Character animation

Details Character animation Character animation

2003-02-11

2004-08-03

Knepper, David D. (Department: 2654)

Data processing: speech signal processing, linguistics, language

Speech signal processing

Application

C704S235000

Reexamination Certificate

active

06772122

ABSTRACT:

FIELD OF THE INVENTION
The present invention relates to a method of generating an animated character representation using a processing system and apparatus for generating an animated character representation. In particular, the present invention relates to a system that uses input data, comprising content data and presentation data, to animate a character representative of a person, such as a news reader, announcer, presenter, or the like.
BACKGROUND TO THE INVENTION
Character animation has previously been achieved in a number of ways. The most basic system is standard hand drawn animation, achieved by drawing a number of sequential frames and then displaying the frames at a high rate to generate the appearance of character movement. However, the production of such frames is time consuming and requires great skill in producing the desired appearance.
More recently, character animation has been achieved using computer-based systems. However, in systems like this, again the animation is predetermined by an artist, requiring great skill and work to produce the desired appearance.
Automated character animation has also been produced which operates by converting a standard text file into speech and then using visemes to animate a character. However, these systems suffer from the drawback that the range of movement presented by the character is limited and in particular is normally limited to the movement required to present the visemes. Any additional character movement must be added in manually at a later date and cannot be incorporated automatically. Furthermore, the characters can only demonstrate a very limited linear response to the text. Accordingly, each time the character reads the text the appearance of the character is identical. An example of such a system is described in U.S. Pat. No. 5,657,426.
This therefore does not present a very human appearance in which the specific movement of the character would vary each time the text is read out. Furthermore, when no text is being read the character is motionless, again contributing to the lack of human appeal or characterization of the character.
SUMMARY OF THE INVENTION
In accordance with a first aspect of the present invention, we provide an apparatus for generating an animated character representation, the apparatus comprising a processing system having:
an input for receiving marked-up input data including:
i content data representing speech to be presented; and,
ii presentation data representing the manner in which the speech is presented;
a processor coupled to the input for generating data according to a defined time-base, the data including:
i phoneme data generated in accordance with the content data; and,
ii viseme data generated in accordance with the phoneme data and the presentation data;
the processor being further adapted to:
iii generate audio data in accordance with the phoneme data;
iv generate image data in accordance with the viseme data; and,
v synchronise the output of the audio and image data in accordance with the defined time-base.
In accordance with a second aspect of the present invention, we provide a method of generating an animated character representation using a processing system, the method comprising:
receiving marked-up input data including:
i content data representing speech to be presented; and,
ii presentation data representing the manner in which the speech is presented;
generating data according to a defined time-base, the data including:
i phoneme data generated in accordance with the content data; and,
ii viseme data generated in accordance with the content data;
generating audio data in accordance with the phoneme data;
generating image data in accordance with the viseme data and the presentation data; and,
synchronising the output of the audio and image data in accordance with the defined time-base.
The present invention provides a method and apparatus for generating an animated character representation. This is achieved by using marked-up data including both content data and presentation data. The system then uses this information to generate phoneme and viseme data representing the speech to be presented by the character. By providing the presentation data this ensures that at least some variation in character appearance will automatically occur beyond that of the visemes required to make the character appear to speak. This contributes to the character having a far more lifelike appearance.
The marked-up data input to the system may be manually entered, for instance by typing text at a terminal, or may be derived from a data source. This allows the system to be used for the automated presentation of information from news and data sources and the like.
The processor usually includes:
a text-to-speech processor for generating the phoneme data and the audio data;
an animation processor for generating the viseme data and the image data; and,
a parser for:
parsing the received marked-up data;
detecting predetermined content data which is to be presented in a predetermined manner;
generating presentation data representative of the predetermined manner; and,
modifying the received marked-up data with the generated presentation data.
The use of specialised text-to-speech and animation processors allows the system to generate the audio and image data in real time, thereby speeding up the character animation process. The audio and image data can be generated at the same time or at different times, and/or in different locations, as required. It will be appreciated, that the text-to-speech and animation processors may be implemented as software within a single processor, or may alternatively be implemented as separate hardware components.
Parsing the received marked-up data allows presentation data to be added, which in turn allows data which has only minimal or no mark-up to be processed by the present invention. This also allows predetermined content to be represented in a predetermined manner. Furthermore, this allows the animated character to stress certain words, such as numbers, names, nouns and negatives, although this is not essential to the present invention.
The processing system will usually include a store for storing data, the parser being coupled to the store to obtain an indication of the predetermined content data therefrom. This allows information concerning the mark-up to be added to be stored centrally so that it can be accessed directly by the parser. Alternatively, the information may be obtained via a communications system, such as a LAN (Local Area Network) or the like, from a remote store.
Typically the apparatus includes a linguistic processor adapted to:
parse the content data;
determine the phonemes required to represent the content data; and,
generate phoneme time references for each of the phonemes, the phoneme time reference indicating the time at which the respective phoneme should be presented with respect to the time base.
It is preferable to use phonemes to generate the audio data to be presented by the animated character as this allows a small number of elemental sound units to represent the majority of sounds that would need to be made by the character to present the speech. Additionally, processing systems for determining phonemes from text are well known and readily implementable.
Furthermore, the generation of the phoneme time references allows the temporal location of each of the phonemes to be maintained as well as enabling the synchronization of remaining steps in the procedure.
Typically the linguistic processor is further adapted to:
parse the presentation data;
generate a number of tags representing the presentation data; and,
generate tag time references for each of the tags, the tag time reference indicating the time at which the respective tag should modify the manner of presentation with respect to the time base.
The use of tag time references allows the temporal position of the presentation data to be maintained relative to the phoneme data. Alternatively, other synchronisation techniques, could be used.
Usually the linguistic processor is co

Affiliated with

Also associated with

Rating

Character animation does not yet have a rating. At this time, there are no reviews or comments for this patent.

If you have personal experience with Character animation, we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Character animation will most certainly appreciate the feedback.

Rate now

Comments { 0 }

Profile ID: LFUS-PAI-O-3332514