22.2433GPPRelease 17Speech recognition framework for automated voice servicesStage 1TS
A 3GPP speech recognition framework enables the use of conventional codecs (e.g. AMR) or DSR optimized codecs to distribute in the network the speech engines that process speech input or generate speech output. It includes:
– Default uplink and downlink codec specifications.
– A stack of speech recognition protocols to support:
– Establishment of uplink and downlink sessions, along with codec negotiation
– Transport of speech recognition payload (uplink) with conversational QoS
– Support of transport (also at conversational QoS) of meta-information required for the deployment of speech recognition applications between the terminal and speech engines (meta-information may include terminal events and settings, audio sub-system events, parameters and settings, etc.).
IMS provides a protocol stack (e.g. SIP/SDP, RTP and QoS), that may advantageously be used to implement such capabilities.
It shall be possible to recommend a codec to be supported by default to deploy services that rely on the 3GPP speech recognition framework. To that effect, the specifications will consider either conventional speech codecs (e.g. AMR) or DSR optimized codecs.
ETSI has published DSR optimized codecs specifications (ETSI ES 201 108 & ETSI ES 202 050 [7, 10]) and a payload format for transport of DSR data over RTP (IETF AVT DSR).
The following list gives the high level requirements for the SRF-based automated voice services: .
- Users of the SRF-based automated voice service shall be able to initiate voice communication, access information or conduct transactions by voice commands using speech recognition. Examples of SRF-based automated voice services are provided in Appendix A.
The speech recognition framework for automated voice services will be offered by the network operators and will bring value to the network operator by the ability to charge for the SRF-based automated voice services.
This service may be offered over a packet switched network; however in general this requires specification of a complete protocol stack. When this service is offered over the IMS, the protocols used for the meta information and front-end parameters (from terminal to server) and associated control and application specific information can and shall be based on those in IMS.
It shall be possible for a user to initiate a connection to the SRF-based automatic voice services by entering the identity of the service. Most commonly, when used as a voice service, this will be performed by entering a phone number. However, particular terminals may offer a user agent that accepts other addressing schemes to be entered by the user: IP address, URI, e-mail address possibly associated to a protocol identifier. This is particularly important for multi-modal usages.
In all cases, the terminal will convert the address entered by the user to initiate a session via the SIP IMS session initiation protocol and establish the different SRF protocols. During this initiation of the SRF session, it shall be possible to negotiate the uplink and downlink codecs. The terminal shall support a codec suitable for speech recognition as a default uplink codec.
4.2 Information during the speech recognition session
Codec negotiation during a SRF session should be optionally supported.
This may be motivated by the expected or observed acoustic environment, the service package purchased by the user, the user profile (e.g. hands-free as default) or service need. The user speaks to the service and receives output back from the automated voice service provider as audio (recorded ‘natural’ speech) or Text-to-Speech Synthesis. The output from the server can be provided in the downlink as a streaming service or by using conversational speech codec.
Additional control and application specific information shall be exchanged during the session between the client and the service. Accordingly some terminals shall be able to support sending additional data to the service (e.g. keypad information and other terminal and audio events) and receiving data feedback that shall be displayed on the terminal screen.
Dynamic payload switches within a session may be considered to transport meta-information.
It shall be possible to use SRF sessions in order to provide access to SRF-based automated voice services. For example applications might use a SRF session to access and navigate within and between the various SRF-based automated voice services by spoken commands or pressed keypads.
It shall be possible for network operators to control access to SRF-based automated voice services based on subscription profile of the callers.
4.4 User Perspective (User Interface)
The user’s interface to this service shall be via the UE. User can interact by spoken and keypad inputs. The UE can have a visual display capability. When supported by the terminal, the server-based application can display visual information (e.g., stock quote figures, flight gates and times) in addition to audio playback (via recorded speech or text-to-speech synthesis) of the information. These are examples of multimodal interfaces. SRF enables distributed multimodal interfaces as described in .