function [Obj,results]=ziegelwanger2014(Obj,estimation,outlierDetection,model,p0_onaxis)
%ZIEGELWANGER2014 Time of arrival estimates
%   Usage: [Obj,results]=ziegelwanger2014(data,estimation,outlierDetection,model,p0_onaxis) 
%
%   Input parameters:
%       Obj: SOFA object
% 
%       estimation: (optional) select one of the estimation methods (1: Threshold-Detection, 2: Centroid of squared IR, 3: Mean Groupdelay, 4: Minimal-Phase Cross-Correlation (Max) (default), [TOAest]: pre-estimated TOAs)
%
%       outlierDetection: (optional) detect outliers in estimated TOAs (0: off, 1: on (default values: [0.05;0.01]), [alpha r]: rejects outliers using the extreme Studentized deviance test with the significance level of ALPHA and upper bound of outlier rate R. )
%
%       model: (optional) correct estimated toa, using geometrical TOA-Model (0: TOA estimated, 1: off-axis TOA modeled (default), 2: on-axis TOA modeled)
%
%       p0_onaxis: (optional) startvalues for lsqcurvefit
% 
%   Output parameters:
%       Obj: SOFA Object
% 
%       results.toa: data matrix with time of arrival (TOA) for each impulse response (IR):
%       results.p_onaxis: estimated on-axis model-parameters
%       results.p_offaxis: estimated off-axis model-parameters
%
%   Estimates the Time-of-Arrival for each measurement in Obj (SOFA) and
%   corrects the results with a geometrical model of the head.
%
%   Requirements: 
%   -------------
%
%   1) SOFA API from http://sourceforge.net/projects/sofacoustics for Matlab (in e.g. thirdparty/SOFA)
% 
%   2) Optimization Toolbox for Matlab
%
%   3) Data in hrtf/ziegelwanger2014
%
%   Examples:
%   ---------
% 
%   To calculate the model parameters for the on-axis time-of-arrival model
%   (p_onaxis) and for the off-axis time-of-arrival model (p_offaxis) for a
%   given HRTF set (SOFA object, 'Obj') with the minimum-phase
%   cross-correlation estimation, use:
%
%       [Obj,results]=ziegelwanger2014(Obj,4,1);
%
%   See also: ziegelwanger2014_onaxis, ziegelwanger2014_offaxis,
%             data_ziegelwanger2014, exp_ziegelwanger2014
%
%   References:
%     H. Ziegelwanger and P. Majdak. Modeling the direction-continuous
%     time-of-arrival in head-related transfer functions. J. Acoust. Soc.
%     Am., 135:1278-1293, 2014.
%     
%
%   Url: http://amtoolbox.sourceforge.net/amt-0.9.9/doc/models/ziegelwanger2014.php

% Copyright (C) 2009-2015 Piotr Majdak and the AMT team.
% This file is part of Auditory Modeling Toolbox (AMT) version 0.9.9
%
% This program is free software: you can redistribute it and/or modify
% it under the terms of the GNU General Public License as published by
% the Free Software Foundation, either version 3 of the License, or
% (at your option) any later version.
%
% This program is distributed in the hope that it will be useful,
% but WITHOUT ANY WARRANTY; without even the implied warranty of
% MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
% GNU General Public License for more details.
%
% You should have received a copy of the GNU General Public License
% along with this program.  If not, see <http://www.gnu.org/licenses/>.

% AUTHOR: Harald Ziegelwanger, Acoustics Research Institute, Vienna,
% Austria

%% ----------------------------convert to SOFA-----------------------------
if ~isfield(Obj,'GLOBAL_Version')
    Obj=SOFAconvertARI2SOFA(Obj.hM,Obj.meta,Obj.stimPar);
end

%% ----------------------------check variables-----------------------------
if ~exist('estimation','var')
    estimation=4;
else if isempty(estimation)
        estimation=4;
    end
end

if ~exist('outlierDetection','var')
    outlierDetection=[0.05;0.01];
else if isempty(outlierDetection) || (isscalar(outlierDetection) && outlierDetection(1)>0)
        outlierDetection=[0.05;0.01];
    end
end
outlierDetection=prod(outlierDetection);

if ~exist('model','var')
    model=1e-6;
else if isempty(model) || model==1
        model=1e-6;
    end
end

if ~exist('p0_onaxis','var')
    p0_onaxis=[[0.0875; pi/2; 0; 0] [0.0875; -pi/2; 0; 0]];
else if isempty(p0_onaxis)
        p0_onaxis=[[0.0875; pi/2; 0; 0] [0.0875; -pi/2; 0; 0]];
    end
end

%% -------------------------initialize variables---------------------------
p0_onaxis=transpose(p0_onaxis);
p_onaxis=zeros(size(p0_onaxis));
p0_offaxis=zeros(2,7);
p_offaxis=p0_offaxis;

toa=zeros(Obj.API.M,Obj.API.R);
toa_onaxis=toa;
toa_offaxis=toa;
indicator=zeros(Obj.API.M,Obj.API.R);

pos(:,1:2)=Obj.SourcePosition(:,1:2);

%% -----------------------estimate time-of-arrival-------------------------
if isscalar(estimation)
    hM=Obj.Data.IR;
    
    toaEst=zeros(Obj.API.M,Obj.API.R);
    switch estimation
        case 1 %---------------------------Threshold---------------------------
            for ii=1:Obj.API.M
                for jj=1:Obj.API.R
                    toaEst(ii,jj)=find(abs(hM(ii,jj,:))==max(abs(hM(ii,jj,:))),1);
                end
            end
        case 2 %---------------------------Centroid----------------------------
            for ii=1:Obj.API.M
                for jj=1:Obj.API.R
                    toaEst(ii,jj)=find(cumsum(hM(ii,jj,:).^2)>(sum(hM(ii,jj,:).^2)/2),1);
                end
            end
        case 3 %---------------------------Groupdelay--------------------------
            for ii=1:Obj.API.M
                for jj=1:Obj.API.R
                    [Gd,F]=grpdelay(transpose(double(squeeze(hM(ii,jj,:)))),1,Obj.API.N*4,Obj.Data.SamplingRate*4);
                    toaEst(ii,jj)=mean(Gd(find(F>1000):find(F>5000)));
                end
            end
        case 4 %---------------------------Minimal-Phase-----------------------
            hMmin=ARI_MinimalPhase(Obj);
            corrcoeff=zeros(Obj.API.M,Obj.API.R);
            for ii=1:Obj.API.M
                for jj=1:Obj.API.R
                    [c,lag]=xcorr(squeeze(hM(ii,jj,:)),squeeze(hMmin(ii,jj,:)),Obj.API.N*4-1,'none');
                    [corrcoeff(ii,jj),idx]=max(abs(c));
                    corrcoeff(ii,jj)=corrcoeff(ii,jj)/sum(hM(ii,jj,:).^2);
                    toaEst(ii,jj)=lag(idx);
                end
            end
    end
else
    toaEst=estimation;
end

%% --------------------Detect-Outliers-in-estimated-TOA--------------------
if outlierDetection>0
    for ch=1:Obj.API.R
        p0_onaxis(ch,4)=min(toaEst(indicator(:,ch)==0,ch))/Obj.Data.SamplingRate;
        p0offset_onaxis=[0.06 pi pi/2 0.001];
        x=pos(:,1:2)*pi/180;
        y=toaEst(:,ch)/Obj.Data.SamplingRate;
        if isoctave
            fprintf('Sorry! Octave is not supported. This model requires MATLAB and the Optimization Toolbox!\n');
        else
            tmp=lsqcurvefit(@ziegelwanger2014_onaxis,p0_onaxis(ch,:),x,y,p0_onaxis(ch,:)-p0offset_onaxis,p0_onaxis(ch,:)+p0offset_onaxis,optimset('Display','off','TolFun',1e-6));
        end
        [~,idx]=deleteoutliers(toaEst(:,ch)-ziegelwanger2014_onaxis(tmp,pos(:,1:2)*pi/180)*Obj.Data.SamplingRate,outlierDetection*Obj.API.M);
        indicator(idx,ch)=ones(length(idx),1);
    end
end
    
%% ----------------------Fit-Models-to-estimated-TOA-----------------------
if model>0
    % Fit on-axis model to outlier adjusted set of estimated TOAs
    for ch=1:Obj.API.R
        p0_onaxis(ch,4)=min(toaEst(indicator(:,ch)==0,ch))/Obj.Data.SamplingRate;
        p0offset_onaxis=[0.06 pi pi/2 0.001];
        idx=find(indicator(:,ch)==0);
        x=pos(idx,1:2)*pi/180;
        y=toaEst(idx,ch)/Obj.Data.SamplingRate;
        if isoctave
            fprintf('Sorry! Octave is not supported. This model requires MATLAB and the Optimization Toolbox!\n');
        else
            [p_onaxis(ch,:),performance.on_axis{ch}.resnormS,performance.on_axis{ch}.residualS,performance.on_axis{ch}.exitflag,performance.on_axis{ch}.output]=...
                lsqcurvefit(@ziegelwanger2014_onaxis,p0_onaxis(ch,:),x,y,p0_onaxis(ch,:)-p0offset_onaxis,p0_onaxis(ch,:)+p0offset_onaxis,optimset('Display','off','TolFun',1e-6));
            toa(:,ch)=ziegelwanger2014_onaxis(p_onaxis(ch,:),pos(:,1:2)*pi/180)*Obj.Data.SamplingRate;
        end
        performance.on_axis{ch}.resnormS=sqrt(performance.on_axis{ch}.resnormS/(Obj.API.M-sum(indicator(:,ch))));
        performance.on_axis{ch}.resnormP=norm((toaEst(:,ch)-toa(:,ch))/Obj.Data.SamplingRate)/sqrt(Obj.API.M);
    end
    toa_onaxis=toa;

    % Fit off-axis model to outlier adjusted set of estimated TOAs
    if model~=2
        for ch=1:Obj.API.R
            idx=find(indicator(:,ch)==0);
            x=pos(idx,1:2)*pi/180;
            y=toaEst(idx,ch)/Obj.Data.SamplingRate;
            p0_offaxis(ch,:)=[mean(p_onaxis(:,1)) 0.001 -diff(p_onaxis(:,1))/2 0.001 mean(p_onaxis(:,4)) p_onaxis(ch,2) p_onaxis(ch,3)];
            p0offset_offaxis=[abs(diff(p_onaxis(:,1))/4) 0.1 0.1 0.1 0.001 pi/4 pi/4];
            if isoctave
                fprintf('Sorry! Octave is not supported. This model requires MATLAB and the Optimization Toolbox!\n');
            else
                [p_offaxis(ch,:),performance.off_axis{ch}.resnormS,performance.off_axis{ch}.residualS,performance.off_axis{ch}.exitflag,performance.off_axis{ch}.output]=...
                    lsqcurvefit(@ziegelwanger2014_offaxis,p0_offaxis(ch,:),x,y,p0_offaxis(ch,:)-p0offset_offaxis,p0_offaxis(ch,:)+p0offset_offaxis,optimset('Display','off','TolFun',model));
                toa(:,ch)=ziegelwanger2014_offaxis(p_offaxis(ch,:),pos(:,1:2)*pi/180)*Obj.Data.SamplingRate;
            end
            performance.off_axis{ch}.resnormS=sqrt(performance.off_axis{ch}.resnormS/(Obj.API.M-sum(indicator(:,ch))));
            performance.off_axis{ch}.resnormP=norm((toaEst(:,ch)-toa(:,ch))/Obj.Data.SamplingRate)/sqrt(Obj.API.M);
        end
        toa_offaxis=toa;
    end
else
    toa=toaEst;
    p_offaxis=p0_offaxis;
end

%Save to output variables
performance.outliers=indicator;
for ii=1:size(indicator,2)
  performance.outlierRate(ii)=sum(indicator(:,ii))/Obj.API.M*100;
end
results.toa=toa;
results.toaEst=toaEst;
results.toa_onaxis=toa_onaxis;
results.toa_offaxis=toa_offaxis;
results.p_onaxis=transpose(p_onaxis);
results.p_offaxis=transpose(p_offaxis);
results.performance=performance;
if exist('corrcoeff','var')
    results.performance.corrcoeff=corrcoeff;
end

end %of function

function hMmin=ARI_MinimalPhase(Obj)
    hM=Obj.Data.IR;
    hMmin=hM;

    for jj=1:Obj.API.R
        for ii=1:Obj.API.M
            h=[squeeze(hM(ii,jj,:)); zeros(4096*4-size(hM,3),1)];
            amp1=abs(fft(h));
            amp2=amp1;
            an2u=-imag(hilbert(log(amp1)));
            an2u=an2u(1:floor(length(h)/2)+1);
            an3u=[an2u; -flipud(an2u(2:end+mod(length(h),2)-1))];
            an3=an3u-round(an3u/2/pi)*2*pi;
            amp2=amp2(1:floor(length(h)/2)+1);
            amp3=[amp2; flipud(amp2(2:end+mod(length(h),2)-1))];
            h2=real(ifft(amp3.*exp(1i*an3)));
            hMmin(ii,jj,:)=h2(1:Obj.API.N);
        end
    end
end