How can I generated an audio sequence from segmentes of an audio sample?

Hi. Please help me!

I need generated a sequence of audio from segments of an segmented sample audio. This sequence has not a limit of time. The duration of the sequence is determinated from the user.

I have a sample of 10 seconds, that was segmented in sub samples. This sub clips has to be reproduce in a sequence of undeterminated duration.

clear all clc close all

%load sample sample = miraudio('lluvia.wav');

%Descompose the sample in frames frames = mirframe(muestra, 0.032);

%Spectral Analysis spectrum = mirspectrum(frames);

%Analysis of timbre for each frame mfcc = mirmfcc(spectrum);

%Similarity of each frame disimilitud = mirsimatrix(mfcc, 'Dissimilarity'); similitud = mirsimatrix(disimilitud, 'Similarity', 'oneminus'); %Novelty curve of similarity novedad = mirnovelty(similitud); %Select peaks picos = mirpeaks(novedad); %segment sample segmentos = mirsegment('lluvia.wav',picos); %get data from mirobject t = get(segmentos,'Data'); D = t{:}; %store each segment in a variable s1 = D{1:1}; s2 = D{2:2}; s3 = D{3:3}; s4 = D{4:4}; s5 = D{5:5}; s6 = D{6:6}; s7 = D{7:7}; s8 = D{8:8}; vector = [s1; s2; s3; s4; s5; s6; s7; s8;];

[Puneet Rana]

Can you explain more about what you want to do in reference to the code that you have posted? Looks like you get a vector of samples in 'vector', but what do you want instead?

[Jan]

If you use the "{} Code" button to format your code, we could read it.

Answer 1

[Jan]

Do not create a bunch of variables with an index hidden in the name. Using indices as indices is much easier and better.

t = get(segmentos,'Data');

A "sequence of undeterminated duration" is not possible - how could a deterministric programming language create something undeterminated? Perhaps you want:

Result = cat(1, t{randi([1, length(t), 100)})

Now you get a concatenated vector of 100 random parts. If you want something else, explain this.