Abstract:
PURPOSE: A device and a method for removing broadband acoustic echoes having adaptive tail length mounted in an embedded system are provided to possessively determine the tail length according to a result of a monitoring module estimating an environment capable of changing the toil length. CONSTITUTION: A device(100) for removing broadband acoustic echoes having adaptive tail length mounted in an embedded system comprises a delay length calculation unit(110), an adaptive filter implementation unit(120), an error calculation unit(130), a monitoring module(140), and an adaptive filter control unit(150). The delay length calculation unit calculates the delay length of an echo path by using close speaker signals and remote speaker signals. The adaptive filer implementation unit implements an adaptive filer by using a coefficient set based on the calculated delay length. The error calculation unit finds three sections which impulse responses with respect to the entire section of a tail of the adaptive filter are largest, thereby calculating the errors during the set coefficient. The monitoring module estimates a condition capable of changing the tail length. The adaptive filter control unit changes the tail length based on the estimated result. [Reference numerals] (100) Broadband acoustic echo removing device; (110) Delay length calculation unit; (120) Adaptive filter implementation unit; (130) Error calculation unit; (140) Monitoring module; (150) Adaptive filter control unit
Abstract:
본 발명은 CELP 기반 코덱으로 오디오 신호를 처리할 때, 오디오 신호의 처리를 단말기에 장착되는 코어 프로세서와 코어 프로세서에 부가되는 하드웨어가 분담하여 처리하도록 함으로써 비용대비 효과를 최대화하는 분산형 CELP 기반 가변비트 코덱 처리장치에 대한 것이다. 이를 위해 본 발명은 제1인코더, 및 제2인코더를 이용하는 CELP 기반 광대역 가변비트 코덱 처리장치에 있어서, 제1인코더의 기능을 소프트웨어 처리하여 제1인코딩값을 생성하는 코어 프로세서, 및 코어 프로세서가 제공하는 오디오 신호를 인코딩하여 제2인코딩값을 생성하는 분산 코덱 처리부를 포함하며, 코어 프로세서는, 제1인코딩값과 제2인코딩값을 멀티플렉싱하여 최종 인코딩값을 생성한다. CELP 코덱, G.729.1, 멀티플렉싱
Abstract:
A method and an apparatus for direct memory access control are provided to minimize frequency of a process accessing an external memory for processing interrupt by minimizing frequency of interrupt in a direct memory access controller regardless of size of a storage medium of a voice output device, thereby improving performance of a system. If DMAC(Direct Memory Access Control) operation is requested, data to transmit is acquired and stored inside(S201-S203). The data is acquired as much as initial burst size to transmit the data to a storage medium for output(S204-S206). Size of burst is reset in accordance with a state of a data destination and the data is acquired as much as the reset burst size to transmit the data to the data destination(S208-S210). If all data are transmitted, interrupt is generated(S211). If not, it is judged whether a retransmission step is operated.
Abstract:
본 발명은 무선 인터넷 전화용 통신 단말 장치에 관한 것이다. 본 발명에 따른 인터넷 망을 통하여 데이터를 송수신하는 통신 단말 장치는 인터넷 망을 통하여 데이터를 송수신하는 통신 단말 장치로서, 상기 단말 장치의 상태 정보나 운영 정보 등을 표시하는 표시부; 상기 통신 단말 장치를 운용하기 위한 운용 프로그램을 저장하는 메모리; 상기 통신 단말 장치를 상기 인터넷 망의 액세스 포인트에 등록시키고, 상기 액세스 포인트를 통하여 데이터를 송수신하며, 상기 송수신할 데이터가 임계 시간 동안 발생되지 않는 경우에 클락 신호의 주파수를 감소시키는 단말 제어부; 및 음성 데이터를 상기 단말 제어부로 전송하거나, 상기 단말 제어부로부터 출력된 음성을 수신하여 전송하는 음성 전송부를 포함한다.
Abstract:
PURPOSE: A method for deciding a bus master priority of a DMAC(Direct Memory Access Controller) using a weight parameter is provided to rearrange a priority value of a module finishing the data I/O(Input/Output) by using the weight parameter applying a time restriction degree of each function module. CONSTITUTION: The weight parameter corresponding to the time restriction degree is given to each module. When the data I/O is requested through a channel of each module, it is judged that the channel requesting the data I/O is more than one. In case of more than one channel, the highest priority channel is decided depending on a size of the weight parameter by comparing the weight parameter of each module(S150). The data I/O is performed through the decided highest priority channel and the weight parameter of other channel is rearranged(S160).
Abstract:
PURPOSE: A transition density calculating method is provided to definitely estimate a dynamic power consumption of a system applied to mobile communication electronics. CONSTITUTION: A logic equation of a circuit is input for measuring a power consumption, a logic diagram is constructed based on the logic equation, and the total number of internal nodes and output nodes in the circuit are counted(101, 102). Obtained is a probability that an input node of a gate element within a circuit is 1, and the transition density of the input node is detected(103, 104). It is determined whether obtained is a probability that an output node of a gate element within a circuit is 1, and if the probability is not obtained, the probability is calculated based on a BDD(Binary Decision Diagram) method(105, 106). An input transition condition for making a transition of an output of a gate element is detected(107). The output transition density of a gate element is calculated by using an occurrence probability of an input transition condition and a probability of an input node is 1(108).