0
$\begingroup$

Is there a name for the result of summing the bins of a DFT?

I don't mean to sum the squares of the bins, but to simply add the magnitude of the frequency bins together to get a single result.

Is there a special name that describes how this result would relate to the signal in the time-domain?

The best I can think of would be the total power across the DFT bandwidth, but I believe that is calculated by taking the root of the sum of the squares of the bins, per Parseval's theorem.

$\endgroup$
1
  • 1
    $\begingroup$ I'm afraid you'll be disappointed. $\endgroup$ Feb 21 '19 at 23:19
2
$\begingroup$

It has the name "first element of the time domain vector times the length of the vector".

reason:

With the DFT of a $N$ long sequence $y[n]$ being

$$Y[k] = \sum_{n=0}^{N -1}y[n]e^{-j2\pi k\frac{n}N}\text,$$

your sum of FFT bins is

\begin{align} \sum_{k=0}^{N -1}Y[k] &= \sum_{k=0}^{N -1}\sum_{n=0}^{N -1}y[n]e^{-j2\pi k\frac{n}N}&\text{finite sums can be exchanged in order}\\ &=\sum_{n=0}^{N -1}\sum_{k=0}^{N -1}y[n]e^{-j2\pi k\frac{n}N} &\text{$y[n]$ is constant over all $k$}\\ &=\sum_{n=0}^{N -1}y[n]\underbrace{\sum_{k=0}^{N -1}e^{-j2\pi k\frac{n}N}}_{=\begin{cases}N\cdot1 &n=0\\0&\text{else}\end{cases}}& \text{sum-orthonogonality of the DFT}\\ &=y[0]N \end{align}

$\endgroup$
1
  • $\begingroup$ Thank you Marcus. This would explain the result I'm getting. $\endgroup$
    – Pol DSP
    Feb 21 '19 at 23:44

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.