公开(公告)号：US20080071549A1

公开(公告)日：2008-03-20

申请号：US11629135

申请日：2005-06-28

申请人： Kok Chong ,  Naoya Tanaka ,  Sua Neo ,  Mineo Tsushima

发明人： Kok Chong ,  Naoya Tanaka ,  Sua Neo ,  Mineo Tsushima

IPC分类号： G10L19/00

CPC分类号： G10L19/008 ,  G10L19/24

摘要： In the conventional art inventions for coding multi-channel audio signals, three of the major processes involved are: generation of a reverberation signal using an all-pass filter; segmentation of a signal in the time and frequency domains for the purpose of level adjustment; and mixing of a coded binaural signal with an original signal coded up to a fixed crossover frequency. These processes pose the problems mentioned in the present invention. The present invention proposes the following three embodiments: to control the extent of reverberations by dynamically adjusting all-pass filter coefficients with the inter-channel coherence cues; to segment a signal in the time domain finely in the lower frequency region and coarsely in the higher frequency region; and to control a crossover frequency used for mixing based on a bit rate, and if the original signal is coarsely quantized, to mix a downmix signal with an original signal in proportions determined by an inter-channel coherence cue.

摘要翻译： 在用于编码多声道音频信号的传统技术发明中，涉及的三个主要过程是：使用全通滤波器产生混响信号; 用于时间和频域的信号分段，以进行电平调整; 并将编码的双耳信号与编码为固定的交越频率的原始信号进行混合。 这些处理构成了本发明中提到的问题。 本发明提出以下三个实施例：通过使用信道间相干线索动态调整全通滤波器系数来控制混响的程度; 以在较低频率区域精细地分割时域中的信号，在较高频率区域中粗略地分割信号; 并且基于比特率来控制用于混合的交叉频率，并且如果原始信号被粗量化，则以通过信道内相干提示确定的比例将下混合信号与原始信号混合。

公开(公告)号：US20070271095A1

公开(公告)日：2007-11-22

申请号：US11659949

申请日：2005-08-18

申请人： Shuji Miyasaka ,  Yoshiaki Takagi ,  Naoya Tanaka ,  Mineo Tsushima

发明人： Shuji Miyasaka ,  Yoshiaki Takagi ,  Naoya Tanaka ,  Mineo Tsushima

IPC分类号： G10L19/00 ,  G10L19/02 ,  H03M7/30

CPC分类号： G11B27/022 ,  G10L19/008 ,  G11B20/00007

摘要： An audio encoder, which is capable of encoding multiple-channel signals so that only a downmixed signal is decoded and of further generating specific auxiliary information necessary for dividing the downmixed signal, is provided. An audio encoder (10), which compresses and encodes audio signals of N-channels (N>1), includes a downmixed signal encoding unit (11) which encodes the downmixed signal obtained by downmixing the audio signals, and an auxiliary information generation unit (12a) which generates auxiliary information necessary for decoding the downmixed signal encoded by the downmixed signal encoding unit (11) into N-channel audio signals. The auxiliary information generation unit (12a) includes transformation units (121) and (122) which transform audio signals respectively into frequency domain signals, a detection unit (123) which detects phase difference information and gain ratio information each indicates a degree of difference between frequency domain signals, and a quantization unit (125) which quantizes, for each frequency band, the phase difference information and gain ratio information detected by the detection unit (123) using the quantization precision setting table (124). The quantization precision setting table (124) functions as a division unit which divides a frequency band of a frequency domain signal into plural sub-bands.

摘要翻译： 一种音频编码器，其能够对多声道信号进行编码，使得只下降混合信号被解码，并且进一步产生分割下混合信号所需的特定辅助信息。 对N声道（N> 1）的音频信号进行压缩编码的音频编码器（10）包括下混合信号编码单元（11），其对通过降低音频信号而获得的下混合信号进行编码，以及辅助信息生成单元 （12a），其产生将由下混合信号编码单元（11）编码的下混合信号解码成N声道音频信号所需的辅助信息。 辅助信息生成单元（12a）包括将音频信号分别变换为频域信号的变换单元（121）和（122），检测单元（123），其检测相位差信息，增益比信息各自表示差异度 和量化单元（125），其对于每个频带，使用量化精度设定表（124）对由检测单元（123）检测的相位差信息和增益比信息进行量化。 量化精度设定表（124）作为将频域信号的频带划分为多个子带的分割单位起作用。

公开(公告)号：US20070160236A1

公开(公告)日：2007-07-12

申请号：US10589818

申请日：2005-07-01

申请人： Kazuhiro Iida ,  Mineo Tsushima ,  Yoshiaki Takagi ,  Naoya Tanaka

发明人： Kazuhiro Iida ,  Mineo Tsushima ,  Yoshiaki Takagi ,  Naoya Tanaka

IPC分类号： H04B1/00 ,  H04B1/20

CPC分类号： H04S3/00

摘要： An audio signal encoding device includes a downmix signal encoding unit 203 and an auxiliary information generation unit 204. The downmix signal encoding unit 203 generates a downmix signal acquired by adding input signals each other using a predetermined method, encodes the downmix signal, and outputs downmix signal information 206. The auxiliary information generation unit 204 generates auxiliary information 205 using the downmix signal and the downmix signal information 206 generated by the downmix signal encoding unit 203. The auxiliary information generation unit 204 efficiently quantizes the auxiliary information 205 using human's characteristics of a perceptual direction of a sound source, a perceptual broadening, and a perceptual distance.

摘要翻译： 音频信号编码装置包括下混信号编码单元203和辅助信息生成单元204。 下混信号编码单元203通过使用预定方法相加输入信号而产生下混信号，对下混信号进行编码，并输出下混合信号信息206。 辅助信息生成单元204使用由混合信号编码单元203生成的下混信号和下混信号信息206来生成辅助信息205。 辅助信息生成单元204使用人的声源的感知方向，感知加宽度和感知距离的特性来有效地量化辅助信息205。