Spire Light LogoSpire Lightspirelight.ai

Get paid to read aloud for speech technology

You come in, read sentences aloud in your normal voice, and get paid by bank transfer afterwards. The recordings help build technology for people who have lost the ability to speak.

We run the sessions for two companies. Altavo records in Berlin, in English and German. Acapela Group records in Denmark, in US English. Pick where you can get to and see what fits.

Berlin, Roskilde and LyngbyEnglish and German speakersPaid by bank transfer

Find your study

BerlinAltavo, at Mollstr. 32, 10249 Berlin
Sensor study, in English
130 EUR, two sessions of 2.5 hours
Sensorstudie, auf Deutsch
130 €, zwei Termine à 2,5 Stunden
DenmarkAcapela Group, at Roskilde University or DTU in Lyngby
Voice study, in English
500 DKK, one session of about 1.5 hours

Already booked?

Your confirmation email has a link to reschedule or cancel. If your booking needs a voice sample, it also has the link to record it.

Who we are

Spire Light is a small Danish company that collects speech recordings. When a company building voice technology needs a particular kind of speaker, say a native German speaker or someone with a Michigan accent, we find those people, book them in, run the sessions and pass the recordings on.

Our office is at DTU Skylab in Lyngby, north of Copenhagen, which is also one of the places you can record. There is no account to create and no app to install. You book a time on a study page, and from then on you deal with one named person from our team who you can write to or call.

Read more about us at spirelight.ai

A researcher at a desk in a sound-treated room, wearing the sensor unit used in the Altavo sessions, with recording hardware on the desk
The sensor hardware worn in the Altavo sessions. It sits on the outside of the cheek, and nothing is placed inside your mouth. The Acapela session uses a microphone and nothing else.

Who the recordings are for

Every study on this page is for one of two companies. They build different things, so their sessions ask different things of you.

Altavo

Altavo builds speech-assistive technology for people who have lost their voice after illness, injury or surgery. Their sensors sit on the outside of the cheek and pick up the muscle movement that goes with speaking, so a device can learn which words someone is trying to say.

That needs recordings of people speaking while wearing the sensors, which is why the sessions are long and filmed. Neither study will make a copy of your voice: your recordings are not used to create a synthetic voice that sounds like you, and they are never used to identify or verify you.

Two studies in Berlin, one in English and one in German

Acapela Group

Acapela Group makes the computer voices that read text aloud, in more than 30 languages. Many people who cannot speak use them every day in communication devices, and for them it matters a lot that the voice sounds like a person and not a machine.

Voices like that are built from recordings of many real people reading aloud. That is the Denmark study: 1.5 hours of everyday sentences, audio only. No voice is ever made to sound like you.

One study in Denmark, in US English, at Roskilde University or DTU in Lyngby

How it works

The steps are the same for every study. What differs is how long you come in for and what is recorded.

  1. 1

    Book a time

    Open the study page, pick a date that is still free and fill in a short form. You confirm your email with a code, and that is the booking made.

  2. 2

    Record a voice sample, if your booking needs one

    The Denmark study asks everyone for one, and the Berlin English study asks every US speaker. It takes a couple of minutes at home, on a computer or a phone. Your slot is held until someone has listened to it, and you hear back by email either way. The German study skips this step, and so do UK speakers.

  3. 3

    Come in and read

    In Berlin you come twice, for 2.5 hours each time, with sensors on your cheek. In Denmark you come once, for about 1.5 hours, and only your voice is recorded. Your contact person tells you where to go beforehand.

  4. 4

    Get paid

    The money goes to your bank account by transfer. 130 EUR once both Berlin sessions are done, or 500 DKK after the session in Denmark.

The studies

Grouped by where you record. Each card opens the study page, which has the full detail and the dates that are still open.

Berlin

Two sensor studies for Altavo, one read in English and one in German. They are the same study in two languages, so you take part in the one that matches the language you grew up speaking.

Altavo

Sensor study, in English

130 EUR
for two sessions
  • Two appointments of 2.5 hours each, on separate days.
  • You read sentences aloud while non-invasive sensors sit on the outside of your cheek.
  • The session is recorded in audio and video.
  • After booking, US speakers record a short voice sample, which takes about two minutes, and their slot is held until it has been reviewed. UK speakers record none.

Who it is for

You are a native US or UK English speaker, with any accent, and can attend both appointments in Berlin

Your voice will not be cloned.

Read more and book
Altavo

Sensorstudie, auf Deutsch

130 €
für beide Termine
  • Zwei Termine von je 2,5 Stunden, an verschiedenen Tagen.
  • Du liest Sätze vor, während nicht-invasive Sensoren außen an der Wange sitzen.
  • Die Aufnahmen erfolgen mit Bild und Ton.

Für wen

Du bist deutsche:r Muttersprachler:in und kannst zu beiden Terminen nach Berlin kommen

The same study as the English one, run in German.

Deine Stimme wird nicht geklont.

Mehr erfahren und buchen

Denmark

One voice study for Acapela Group, read in US English. You choose between the two sites when you book, and both run the same 1.5-hour session for the same pay.

Acapela Group

Voice study, in English

500 DKK
for one session
  • One session of about 1.5 hours, at Roskilde University or at DTU Skylab in Lyngby.
  • You read everyday sentences aloud in your ordinary speaking voice.
  • Audio only. Nothing is filmed, and you do not wear anything.
  • After booking you record a short voice sample at home, and your place is confirmed once Acapela have listened to it.
  • You sign a participant agreement before you record.

Who it is for

You are a native US English speaker, with any American accent, and can attend the appointment at Roskilde University or at DTU in Lyngby

No voice is made to sound like you.

Read more and book

What you are paid, and when

Sensor study, in English
130 EURAfter both sessions
Sensor study, in German
130 EURAfter both sessions
Voice study, in Denmark
500 DKKAfter the session

Payment is by bank transfer once you have recorded. The Berlin studies pay in euro and the Denmark study pays in Danish kroner.

In Berlin, the 130 EUR covers both sessions together.

It is not 130 EUR per appointment, and both appointments have to be completed before it is paid.

What happens to your recordings

Each company makes its own promises, and each study page states them in full. You agree to them when you book.

The Altavo sessions, in Berlin

  • The purpose is to match spoken words with cheek muscle movement, and that relationship is what the training uses.
  • Your voice will not be used to create a synthetic voice that sounds like you.
  • The sensors are non-invasive and placed externally on the cheek.
  • The session is recorded in audio and video.
  • You can ask questions before the session begins, and you can cancel or reschedule if you need to.

We are not building a voice clone of you.

The Acapela session, in Denmark

  • Your recordings are used to develop and train speech technology, including the synthetic voices used in speech-assistive devices: this is commercial development work.
  • No synthetic voice will be created to imitate or reproduce your individual voice.
  • Your recordings will never be used to identify or verify you.
  • The recordings are never sold or published as an audio dataset.
  • You sign a participant agreement before recording, and you keep a copy.
  • You can ask questions before the session begins.
  • You can cancel or reschedule if needed.

New voices, not a copy of yours.

The voice sample, where a study asks for one

  • It is a short recording of a few sentences, made at home on your own device, and it exists to check that your accent and your recording conditions suit the study.
  • It is listened to by the people reviewing samples for that study and used for that decision only.
  • Your slot is held rather than confirmed until it has been reviewed, and you are emailed either way.
  • If it is not accepted, the slot is released and nothing is charged. It is a judgement about accent fit, not about you.
  • The sample is not published, not sold, and not used to build a voice that sounds like you.

The sample decides your place, not your worth.

Talk to a person

Each study has one person responsible for it. Write or call before you book if something here is unclear.

EM

Emil·Berlin studies

emil@spirelight.net+45 53 34 40 23
GU

Gustav·Denmark study

gustav@spirelight.ai+45 21 22 29 38

Questions before you book

Which study is right for me?

Start with where you can get to and the language you grew up speaking. In Berlin, native German speakers book the German study, which has no accent requirement, and native US or UK English speakers book the English one. In Denmark there is one study, in US English. The Berlin English study takes any US or UK accent, and the Denmark study takes any American accent.

What is the difference between the Berlin and Denmark studies?

They are for different companies and ask different things. In Berlin you come twice for 2.5 hours, wear sensors on your cheek and are filmed as well as recorded, for 130 EUR. In Denmark you come once for about 1.5 hours and only your voice is recorded, for 500 DKK.

Do I need to live in Germany or Denmark?

No. Where you live is not a condition of any study. What matters is that you can get to the place you book: Mollstr. 32, 10249 Berlin for the Berlin studies, or Roskilde University or DTU Skylab in Lyngby for the Denmark study.

My accent is not in the list. Can I still take part?

Yes. Both English studies take any American accent: if yours is not in the list, pick Other when you book and type it in. Every US speaker records a short voice sample after booking. UK English speakers book the Berlin English study too, and are asked for neither an accent nor a voice sample. The Denmark study is for US English speakers only.

What is the voice sample, and why is my slot only held?

After you book the Denmark study, or the Berlin English study as a US speaker, you record a few sentences on your own device, which takes a couple of minutes. It is reviewed to confirm your accent suits the study, and your session is confirmed once it has been accepted. We email you either way. Recording it early is the surest way to keep the slot you picked.

Can I take part in more than one study?

Not in both Berlin studies. They are the same study run in two languages, so you take part in the one that matches your native language. The Denmark study is a separate project for a different company, so if you would like to do it as well, write to Gustav first and we will check.

I have already booked. How do I change it or record my sample?

Use the links in your confirmation email. One lets you reschedule or cancel, and if your booking needs a voice sample, another takes you to it. If you cannot find the email, write to the contact person for your study and they will sort it out.

Pick the study that fits you

Each study page has the full detail and the dates that are still open.

English sensor study, 130 EURDeutsche Sensorstudie, 130 €Voice study in Denmark, 500 DKK
Paid Speech Studies in Berlin and Denmark | Spire Light