I have a sound sample, and by applying window length 0.015 and time step 0.005, I have extracted 12 MFCC features for 171 frames directly from the sample using a software tool called PRAAT.
Now I have all 12 MFCC coefficients for each frame. I want to process them further, making a 39-dimensional matrix by adding energy features and delta-delta features and applying dynamic time warping. How do I deal with the coefficients and how do I make delta-delta coefficients?
It depends on what you need these features. For example you can calculate the average value for each MFCC coefficient . Than you will have 13 features. To calculate delta (first derivative) or delta delta (second derivative) you have to calculate the rate of change and acceleration changes basing on MFCC. Than you can also calculate the averge.
You can try using a tool such as HTK, it automatically extracts MFCCs, delta, acceleration and energy coefficients
I think this will be a very nice link, appropriate to your question.
http://practicalcryptography.com/miscellaneous/machine-learning/guide-mel-frequency-cepstral-coefficients-mfccs/
Of course, as suggested by the others, HTK is commonly used for automatically doing this task.
http://practicalcryptography.com/miscellaneous/machine-learning/guide-mel-frequency-cepstral-coefficients-mfccs/#implementations