Giter VIP home page Giter VIP logo

auxiva-iss's People

Contributors

d-kitamura avatar

Stargazers

 avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar

Watchers

 avatar

auxiva-iss's Issues

Why is the speech signal not convolutional mixing in main. m

Hello, d-kitamura, thanks for your amazing work!
I have just started exploring BSS and I have a question. Why is the signal not convolutional mixing in main. m? The reason why I raised this question is because I used RIR simulation to convolve mixed speech signals, but the performance of IVA was very poor. I want to know why the performance is so poor. Can you provide some suggestions?
following is the matlab code:

roomDimensions = [7 5 2.75]; 
microLocation = [3 1 1.5; 3.5 1 1.5; 4 1 1.5];
sourceLocation = [2 3.5 1.5; 4 3.5 1.5; 6 3.5 1.5];      

c = 343;                    % Sound velocity (m/s)
fs = 44100;                 % Sample frequency (samples/s)
beta = 0;                 % Reverberation time (s)
n = 4096;                   % Number of samples
mtype = 'omnidirectional';  % Type of microphone
order = -1;                 % -1 equals maximum reflection order!
dim = 3;                    % Room dimension
orientation = 0;            % Microphone orientation (rad)
hp_filter = 1;              % Enable high-pass filter
h = rir_generator(c, fs, microLocation, sourceLocation, roomDimensions, beta, n, mtype, order, dim, orientation, hp_filter);

mixSig(:,1) = conv(srcSig(:,1),h(1,:), 'same') + conv(srcSig(:,2),h(1,:), 'same') + conv(srcSig(:,3),h(1,:), 'same');
mixSig(:,2) = conv(srcSig(:,1),h(2,:), 'same') + conv(srcSig(:,2),h(2,:), 'same') + conv(srcSig(:,3),h(2,:), 'same');
mixSig(:,3) = conv(srcSig(:,1),h(3,:), 'same') + conv(srcSig(:,2),h(3,:), 'same') + conv(srcSig(:,3),h(3,:), 'same');

%% using AuxIVA
% Set parameters
seed = 1; % pseudo random seed
refMic = 1; % reference microphone for back projection
nSrc = 3; % number of sources
fftSize = 4096; % window length in STFT [points]
shiftSize = 2048; % shift length in STFT [points]
windowType = "hamming"; % window function used in STFT
nIter = 30; % number of iterations (define by checking convergence behavior with drawConv=true)
applyWhitening = false; % true or false (true: apply whitening to the observed multichannel spectrograms)
drawConv = true; % true or false (true: plot cost function values in each iteration and show convergence behavior, false: faster and do not plot cost function values)
% Fix random seed
RandStream.setGlobalStream(RandStream('mt19937ar','Seed',seed))
% Blind source separation based on AuxIVA-ISS
[estSig_auxiva, cost] = AuxIVAISS(mixSig, nSrc, targetFs, fftSize, shiftSize, windowType, nIter, refMic, applyWhitening, drawConv);

implicit expansion on matlab script

hi, kitamura,

Thanks for your sharing on BSS scripts which help me a lot on understanding BSS algorithms.
Here is a question.
There are many implicit expansions at matlab script, where I counter error when running it at Matlab R2013a.
for example, in AuxIVAISS.m
YY = Y .* conj(Y(n,:,:)); % N x J x I, using implicit expansion (NxJxI .* 1xJxI)

  1. do you use other version of Matlab, so you can run the scripts without error?
  2. I'm not sure what this multiplication with expansion really mean?

Can it be used to separate music?

Hello, thank you very much for sharing. I saw that there are two music signals in your input folder. As far as I know, its objective function is based on speech statistical models such as Laplacian distribution. So, is it applicable to music signals ?

How to change AuxIVA-ISS to online version

Hi d-kitamura,
The AuxIVA-ISS you provided is very helpful. When I tried to make this program into an online version, I found it couldn’t converge, it is no problem with AuxIVA-IP. Is this related to the update method?

Recommend Projects

  • React photo React

    A declarative, efficient, and flexible JavaScript library for building user interfaces.

  • Vue.js photo Vue.js

    🖖 Vue.js is a progressive, incrementally-adoptable JavaScript framework for building UI on the web.

  • Typescript photo Typescript

    TypeScript is a superset of JavaScript that compiles to clean JavaScript output.

  • TensorFlow photo TensorFlow

    An Open Source Machine Learning Framework for Everyone

  • Django photo Django

    The Web framework for perfectionists with deadlines.

  • D3 photo D3

    Bring data to life with SVG, Canvas and HTML. 📊📈🎉

Recommend Topics

  • javascript

    JavaScript (JS) is a lightweight interpreted programming language with first-class functions.

  • web

    Some thing interesting about web. New door for the world.

  • server

    A server is a program made to process requests and deliver data to clients.

  • Machine learning

    Machine learning is a way of modeling and interpreting data that allows a piece of software to respond intelligently.

  • Game

    Some thing interesting about game, make everyone happy.

Recommend Org

  • Facebook photo Facebook

    We are working to build community through open source technology. NB: members must have two-factor auth.

  • Microsoft photo Microsoft

    Open source projects and samples from Microsoft.

  • Google photo Google

    Google ❤️ Open Source for everyone.

  • D3 photo D3

    Data-Driven Documents codes.