How to use accents when you convert text to speech?

Do you think the most important element of choosing a computer text to speech voice is how realistic the voice is? Actually, it’s the accent. The wrong (or right) accent can kill or amplify your message.

Do you know how many accents there are in English? 160! German? 250! In the Netherlands, residents of neighboring towns may find it difficult to understand each other. But they have only 5 dialects.

Sometimes accent differences are subtle. But they can still complicate communication. You have probably found yourself in a situation where it is impossible to understand a person because of his accent, or perhaps you have not been understood because of your accent or pronunciation.

Sometimes this results in a humorous situation, as in this video, where the famous tennis player could not understand the journalist's question.

However, in this guide, we address the issues of correct voice acting, and our goal is to make sure that after hearing your message, your audience correctly understands you.

When you are recording a message for your audience, the accent is very important. If your audience senses a foreign accent, they may very well ignore the message, unless you have selected that accent for a reason. It is very important to choose the type of accent that will be close and authoritative for your listeners.

It is no coincidence that the developers of voice synthesizers are constantly adding new accents to languages. On the Kukarella website, you will find several options for English: American, British, Canadian, Australian, Irish, Indian. And you can also convert your text using voices from other languages to get a sense of how it sounds with that accent.

What the accent says

Wherever you live today, you’ve probably received calls informing you that you’ve committed a tax crime and you urgently need to call the indicated number and settle the issue or face a large fine or even imprisonment. Sound familiar? Over the past few years, scammers have called perhaps every inhabitant of the planet over 16.


Many people were caught up in the scam and ended up paying large sums of money. What convinced them to that they were really receiving calls from the tax office?

The text of these messages is simple, formal and short. The phrases are similar to texts from a TV show about court hearings, and in principle, they inspire confidence. Of course, the terminology is vague, we don't understand a lot, but in general, the message is perceived as authoritative.

But there is another nuance of these messages. Guess which one? Yes, accent.

Have you paid attention to what accent was used in the audio messages received on your phone? Probably not, because the accent was “normal.” In other words, typical for the country or region where you live.

For calls to America and Canada, scammers use a middle American accent, in the UK they use a British English accent, and in Australia they use Australian voices only. It doesn’t matter that if you make the call after that initial message you’ll be talking to a person with a strong accent. You know that lots of companies have their call centers in India. What matters is that the first message has exactly the same accent that is used in your territory.

When creating these messages, scammers know that accent is a radar that quickly identifies who's theirs and who's a stranger.

How people react to accent

The British often frown when they hear American English. Although, it would be more correct to say they used to frown. As a global financial and cultural center, America exports its blockbuster movies and the songs of American artists fill the radio waves. So it's hard to laugh at an accent when a significant part of your life is permeated by it.

It's another matter when you are within your own cultural circle. Your own community. When you take a subway train or go to a local shop or bar. Remember ‘Inglourious Basterds’ German Accent Scene?

Voice not only carries information, but it also creates an atmosphere. If your potential audience is not the entire global network of the Internet, and you address a certain group of people and expect to inspire their trust, then the accent familiar to this area will assist you. It will inspire a connection with a person who will feel “at home” in this atmosphere and will consider that the creators of the message thought about them and their needs.

Example, in Vienna, the issue is not only what languages to use (English and Austrian German), but what accent. Angela Schneider has been the voice of the transport announcements since 2012. Schneider was chosen, in part, for her Viennese accent – had it been German as spoken in Germany, “people in Vienna would go crazy”, says Dominik Gries from Wiener Linien, the company running most of Vienna’s public transit network in an interview with The Guardian.

“We want to give people a sort of sense of being at home in our public transport system – and that’s hearing a voice that sounds like she’s grown up with you, not in Hamburg.”

Unusually for a German-speaking country, Schneider also does the English announcements, to maintain that “feeling that we are in Austria” (while still stressing the need for correct English and being “welcoming hosts”). The effects of the speaker's dialect on multimedia learning was examined by researchers, who found that learners who listened to the multimedia presentation with a standard-accented human voice (American English) had better transfer performances and gave higher social ratings for the speaker than did learners who listened to the multimedia presentation with a Russian accent.


Ahn (2010) compared four voices ranging from slight to heavy accents and found that they did not differently affect learning outcomes. However, it was noted that the learning outcomes were affected only when the learners had indicated beforehand that they did not like a certain accent. This observation indicates that the perceptions of learners regarding the appeal of social cues can actually affect the learning process.

Accent and language norms

Along with the accent, there are also the language norms adopted by a group of people or within a given area. In large metropolitan areas and cultural capitals, speech literacy and generally accepted language norms are still considered the most important criterion.

In bustling London, for example, passengers prefer short, crisp announcements over conventional designs. They understand simple “149” over “route 149” better. The transportation authorities explain that, while it is not their intention to promote poor grammar, their top priority is brevity and clarity.

Emma Hignett, the voice of London buses, was hired after being judged on a range of criteria: tone, clarity, image, appropriateness for the bus operating environment, and the type of information the voice is required to deliver. She admits that she doesn’t think about accent when recording, and more about “putting a smile into it, so it sounds warmer”.

In Vienna, where the pace of life is slower, the London rule does not apply. They once tried to shorten the announcement about the next stop, but passengers complained that they felt that it was not good German, even though it's a bit shorter.

Even countries as close as Sweden and Norway have different priorities. In Sweden, names are pronounced as they are locally, while in Norway they are spoken as a foreigner would probably read them.

How to choose the right accent for your message

Try to voice a fragment of your message by choosing different accents on the Kukarella website. Here, in addition to the American version of English, you will find British, Irish, Canadian, Australian, and Indian. But you will also find different versions of other languages. When you listen to the message in several versions, you will probably want to tweak something. So, for example, writers, when they hear their text, they notice grammatical errors in the text and some oversights that they would otherwise have missed. You can continue experimenting. Try using voices from other languages. For example, enter text in English, French, or German and try voices from Spain, Russia, or Japan. It's a fun experiment. It can cheer you up, but it can also help you better understand your message and even see hidden meanings. Consider Trevor Noah, who is a master of the nuances in different accents:

Do you wonder how to pick the right accent for your message? Here are three recommendations to help you get started:

  • Define your audience: In which country and in which city do they live? What language do they speak?

  • Find out which accent is popular in everyday life, and which one is used for business communication (it can be one accent, or it can be two or more);

  • Decide if you want your message to sound official or casual.

Ask yourself these simple questions and it is likely that the answers will give you a new perspective on your audience and help you find common ground more easily.