From what I have learned, I think you can use SVM for two or three dimension vectors for classification (common cases). Meaning you can use a combination of features (i.e. five vectors of two dimensions containing [Volume,Pitch] or a more important characteristic or three dimensional [Volume, Pitch, MFCC] vectors or a combination by applying dimension reduction by PCA or other methodology (i.e. mean, sd,...), you will need to find a suitable set of features). You will also need to have a vector containing the labels of your samples (say 'high frequency', 'low frequency' or the characteristic you would like to categorize your samples). You can download and use the support vector machine toolbox for MATLAB:

https://sourceforge.net/projects/svmclassmatlab/

In the case you have two dimensional vectors and you have a vector xdata containing your training sample features and a vector group containing your labels, you will have to train your classifier first:

svmStruct = svmtrain(xdata ,group,'ShowPlot',true);

Then, in order to test the accuracy of the classifier, you can define a set of vectors with the same dimension and characteristics from your training samples:

test_vector; %Assume two samples for test purposes

test_labels= svmclassify(svmStruct, test_vector, 'ShowPlot', true) %Classifying two test vectors

plot(test_vector(:1), test_vector(:2), 'ro', 'MarkerSize', 12); %Plotting the results in previous graph (Assuming hold on)

Again, your classification might improve depending on the search of good features but also on the number of samples you can use (you can create clusters or neighborhoods, I think the number of samples is very small for a good classification, but the way you use the features might create a good division for the characteristic you would like to use for classification). Also, you will need to think on test samples (or use three for classification, two for test, or 4:1, up to you..)

I hope this was useful.

P.S. This site can be used for further reference

http://mathworks.com/help/stats/svmclassify.html