Ricoh Company, Ltd.

Junichiroh Fujimoto

Tetsuya Muroi

Seigou Yasuda

Tomofumi Nakatani

Toshiki Kawamoto

Seigo Yasuda

After sampling a voice signal by a microphone (31), a voice interval is detected by a voice interval detector (32). Then, the voice signal is processed by a frequency analyzer (33) having a predetermined number of channels at a predetermined time interval, whereby a corresponding portion of the voice signal is quantized at each channel. Then, quantized data obtained from these channels are binary converted (34) to thereby form a frame comprised of a series of binary data. Each data of this frame corresponds to one of the channels, and preferably it is set to be an integral multiple of a computer calculation unit (e.g., 4 bits, 8 bits, etc.). When forming a combination frame by superimposing two or more of such frames, the combination frame is divided into layers, whereby it is so structured that each bit is represented by a binary number in each layer. On the other hand, it may also be so structured that one frame is divided in a plurality of sub-frames and a preliminary comparison process is carried out using each sub-frame.

Voice recognition system and method involving registered voice patterns

A pattern matching system for a speech recognition apparatus includes a dictionary for storing standard speech patterns, where each of the standard speech patterns are stored in the dictionary with two or more different pattern lengths, a converting part for converting an input speech pattern which has an arbitrary pattern length into at least one converted speech pattern which has one of the pattern lengths with which the standard speech patterns are stored in the dictionary depending on the arbitrary pattern length, and a collating part for collating the converted speech pattern with the standard speech patterns stored in the dictionary to obtain a standard speech pattern which has a largest degree of similarity to the converted speech pattern.

Pattern matching system for speech recognition system, especially useful

A voice pattern identifying system includes a library for storing therein a plurality of registered voice patterns and a register array into which a sampled voice pattern is temporarily stored for comparison with said registered voice patterns one by one. The voice pattern is represented by a plurality of frames which are sampled at a predetermined time interval and each of which includes a predetermined number of data at different frequency ranges. The registered voice patterns are stored as compressed in the library so that each frame is accompanied by a compression data indicating how many times the corresponding frame should be repeated with time. In accordance with the present invention, there is provided a pattern compander which only modifies the compression data of either of the registered or sample voice pattern so as to make the number of frames identical before determination of the degree of similarity the two voice patterns.

Identification of voice pattern

A method for forming a reference voice pattern to be stored in a library for use in identifying an unknown input sound. The method comprises the steps of inputting a first sound pronounced for a particular letter combination by a speaker to form a first voice pattern, the first voice pattern represented by a sound energy level as a function of time; examining the first voice pattern to determine whether or not an intermediate voiceless interval is present within a predetermined time period at least from either a beginning or terminating end of the first voice pattern; repeating the inputting and examining steps for a second sound; combining the first and second voice patterns to define the reference voice pattern, wherein, if either of the first and second voice patterns has a voiced portion in the predetermined time period and the other of the first and second voice patterns does not have a corresponding voiced portion in the predetermined time period, the combining step comprises adding the voiced portion to said other of the first and second voice patterns.

Methods for forming reference voice patterns, and methods for comparing

A voice or sound recognition system including a microphone for converting a voice into an electrical voice signal, a frequency analyzer for generating a voice pattern in the form of a time-frequency distribution, and a matching unit for matching the voice pattern with registered voice patterns. A voice pattern sometimes contains a bass bar section starting from the beginning of the voice pattern over a time period. In one form of the present invention, the bass bar section is detected and/or eliminated before processing the voice pattern for the purpose of registration or matching. The voice level is sometimes too strong or too weak for appropriate processing. In another form of the present invention, a voice recognition system including a voice input cancelling function is provided. In addition, a system for detecting a voice interval for use in voice data processing, including a device for determining a start point of the voice interval is provided.

Voice recognition system

A voice processing method and system using voice power is provided. A voice power signal is produced in analog format and this voice power signal is sampled at a predetermined time interval to determine its current envelope or amplitude, which is then compared with a predetermined number of voice power levels, whereby '1' is assigned to at least one of the predetermined number of voice power levels which corresponds in level to the amplitude thus determined and '0' is assigned to the rest of the predetermined number of voice power levels, thereby converting the analog voice power signal into a binary voice power pattern. In the preferred embodiment, the voice signal is also processed through a frequency analyzer including a plurality of band-pass filters different in frequency range and a binary converter to form a binary time-frequency voice distribution pattern which is then combined with the binary voice power pattern to form a combined voice pattern. As a modification, a voice power range is determined by multiplying a predetermined threshold to the amplitude, and the voice power pattern is compared with the predetermined number of voice power levels.

Voice recognition system using voice power patterns

A sound or speech recognition system is made faster by first using data-reduced patterns to find candidates in preliminary testing, and then making a final decision by testing against full-data patterns. Specifically, a signal is transformed to a time-sequence of spectral vectors, which forms a full-data pattern. Then, a predetermined number of vectors representing the beginning or front-end of the signal are duplicated and summed to form a sum spectral vector which is used as the data-reduced pattern for preliminary testing.

Dual-step sound pattern matching

A voice recognition system includes a microphone for converting a voice in the form of pressure waves into an electrical voice signal and a keyboard for inputting the phonetic spelling of the voice. A check mechanism is connected to the keyboard to determine whether or not the voice has a predetermined sound element, which is difficult to be detected, such as /u/, /i/, /n/ and explosives like /p/, /t/, /k/ and /b/, at or near the end thereof. If the check mechanism has determined the presence of such a predetermined sound element, then it checks the corresponding voice pattern to see whether the corresponding component is present in the pattern or not. If not, a warning may be issued to request reinputting of the voice, thereby insuring the fact that the voice pattern to be registered as a reference voice pattern is always accurate.

Growing community of inventors

Yokohama, Japan

Junichiroh Fujimoto