Background
When developing software there is a great need for testing of all functionality and all use cases after each addition and change to the code. In order to save time for the developer during testing it is desired to get as much as possible to be carried out automatically. During automatic testing the developer is available for other tasks and the amount of time required to execute an automatic test should also be known which simplifies resource planning.
In some areas of software development it is hard to implement automated testing since the tests actually requires human supervision on some parts, e.g. graphics and audio are such areas where someone has to observe or at least analyze the result.
Problem statement
In software (SW) development every change to the code has to be tested thoroughly before being released and since large scale tests and long duration tests requires a large amount of resources. As the scope for testing the software prior to release has to be limited since a resource is required for monitoring the tests, in the case of audio testing it usually requires someone to listening to the output.
By replacing the human resource with a computer with an electric ear and some rules for judging whether the sound quality was good enough or not more tests can be executed and also freeing up resources for other work. In order to execute the correct test case some sort of communication between the device under test (DUT) and the test suite SW has to be established.
Objective
Sample output from the DUT and analyze samples in real-time or after test case completion and achieve a benchmark on the audio quality.
Aid the software developing process by simplifying software testing.
Suggested scope
- Implement system for acquiring audio data from DUT.
- Present different ways of communicating with DUT, but only implement
one.
- Decide which is preferable acoustic or electrical connection to DUT,
electrical connection could be achieve by using the signal that is
feeding for example a loudspeaker.
- Set up a set of rules for deciding when the audio quality is good enough to pass a test.
- Is it possible to measure and judge audio quality for any type of signal or will there be requirements on the test content?
- Create or locate a base of test content that can be used in generic audio testing.
Candidate competencies
- A strong knowledge in SW development under Linux or Windows.
- Knowledge of audio signal processing, mainly in the digital domain.
- Knowledge of statstics for setting up audio quality judgement rules.
Suggested courses
- Signal Theory.
- Digital Signal Processing.
- Project Course in Signal Processing and Digital Communications.
Useful courses
- Sound Perception.
- Pattern Recognition.