MPAI-PAF V1.6 AIMs Speaking Avatar Formation

Function
Ref. Model
I/O Data
SubAIMs
JSON MData
Profiles
Ref. Software
Conformance
Performance

1 Functions

The Speaking Avatar Formation (PAF‑SAF) AIM forms the Speaking Avatar of an input Avatar Model, animated as it utters the input Speech. Face and Body Descriptors are generated from the Speech and the Text, expressive when a Face or Gesture Personal Status is supplied, and packaged with the Avatar and the Speech, time-aligned and ready to play:

Receives	Speech	The Speech the Avatar utters.
	Text	The Text of the Speech, used to generate Face and Body Descriptors.
	Face Personal Status	The Personal Status the Face expresses (optional).
	Gesture Personal Status	The Personal Status the Gesture expresses (optional).
	Avatar Model	The Avatar whose Model is animated.
Produces	Speaking Avatar	The Avatar, the Speech and the Face and Body Descriptors, ready to play.

2 Reference Model

Figure 1 depicts the Reference Model of the Speaking Avatar Formation (PAF‑SAF) AIM.

Figure 1 – The Speaking Avatar Formation (PAF‑SAF) AIM

3 I/O Data

Table 1 specifies the Input and Output Data of the Speaking Avatar Formation (PAF‑SAF) AIM.

Table 1 – I/O Data of the Speaking Avatar Formation (PAF‑SAF) AIM

Input	Description
Speech	The Speech the Avatar utters.
Text	The Text of the Speech, used to generate Face and Body Descriptors.
Face Personal Status	The Personal Status the Face expresses (optional).
Gesture Personal Status	The Personal Status the Gesture expresses (optional).
Avatar Model	The Avatar whose Model is animated.
Output	Description
Speaking Avatar	The Avatar, the Speech and the Face and Body Descriptors, ready to play.

4 SubAIMs

4.1 Reference Model

A Speaking Avatar Formation (PAF‑SAF) AIM may be implemented as a Composite AIM depicted in Figure 2.

Figure 2 – The Speaking Avatar Formation (PAF‑SAF) Composite AIM

4.2 Operation

Generative Face Description produces the Face Descriptors from the Speech and the Text, expressive of the Face Personal Status when one is supplied. Generative Body Description produces the Body Descriptors from the Text, expressive of the Gesture Personal Status when one is supplied. Speaking Avatar Synthesis packages the Avatar Model, the Speech and the Face and Body Descriptors into the Speaking Avatar, time-aligned and ready to play.

4.3 Functions of AI Modules

Table 2 specifies the Function of the AI Modules.

Table 2 – Functions of End AI Modules

Generative Face Description	Generates the Face Descriptors of the Avatar from Speech and Text.
Generative Body Description	Generates the Body Descriptors of the Avatar from Text.
Speaking Avatar Synthesis	Packages Avatar, Speech, and Face and Body Descriptors into the Speaking Avatar.

4.4 I/O Data of AI Modules

Table 3 provides the links to the AIM specifications and JSON schemas. AIM1 indicates the Composite AIM and AIM2 its SubAIMs.

Table 3 – AIMs and JSON Metadata of the Speaking Avatar Formation (PAF‑SAF)

AIM1	AIM2	Name	JSON
PAF‑SAF		Speaking Avatar Formation	X
	PAF‑GFD	Generative Face Description	X
	PAF‑GBD	Generative Body Description	X
	PAF‑SAS	Speaking Avatar Synthesis	X

5 JSON Metadata

https://schemas.mpai.community/PAF/V1.6/AIMs/SpeakingAvatarFormation.json

6 Profiles

No Profiles.

7 Reference Software

Not part of this specification.

8 Conformance Testing

Table 4 provides the Conformance Testing Method for the PAF‑SAF AIM.

If a schema contains references to other schemas, conformance of data for the primary schema implies that any data referencing a secondary schema shall also validate against the relevant schema, if present, and conform with the Qualifier, if present.

Table 4 – Conformance Testing Method for the PAF‑SAF AIM

Receives	Speech	Shall validate against Speech Object schema.
	Text	Shall validate against Text Object schema.
	Face Personal Status	Shall validate against Personal Status schema.
	Gesture Personal Status	Shall validate against Personal Status schema.
	Avatar Model	Shall validate against Avatar schema.
Produces	Speaking Avatar	Shall validate against Speaking Avatar schema.

9 Performance Assessment