% SUpDEq - Spatial Upsampling by Directional Equalization
function [HRTF_L, HRTF_R] = supdeq_getArbHRTF(HRIRs_sfd,samplingGrid,mode,channel,transformCore)
This function returns a HRTF (separately for the left/right ear) for an
arbitrary direction, based on interpolation in the SH-domain and inverse
spherical Fourier transform.
Output:
HRTF_L / R - Result of ISFT in Fourier domain (single-sided complex
spectrum). HRTF for L/R channel at each spatial sampling point
Input:
HRIRs_sfd - Struct with HRIRs in spatial Fourier domain (sfd) / SH-domain
samplingGrid - Spatial sampling grid (Q x 2 matrix), where the first
column holds the azimuth and the second the
elevation (both in degree).
Azimuth in degree (0=front, 90=left, 180=back, 270=right)
(0 points to positive x-axis, 90 to positive y-axis)
Elevations in degree (0=North Pole, 90=front, 180=South Pole)
(0 points to positive z-axis, 180 to negative z-axis)
mode - 'DEG' or 'RAD' (radiant or degree).
Default: 'DEG'
channel - Integer with 0, 1, or 2
0 - Only left channel
1 - Only right channel
Default: 2 - Both channels (stereo)
transformCore - String to define method to be used for the inverse
spherical Fourier transform.
'sofia - sofia_itc from SOFiA toolbox
'ak' - AKisht from AKtools
The results are exactly the same, but AKisht is faster
with big sampling grids
Default: 'sofia'
Dependencies: SOFiA toolbox, AKtools
Reference:
Bernschütz, B. (2016). Microphone Arrays and Sound Field Decomposition
for Dynamic Binaural Recording. TU Berlin.
(C) 2018 by JMA, Johannes M. Arend
TH Köln - University of Applied Sciences
Institute of Communications Engineering
Department of Acoustics and Audio Signal Processing0001 %% SUpDEq - Spatial Upsampling by Directional Equalization 0002 % 0003 % function [HRTF_L, HRTF_R] = supdeq_getArbHRTF(HRIRs_sfd,samplingGrid,mode,channel,transformCore) 0004 % 0005 % This function returns a HRTF (separately for the left/right ear) for an 0006 % arbitrary direction, based on interpolation in the SH-domain and inverse 0007 % spherical Fourier transform. 0008 % 0009 % Output: 0010 % HRTF_L / R - Result of ISFT in Fourier domain (single-sided complex 0011 % spectrum). HRTF for L/R channel at each spatial sampling point 0012 % 0013 % Input: 0014 % HRIRs_sfd - Struct with HRIRs in spatial Fourier domain (sfd) / SH-domain 0015 % samplingGrid - Spatial sampling grid (Q x 2 matrix), where the first 0016 % column holds the azimuth and the second the 0017 % elevation (both in degree). 0018 % Azimuth in degree (0=front, 90=left, 180=back, 270=right) 0019 % (0 points to positive x-axis, 90 to positive y-axis) 0020 % Elevations in degree (0=North Pole, 90=front, 180=South Pole) 0021 % (0 points to positive z-axis, 180 to negative z-axis) 0022 % mode - 'DEG' or 'RAD' (radiant or degree). 0023 % Default: 'DEG' 0024 % channel - Integer with 0, 1, or 2 0025 % 0 - Only left channel 0026 % 1 - Only right channel 0027 % Default: 2 - Both channels (stereo) 0028 % transformCore - String to define method to be used for the inverse 0029 % spherical Fourier transform. 0030 % 'sofia - sofia_itc from SOFiA toolbox 0031 % 'ak' - AKisht from AKtools 0032 % The results are exactly the same, but AKisht is faster 0033 % with big sampling grids 0034 % Default: 'sofia' 0035 % 0036 % Dependencies: SOFiA toolbox, AKtools 0037 % 0038 % Reference: 0039 % Bernschütz, B. (2016). Microphone Arrays and Sound Field Decomposition 0040 % for Dynamic Binaural Recording. TU Berlin. 0041 % 0042 % (C) 2018 by JMA, Johannes M. Arend 0043 % TH Köln - University of Applied Sciences 0044 % Institute of Communications Engineering 0045 % Department of Acoustics and Audio Signal Processing 0046 0047 function [HRTF_L, HRTF_R] = supdeq_getArbHRTF(HRIRs_sfd,samplingGrid,mode,channel,transformCore) 0048 0049 if nargin < 3 || isempty(mode) 0050 mode = 'DEG'; 0051 end 0052 0053 if nargin < 4 || isempty(channel) 0054 channel = 2; 0055 end 0056 0057 if nargin < 5 || isempty(transformCore) 0058 transformCore = 'sofia'; 0059 end 0060 0061 %Get angles 0062 az = samplingGrid(:,1); 0063 el = samplingGrid(:,2); 0064 0065 %SOFiA needs RAD 0066 if strcmp(transformCore,'sofia') 0067 if strcmp(mode,'DEG') 0068 az = az*pi/180; 0069 el = el*pi/180; 0070 end 0071 end 0072 0073 %AK needs DEG 0074 if strcmp(transformCore,'ak') 0075 if strcmp(mode,'RAD') 0076 az = az*180/pi; 0077 el = el*180/pi; 0078 end 0079 end 0080 0081 %Check range 0082 if sum(az < 0) > 0 || sum(el < 0) > 0 0083 error('Only positive azimuth/elevation values allowed!'); 0084 end 0085 0086 %% Perform transform with sofia_itc 0087 0088 if strcmp(transformCore,'sofia') 0089 %Perform inverse spherical Fourier transform 0090 if channel == 0 %Only left channel 0091 0092 HRTF_L = sofia_itc(HRIRs_sfd.Hl_nm, [az el]); 0093 HRTF_R = nan; 0094 0095 elseif channel == 1 %Only right channel 0096 0097 HRTF_L = nan; 0098 HRTF_R = sofia_itc(HRIRs_sfd.Hr_nm, [az el]); 0099 0100 elseif channel == 2 %Default - Stereo 0101 0102 HRTF_L = sofia_itc(HRIRs_sfd.Hl_nm, [az el]); 0103 HRTF_R = sofia_itc(HRIRs_sfd.Hr_nm, [az el]); 0104 0105 end 0106 end 0107 0108 %% Perform transform with AKisht 0109 0110 if strcmp(transformCore,'ak') 0111 %Perform inverse spherical Fourier transform 0112 if channel == 0 %Only left channel 0113 0114 HRTF_L = AKisht(HRIRs_sfd.Hl_nm,false,[az el],'complex'); 0115 HRTF_L = HRTF_L.'; 0116 HRTF_R = nan; 0117 0118 elseif channel == 1 %Only right channel 0119 0120 HRTF_L = nan; 0121 HRTF_R = AKisht(HRIRs_sfd.Hr_nm,false,[az el],'complex'); 0122 HRTF_R = HRTF_R.'; 0123 0124 elseif channel == 2 %Default - Stereo 0125 0126 HRTF_L = AKisht(HRIRs_sfd.Hl_nm,false,[az el],'complex'); 0127 HRTF_L = HRTF_L.'; 0128 HRTF_R = AKisht(HRIRs_sfd.Hr_nm,false,[az el],'complex'); 0129 HRTF_R = HRTF_R.'; 0130 0131 end 0132 end 0133 0134 end 0135 0136 0137 0138