Signing Up for the CereVoice Cloud

To sign up for the CereVoice Cloud:

1. You must first be a registered CereProc website user. To become a CereProc user fill in the User Registration Form
2. Once you have been registered as a CereProc user, navigate to the CereVoice Cloud Registration page in your account by clicking on the CereCloud tab. Then you will be automatically registered for this service.

Migrating CereVoice Cloud Account from Version 1

If you have been using CereVoice Cloud v1 and would like to migrate your account to be able to use CereVoice Cloud v2, navigate to the CereCloud tab on your account page on the Cereproc website. The migration then will be done automatically.

CereVoice Cloud Credits

CereVoice Cloud credit is required to generate TTS output, 1 credit = 1 character of text input. A free tier of credit is provided to all registered CereVoice Cloud users, providing 10,000 characters of text input per month. Additional credit can be purchased from the CereProc Store.

• CereVoice Cloud 1M Credit - £12.99 (1,000,000 credits)

Credit Threshold Alarm

You can set up a notification email once the amount of credits on your account go under the credit threshold. To set up the credit threshold use /credit-threshold endpoint. The email notifications will occur daily.

Cerevoice Cloud API Authorization

In order to use any of the API endpoints, an access token must be provided in the Authorization header as Bearer token. The token can be acquired using /auth endpoint by providing your account credentials (email address and password which are used to access our website) in Authorization header as basic authentication type.

Using Multiple Voices

The CereVoice Cloud allows the use of multiple voices during synthesis. By placing tags around the input text, the user is able to assign specific text to specific voices. For example, if the following call was sent to the CereVoice Cloud:

curl -X POST "{url}speak?voice=Stuart" -H  "accept: application/json" -H  "Content-Type: text/xml" -H  "Authorization: Bearer <access_token>" -d "<doc>Hello. My name is Stuart. This is my CereProc sister, Heather.<voice name='Heather'>Hello, my name is Heather.</voice></doc>"

The text - Hello, my name is Heather. is spoken by Heather, whilst the rest is spoken by Stuart.

CereProc Tag Set

CereProc has implemented additional TTS functionality that is not part of the SSML specification.

Variant Tags

The variant tag allows the user to request a different version of the synthesis for a particular section of speech. This is a very useful tag that can be used to make sections of speech sound more appropriate, or to vary otherwise repetitive content. The variant number can be increased to produce different versions of the speech. The original version is equivalent to variant 0. For example, to change the version of the word test in This is a test sentence, use:

<s>
    This is a <usel variant="1">test</usel> sentence.
</s>

Setting variant="2" produces another different version, and so on. The variant tag can be used to produce a bespoke rendering of a particular piece of speech. For example, an often-used speech prompt could be tuned to give a different rendering if desired. Please note that the variant tag should mainly be used for creating static prompts(i.e. audio files). The effect of the variant number is different between voices, and may also change when a new version of the same voice is produced (this is because the underlying speech engine is being constantly improved, and the default rendering may change).

Vocal Gestures

Non-speech sounds, such as laughter and coughing, can be inserted into the output speech. The <spurt> tag isused with an audio attribute to select a vocal gesture to include in the synthesis output, for example:

<speak>
    <spurt audio="g0001_004">cough</spurt>, excuse me, <spurt audio="g0001_018">err</spurt>, hello.
</speak>

The <spurt> tag cannot be empty, however the text content of the tag is not read, it is replaced by the gesture.
See the List of vocal gesture IDs for the full list of available gestures.

Emotion Tags

Available in voices with emotional support (for example Adam, Caitlin, Heather, Isabella, Jack, Jess, Katherine, Kirsty, Laura, Sarah, Stuart, Suzanne, William).

Happy Emotion Tag

For example:

<s>
    Today, <voice emotion='happy'>the sun is shining.</voice>
</s>

Sad Emotion Tag

<s>
    The outbreak <voice emotion='sad'>cast a shadow</voice> over the former Victorian holiday resort.
</s>

Calm Emotion Tag

<s>
    The beautiful gardens have been restored to all their <voice emotion='calm'>eccentric Victorian splendour.</voice>
</s>

Cross Emotion Tag

<s>
    When people leave a tip they want to know it will <voice emotion='cross'> not be used</voice> to make up the minimum wage.
</s>

Support

CereProc offers support via email. There are two methods of contacting CereProc Support:

• Support Request: The fastest way to contact CereProc Support is via a support request. First log in to the CereProc website. Registered users can then access the support request form. Please select the appropriate product from the list and submit the support request.
• Direct Email: CereProc support can be emailed at support@cereproc.com. However, queries sent to this address may take longer to reach the appropriate technical support representative than requests sent using the support request form.

List of vocal gesture IDs

These IDs can be used to insert a 'vocal gesture' (non-speech sound) into synthesis.
Note that gesture g0001_035 is available in Scottish voices only.