极性VS降噪相移 [英] polarity vs phase shift in noise reduction
问题描述
我想实现的噪声消除和我有几个可能的基本问题
I am trying to implement noise cancellation and I have a few perhaps basic questions
这是我所试图做的基本信息：
*我输入源的音频在时域上的格式，所以我有离散的真正价值。我猜测，这些都是所谓的极值？
*我从<一个公式href="http://stackoverflow.com/questions/6503562/which-algo-uses-for-noise-canceling-in-earphones">Algorithm噪音消除工作于频域。
*因此，我有FFT时间域转换成频率域
Basic Information on what I am trying to do:
* My input source audio is in time domain format, so I have discrete real values. I am guessing that these are what are known as "polar values"?
* I have a formula from Algorithm for Noise Cancellation which operates in frequency domain.
* Hence I have FFT to convert time domain to frequency domain
问题
-
的基本思想是，以破坏性消除噪声扭转原来的源信号的相位？
Is the basic idea to destructively cancel noise to reverse the phase of the original source signal?
是相移功能，工作在频域信号？也就是说，我只是用180度转向的频率在每个频率上的，我的时域转换到频域的FFT的结果？
Is phase shift a function that operates on signals in frequency domain? That is, do I just shift the frequency by 180 degrees on each frequency in the result of the FFT that converts my time domain to frequency domain?
时改变极性的同样的事情，移相？也就是说，如果我-1（负一）乘以时域的每个值，将那主动的事情是一样的信号转换到频域，然后用180度的移相？
Is reversing the polarity the same thing as shifting phase? That is, if I multiplied each value in time domain by -1 (negative one), will that active the same thing as converting the signal to frequency domain and then shifting the phase by 180 degrees?
最后一个问题，我怎么移相180度？请问我看到上面基本上是积极的链接只是？我的意思是，每个频率值我从我的FFT，什么数学运算上的每个不同的值进行180度相移的相同呢？
Final question, How do I shift a phase by 180 degrees? Does that link I saw above essentially active just that? I mean, on each frequency value I have from my FFT, what mathematical operation on each distinct value performs the equivalent of a 180 degree phase shift?
谢谢 德肖恩 -
推荐答案
没有，噪音消除不只是扭转原始源的阶段。它更表征和估计振幅和噪声信号的相位在空间某些位置在某个时间未来点，等于软件/硬件方法的等待时间，加上来自音频换能器的声音的传播延迟（多个）到取消的点位置（多个），然后合成的音频在此估计（其可以是从输入信号非常不同）的相对相位，同时也补偿音频输出子系统和换能器的频率和相位响应。
No, noise cancellation is not just to reverse phase of the original source. It is more to characterize and estimate the amplitude and phase of the noise signal at some location in space at some future point in time, equal to the latency of the software/hardware processes, plus the sound propagation delay from the audio transducer(s) to the point location(s) of cancellation, then to synthesize audio at the opposite phase of this estimation (which can be very different from the input signal) while also compensating for the frequency and phase response of the audio output subsystem and transducer.
由于在FFT是线性算子，乘以-1输入矢量产生相同的结果为乘以-1的复结果矢量，这是相同的一个180度的相位旋转。
Since the FFT is a linear operator, multiplying the input vector by -1 produces the same result as multiplying the complex result vector by -1, which is the same as a 180 degree phase rotation.
这篇关于极性VS降噪相移的文章就介绍到这了，希望我们推荐的答案对大家有所帮助，也希望大家多多支持IT屋！
