If you don't want aliasing, the sampling rate should be above twice the start of the high frequency stop-band of whatever filter you can find to use before the sampler.
The physics of the environment in which the waveform resides will often provide its own low-pass or band-pass filtering (propagation loses, transducer loses, etc.). If you can determine the cutoff of this natural filter such that its stop-band is below your desired noise floor, you can use greater than 2X this frequency.
If you don't care about aliasing, then any sample rate will do, but the lower the frequency, the smaller the bandwidth of the result and the greater the multiple of layers of aliasing within that bandwidth.
发布评论
评论（2）
如果您不想出现混叠，则采样率应高于您在采样器之前使用的任何滤波器的高频阻带起点的两倍。
波形所在环境的物理特性通常会提供其自身的低通或带通滤波（传播损耗、换能器损耗等）。如果您可以确定该自然滤波器的截止值，使其阻带低于您所需的本底噪声，则可以使用大于该频率 2 倍的频率。
如果您不关心混叠，那么任何采样率都可以，但是频率越低，结果的带宽就越小，并且该带宽内的混叠层数就越大。
If you don't want aliasing, the sampling rate should be above twice the start of the high frequency stop-band of whatever filter you can find to use before the sampler.
The physics of the environment in which the waveform resides will often provide its own low-pass or band-pass filtering (propagation loses, transducer loses, etc.). If you can determine the cutoff of this natural filter such that its stop-band is below your desired noise floor, you can use greater than 2X this frequency.
If you don't care about aliasing, then any sample rate will do, but the lower the frequency, the smaller the bandwidth of the result and the greater the multiple of layers of aliasing within that bandwidth.
我会简单地以不同的采样率聆听 wav 文件，并寻找听起来自然的采样率。
I would simply listen to the wav file at different sampling rates and look for the sampling rate at which it sounds natural.