http://www.k-wave.org/forum/topic/early-artifact-in-3d-transducer-simulation Support for the k-Wave MATLAB toolbox en-US Wed, 13 May 2026 01:31:09 +0000 http://bbpress.org/?v=1.0.2 q http://www.k-wave.org/forum/search.php http://www.k-wave.org/forum/topic/early-artifact-in-3d-transducer-simulation#post-5059 Wed, 29 Apr 2015 16:12:18 +0000 Dean Douthat 5059@http://www.k-wave.org/forum/ <p>Thanks for the quick response.</p> <p>Following your guess, I increased the attenuation in PML and am able to reduce the artifacts to insignificance. Good catch! </p> http://www.k-wave.org/forum/topic/early-artifact-in-3d-transducer-simulation#post-5057 Wed, 29 Apr 2015 09:06:31 +0000 bencox 5057@http://www.k-wave.org/forum/ <p>Hi Dean, </p> <p>It's difficult to say what's causing it without seeing the whole simulation, but - just a guess - you have only 12 points in the z-direction, so maybe the PML absorbing boundary is not working very well, and the signal's wrapping around from the opposite side of the domain?</p> <p>Ben </p> http://www.k-wave.org/forum/topic/early-artifact-in-3d-transducer-simulation#post-5054 Tue, 28 Apr 2015 14:33:08 +0000 Dean Douthat 5054@http://www.k-wave.org/forum/ <p>I'm simulating a transducer used as sensor for photoacoustic signals in a 3D setup. Here is my setup:</p> <p> kWaveGrid with properties:</p> <p> Nx: 304.0000e+000<br /> Ny: 748.0000e+000<br /> Nz: 12.0000e+000<br /> dx: 170.0000e-006<br /> dy: 170.0000e-006<br /> dz: 170.0000e-006<br /> kx_vec: [304x1 double]<br /> ky_vec: [748x1 double]<br /> kz_vec: [12x1 double]<br /> k: [304x748x12 double]<br /> kx_max: 18.4800e+003<br /> ky_max: 18.4800e+003<br /> kz_max: 18.4800e+003<br /> k_max: 18.4800e+003<br /> dim: 3.0000e+000<br /> nonuniform: 0<br /> t_array: [1x1511 single]<br /> x: [304x748x12 double]<br /> y: [304x748x12 double]<br /> z: [304x748x12 double]<br /> kx: [304x748x12 double]<br /> ky: [304x748x12 double]<br /> kz: [304x748x12 double]<br /> x_vec: [304x1 double]<br /> y_vec: [748x1 double]<br /> z_vec: [12x1 double]<br /> kWaveTransducer with properties:</p> <p> number_elements: 128.0000e+000<br /> element_width: 5.0000e+000<br /> element_length: 6.0000e+000<br /> element_spacing: 0.0000e+000<br /> position: [6.0000e+000 55.0000e+000 4.0000e+000]<br /> radius: Inf<br /> active_elements: [128x1 double]<br /> elevation_focus_distance: Inf<br /> transmit_apodization: 'Rectangular'<br /> receive_apodization: 'Rectangular'<br /> sound_speed: 1.5289e+003<br /> focus_distance: Inf<br /> record_start_index: 1.0000e+000<br /> active_elements_mask: [304x748x12 uint8]<br /> indexed_active_elements_mask: [304x748x12 uint8]<br /> all_elements_mask: [304x748x12 uint8]<br /> indexed_elements_mask: [304x748x12 uint8]<br /> mask: [304x748x12 uint8]<br /> transducer_width: 640.0000e+000<br /> number_active_elements: 128.0000e+000<br /> steering_angle_max: 'auto'<br /> appended_zeros: 'auto'<br /> beamforming_delays_offset: 'auto'<br /> steering_angle: 0.0000e+000 x_size: 51.6800e-003<br /> z_size: 2.0400e-003<br /> y_size: 127.1600e-003<br /> dt: 33.3575e-009<br /> Nt: 1.5110e+003<br /> total_grid_points: 2.7287e+006</p> <p>medium = </p> <p> sound_speed: 1.5289e+003<br /> alpha_coeff: 750.0000e-003<br /> alpha_power: 1.5000e+000</p> <p>&gt;&gt; find(source.p0)</p> <p>ans =</p> <p> 1250594</p> <p>&gt;&gt; [x y z]=ind2sub(size(source.p0),find(source.p0))</p> <p>x =</p> <p> 242</p> <p>y =</p> <p> 374</p> <p>z =</p> <p> 6</p> <p>An initial value of 1 KPa is applied to the source grid point. The leading edge of the wave should arrive at the farthest sensor in ~43.5 usec and the time trace of that element indeed shows that pulse as expected with a peak ~0.3 Pa.</p> <p>But, there is another, small blip at ~36.1 usec with amplitude ~3.5% of the expected pulse.</p> <p>Am I doing something wrong? Is a source with only one grid point not allowed?</p> <p>I have run this both with/without GPU and in single and double precision all with the same results. </p>