Automatic baseband (postprocessing) verification of multi-channel audio on MSP-M
Developed AP based test suite for baseband audio testing presents easy to run, adapt and expand automatic solution for complex and large in scope and scale audio tests with intensive frequency of repetition.
Main Features
- Test performed on all digital and analog interfaces
- Wide spectrum of measurements of the audio signal quality available by AP
- Run time and post run analysis possible
- Various mechanisms for definition of expected results, limits, and areas of interest
- Automatic verification of the signal quality based on previously set references and tolerances
- Powerful APWin library for automatic verification of the graph results obtained in measurements
- Simple graphic user interface (GUI)
- Support for special mode when measurements are stored as references and later used for verification of further measurements
- Possibility for simple implementation and integration of libraries for dedicated signal analysis (like steering const or crosstalk in DPLII)
- Results in Word documents
- Summary report available for the overall testing
Single value measurements
Single measurement refers to AP measurements that produce one value as a result, like DC component of the signal, THD, SNR and other values supported by Analog Precision device. Obtained values can be immediately compared against expected value given in advance in form of a referent values within areas of tolerance.
Complex measurements
Complex measurements refer to all tests that produce output which is indented to be post analyzed either by a tester who approves the correct behavior of DUT during performed test, or automatically by dedicated procedures. Results of these tests are usually presented in form of graphs. The suite implements various mechanisms for verification of graphic results thus decreasing role of a tester to a minimum.
Definition of pass/fail criterion
In single value measurement pass/fail criterion is simply stated as a range of acceptable result values.
In case of complex measurements, where it is required to post analyze mainly graphic output, pass/fail criterion defines in different ways depending on what needs to be verified. There are three categories of these analyses:
- Analysis which requires verification of all singular values to be within certain range, typical example THD measurement during frequency sweep where all singular values need to be below previously defined level, or FFT analysis which needs to isolate frequencies of expected strength, and the other parts of the spectra need to be below defined level.
- Analysis which requires calculation of certain parameters from captured graph, typical example DPLII tests, steering time constants and crosstalk.
- Analysis of various waveforms compared against a referent waveform. For this purpose, dedicated procedures are developed depending on the shape of the waveform.
In the above example, during transition period around t0 point, it is allowed that signal varies in amplitude around certain value, while it is more important that after a transition period, measured signal behaves very close to a reference. For these purposes, a slope verification procedure which modifies a defined tolerance L to Lcomp according to an angle enclosed by a tangent on the slope in a given point and x axis, is implemented.
Example 2
In the above example, verification of the graph output can be performed by comparison of the two envelopes, of the captured and referent graphs. The envelopes are obtained by filtering of the absolute values of the captured signals.
Despite developed methods and procedures, there are always some graphs that need to be verified by a tester. Typical examples are captures of noise. The intention is to keep the amount of such tests which require assistance of a tester to a minimum.
Working with Word
Work in Word is single document based VBA for APWin. Word documents are prepared as templates, more or less complex depending on the way the results need to be presented. The concept is based on the substitution of the key words from the template by actual numeric results or graphics. Each Word result document is usually used for presentation of results of a given feature or module under test. A Word summary overview is also created for a quick overview of the overall results and the course of the testing.
Test procedures for Dolby Pro Logic II
Reference Level/Cardinal Point separation
- level checks against limits set according to expectation
- measured DC component of the signals, THD+N, and S/N ratio in each test
Center Width Control
- level checks against limits set according to expectation
Dimension Control
- level checks against limits set according to expectation
Panorama Mode Characteristics
- level checks against limits set according to expectation
Auto-balance Characteristics
- level checks against limits set in a form of slope deviation from a reference
Surround Output Peak Level
- level checks against limits set in a form of slope deviation from a reference
Maximum Crosstalk Pan
- level checks against limits set in a form of slope deviation from a reference
- measured crosstalk from the graph
Steering Time Constants
- steering burst shape checked by "slope comparison" of its envelope with the envelope of the referent steering burst record
- decay or steering constant value calculated from the graph of the burst
Frequency response
- level checks against absolute limits around expected value
Output configuration
- level check against absolute limits around expected value
Surround channel filter characteristics
- level checks against limits set in a form of slope deviation from a reference
Surround channel THD+N
- level checks against limits set in a form of slope deviation from a reference
Dynamic Range
- level check against absolute limits around expected value
Surround Channel Delay
- shapes not compared automatically
- delay calculated from the graph
Surround Channel Polarity
- level checks against limits set in a form of slope deviation from a reference
- phase offset between Ls and Rs calculated from the graph
| < Prev |
|---|