k-Wave

1. ffarias
   

   Member
   Joined: May '14
   Posts: 9

   Hello,
   I have been working on simulating a parametric loudspeaker, so I should expect a sound wave from 2 ultrasonic waves (the difference frequency) and also the harmonics of the original waves and the sum of them. I made the grid so it allows frequencies over the harmonics, but I can't get them clear.
   I have two frequencies in the source (40 and 42kHz), so every time I got a 2 kHz wave which is what I expect, but I don't get 80 ,82 and 84 kHz waves.

   I used a sampling frequency of 192 kHz to generate the toneburst (and to graph the FFT later), but I don't get any harmonic, instead I get a 0 Hz component.
   Then, as I thought I didn't have any harmonics, I used 96 kHz for the toneburst but then I got 14 kHz and 12 kHz which are clearly the aliased harmonics (I used different frequencies in other cases, but I got similar results including the third expectid high frequency wave, so I am sure these are the aliased harmonics).

   In both cases the grid allows up to 90 kHz, and the original signals (40 and 42 kHz) and the difference (2kHz) are ok, so it's just about the harmonics.

   Apparently this is a problem about the sampling frequency, should I change it? If so, which one chould I choose?

   Thanks in advance.

   Felipe.

   Posted 10 years ago #

2. Bradley Treeby
   

   Developer
   Joined: Mar '10
   Posts: 1,643

   Hi Felipe,

   There are two sampling frequencies to define, one is spatial and the other is temporal. In most cases, due to the CFL condition, the temporal sampling frequency will be higher than the equivalent spatial sampling frequency. To avoid trying to input frequency content from time varying sources that cannot be represented spatially on the grid, I'd suggest using the function filterTimeSeries. If you use 96 kHz for your toneburst, and the grid only supports 90 kHz, then you will see aliasing. I would suggest increasing the maximum frequency supported by the grid by reducing the size of dx/dy.

   Hope that helps,

   Brad.

   Posted 10 years ago #

3. ffarias
   

   Member
   Joined: May '14
   Posts: 9

   I got some low frequency components when I used 192 kHz as sampling freq. but no when using 96 kHz. Apparently it was always correct, I compared it to actual measurements.
   Thanks for your answer,

   Felipe.

   Posted 10 years ago #

RSS feed for this topic

Reply

You must log in to post.