SGN-24006 Analysis of Audio, Speech and Music Signals, 5 cr

Additional information

Suitable for postgraduate studies.

Person responsible

Annamaria Mesaros

Lessons

Implementation Period Person responsible Requirements
SGN-24006 2016-01 4 Annamaria Mesaros
Final exam and project work.

Learning Outcomes

After completing this course, the student will -be able to implement common mid-level data representations used in the analysis of audio signals. He or she will understand how structural regularities of audio signals can be modeled to facilitate their analysis. -be able to implement some some widely-used audio feature extraction techniques and signal analysis algorithms such as spectrogram factorization and multi-pitch analysis. -understand the basic techniques used in source separation and will be able to implement some source separation algorithm. -understand the basic techniques used in audio classification and speech recognition. He or she will be able to implement the front-end used for extracting relevant information from audio or speech signal and understand the the mathematical principles and application of hidden Markov models that are used to model the feature sequences.

Content

Content Core content Complementary knowledge Specialist knowledge
1. Mid-level representations of acoustic signals for their content analysis. Modelling of structural regularities of audio signals for analysis purposes.     
2. Source separation (one channel and multichannel). Spectrogram factorization and other unsupervised learning techniques. Pitch analysis and music transcription.     
3. Acoustic feature extraction and audio classification. Automatic speech recognition. Acoustic feature extraction and hidden Markov models.     

Instructions for students on how to achieve the learning outcomes

The course is marked based on the exam. The highest mark is given for correct answers that cover the depth and breadth delivered at the lectures and exercises. The threshold for passing the course is at about half of the maximum amount of points. Bonus points worth a maximum of one mark are given by active participation in weekly exercises. An acceptable project work has to be returned by the deadline.

Assessment scale:

Numerical evaluation scale (0-5)

Study material

Type Name Author ISBN URL Additional information Examination material
Book   Speech and Audio Signal Processing: Processing and Perception of Speech and Music   B. Gold, N. Morgan, D. Ellis         No   
Book   Spoken Language Processing   X. Huang, A. Acero, H.-W. Hon         No   
Lecture slides             Yes   
Online book   Lecture Notes for Audio Engineering   University of Illinois Urbana-Champaign         No   

Prerequisites

Course Mandatory/Advisable Description
SGN-13000 Johdatus hahmontunnistukseen ja koneoppimiseen Advisable   1
SGN-13006 Introduction to Pattern Recognition and Machine Learning Advisable   1
SGN-14006 Audio and Speech Processing Mandatory    

1 . Either SGN-13000 or SGN-13006 is advisable.



Correspondence of content

Course Corresponds course  Description 
SGN-24006 Analysis of Audio, Speech and Music Signals, 5 cr SGN-4106 Speech Recognition, 5 cr +
SGN-4227 Digital Audio Processing and Analysis, 6 cr
 
SGN-24006 Analysis of Audio, Speech and Music Signals, 5 cr SGN-24007 Advanced Audio Processing, 5 cr  

Updated by: Värri Alpo, 08.04.2016