# tensorplay.signal

Source: https://www.tensorplay.cn/docs/signal.html

Window functions for spectral analysis and filter design. A window is a
length-M sequence of weights that tapers a finite segment of a signal,
reducing the spectral leakage that comes from cutting the segment off sharply.

All window functions share a common shape of keyword arguments:

- M — the window length, a positive integer.

- sym — whether the window is symmetric (True, the default) or periodic (False). Symmetric windows are used in filter design; periodic windows are used for spectral analysis, where the window is applied to one period of a sampled signal and the first and last samples are not duplicates.

- dtype, layout, device, requires_grad — the usual tensor construction arguments, so window tensors are created directly on the desired device with the desired data type.

```
import tensorplay as tp
from tensorplay.signal import windows

print(windows.hann(8))
print(windows.blackman(8, sym=False))
# window tensors are ordinary tensors, so they can be moved or differentiated
w = windows.kaiser(64, beta=8.0)
```

## Window functions

| tensorplay.signal.windows.bartlett |Computes the Bartlett window. |
| --- | --- |
| tensorplay.signal.windows.blackman | Computes the Blackman window. |
| tensorplay.signal.windows.cosine | Computes a window with a simple cosine waveform, also known as the sine window. |
| tensorplay.signal.windows.exponential | Computes a window with an exponentially decaying waveform, also known as the Poisson window. |
| tensorplay.signal.windows.gaussian | Computes a window with a Gaussian waveform. |
| tensorplay.signal.windows.general_cosine | Computes the general cosine window, a weighted sum of cosines whose frequencies are integer multiples of the fundamental. |
| tensorplay.signal.windows.general_hamming | Computes the general Hamming window, the two-term member of the general cosine family. |
| tensorplay.signal.windows.hamming | Computes the Hamming window. |
| tensorplay.signal.windows.hann | Computes the Hann window. |
| tensorplay.signal.windows.kaiser | Computes the Kaiser window. |
| tensorplay.signal.windows.nuttall | Computes the minimum 4-term Blackman-Harris window described by Nuttall. |

## Where to go next

- [FFT functions](/docs/fft.html) — windowing is typically a preamble to a spectral transform; the FFT page covers the transform side.

- [Tensors](/docs/tensorplay.html) — the tensor APIs the window functions are built on.