Free Text-To-Speech and Text-to-MP3 for Cantonese

ttsMP3 converts your Cantonese text to speech for free: type or paste up to 3,000 characters a day, listen online and download the audio as MP3 – no account, no sign-up. Choose from 1 Cantonese Amazon Polly voice with a native accent. You may use the MP3 files in commercial and non-commercial projects such as YouTube videos, presentations, e-learning and apps.

Today's use: 0 / 3,000 characters

Hint: If you finish a sentence, leave a space after the dot before the next one starts for better pronunciation.

How to convert Cantonese text to MP3

  1. Type or paste your text into the box above (up to 3,000 characters per day for free).
  2. Choose a voice from the list.
  3. Click Read to listen in your browser, or Download as MP3 to save the audio file.

No account, no app and no AWS account are needed. Longer texts are processed in the background; please keep the page open until the audio is ready.

Free limit and premium plans

The free converter allows 3,000 characters per day (about 375 words). Your usage today is shown below the converter and resets every day.

Need more? Premium plans range from a 24-hour pass to monthly and yearly plans with up to 10 million characters. Every yearly plan includes access to the text-to-speech API.

Commercial use of the MP3 files

You may use the MP3 files created with ttsMP3 in commercial and non-commercial projects – for example YouTube voiceovers, podcasts, presentations, e-learning courses, websites and apps. This applies to free and paid usage, and you do not have to credit ttsMP3 (a mention is always appreciated). See commercial use of text to speech for details.

Cantonese voices

For Cantonese you can choose between these voices: Hiujin. All of them are Amazon Polly voices and pronounce your text with a native Cantonese accent. In total ttsMP3 offers 74 regular voices in 41 languages and accents.

Pauses, emphasis, speed and pitch (SSML)

You can control the speech with SSML tags directly in your text. Here are the most useful ones:

Add a break

Mary had a little lamb <break time="1s"/> Whose fleece was white as snow.

Emphasizing words

I already told you I <emphasis level="strong">really like</emphasis> that person.

Speed

For dramatic purposes, you might wish to <prosody rate="slow">slow down the speaking rate of your text.</prosody>
Or if you are in a hurry <prosody rate="fast">you may want to speed it up a bit.</prosody>

Pitch

Do you like synthesized speech <prosody pitch="high">with a pitch that is higher than normal?</prosody>
Or do you prefer your speech <prosody pitch="-20%">with a somewhat lower pitch?</prosody>

Whisper

<amazon:effect name="whispered">If you make any noise,</amazon:effect> she said, <amazon:effect name="whispered">they will hear us.</amazon:effect>

Conversations

You can switch between speakers within one text:

[speaker:Brian] Hello Emma
[speaker:Emma] Hey Brian
[speaker:Brian] How are you doing?
[speaker:Emma] I am fine. May I invite you to a cup of tea?

Please remove diacritical signs from the speaker names: Léa = Lea, Penélope = Penelope.

Need more effects? See our SSML guide with examples or the Amazon SSML tags for Amazon Polly.

Facts about the Cantonese language:

Cantonese is a Chinese language spoken by about 85 million people, mainly in the Guangdong province of southern China, in Hong Kong and Macau, and in Chinese communities around the world, from Malaysia to Canada and the United States. In Hong Kong and Macau it is the language of daily life, schools, business and the media.

Cantonese and Mandarin are written with Chinese characters but are not mutually intelligible when spoken. Hong Kong and Macau use traditional characters, and written Cantonese has some characters of its own for words that do not exist in standard written Chinese, such as 係 (to be), 嘅 (the possessive particle) and 唔 (not).

Cantonese is a tonal language with six main tones, more than the four tones of Mandarin. The same syllable can have completely different meanings depending on the tone, so listening carefully is essential for learners. Cantonese has also kept final consonants such as p, t and k that disappeared in Mandarin, which is why many classical Chinese poems still rhyme better when read in Cantonese.

The most common romanisation system for learners is Jyutping, developed by the Linguistic Society of Hong Kong. Everyday Cantonese is known for its lively expressions and its many final particles, such as 啦, 喎 and 囉, which add emotion and nuance to a sentence.

Cantonese is the language of Hong Kong cinema, Cantopop and a large part of the Chinese diaspora. A Cantonese voice helps learners practise the tones and is the right choice for content aimed at listeners in Hong Kong, Macau and Guangdong.

Supported voice languages:

Arabic
Australian English
Austrian German
Belgian Dutch
Belgian French
Brazilian Portuguese
British English
Canadian French
Cantonese
Castilian Spanish
Catalan
Chinese Mandarin
Czech
Danish
Dutch
Finnish
French
German
Gulf Arabic
Icelandic
Indian English
Irish English
Italian
Japanese
Korean
Mexican Spanish
New Zealand English
Norwegian
Polish
Portuguese
Romanian
Russian
Singaporean English
South African English
Swedish
Swiss Standard German
Turkish
US English
US Spanish
Welsh
Welsh English

Frequently asked questions

Is ttsMP3 really free?

Yes. You can convert up to 3,000 characters per day for free, listen online and download the result as MP3. You do not need an account or a sign-up.

Can I use the MP3 files commercially?

Yes. MP3 files created with ttsMP3 may be used in commercial and non-commercial projects such as YouTube videos, presentations, apps, websites and e-learning, for free and paid usage. A credit is not required.

Which voices are used?

The regular voices are Amazon Polly voices: 74 voices in 41 languages and accents. You can use them directly in your browser, without an AWS account.

How do I add pauses?

Insert an SSML break tag into your text, for example <break time="1s"/>. Emphasis, speed, pitch and whispering work the same way.

What if I need more than 3,000 characters?

Paid plans range from a 24-hour pass to yearly plans with up to 10 million characters. Every yearly plan includes API access.

More questions and answers in our FAQ

ConvertSpeech Need the opposite — Speech to Text?
Upload any audio or video and get an accurate transcript in minutes.
Free to try on our sister site ConvertSpeech.com →

Current Limit: ~375 words or 3,000 characters / day | Powered by AWS Polly

Get in touch? Mail us: [email protected]

Need to convert more text to speech? Register here for a 24 hour premium access.