Overview
ISO/IEC 24661:2023 - Information technology - User interfaces - Full duplex speech interaction defines a framework for designing user interfaces (UIs) that support full duplex (FDX) speech interaction. The standard specifies an FDX speech interaction model, main features, functional components, functional requirements and privacy considerations to enable natural, session‑oriented conversational interfaces between humans and machines. It applies to UIs and communication protocols for setting up session‑oriented FDX interactions, but it does not define specific speech engines, devices or implementation details.
Key topics and technical requirements
ISO/IEC 24661:2023 covers the technical building blocks and quality requirements for FDX speech interaction UIs, including:
- Main characteristics of FDX: continuous, natural, adaptable, initiative‑capable, context‑based, knowledge‑based and model‑based interactions.
- Reference architecture and functional components:
- Acoustic acquisition (microphone arrays, voice activity detection - VAD)
- Speech recognition (ASR)
- Conversation processing (NLU, dialogue management, context handling)
- Speech synthesis (TTS / NLG)
- Resources (knowledge bases, data resources) and computing infrastructures (cloud/edge, AI/ML systems, network requirements)
- Functional requirements and recommendations: general, component‑level (acoustic, ASR, conversation processing, synthesis), resource and infrastructure requirements - focused on session continuity, latency, robustness and adaptability.
- Processes: engineering and interaction processes for designing, deploying and operating FDX speech UIs.
- Security and privacy: considerations for data handling, session management and user privacy in continuous, bidirectional speech sessions.
Practical applications and users
Practical applications include:
- Smart speakers and home assistants
- Contact center agents and interactive voice response (IVR) systems upgraded to session‑based interaction
- In‑vehicle voice assistants and hands‑free systems
- Conversational agents in healthcare, retail, banking and industrial settings
Who uses this standard:
- System designers and UI architects building conversational interfaces
- Speech and NLP developers integrating ASR/NLU/TTS components
- Service providers and platform operators deploying session‑oriented voice services
- Device makers and integrators implementing microphone arrays, edge/cloud connectivity
- Security and privacy officers assessing risks in voice interactions
Related standards (if applicable)
ISO/IEC 24661:2023 helps align development of FDX speech interaction UIs with standardized models, ensuring interoperability, privacy awareness and robust, natural conversational experiences. Keywords: ISO/IEC 24661:2023, full duplex speech interaction, FDX, speech interaction UI, ASR, TTS, VAD, conversational interfaces, reference architecture.