Azure Bot Directline channel vs Directline speech channel

Comparing two communication channels for Azure Bot Service

Azure Bot Service is a cloud platform that allows developers to build and deploy conversational agents (also known as bots) that can interact with users across a variety of channels, such as web, mobile, or messaging applications. One of the key features of Azure Bot Service is the ability to connect bots to multiple channels, such as Skype, Facebook Messenger, Slack, or Teams, using a single bot registration and code base.

However, not all channels offer the same level of functionality and user experience. Some channels, such as web chat or direct phone, allow you to customize the look and feel of your bot, and the mode of communication, such as text, voice, or both. Other channels, such as Skype or Teams, have predefined user interfaces and support only text-based communication.

In this blog post, we compare two communication channels that can use voice as an input and output modality for your bot: the direct line channel and the direct line voice channel. We will explain what the channels are, how they differ, how to use them, and what the benefits and limitations of each are.

A direct line channel is a REST API that allows any client application that can send and receive HTTP requests to communicate with your bot. You can use a direct line channel to create a custom user interface for your bot or integrate your bot with other services or applications. For example, you can use a direct line channel to embed your bot in a web page, mobile app, or voice assistant.

The direct line channel supports both text and speech as input and output modalities for your bot. However, if you want to use speech, you need to use an additional service, such as the Cognitive Services Speech SDK or Web Chat Speech Services, to handle speech recognition and synthesis. You also need to handle audio streaming and encoding between your client application and the speech service.

Direct line speech channel is a WebSocket-based API that allows you to communicate with your bot using voice as the primary input and output mode. You can use Direct line speech channel to create a voice-enabled user interface for your bot, such as a voice assistant, smart speaker, or voice bot. For example, you can use Direct line speech channel to connect your bot to Cortana, Alexa, or Google Assistant.

Direct line speech channel simplifies voice integration in your bot by handling speech recognition and synthesis, audio streaming, and encoding. You don’t need to use any additional services or SDKs to use voice in your bot. Just use the Direct line speech channel SDK in your client application, and it is available in C#, JavaScript, and Java.

Both the Direct Line Channel and the Direct Line Voice Channel allow you to connect your bot to any client application that can communicate with it. You can also use voice as an input and output modality for your bot, and customize your bot’s user interface and user experience.

However, there are some key differences between them:

• Direct line channels support both text and voice, while direct line voice channels support only voice.
• Direct line channels require additional services or SDKs to use voice, whereas direct line voice channels do not.
• Direct line channels use a REST API, while direct line voice channels use a WebSocket API.
• Direct line channels must handle audio streaming and encoding themselves, whereas direct line voice channels do not.
• Direct line channels offer greater flexibility and control over voice settings such as language, voice, speed, and pitch, while direct line voice channels offer less flexibility and control.
• Direct line channels have higher delay and overhead for voice communications, whereas direct line voice channels have lower delay and overhead.

To use the voice channel directly from your bot, follow these steps:

1. Create a bot using Azure Bot Service and enable the call channel directly in the bot settings. How to add a voice channel directly to Azure Bot
2. Generate a direct call key and endpoint for your bot.
3. Create a client application using the Direct Voice Channel SDK for your platform (C#, JavaScript, or Java). Directline Voice Channel SDK
4. Initialize a direct call channel client using the direct call key and endpoint. Directline Voice Channel Client
5. Start a conversation with your bot using the direct call channel client.
6. Send and receive voice messages from your bot using the direct voice channel client.
7. End a conversation with a bot using the direct call channel client.

Here’s an example of how to use the voice channel directly in C#.

reference : https://github.com/microsoft/BotFramework-WebChat/tree/main/samples/03.speech/a.direct-line-speech

To use a direct call channel with a bot, follow these steps:

1. Create a bot using Azure Bot Service and enable direct line channels in the bot settings. How to activate a direct line channel
2. Generate a direct call password or token for your bot.
3. Create a client application that can send and receive HTTP requests.
4. Initializes a direct line channel client using a direct line secret or token.
5. Start a conversation with your bot using the direct channel client.
6. Send and receive text or voice messages from your bot using the direct channel client.
7. End a conversation with a bot using the direct channel client.

Here's an example of how to use a direct line channel using C# and the Cognitive Services Speech SDK.

reference : https://github.com/microsoft/BotFramework-WebChat/tree/main/samples/03.speech/f.web-browser-speech

To use direct channel connections from your bot, you will need:

• Azure subscription.
• Azure Bot Service resource.
• A direct contact secret or token for the bot.
• A client application that can send and receive HTTP requests.
• Optionally, there are voice services or SDKs available to enable voice in your bot.

To use voice channels directly from your bot, you will need:

• Azure subscription.
• Azure Bot Service resource.
• The bot's direct call key and endpoint.
• A client application that can use the Direct Voice Channel SDK for your platform (C#, JavaScript, or Java).

For more information and examples of how to use Direct Line Channels and Direct Line Voice Channels in your bot, see the following resources:

To use a direct line channel with a cognitive service like speech, you must handle speech recognition and synthesis for your bot using the Cognitive Services Speech SDK or Web Chat Speech Services. You must also handle audio streaming and encoding between your client application and the speech service.

The Cognitive Services Speech SDK is a cross-platform library that lets you use speech as an input and output modality for your bot. It supports a variety of languages, voices, and voice settings. You can use the Speech SDK in any client application that can communicate with a direct line channel, such as a web, mobile, or desktop app.

Web Chat Voice Service is a web-based service that allows you to use voice as an input and output modality for your bot. It supports a variety of languages, voices, and voice settings. You can use the Web Chat Voice Service in any web-based client application that can embed a web chat component, such as a web page or web app.

To use the Direct Line Voice Channel with cognitive services such as speech, you do not need any additional services or SDKs, as the Direct Line Voice Channel handles speech recognition and synthesis. You only need to use the Direct Line Voice Channel SDK in your client applications, which is available in C#, JavaScript, and Java.

The Direct Line Speech Channel SDK is a cross-platform library that enables you to use speech as the primary input and output modality for your bot. It supports a variety of languages, voices, and voice settings. The Direct Line Speech Channel SDK can be used with any client application that can communicate with the Direct Line Speech Channel, such as a web, mobile, or desktop app.