k-Wave

1. mrtm3050
   

   Member
   Joined: May '12
   Posts: 2

   Hi,

   I'm new to k-Wave so forgive my naive questions. Perhaps this has been answered many times, but I was wondering if there is any code available to show how you steer a 2D planar (or cylindrical) array very similar to the 1D case:

   http://www.k-wave.org/documentation/example_tvsp_steering_linear_array.php

   In particular, is it possible using k-Wave to steer a beam given that the input is a 3D point, say in a 1000, 1000, 1000 volume, the point that I would like to steer the beam to is at (400, 400, 400).

   Posted 12 years ago #

2. mrtm3050
   

   Member
   Joined: May '12
   Posts: 2

   Just wanted to add what I'm trying to do:

   I'm trying to simulate a 2D phase array of transducers that will emit ultrasonic waves in air. I have to form a tight beam by having an array of transducers as close as possible. Does k-Wave anything model for the non-linearity of the air (this is not completely necessary). Due to the non-linearity of the air, the air will cause the sum and difference of the ultrasonic frequencies transmitter to change, and in this way, we will be able to demodulate the audio signal after a certain distance. I would like this beam to focused at a specific point in 3D space.

   Posted 12 years ago #

3. Bradley Treeby
   

   Developer
   Joined: Mar '10
   Posts: 1,643

   Hi mrtm3050,

   Thanks for your question. You can certainly use k-Wave to model a 2D phased array. One simple approach would be to define a 2D binary source mask within a 3D grid, and then for each source element, delay the corresponding time series by the appropriate offset (for example, based on the distance to the focus point and the speed of sound in the medium). You may have to be careful with how the 2D array and the input time series are indexed.

   You should note that a simulation using a grid size of 1000 x 1000 x 1000 will likely take a very long time to compute and also consume a lot of memory (> 100 GB). If you get some working code and would like to try bigger examples, we have some compiled c++ code that links to MATLAB that you would be welcome to try. This is faster and more memory efficient than the pure MATLAB code.

   At the moment, k-Wave can model convective and material nonlinearities (these are cumulative effects). Is this the sort of nonlinearity you mean?

   Brad.

   Posted 12 years ago #

4. jwjs
   

   Member
   Joined: Jun '11
   Posts: 14

   Hi Brad,

   I would be very interested in having access to some C++ for simulation of larger grids. Currently I run into memory issues for a lot of simulations I want to perform, so something a little bit more memory efficient (even at the sacrifice of time) would be beneficial.

   Posted 12 years ago #

5. Bradley Treeby
   

   Developer
   Joined: Mar '10
   Posts: 1,643

   Hi jwjs,

   We have a pre-release version of compiled C++ code (for 64-bit linux) packaged as a "patch" for use with B.0.5. Would you be interested in trying this out?

   Brad.

   Posted 12 years ago #

6. jwjs
   

   Member
   Joined: Jun '11
   Posts: 14

   Yes I would be interested in trying this out, but I'm not running Linux unfortunately. Why the dependency? I'm guessing you're generating mex files for the backend computation? Has it been tried on OSX?

   Posted 12 years ago #

7. Bradley Treeby
   

   Developer
   Joined: Mar '10
   Posts: 1,643

   Unfortunately it's restricted to Linux at the moment due to compiler dependencies (GCC 4.4.X). We are working hard on a release version of the code at the moment that should hopefully be more general. It should be ready by October at the latest.

   Posted 12 years ago #

8. jwjs
   

   Member
   Joined: Jun '11
   Posts: 14

   There is of course the option of compiling against the GNU compiler using MinGW. I seamlessly build "linux" flavored code on Windows using the above approach quite regularly. If you are willing I would like to give it a try. That is, building and running this on a Windows machine under MinGW.

   Posted 12 years ago #

9. Bradley Treeby
   

   Developer
   Joined: Mar '10
   Posts: 1,643

   Thanks for the offer, it's certainly appreciated. We are tidying up the OpenMP C++ source code at the moment ready for the first open-source release. It's not quite ready yet (we are still working on finalising the i/o file-format, input options, etc), but I'll definitely get in touch closer to release date if we're having trouble with windows makefiles. The plan at the moment is to distribute binaries for both 64-bit Linux and Windows along with the source codes and makefiles.

   Posted 12 years ago #

10. Chesteta
    

    Member
    Joined: Mar '11
    Posts: 4

    I was wondering if there has been any progress on the windows code? Please 'add me to the list' of interested parties :) Thank you! (looking for Win x64)

    Posted 12 years ago #

11. Bradley Treeby
    

    Developer
    Joined: Mar '10
    Posts: 1,643

    Hi Chesteta,

    We have now created both Linux and Windows versions. The compiled binaries along with the source codes will be released with the next version of k-Wave, which should be within the next few weeks.

    Brad.

    Posted 12 years ago #

RSS feed for this topic

Reply

You must log in to post.