Setting Up AI Voice Links In Myriad Playout - requires Myriad Playout v6.3.1 or higher

Modified on Mon, 30 Sep at 11:52 AM

Overview & Pre-requisites

This article will guide you through the process of setting up Myriad Playout (6.3 and higher) to use AI Voice Links to add automated vocal elements to your scheduled station output.

To use AI Voice Links you will need the following:

Myriad Playout v6.3 or higher.
Local Windows voice (built in)

Optional extra elements to enhance the service:

3rd Party AI Text Service (OpenAI (ChatGPT) or AzureOpenAI ) - This can be used to add extra variation to your AI Voice Link scripts.
3rd Party AI Text To Speech (TTS) Service (Azure AI Voices, Google TTS Service or ElevenLabs) - This is used to generate high quality, natural sounding voice links.

For more information on these services, see - AI Text & Speech Synthesis Services Supported By AI Voice Links In Myriad 6 Playout

Setting Up AI Voice Links In Myriad Playout

Once you have decided on which optional 3rd party services you wish to use (if any), setup your accounts and received your API keys and/or config files, it is time to setup Myriad Playout to use them to generate AI Voice Links.

Firstly, make sure you are on Myriad Playout 6.3 or higher.
Login with appropriate access rights to edit settings.
Go to Station Menu > Station Settings.
Switch to the AI Voice Links tab.

AI Voice Link script options:

The top section allows you to set some basic information about your station as well as connect with an optional AI Script service such as OpenAI (Chat GPT) or Azure OpenAI.

Station Name - Set the name you want to use for your Station in AI Voice Links.
Station Tagline - Set a tag line to use in AI Voice Links.
Station Phrases - This advanced settings allows you to change the way the internal Myriad script AI works to generate AI Voice Links. You should only adjust these settings if asked to by Broadcast Radio technical support.
Chat GPT Script Engine - Myriad will generate scripts internally but if you would like to enhance scripts further, you can set access to a 3rd party AI Script service - OpenAI (ChatGPT) or Azure OpenAI - to provide infinite variation. Use the drop down list to select your preferred service.

The Broadcast Radio recommendation is currently to use Azure Open AI and create a "Deployment" that uses "GPT-4 Large" as this has given the best results in testing. It is also signifcantly faster than ChatGPT from OpenAI.

Azure OpenAI options:

Script Creativity - Use the slider to set how 'creative' the AI service can be. Setting this to 'More Precise' will make sure the re-written script sticks more closely to the original script. Setting more towards the 'More Creative' side will allow the AI to be more creative with the script (but may introduce less accuracy!). This is applied to all AI Voice Links.
Presentation Style Keywords - These words are passed to the AI along with the original script, to allow the service to determine the 'style' of the returned script. For example, if you added 'pirate' to this list, the returned script may include references like 'shiver my timbers'! Please note these are the default styles, you can add extra style keywords on each individual AI Voice Link.
Download The Latest GPT Optimisations - Click this button to download the latest optimised list of keywords from the Broadcast Radio cloud servers.
Azure OpenAI Endpoint - Paste in the Azure OpenAI endpoint URL supplied by Azure when you created your Azure OpenAI account.
Azure OpenAI API Key - Paste in the API key provided by Azure.
Azure Deployment Name - Paste or type in the Deployment Name you used when creating your Azure OpenAPI GPT deployment.
Manage Deployments - This opens the Azure portal so you can check your settings and get the connection details.

OpenAI (ChatGPT) options:

Script Creativity - Use the slider to set how 'creative' the AI service can be. Setting this to 'More Precise' will make sure the re-written script sticks more closely to the original script. Setting more towards the 'More Creative' side will allow the AI to be more creative with the script (but may introduce less accuracy!). This is applied to all AI Voice Links.
Presentation Style Keywords - These words are passed to the AI along with the original script, to allow the service to determine the 'style' of the returned script. For example, if you added 'pirate' to this list, the returned script may include references like 'shiver my timbers'! Please note these are the default styles, you can add extra style keywords on each individual AI Voice Link.
Download The Latest GPT Optimisations - Click this button to download the latest optimised list of keywords from the Broadcast Radio cloud servers.
OpenAI API Key - Paste in the API key provided by OpenAI.

AI Voice Link options:

The middle section allows you to set the options for generating the audio for the AI Voice Links.

Voice Source - This allows you to select the service you want to use to generate your AI Voice Links.

Disabled - Myriad will not generate an AI Voice Links.
Local - Myriad will use the built voice service in Windows.
Azure - Myriad will use Azure AI Voices service to generate AI Voice Links.
ElevenLabs - Myriad will use ElevenLabs to generate AI Voice Links.
Google - Myriad will use Google TTS services to generate AI Voice Links.

"Local" options:

When the Voice Source is set to Local, Myriad will list all the Windows built in Text To Speech services plus any additional voices you have installed.

More Information About Voices - This opens a help article listing supported voice services.
Language - Select the language you want to use.
Voice - Select the installed voice you want to use to generate AI Voice Links.

"Azure" options

Setting the Voice Source as Azure allows you to set the connection settings for your Azure AI Voice account as well as select the default Language and Voices.

More Information About Voices - This opens a help article listing supported voice services.
Azure Subscription Key - Paste in the Subscription Key for your Azure AI Voice account.
Azure Region - Set the region you setup your Azure AI Voices subscription to use (example: westeurope).
Language - Select the language you want to use for the default voice when generating AI Voice Links. Please note this will dictate the available voices as if you selected English US then only US voices will be listed.
Voice - Select the voice you want to use for all default AI Voice Links. Please note that you can set different voices on individual AI Voice Links to override this setting. Azure offers a high number of voice options. We recommend trying several options until you find the one that suits your station. Also note that some voices are more costly than others so please check with your subscription pricing when selecting the default voice for your station.
Update The List Of Voices From Azure - Clicking this option will force Myriad Playout to update the list of available voices from Azure.

"ElevenLabs" options:

ElevenLabs offers a dedicated AI voice generation service including lots of high quality voices as well as the option to create new 'cloned' voices from a short recording of a source voice. They also offer a much simpler signup and process and clear pricing packages.

More Information About Voices - This opens a help article listing supported voice services.
ElevenLabs API Key - Paste in the Subscription API Key for your ElevenLabs account.
Language - Select the language you want to use for the default voice when generating AI Voice Links. Please note this will dictate the available voices as if you selected English US then only US voices will be listed.
Voice - Select the voice you want to use for all default AI Voice Links. Please note that you can set different voices on individual AI Voice Links to override this setting. ElevenLabs offers a high number of voice options. We recommend trying several options until you find the one that suits your station.
Download The List Of Voices From ElevenLabs - Clicking this option will force Myriad Playout to update the list of available voices from ElevenLabs.

"Google" options:

Google Text To Speech (TTS) services offers a range of high quality voices that can be used by Myriad to generate AI Voice Links. Unlike the other services in this list, Google TTS services provides a configuration file that you can download from your Google TTS portal and add to Myriad to setup the connection credentials.

Google Credentials - Click on the Import A Google Credentials File button to browse for your Google TTS credentials file. Once selected, the green tick indicates that the connection credentials have been applied.
Language - Select the language you want to use for the default voice when generating AI Voice Links. Please note this will dictate the available voices as if you selected English US then only US voices will be listed.
Voice - Select the voice you want to use for all default AI Voice Links. Please note that you can set different voices on individual AI Voice Links to override this setting. Google TTS offers a high number of voice options. We recommend trying several options until you find the one that suits your station.
Download The List Of Voices From Google - Clicking this option will force Myriad Playout to update the list of available voices from Google.

AI Voice Link Creation & Placement Options

The final section of the AI Voice Links tab allows you to set some more generalized settings regarding the creation and placement of AI Voice Links:

How many seconds in advance should the log automatically create voice links - This setting controls how far ahead Myriad will check the log and automatically create voice links. This work is done by the On-Air Myriad workstation and only checks the items that have been loaded into memory based on the setting in the 'Scheduled Log' tab. By default this is set to 300 seconds (5 minutes) but if you wanted to increase this duration then you may also need to increase the "Additional number of log items ahead to cache" on the Scheduled Log tab. For example, if you wanted to precreate voice links half an hour before they were needed, then you would set this setting to 600 seconds, and then likely set the "Additional number of items... " to 10 items to make sure that Myriad is loading the next 10 items into memory when it plays the log so it can check those to see if there are any AI Voice Links that need to be created.
Voice speech rate - This settings allows you to provide a percentage speed increase of decrease that will be applied to the generated AI Voice Links. We recommend setting this to around 5% for most stations as it helps make the links sound more 'upbeat'.
When playing AI Voice Links, duck other Items to - Set the audio 'duck' level (in dB) for Songs and Links playing at the same time as AI Voice Links. Due to the nature of artificially generated speech, this usually needs to be set quite low, for example -10dB or -12dB
Maximum number of seconds to overlay Voice Link over previous item - Myriad will automatically play AI Voice Links over the end of the previous item so that the item continues to play 'underneath' the voice link. This setting allows you to set a maximum number of seconds that Myriad use in the previous track. If the AI Voice Link length exceeds this maximum (and the Intro on the next item) then a section of the AI Voice Link will be played 'dry'.
Number of seconds to leave between the end of the Voice Link and the intro of the next item - This setting allows you to set a short gap between when the AI Voice Link ends and the Intro of the next item ends. This allows you to avoid the AI Voice Link being unnaturally close to the Intro end which in turn makes the link sound more 'real'.
Replacement Word List - The Replacement Word List allows you to 'tune' AI Voice Links to pronounce words correctly or remove strings of characters entirely.
To view and edit the Replacement Word List, click on Edit List.

In the example above, you can see that the common abbreviation "feat." will be replaced with the full word "featuring" so wherever the word "feat." is included in an Artist or Title that is included in an AI Voice Link, the system will replace it with the full word when it creates the audio.

Another example could be if you have a Title or Artist that the AI Voice generation service does not handle correctly, you can correct that in the Replacement list using the Click here to add option.

In this example we have added the artist 'N-SYNC' but replaced it with a phonetically spelled version so that the AI Voice generation service can pronounce it correctly.

AI Voice Links are now configured but you still need to add them to your Clocks in order to have them generated and played as part of your station output. This is covered in a separate article.

TIP - If not already enabled, turn on automatic deletion of old Voice Links - Because AI Voice Links are rendered to Media Items in your Voice Link range on the Media Wall, unless you turn on the option to automatically delete old Voice Links, eventually you will run out of available Media Items to generate AI Voice Links into.

If you have not already enabled this feature - Setting up automatic deletion of old Voice links in Myriad Playout v5

The process is the same for Myriad 5 and Myriad 6.

Attachments (15)