens micro1. You apply and are vetted by micro1, not by this site.

Step 2 — Check the excluded countries against your residence

micro1 states it does not hire residents of: Afghanistan, Belarus, China, Cuba, Democratic Republic of the Congo, Hong Kong, Iran, Iraq, Libya, Macao, Myanmar, North Korea, the Russian Federation, Somalia, South Sudan, Sudan, Syria, Venezuela, Ukraine and Yemen. Snapshot from the date in Sources — check the current list before applying.

This interacts with Step 1: a language may be sought while a particular country of residence is excluded. Check both, separately.

Step 3 — Decide on retention, reuse and payment before you record

Do this before you spend time on setup: if the answers are not acceptable to you, the later steps are wasted.

  • Your voice is the data. The descriptions mention privacy, storage and intended-use requirements without spelling them out, and one asks contributors to acknowledge that recordings may be retained for ongoing research and system improvement. Ask to see those terms: what the recordings are used for, how long they are kept, and whether you can withdraw them.
  • Ask whether your voice can be used to imitate you. The descriptions do not say. If the project’s terms do not answer it clearly, that is a reason to walk away.
  • Be wary of anyone who asks you to pay for equipment, training or a “kit” before you start. That is this guide’s advice; apply through official listings.

On payment: the cards on the opportunities page show an hourly range. None of the voice recording descriptions we read says how pay is calculated — by hour worked, by minute of accepted audio, by sentence or by task — or whether there is a minimum. The unit of payment is unknown, and the range on a card does not settle it. For recording, the unit matters: a session includes setup, retakes and file handling as well as the audio itself, so pay per accepted minute and pay per hour worked can be far apart. Before you agree, ask in writing what the unit of payment is, who accepts a recording and on what criteria, whether rejected takes are paid or can be redone, and what volume is expected each week. Do not assume the hours you spend are the hours you are paid for until the contract says so.

You will not find a rate here; we left it out. A different point: micro1’s referral terms forbid anyone promoting the program from making promises about compensation or acceptance.

Step 4 — Match your equipment to the description

The descriptions differ on this, so read the one you are applying to:

  • Sentence-reading listings specify the place — a stationary car or a quiet office — more than the gear.
  • Audio recording expert listings ask for professional-grade equipment and editing software, and several want a sample made with it.
  • The French listings mention instructions on recording environment, equipment settings and file naming.

This guide suggests starting with changes that cost nothing, before considering purchases. This is a preparation sequence, not a ranking of acoustic effects:

  1. The room. Use the quietest room you have when the brief asks for quiet recording.
  2. Consistent mic position. Fix the distance and do not move between sessions.
  3. The microphone the listing implies. Do not buy anything until a description or a recruiter tells you what is needed.
  4. Headphones for monitoring, so you hear problems while recording instead of afterward.

Step 5 — Rehearse five recording briefs

The examples below are illustrative exercises created for this guide, not published micro1 assignments or interview questions. They show the kind of judgment the work needs; they are not a preview of what a project will hand you.

Each brief names the kind of recording, from Step 1, that it is practice for. Every recording is of your own voice, and the sentences are ones you write yourself or text you are free to reuse. Nothing here practices the studio-grade listings beyond what you already do professionally; for those, the description asks for a sample on your best equipment.

1. A long list of sentences. Practice for: reading provided sentences. Read a hundred short sentences, one file each. Listen to the first ten and the last ten side by side: pace and delivery should match.

2. The same sentences in a parked car. Practice for: the same listings, which name a stationary car as a recording place. Engine off, windows closed. Listen back for what the cabin does to your voice compared with the room.

3. A sound on cue. Practice for: creating specific background sounds on request. Read a sentence, then make one specified sound — a seatbelt click, a door — without talking over it. Stay parked.

4. Two languages, one setup. Practice for: the listings that pair a language with English. Record the same passage in your language and in English without changing microphone or position.

5. Record, then transcribe yourself. Practice for: recording plus transcription and tagging. Record a minute of speech and transcribe it exactly as spoken. Mark every place where you wanted to tidy what you said.

Step 6 — Run a self-check on every batch

A worked checklist, written by this blog as practice — not a micro1 acceptance standard. Each row is judged on playback and its two columns exclude each other. A row you have not listened for is at Fail until you do. A file at Fail gets the action its cell names: some failures mean recording again, a naming slip only means renaming.

CheckFailGood
LevelsClipping is audible, or hiss is audible under the voice. Redo: record the file again after adjusting the levelNo clipping, and no audible hiss under the voice
RoomEcho is audible. Redo: record the file again after treating the roomNo echo is audible
NoiseA background sound the brief did not ask for is audible, or a sound it asked for is missing. Redo: record the file againEvery sound the brief asked for is audible, and no other background sound is
Mouth noiseLip smacks or clicks are audible. Correct the file if the brief allows editing; otherwise Redo itNone are audible
ConsistencyAt least one of the last ten files differs audibly from the first ten in distance or delivery. Redo the files that drifted, not the batch: listen back from the end to find where the drift beganNone of the last ten files differs audibly from the first ten
NamingA file name or order differs from the convention asked for. Correct the name; nothing needs recording againEvery file follows the convention asked for

Step 7 — Take the interview from your recording room

Vetting includes an AI interview (step two of the four micro1 lists, between applying and onboarding), and micro1 states that human recruiters also review profiles and match experts to opportunities. The professional listings add a recorded sample.

  1. Do the interview from the room you will record in, so you hear how that room sounds on a call.
  2. State the locale you wrote down in Step 1, so a recruiter knows which listings you match.
  3. State your English level if you are applying to a bilingual listing.
  4. Have your equipment list ready, and a sample if the description asks for one.

Will a referral make my sample or interview count for more? Nothing we found suggests so. The vetting is done by micro1, without this site, and micro1’s referral terms forbid anyone promoting the program from promising placement, compensation or acceptance. A promise that a referral helps is a warning sign.

Browse all expert opportunities

Opens micro1. You apply and are vetted by micro1, not by this site.

Sources

Programme terms, eligibility rules and reference pages change. Each claim above is dated; verify it on the page it cites before acting on it.