Method for eliminating echo from multi-party call
An echo cancellation, multi-party call technology, applied in the direction of the two-way sound reinforcement telephone system, etc., can solve the problems of echo interference of multi-party calls, and achieve the effect of simple performance
- Summary
- Abstract
- Description
- Claims
- Application Information
AI Technical Summary
Problems solved by technology
Method used
Image
Examples
Embodiment 1
[0053] refer to figure 1 The method described is specific to an embodiment of a two-channel stereo audio system, which is applied to a video conferencing system. In this system, a microphone MIC1 and two speakers SPK1-L and SPK1-R are placed in the far end room, with the participants as the center, SPK1-L is placed on the left, and SPK1-R is placed on the right. distance. MIC1 is correspondingly placed within a range not far from the speaker, and is used to collect the audio signal of the participant speaking. Similarly, place a microphone MIC2 and two speakers SPK2-L and SPK2-R in the near-end room, centering on the participants, place SPK2-L on the left and SPK2-R on the right at the same distance, Corresponding SPK2-L and SPK2-R restore the signal of MIC1 in the distal chamber.
[0054] Accepting references from remote rooms figure 2 After processing the voice signals of multiple participants, while restoring the sound through the near-end SPK2-L and SPK2-R, the near-e...
Embodiment 2
[0057] With [embodiment 1], just adopt 4 microphones MIC to collect voice signal, refer to Image 6 The method shown is converted into a dual-channel stereo signal. The signals collected by 4 microphones need to be processed by 4*2 head-related transfer function HRTF, and then linearly superimposed to synthesize a dual-channel stereo signal and fed to the left and right speakers for playback.
[0058] Using the echo cancellation system of the present invention, the test of 200M audio data shows that not only can effectively eliminate multi-channel echo interference, but also the calculation speed of the FFT part of the echo cancellation algorithm executed by the GPU is about 5.5 times faster than that of the FFT part of the echo cancellation algorithm executed by the CPU times, greatly improving the overall performance of the system.
Embodiment 3
[0060] Same as [Example 1], except that 2 microphones MIC and 2 speakers SPK are set in each independent space. When collecting signals with 2 microphones MIC, 4 pairs of head-related transfer function HRTF are required for processing, and the echo cancellation system requires 4 groups of MDF filters detector and residual echo and background noise post-processing modules form a system.
[0061] Using the echo cancellation system of the present invention, the test of 100M audio data shows that not only can effectively eliminate the multi-channel echo interference, but also the calculation speed of the FFT part of the echo cancellation algorithm executed by the GPU is about 4 times faster than that of the FFT part of the echo cancellation algorithm executed by the CPU times, greatly improving the overall performance of the system.
[0062] By analogy, if there are N microphones MIC and 2 speakers SPK, then there must be N output signals after processing by this system. Since the...
PUM
Login to View More Abstract
Description
Claims
Application Information
Login to View More 