The 2024 Developer Survey results are live! See the results

# Tag Info

### Algorithm to "serialize" impulse responses

You convolve the two impulse responses. $$h_{cascade}[n] = h_1[n]*h_2[n]$$
• 46.4k

### How can the quality of a resampling algorithm be measured?

which resampling algorithm should I look at? Polyphase FIR is the way to go. It's actually pretty east to implement from scratch. Much easier than anything FFT related. And you get a lot of knobs to ...
• 46.4k
Accepted

### Why do we have negative sign dB in spectrum in audacity?

The audio amplitude of 100% is full scale and that will be 0 dB in comparison to the 100% full scale. That's the amplitude of 1.0 in Audacity. As your audio amplitude is not full scale 100% but less, ...
• 2,321

### How to reduce clicking in an audio signal with seek jumps?

Clicking noise on seek is a result of discontinuity in the audio signal. One way of reducing this noise is applying a fade-in effect for few frames post seek.
• 665

### How are the results of a fourier-transform interpolated to cover the entire frequency range?

This is coming from an audio developer (me): Most FFT analyzers in audio plugins are actually very simple. From the 1024-4096 frequency bins that are usually output by the FFT a 2D Path is created. ...

### What measures are used to identify a faulty speaker?

Faultiness in the sense of mechanical damage is commonly referred to as "Rub and Buzz" distortion. It can be caused by loose particles, rubbing of the voice coil on something else, voice ...
• 2,383
Accepted

### Create a minimum phase filter from an elliptic digital filter

The impulse response hmp is computed by convolving the impulse response of the elliptic filter with itself. We know that the elliptic filter is marginally minimum-...
• 90.7k

### What kind of distortion effect is this?

If a sine wave was going in, because of the asymmetric waveshape coming out, it can't be a straight memoryless waveshaper. The half wave coming out looks the impulse response of an overdamped or ...

• 6,555

### What books describe convolutional neural networks for DSP?

To my knowledge there might not be good books on it already. I'll assume you have enough background in classical DSP and CNNs ( and NNs in general): go read the papers that things like WaveRNN, ...
• 31.8k
Accepted

### How to merge audio better?

Your denoising algorithm seems to be nonlinear in what it is leaving behind. The leftover noise is not phase aligned with the original noise or the noise of the next chunk. So, by overlapping and ...
• 2,383
Accepted

### How to identify the middle of the first trough in a waveform?

Compute a moving average or a moving median $p_a[n]$ of the signal power, with a long enough window that it doesn’t get affected by spurious dips in power. Find the point at which $p_a[n]$ dips below ...
• 6,555
Accepted

### Mixer and control systems

If we define "frequency domain" as the Fourier Transform: multiplication in one domain is equivalent to convolution in the other domain. So multiplication in time is the same as convolution ...
• 46.4k

### Why will the output be imprinted with delay effect or echo after convolving two direct dry signals' FFT values (convolution filtering)

The correct way of doing this is overlap add. Neither version of your code looks remotely like a correct overlap add implementation, so chances both are wrong. The proper way to determine this is to ...
• 46.4k

### Why will the output be imprinted with delay effect or echo after convolving two direct dry signals' FFT values (convolution filtering)

The difference in the behavior of the two methods, particularly the delay or echo effect in the first method, is primarily due to the way the kernel is being used in both methods. In your first method,...
• 1,583
Accepted

### Audio EQ Cookbook Stability

Do I need to be concerned about stability? Yes, but not for the reasons you think (we'll get to this). Or are the coefficients that come out guaranteed to keep all poles inside the unit circle? Yes,...
• 46.4k
Accepted

### Laplace transform of this simple parallel RLC circuit? (For audio speaker simulation ...)

Do what r b-j suggests: get a Laplace domain equivalent first, and then transform from $s$ to $z$. The equivalent circuit will be something like: \begin{align} H(s) &= R_1 + sL_1 + \frac{1}{\frac{...
• 25.9k