Home

Audio Expert System Reference & User Guide

1. csv format for used in Microsoft Excel or any other Windows csv compatible application First select the events to export Multiple events are selectable by selecting an event then holding the Shift key while clicking on another event This will select all events between the two selections If the selections are not adjacent you can hold the Ctrl control key while clicking events Next right click anywhere int he Event Table to open the pop up menu and click on the Export Event Table option A Windows Save As dialog will open Enter a file name and select a file location and click on Save A confirmation dialog will open Click OK to close the confirmation dialog If you have not selected an event in the table before exporting a warning to Please select an event row first appears 3 3 Frame Packet and Protocol Analysis Synchronization The Audio Expert System module integrates seamlessly with ComProbe software with common timestamping of Bluetooth protocol data audio events audio waveform display and codec events The audio expert data and results are synchronized and coordinated with the existing ComProbe software data views such as Frame Display Bluetooth Timeline etc to expedite the root cause analysis of Bluetooth protocol related audio issues When a frame is selected in any ComProbe software data views the corresponding audio data associated with those frames is also selected in the Wave Panel Event Timeline and Event Table
2. frontline Debug Com Audio Expert System Reference amp User Guide Revision Date 11 2 2015 Audio Expert System Reference User Guide Copyright O 2015 Frontline Test Equipment Inc All rights reserved Frontline Frontline Test System ComProbe Protocol Analysis System and ComProbe are registered trademarks of Frontline Test Equipment Inc The Bluetooth SIG Inc owns the Bluetooth word mark and logos and any use of such marks by Frontline is under license All other trademarks and registered trademarks are property of their respective owners Audio Expert System Reference User Guide Contents Chapter 1 About thisGuide 2 2 2 eee nao 1 1 1 What You Will AA 1 Chapter 2 Audio Expert System Reference Guide _ __ __ _ 22 22 aoaaa aaa aaa aaao cnn 3 2 1 Audio Expert System Overview occ cs 3 2 1 1 Non Referenced Mode Event Reporting rro 3 2 1 1 1 Low Volume Alarm Event 2 occ cc eee cece eee eee ences 4 2 1 1 2 Clipping EVEN a tn dista 4 2 1 1 3 High Volume Alarm Event coco 5 2 1 1 4 Dropout Event ee ee eee eee eee eee cece aonana 6 A io lees du a ee 7 2 1 2 Referenced Mode Event Reporting 22 2 ieee cece ec cece eee eee eee eee annua 7 De A Test POUNCE CMG 225222 oe hee acosa cotton Posten sts 8 2 1 2 2 Test Script Not Found Event _ 2 22 22 eee ro 9 2 123 WONG test Scrip EVERY 2 5 aag ooo ori rt its 9 2 1 2 4 Synchronization Lost Event
3. 9 2 1 2 5 Unexpected Frequency Event _ 2 2 22 eee ce cee cece cece cece cece cece ee eeceeeeeeees 10 2 1 2 6 Unexpected Level Event _ 2 22 loo cee cc eee ee aLa aLaaa anaana 10 2 1 2 7 Unexpected Duration Event ciunccos ronca ciencia ire rat 10 2 1 2 8 Amplitude Fluctuation Event 11 2 1 2 9 Unexpected Phase Change Event a 11 2 1 2 10 Clipping Event _ 2 2 22 cee cece cee LaL LaLa LLa aLaaa aana 11 2 1 2 11 EXCESS NOISE EVENT cc makaawa esporas DISK ES BKA tea eccierttcenGenducociececdewusutene eee sa 12 2 1 2 12 CVSD HF Level Too High Event ooo cece cece eee eee e eens 12 212 13 End of Test Event cecosciccenissdesobaleecosieconaesecucccssitr Bebeesecnecseseazereestaceseene 13 2 1 3 Audio Expert System Event Type anaana cece cece cece eee cece eee eeeeee 13 2 1 3 1 Event Type Bluetooth Protocol eee cece cee cece cece cece eeceeeeee 13 2 1 3 2 Event Type COQ EC anak a es 14 2 1 3 3 Event Type Audio desig cc cop ccna rr DE AT DRLADLAD ULANG SYN ra ii oe os 16 2 1 4 Supported Codec Parameters ieee eee eee cece ee eee cece eee eee eeeeeeeeees 17 2 2 Referenced Mode Processing a asaan and Sine cierta lit li e ei asii 18 Audio Expert System Reference User Guide 2 2 1 Transition to Referenced Mode cc 18 AA O A A 19 2 2 3 Reference Audio File Structure coc 19 2 2 4 Synchronization _ Aa 19 2 2 5 Minimum and Optimal Amplitude Leve
4. Moves the play cursor to the end of the waveform 33 Audio Expert System Reference User Guide Chapter 3 Audio Expert System User Guide Table 3 2 Global Toolbar Controls continued ME JO OS IAN Loop Loops waveform playback continuously If the Play button is visible it will toggle to the Pause Clicking the Pause button will stop Loop playback Clicking on the Loop button will stop the loop and the playback If there is a selection on the waveform only the selection will loop Horizontal Zoom Out Increases the amount of data that is visible on the screen however less detail is discernable Horizontal Zoom In Decreases the amount of data that is visible on the screen however more detail is discernable Lock Unlock Operational in live mode only Selecting Lock will freeze the waveform display however the Audio Expert System will still continue to analysis new audio data Selecting Unlock will jump to the waveform end and then resume following the waveform Mute Mute will mute unmute audio playback for all Wave Panels Individual Wave Panel Mute control will override the Global Toolbar Mute for that panel only Volume Down Decreases the audio playback volume of all Wave Panels based on the current volume level setting for each individual Wave Panel Volume Down Decreases the audio playback volume of all Wave Panels based on the current volume level setting for each individual Wave Panel Ave
5. Referenced Mode are generally recorded with a maximum tone segment level of 3 dBFS although there are a few exceptions where signal levels may be as high as 1 dBFS Figure 2 10 Example Reference File Waveform 44 1kHz 16 bit Shown above is an exampleof the overall envelope of the Reference File Reference File enables a wide range of tests that includes anumber amplitude changes frequency changes intentional silence and multi frequency tone 26 Chapter 2 Audio Expert System Reference Guide Audio Expert System Reference User Guide segments Its goal is to flush out the audio chain s general ability to convey amplitude frequency silence and duration The ideal calibration for this file is one where the waveform visualization on Frontline s Audio Expert System User Interface Ul looks identical to the one shown below with respect to maximum levels In particular there are three segments in this test whose peaks are at exactly 6 dBFS That is there is zero loss or gain through the chain These 6 dBFS segments are described in Table 2 8 These segments serve as a convenient and quick visual indicator that the levels are appropriate Especially the longer third case which is evident at the 4 999 sec reference time in Figure 2 10 a little over 2 3 of the way through the test Table 2 8 Example ES File 6 dBFS Segments Show in Figure 2 11 is another graphic of just the first 0 5 sec of the exampl
6. 37 Audio Expert System Reference User Guide Chapter 3 Audio Expert System User Guide K Bits Sample 16 tt meo Figure 3 6 Collapsed Wave Panel 3 2 2 3 Audio Waveform Panel The Audio Waveform Panel displays the captured audio waveform If the waveform is stereo both channels are visible in the Wave Panel The user can view the entire waveform or can zoom to view a portion of the waveform in more detail 01 32 37 351 PM 01 32 37 386 PM 01 32 37 420 PM 01 32 37 455 PM 01 32 37 489 PM Figure 3 7 Audio Waveform Panel in the Wave Panel Table 3 4 Global Toolbar Waveform Horizontal Zoom Controls Horizontal Zoom Increases the amount of data that is visible on the screen however less detail is discemible Horizontal Zoom Decreases the amount of data that is visible on the screen however more detail is discernible The audio waveform is plotted as amplitude versus time on the Wave Panel The amplitude scale is located on the left edge of the Wave Panel The waveform s amplitude can be linear or in decibels The linear range is 1 0 to 1 0 The range for the dB scale is O dB for the maximum positive and maximum negative values and silence is negative infinity A toggle switch at the bottom of the amplitude scale will switch between Linear scale and dB scale Moving the switch to the left will display the Linear scale and moving it to the right will display the dB scale 38 Chapter 3 Aud
7. 44 1 KHz 48 KHz 8 KHz 11 025 KHz 12 KHz 16 KHz 22 050 KHz 24 KHz 32 KHz 64 KHz 88 2 KHz 96 KHz e Channels 1 and 2 e Variable Bit Rate and Specified Bit rate Audio Analysis not supported Although user will be able to play back the audio live 17 Audio Expert System Reference User Guide Chapter 2 Audio Expert System Reference Guide Supported Parameters for aptX e Object Types aptX classic aptX LL both content protected and non content protected e Audio Format 16 bit 44 1kHz e Data Rates 352 kbps Supported Parameters for CVSD e Channel Mode Mono e Sampling Rate 64 kHz Supported Parameters for mSBC codec e Channel Mode Mono e Sampling Rate 16 kHz e Allocation method Loudness e Subbands 8 e Block Length 15 e Bitpool 26 2 2 Referenced Mode Processing In Referenced Mode the system operates in a pseudo closed loop test scenario where the user plays a specific Reference Audio file pre recorded audio test file provided by Frontline on the Source DUT The Source DUT negotiates with the Sink DUT to determine the appropriate codec and audio parameters to use and will then process the Reference Audio file accordingly before transmitting the resulting audio via Bluetooth The Sink DUT receives the encoded audio decodes it and processes it for playback In parallel Frontline s system snoops the over the air signal between the Source DUT and Sink DUT and emulates the RF reception and dec
8. Chapter 2 Audio Expert System Reference Guide Audio Expert System Reference User Guide 2 2 2 Test ID All Reference Audio files begin with a unique Test ID and Frontline s Audio Expert System is continuously scanning the received audio in search of a valid Test ID On recognition of a valid Test ID the Referenced Mode state is declared This state then enables the system to perform more thorough testing of the audio data than can be performed in the non referenced mode The system uses the Test ID to determine which test script to use to evaluate the received audio A Test ID is three digits minimum in length representing a dot notation N vv Test Identifier The value N may be any length gt 1 indicating a specific test and vv represents a two digit version Each digit is represented by a tone between 200 and 290 Hz and is followed by either a 1 kHz delimiter tone or a 400 Hz Test ID terminator The digit O is represented by 200 Hz the digit 1 by 210 Hz and so on up to the digit 9 represented by 290 Hz The frequency 400 Hz received following the last digit terminates the digit string There is no encoding of the dot character but test script naming conventions include the string in N vv format asin Test 1 01 44 1kHz 16Bit xm 2 2 3 Reference Audio File Structure The Reference Audio files typically have a structure such as listed in the example below Such a
9. ID Found Event on page 8 Test Script Not See 2 1 2 1 Test ID Found Event on page 8 Found Invalid Test Script See Invalid Test Script Event on page 9 Synchronization See 2 1 2 4 Synchronization Lost Event on page 9 Lost 16 Chapter 2 Audio Expert System Reference Guide Audio Expert System Reference User Guide Table 2 5 Event Type Audio continued Error Unexpected gt 1 2 5 Unexpected Frequency Event on page Frequency Eror Error Unexpected Level Unexpected Level See 2 1 2 6 Unexpected Level Event on page 10 Error Unexpected See 2 1 2 7 Unexpected Duration Event on page 10 Duration Error Amplitude See 2 1 2 8 Amplitude Fluctuation Event on page 11 Fluctuations Error Unexpected Phase See 2 1 2 9 Unexpected Phase Change Event on Change page 11 See 2 1 2 11 Excess Noise Event on page 12 Clipping See 2 1 2 10 Clipping Event on page 11 Error CVSD HF Level See 2 1 2 12 CVSD HF Level Too High Event on Too High page 12 End of Test Event See 2 1 2 13 End of Test Event on page 13 2 1 4 Supported Codec Parameters Supported Parameters for SBC Codec e Sampling Frequencies 16 KHz 32 KHz 44 1 KHz 48 KHz e Channel Modes Mono Dual Channel Stereo Joint Stereo e Block Length 4 8 12 16 e Number of subbands 4 8 e Allocation Method SNR Loudness e Minimum Bitpool Value 2 e Maximum Bitpool Value 53 Supported Parameters for MPEG 2 4 AAC e Object Types MPEG 4 AAC LC e Sampling Frequencies
10. Low Volume Threshold see Figure 2 1 has been detected After this condition is satisfied a Low Volume Alarm event is generated when the detected peak volume drops below the Low Volume thresholds listed below for 10 consecutive 0 5 sec measurement intervals i e 5 sec total This event will not be repeated until the detected audio volume once again rises above the Low Volume Threshold and then falls below this threshold again for 10 more consecutive 0 5 sec measurement intervals e Speech Low Volume Threshold 15 dBFS e Music Low Volume Threshold 20 dBFS Note that if the volume level for any audio sample exceeds the Low Volume Threshold during any 0 5 sec measurement period the Low Volume Alarm state machine resets and begins looking again The information reported with the Low Volume Alarm event is listed below e Info1 Average RMS level just prior to alarm event This is the average level over the last 0 5 sec interval just prior to issuing the alarm e Info2 The time at when the system began detecting low volume i e 5 sec before the event was reported e Msg Not used 15 dBFS Low Volume Threshold for Speech 20 dBFS Low Volume Threshold for Music fF Figure 2 1 Low Volume Threshold Example 2 1 1 2 Clipping Event The purpose of the Clipping event is to report the presence of probable distortion that occurs when the amplitude of a signal exceeds a digital system s ability to represent it accurately Clip
11. Referenced Mode the system is typically processing audio of completely unknown program content e g arbitrary music or speech content Since the system does not have any prior knowledge of the audio being analyzed the types of audio analysis that can be performed is limited The following events are reported whenever the system is operating in Non Reference mode These are the meaningful audio analysis that the system can perform without reporting too many false positive results 1 Volume Level Low Volume or High Volume Warning Reported if the average volume level is not in a range conducive to performing meaningful audio analysis 2 Clipping Error Amplitude distortion due to a signal amplitude exceeding the maximum value that can be represented by the digital system Audio Expert System Reference User Guide Chapter 2 Audio Expert System Reference Guide 3 Dropout Warning Abrupt and very short duration intervals of silence 4 Glitch Error Extremely large sample to sample audio amplitude transitions that have little probability of occurring within natural speech or music 2 1 1 1 Low Volume Alarm Event The purpose of the Low Volume Alarm event is to warn the user that the volume level of the detected audio is below the level needed for performing meaningful audio analysis This event is only reported in Non Referenced Mode The system will not start reporting Low Volume Alarm events until a volume level above the
12. WO sc soe eee coisa ANA E BAY INE GANDA Em kaaa E 36 3222 POCA CONOS AA AA APA 37 3 2 2 3 Audio Waveform Panel 2 2 2 cece eee eee ee ee eee eee ee eee ence ees 38 32 24 Event Timeline RA AO 40 SEVEN TIDO eare E e R passe A E KANG E NG EE E 42 iV Audio Expert System Reference User Guide 3 2 4 Wave Panel amp Event Table Pop up Menu 2 2 2 2 2 ee cece cece ee cece cece eee eeeees 44 3 24 1 Export Audio Dala ecommerce banevacsiudeedieeseuecdex 45 3 2 4 2 Export Event Table occ cece cee cee eee c eee eee e cece ee eeeeeeeeeeeeeee 47 3 3 Frame Packet and Protocol Analysis Synchronization _ 2 2 a 47 3 4 Referenced Mode Testing Processes _ _ 2 2 oe cece eee ce cece ec eee cece cece cece eeeeeeeeeeeeeee 48 3 5 Air Sniffing Positioning Devices ooo cece ee eee cee cee cece cece cece eeceeceeceeceeeeeeees 50 Chapter 3 Contacting Technical Support oe ee ec o 55 List of Figures Figure 2 1 Low Volume Threshold Example o eee cece cece cece cece ccecccceccececcecceeeees 4 Figure 2 2 High Volume Threshold Example _ 22 20 2 eee cece cece cece ccc c cece cece eeeeeceeeees 6 Figure 2 3 Measurement And Silence Threshold 0 0 20 ccc ccc ce cece cece eee ceceeeececeeeees 6 Figure 2 4 Dropout Examples unnocosrocos rosca colosal de diria 7 Figure 2 5 Typical Topology for Voice Scenarios cnn 24 Figure 2 6 Typical Topo
13. a baseline capture Conductive testing isolates the DUTs and analyzer from environmental effects The next step is to ensure that the testing environment is as clutter free as possible e Line of sight obstructions should be eliminated between the ComProbe hardware and the DUTs because they cause a reduction in signal strength Obstructions include but are not limited to water bottles coffee cups computers computer screens computer speakers and books A clear unobstructed line of sight is preferred for DUT and ComProbe hardware positioning e If using an analyzer connected to a computer position the computer on an adjacent table or surface away from the analyzer and DUTs taking advantage of the cables length If this is not possible position the computer behind the analyzer as far away as possible If using the ComProbe FTS4BT which is a dongle either use an extension USB cable or position the computer such that the dongle is positioned towards the DUTs e The preferred placement is positioning the DUTs and the ComProbe hardware at the points of an equilateral triangle in the same horizontal plane i e placed on the same table or work surface The sides of the triangle should be between 1 and 2 meters for Bluetooth transmitter classes 1 and 2 The distance for transmitter class 3 should be 1 2 meter seis Audio Expert System Reference User Guide Chapter 3 Audio Expert System User Guide Figure 3 17 Devices Equally
14. and vice verse Protocol analysis tools synchronized to the Audio Expert System include e Frame Display e Coexistence View 47 Audio Expert System Reference User Guide Chapter 3 Audio Expert System User Guide e Bluetooth Timeline e Message Sequence Chart e Packet Error Rate Statistics When a portion of the waveform is selected in the Wave Panel all frames within the selection will be highlighted in the Frame Display Coexistence View and Bluetooth Timeline Note If the Frame Display is filtered to show non audio events then the frames associated with H selected audio events may not show 3 4 Referenced Mode Testing Processes In the Referenced mode the devices under test use a specific audio file called reference file or test file provided by Frontline whose contents are already known to the ComProbe software The software compares the parameters of the received audio data against its parameters and presents analysis for the user Commonly in Bluetooth technology the music sent via A2DP and speech sent via HFP There are a few ways users can conduct referenced mode testing depending upon what profile they are using The figure 17 shows the source of the audio and the medium through which it can be accessed by Source device to send to sink device via Bluetooth id Using HFP A file stored Play the locally stored file on the Play using the third party App on the audio source device that transmits music data on dev
15. event is listed below e Info1 An integer value representing the absolute value of the sample delta This can be converted to a change in decibels by the relationship 20Log p Info1 e Info2 Not used e Msg Not used 2 1 2 Referenced Mode Event Reporting In Referenced Mode the system operates in a pseudo closed loop test scenario where the user plays specific Reference Audio files pre recorded audio test files provided by Frontline on the Source DUT Device Under Test See Referenced Mode Processing on page 18 for a more complete description The events listed in the following sections are reported when the system is operating in Referenced Mode The following events are reported whenever the system is operating in Reference mode Audio Expert System Reference User Guide Chapter 2 Audio Expert System Reference Guide 1 Test ID Found Event Information Reported when a valid Test ID has been recognized 2 Test Script Not Found Event Error Reported when a valid Test ID was found but the system cannot locate the script for that Test ID 3 Invalid Test Script Event Error Reported when an error occurs while the system is verifying information in a test script 4 Synchronization Lost Event Error Reported when the system is in Referenced Mode and encounters unexpected frequencies or durations of audio segments while analyzing a received Reference Audio file 5 Unexpected Frequency Event Error Rep
16. file structure provides known variations in amplitude and frequency that enable a variety of measurements including the detection of changes to amplitude frequency and duration lost audio substituted audio e TestID segment e a300 Hz tone segment at 12 dBFS for 100 ms e a600 Hz tone segment at O dBFS full amplitude for 50 ms e adual tone segment consisting of equal amplitude 880 and 240 Hz tones at 6 dBFS below maximum for 200 ms e asilence segment for 100 ms e 600 Hz tone segment which fades out over 100 ms 2 2 4 Synchronization In Referenced Mode the synchronization of the received audio from the Reference Audio files versus the internal test script is achieved based on changes in frequency of the tones in the Reference Audio file Frequency changes are used because this parameter is relatively immune to the configuration and typical impairments of the network By using changes in frequency the system is able to continuously re established synchronization with the Reference Audio file at every tone segment That is as each next frequency in the script is encountered the exact time position in the script is re established the expectation of the duration of the new frequency is known and any significant variation before the next frequency change occurs will be reported as an Unexpected Duration event 2 2 5 Minimum and Optimal Amplitude Level Adjustment Whena Test ID Terminator Tone is recognized its amplit
17. four sections 1 Audio Stream Info that provides users with information such as sample rate bit sample codec and DUT Device Under Test addresses 2 Local Controls include audio volume controls and Indicators Mute Vertical Zoom and Collapse Expand 3 An Audio Waveform which is plotted as amplitude linear or dB versus time and an interactive play cursor The play cursor appears as a white vertical line across the waveform 4 Event Timeline that shows color coded Bluetooth ES Codec fa gH and Audio events Details of these events are listed in the Audio Expert System Event Table 35 Audio Expert System Reference User Guide Chapter 3 Audio Expert System User Guide 3 2 2 1 Audio Stream Info The Audio Stream Info displays Audio Bluetooth and Codec information left to right in the image below about the audio waveform displayed in the panel This information is discovered during AVDTP signaling when the devices under test DUT negotiate audio streaming parameters Sample Rate 44100 DUTI A0 EF 40 82 96 6E Mono Stereo Bits Sample 16 AU B0 79 94 80 12 1E Figure 3 3 Audio Stream Info in the Wave Panel Table 3 3 Audio Stream Info Tags Name Description A system assigned index number that represents an audio waveform between a pair of Bluetooth devices This number appears in the Event Table for easy cross referencing Sample Rate Displays the sampling frequency
18. le Ma gt sen Her Win rr AA Y DN PN armada alak Arpa A Tal il rac NAAN PEEPS N 7 E j F Lil mi l i FATT NAT addii Level falls below Silence Threashold Figure 2 4 Dropout Examples The two events of 2 5 ms and 25 ms are reported as Dropout events while the 57 ms interval at the center is not Although 57 ms interval is of sufficient duration and the level remains below Silence Threshold for the duration the interval is not immediately preceded by sustained energy above the Measurement Threshold The info reported with the Dropout event is listed below e Info1 The number of samples over which the dropout persisted The value can be converted to time duration sec by dividing by sample rate Likewise the Start Time of the dropout can be calculate by subtracting the duration from the event s sample number time stamp parameter and converting to time e Info2 Not used e Msg Not used 2 1 1 5 Glitch Event Glitch events are extremely large sample to sample audio amplitude transitions that have little probability of occurring within natural speech or music Such dramatic changes would typically happen only in situations of dropped samples The system uses different thresholds for sample to sample transitions for music versus speech as shown below e Speech greater than 40 dB change from sample to sample e Music greater than 90 dB change from sample to sample The info reported with the Glitch
19. used to digitize the original audio Mono Stereo Indicates if the audio data is monaural or stereophonic Bits Sample Displays the number of bits per sample of the audio data DUT1 Bluetooth address of one device in the connection Can be either sending or receiving the audio data DUT2 Bluetooth address of the other device in the connection Can be either sending or receiving the audio data Displays the Codec type used by the captured audio stream The supported codecs include SBC AAC aptX mSBC and CVSD SBC Codec Information Pop up When you hover over the Codec tag and the Codec SBC a pop up will appear that shows additional information about which SBC parameters can be used The pop up is visible as long as the cursor hovers over the Codec tag AA Channel Modes Mama Figure 3 4 SBC Codec Information Pop Up on Cursor Hover Over 36 Chapter 3 Audio Expert System User Guide Audio Expert System Reference User Guide 3 2 2 2 Local Controls The Local Controls in each Wave Panel provide the user with indicators and controls for waveform display and audio play back QU Al Figure 3 5 Wave Panel Local Controls Waveform Play Back Volume The volume slider controls the playback volume for the audio in each Wave Panel Audio Volume Indicator PAA The volume indicator shows the relative audio volume at the waveform display play cursor When the green bars completely fill the
20. values associated with the term O dBFS which means O dB Full Scale It is an arbitrary reference point that simply means the maximum possible volume that can be represented for a given number of bits per sample resolution Therefore the number of bits per sample determines the maximum values that can be represented and the maximum values represent the maximum volume The lowest level that can be represented is the minimum possible non zero number that can be represented and that of course in all cases is the value 1 or 1 min AY Chapter 2 Audio Expert System Reference Guide Audio Expert System Reference User Guide The difference between the maximum and minimum representable values expressed in decibels is defined as the dynamic range For any given sample value Level in dBFS 20 LOG level max level For example for 8 bit samples the minimum level in decibels is 20L09 10 42 dBFS leo 2 7 Bit Dynamic aie in Pi olution a 8 388 607 138 dB Therefore for 8 bit samples the dynamic range is 42 dB and Table 2 7 indicates the dynamic range for all number of bits per sample resolution discussed To help give these numbers some meaning the human ear has a dynamic range of about 120 dB and since Compact Disks CDs use 16 bit samples they inherently have a dynamic range of 90 dB The legacy telephone network uses 8 bit samples and thus has a dynamic range of 42
21. voice material and 44 100 samples sec is typically the minimum for music Each sample thus represents a signal s instantaneous amplitude at a point in time in units equivalent to voltage 2 4 3 Precision and Dynamic Range The precision of digitally recorded values depends largely on the resolution of the conversion device known as an Analog to Digital A D Convertor Typical A D convertors produce samples of 8 16 or 24 bit resolution The sample values are integers with maximum and minimum values that depend on the number of bits per sample as related in Table 2 6 One of the characteristics of audio that is most affected by the number of bits per sample is dynamic range It is a term that describes the range of loud to soft audio levels that can be accommodated The loudest full volume audio levels are represented by the maximum value that can be represented for a given number of bits per sample These maximum values equate to the voltage levels of analog amplifiers at which the output voltages reach the limits of the power supplies being used If a 30 volt power supply is being used and ignoring output device losses the output voltage can never exceed 30 volts If the input signal is increased beyond this the waveform that would have been above the maximum limit is said to have been clipped A clipped sine wave begins to look like a square wave at extremely over driven levels These maximum values are also the
22. 1 volume accordingly and repeat this step Optimally it would be on or just below the 6 dB gridline but not above The peak should never hit the maximum positive or negative limits of the display 10 Find the Test ID Found event in the Event Table to verify that the system has transitioned to Referenced Mode and verify that the value for Channel Gain or Level as implemented in the Ul is within the range of values specified in Table 2 9 If the observed captured waveforms do not reasonably conform to the above graphic for Test 1 02 or the Test ID Found event is not reported there is a problem along the audio chain This could be as simple as a configuration setting or more subtle such as an encoder decoder incompatibility 29 Audio Expert System Reference amp User Guide Chapter 2 Audio Expert System Reference Guide 30 Chapter 3 Audio Expert System User Guide This section provides information on the Audio Expert System user interface and controls This information can also be found in the ComProbe software Help Also this information is contained in the ComProbe BPA 600 and ComProbe Sodera On line Help and in the ComProbe BPA 600 and ComProbe Sodera User Manual PDF document 3 1 Starting the AudioExpert System Sodera and BPA 600 only 31 3 2 Audio Expert System Window o o 32 3 3 Frame Packet and Protocol Analysis Synchronization o ococccccoccocccccccccccncccncnoooo 47 3 4 Ref
23. 23 B6 2E and 00 07 62 0F 00 00 for stream 1 using codec SBC Mar 31 2014 12 53 05 446738 PM N A A2DP resumed between devices 98 0D 2E 23 B6 2E and 00 07 62 0F 00 00 for stream 1 using codec SBC Mar 31 2014 12 53 05 474864 PM N A Packet retransmission for unknown CID Mar 31 2014 12 53 07 712976 PM N A AVDTP packet loss detected based on missing packet sequence number Mar 31 2014 12 53 13 742943 PM N A issing packet sequence number Mar 31 2014 12 53 15 385434PM T gt DENIA an ve o we a a oe 90 S 13 oe oe oe 06 16 o 1 1 1 0 2 3 2 3 2 3 1 1 0 1 1 Figure 3 14 Event Table Several events can be selected by clicking and dragging over the events or by holding down the Shift key and clicking on events To select events that are not adjacent hold down the Ctrl key and click on the events When selecting multiple events the Wave Panels will not scroll to the selected events 42 Chapter 3 Audio Expert System User Guide Audio Expert System Reference User Guide The Event Table contains eight columns Table 3 6 Event Table Columns System generated sequential numbering of events Information provides information of interest but does not indicate a problem event Warning identifies a potential problem where further investigation may be appropriate Error identifies a definite problem A system generated ID that is assigned in the order that the audio streams are det
24. 3 4 Crest Factor The difference between peak and average levels is commonly referred to as the Crest Factor There are some well characterized relationships between peak and average levels for specific types of audio and variations from the norm can reveal more subtle aspects of how the audio was processed or what it was subjected to When the system is in Referenced Mode it is receiving audio from which the source is one of Frontline s Reference Audio files and the majority of tones in these files are single sine waves For a sine wave the RMS level is 0 707 times the peak level corresponding to a 3 dB difference Therefore any difference greater than 3 dB can indicate an abnormal increase in amplitude such as the presences of spikes or other noise or the presence of multiple frequencies Note Some test segments are intentionally dual tone segments While Crest Factor for a single H sine wave is 3 dB for 2 sine waves of equal amplitude it is 6 dB When the system is in Non Referenced Mode there is no pre defined relationship between peak and average levels However there are well characterized relationships established for speech and audio as listed below These numbers represent typical numbers and absolute numbers depend on the original audio content e Speech A typical Crest Factor is 12 dB assuming a signal captured with no dynamic range processing algorithms such as compression or limiting e Music A typical Cre
25. Any issues or questions should be directed to the app developer 1 In the following steps Device Under Test 1 DUT1 is the device sending the reference test file to DUT2 2 Download the third party app to DUT1 and follow the app vendor s instructions for installation and use 3 Load the Audio Expert System reference test file Test 1 02 64 1kHz 16Bit wav on DUT1 The test file is stored on the users computer In the directory Frontline ComProbe Protocol Analysis System Development Tools Audio Expert Test Files delivered with your latest Frontline ComProbe software version may have changed Note Reference test files are periodically updated Shown here is an example Files Contact Frontline Technical Support for information on the latest reference file versions 4 With the BPA 600 or Sodera connected to the computer configure the datasource and follow procedures to capture data 5 Launch Audio Expert System by clicking on the Control window KS 6 Turn on Bluetooth on your DUTs DUT1 and DUT2 Turn on the third party Bluetooth app for routing the reference file over A2DP or HFP by following the vendor s directions 7 Send the reference test file from DUT1 to DUT2 via the third party app 49 Audio Expert System Reference User Guide Chapter 3 Audio Expert System User Guide 8 Observe the events in the Audio Expert System Events Table Look for an event Description TestIDFound REF Test ID 1 02 Channel Ga
26. Spaced in the Same Horizontal Plane Finally eliminate other RF sources e Wi Fi interference should be minimized or eliminated Bluetooth shares the same 2 4 GHz frequency bands as Wi Fi technology Wi Fi interference can cause loss of packets and poor captures In a laboratory or testing environment do not place the DUTs and ComProbe hardware in close proximity with Wi Fi transmitting sources such as laptops or routers Turning off Wi Fi on the computer running the ComProbe software is recommended Positioning for wideband capture Frontline s Wideband Bluetooth Protocol Analyzer Sodera can capture from multiple devices which requires a different approach to position the DUTs and the analyzer When testing more than two devices arrange the DUTs on the perimeter of a circle 1 2 meters in diameter for Bluetooth transmitter Class 1 and 2 devices For transmitter Class 3 DUTs the circle should be 1 2 meter in diameter Equally space the DUTs on the perimeter Place the Sodera in the center of the circle If not using the Sodera Excursion mode connect the computer and place it outside the circle as far away from the DUTs as possible y NG Chapter 3 Audio Expert System User Guide Audio Expert System Reference User Guide Figure 3 18 Wideband Capture Devices Equally Spaced in the Same Horizontal Plane Positioning for audio capture The Bluetooth Audio Expert System provides analysis of audio streams and can assist in identifying
27. al brief silence period where the brief silence is preceded and followed by normal audio levels A typical definition of Dropout is the short dramatic loss of volume typically caused by lost digital information Root causes include transmission system errors resulting in lost data packets transmission channel reconfigurations bad sections of memory processor overloads that temporarily interrupt the flow of information and so on The challenge is to detect real dropout events while at the same time honoring natural silence For music there are natural musical rests that occur Speech includes natural inter syllable inter word and pause silence Dropout events are reported when the average audio level RMS is initially above the Measurement Threshold then falls below the Silence Threshold and then quickly rises above the Measurement Threshold again see Figure 2 3 This approach largely disqualifies the natural inter syllable silence and pauses that occur in natural speech but will detect gaps caused by dropped data Note that the system does not report dropouts that begin at very low energy levels Measurement Threshold Silende Threshold Figure 2 3 Measurement And Silence Threshold Shown in Figure 2 4 are examples to illustrate the Dropout event Chapter 2 Audio Expert System Reference Guide Audio Expert System Reference User Guide 2 58 Dropout Sms 25ens Dropout Measurement Threshold lena 1
28. as on a smartphone often controls the level of the audio sent over the Bluetooth channel Indirectly as a result it also effectively controls the listening level In contrast a volume control on the Bluetooth sink device DUT2 often controls only the volume level delivered to the speakers The result is that changing the volume setting may affect only the level delivered to the speakers and has no effect on the level sent over the air in the Bluetooth channel As an example shown in Figure 2 7 is an illustration of a smartphone playing a stored audio file to a Bluetooth speaker system Both phone and speakers have their vertical slider volume controls set to maximum This represents the loudest but probably distorted volume the complete system is capable of delivering Bluetooth Audio Channel LA 443 Figure 2 7 Example 1 Shown in Figure 2 8 is second example with the volume control of the speakers set to half volume and this results in the level to the listener being reduced to a desirable level However note that in this example the volume control of the phone has been left at maximum thus the over the air volume through the Bluetooth channel remains at maximum and may be distorted Js Audio Expert System Reference User Guide Chapter 2 Audio Expert System Reference Guide e p Fi Ki i E A i f f F 4 j i fe fiir HC BRAGAS 1 Li LAGAY YA AA 1 Ng A A NG N 1 Figure 2 8 Example 2 Bluetooth Au
29. ata File Name C Users jwt Desktop audiowebinar_1 wav 1 044100 Hz 16 bits stereo Fl Encoded Audio Data raw 2 116000 Hz 16 bits mono 3 1 16000 Hz 16 bits mono Decoded Audio Data wav E Event Table Data csv Only Selected Wave Area Figure 3 15 Export Audio Data dialog The Save As window will close and the file name will appear in the File Name field in the Export Audio Data window Should the file name need to be changed click on the Select button and the Windows Save As dialog will open By default the wav file extension is used in the file name In the window below File Name will appear a list of Stream Ids with a description from the Audio Stream Info If opening from the Audio Expert System Global Toolbar all Stream IDs are checked by default If opening from a Wave Panel the Stream ID where the export dialog was opened is automatically checked You can check each stream that is to be exported For convenience checking Select all below the stream list window will place checks in all streams Export Options After selecting the streams to export select the desired formats to export Table 3 9 Export Audio Data Format Options Encoded Audio Data Exports the selected files as raw format The audio data is in an encrypted format and user will need a codec to decode it Decoded Audio Data Exports the selected files as wav format that can be played ona wide variety of media players Event Tabl
30. ate that the system has completed processing a test script for a Reference Audio file and that the system has exited Reference Mode This event is generated when the elapsed time from the start of test is equal to or greater than the scripted duration of a test It is reached when the number of samples processed equals the number of samples associated with the test duration The information reported with the End of Test event is listed below e Info1 Unused e Info2 Unused e Msg Unused 2 1 3 Audio Expert System Event Type The following tables list the Audio Expert System Bluetooth Codec and audio events with description Included in the tables is the event severity that can have three values Information Warning and Error The event severity will appear as icons and text in the Audio Event System once an audio streams has been captured Refer to 3 2 3 Event Table Event Table Columns on page 43 for an explanation of the severity types 2 1 3 1 Event Type Bluetooth Protocol Table 2 3 Event Type Bluetooth Protocol Profile TX PDUs larger than available bandwidth for active A2DP Streaming interval Error Error Bitpool Bitpool value does not match configured bitpool range does not match Bitpool value does not match configured bitpool range bitpool range 143 Audio Expert System Reference User Guide Chapter 2 Audio Expert System Reference Guide Table 2 3 Event Type Bluetooth Protocol continued 2 1 3 2 Eve
31. dB 2 4 4 Measurement Accuracy The dynamic range of the audio path is important to understand because it has a direct impact on measurement accuracy It is perhaps easiest to grasp the effect of level if we consider accuracy at the extreme loudest and lowest volumes Consider first the accuracy at maximum levels Looking at the difference in dB between samples of value N and N 1 for several different values of N provides some insight For 8 bit resolution the value of N 127 is the max possible value and thus 20109 10 J O dBFS However for the value N 1 20Log 10 ai zJ 0 07 dBFS Therefore at maximum volume the difference between N and N 1 is 0 0 07 which means we can resolve to within 0 07 dB This is fairly good At half volume N 63 and the level in dBFS 20Log tol j 6 09 dBFS For N 1 62 62 20109 10 6 29 dBFS The difference in this case between N and N 1 is 6 09 6 29 0 14 dB 23 Audio Expert System Reference User Guide Chapter 2 Audio Expert System Reference Guide This is still not so bad but it becomes apparent that at half volume we can resolve to 7 the accuracy as at full volume and it follows that at volume the accuracy will be that of full volume and so on For completeness consider now taking it to the lower limit by looking at the difference in dB for N and N 1 at their lowest possible non zero values These are of course N 2 and N 1 1 and from experienc
32. de The information reported with the Synchronization Lost event is listed below e Info1 An integer reason code such as 12 Unexpected Frequency e Info2 Unused e Msg Contains a diagnostic message containing a code point indicator the SegmentID of the script that was expected when the error was declared and the zero based index into the Segment Array Audio Expert System Reference User Guide Chapter 2 Audio Expert System Reference Guide 2 1 2 5 Unexpected Frequency Event In Referenced Mode the system knows the Reference Audio file that is being played on the Source DUT therefore the system knows which frequencies tones to expect at a given time The Unexpected Frequency event is reported when a measured frequency deviates from an expected frequency by a specific percentage determined by the negotiated parameters of the over the air audio stream The information reported with the Unexpected Frequency event is listed below e Info1 The measured frequency Hz e Info2 The expected frequency Hz e Msg SegID N where N is a Segment Identifier indicating the expected segment 2 1 2 6 Unexpected Level Event In Referenced Mode the system knows the Reference Audio file that is being played on the Source DUT therefore the system knows which amplitude level to expect at a given time The Unexpected Level event is reported when the measured level at the start of a tone segment is not within tolerance The tolerance
33. dically updated Shown here is an example Files 4 delivered with your latest Frontline ComProbe software version may have changed Contact Frontline Technical Support for information on the latest reference file versions 2 Before establishing the Bluetooth connection play the file while listening to it on the DUT1 device itself and become familiar with the overall sound quality generally ignoring exact volume 3 Set the playback volume at DUT1 to maximum 4 Set the playback volume at DUT2 to minimum 5 Establish the Bluetooth connection and begin playback of the file on DUT1 if possible in Loop or Repeat mode to avoid having to continuously restart 6 Slowly increase the volume on DUT2 until it is at a comfortable level 7 If the audio sounds distorted reduce the playback volume at DUT1 and repeat Step 6 8 When the clarity of the audio is comparable to that heard when listening to the DUT1 device proceed with using Frontline s ComProbe Analyzer with Audio Expert System enabled to capture and analyze the Bluetooth data 9 Visually observe the waveform in the Audio Expert System Wave Panel comparing it to the image above Figure 2 10 If the level of the 6 dB 0 9 sec duration 400 Hz tone a little over 2 3 of the way through the 90 s Chapter 2 Audio Expert System Reference Guide Audio Expert System Reference User Guide test is grossly above or below the 6 dB 50 volume grid line adjust the DUT
34. dio Channel Shown in Figure 2 9 is athird example In this case in order to reduce the listening volume the slider on the phone has been reduced while the slider on the speakers has been left at maximum This achieves the same objective as example 2 as far as the listener is concerned However reducing the volume at the source DUT1 also reduces the volume level sent over the air through the Bluetooth channel In effect this reduces the dynamic range that the audio channel is normally capable of accommodating Bluetooth Audio Channel iy Figure 2 9 Example 3 Since the Frontline s Audio Expert System only sees the over the air audio it is important that the over the air volume be adjusted to a sufficiently high level but not so high that the signal is distorted Too much reduction of volume through the over the air channel can result in audio levels falling below required Measurement Threshold levels see Low Volume Alarm Event on page 4 and High Volume Alarm Event on page 5 and thus making useful audio analysis measurements are not possible 2 6 2 The General Process The objective is to achieve settings at the Bluetooth Source device DUT1 that bring the PCM sample levels of tones in the Reference Audio files for Referenced Mode operation sent over the air as close as possible to the levels at which they were created without exceeding them Test ID tones and the tones in test file sequences for
35. e Data Exports a text csv file of all the detected events 46 Chapter 3 Audio Expert System User Guide Audio Expert System Reference User Guide Table 3 9 Export Audio Data Format Options continued Exports the Encoded Decoded or Event Data for the selected waveform This option is only active if a selection has been made in one of the Wave Panels Click on OK to save the waveform The dialog will close and a series of progress bars will Ps audiowebinar L raw 439 KB RAW File appear Each progress bar is associated with a audiowebinar 1 wav 18 652 KB Wave Sound file for each export option The exported files will have the following syntax lt filename gt _ n lt filetype gt where lt filename gt the name entered into the File Name field n the audiowebinar 2 wav 22 KB Wave Sound stream id number 1 2 3 and lt filetype gt raw wav and csv The image shows an example where the user exported Stream Id s 1 and 2 in Encoded Audio Decoded Audio and Event Table data to filename audiowebinar EL audiowebinar 1 csv 9 KE Microsoft Excel C audiowebinar_2 csw 19 KE Microsoft Excel C Ps audiowebinar 2 raw 299 KE RAW File Click on Cancel to close the window without exporting 3 2 4 2 Export Event Table Right clicking in the Event table will open a pop up menu with the option to Export Event Table This option will export selected events in the in comma separated variable
36. e Reference File which contains the Test ID value 1 02 see Test ID Found Event on page 8 and Test ID on page 19 A MAMAW a HIN HH i Hl ii KN Ji Hi RN AA Figure 2 11 Test 1 02 Test ID Segment The three digits 1 0 and 2 are represented by the low frequencies 210 Hz 200 Hz and 220 Hz respectively which are 100 ms in duration and are separated by 1 kHz digit delimiters of 50 ms duration The final tone is a 100 ms segment at 400 Hz defined as a Test ID Terminator Note that the levels of all of these tones are exactly 3 dBFS The peak level of the Test ID Terminator is measured and reported in the Test ID Found event s Info1 parameter see 2 1 2 1 Test ID Found Event on page 8 It is also serves to calculate the overall gain of the audio channel up to and including the level of audio on the configured Bluetooth audio channel The value in the Info1 parameter of the Test ID Found event is optimally the value 23196 and may be converted to dBFS by the relationship BFS 20Log xo Optionally the value can be converted to Channel Gain via the relationship nfo1 dB 20Log 107505 27 Audio Expert System Reference User Guide Chapter 2 Audio Expert System Reference Guide Table 2 9 Test ID Found Event info1 Table 2 9 indicates the maximum and minimum acceptable Maximum and Minimum Values levels for the Test ID Found Info1 parameter in int
37. e the factor of two is apparent which translates to a difference of 6 dB Just for reference the difference in dB for the values N 3 and N 1 2 is 3 5 dB for 8 bit resolution So for 8 bit when a low level analog value close to 2 is sampled it will be quantized to an integer value of 1 2 or 3 depending on round off or truncation logic in the A D converter and subject to any noise in the analog signal The point is that a sample value of 2 in dB can only be stated as 36 dBFS accurate to within 3 5 or 6 dB In other words because the accuracy is at best within 3 5 dB and at worst within 6 dB any measurement that seeks to detect changes accurately of say 3 dB is not possible Note Especially for 8 bit measurements it is essential that the signal level be as high as possible H without exceeding maximum supported values 2 5 Audio Chain Network Technologies Audio data sourced at one device referred to as DUT1 can pass through a variety of audio processing components and transmission channels before it is finally delivered to headphones or speakers at a playback device referred to as DUT2 Figure 2 5 shows an example where a land line phone connects to the Public Switched Telephone Network PSTN either directly via copper or often through a Private Branch Exchange PBX The call is handed off to a Public Land Mobile Network PLMN and terminates on a smartphone In turn the smartphone establishes a BT connect
38. ected The ID is not maintained between captures for the same device with the same audio It identifies the Wave Panel where the event can be viewed The ID appears in the Audio Stream Info of the Wave Panel Bluetooth Events generated by analyzing Bluetooth protocol activities Codec Events generated from analyzing the audio coding decoding activities Audio Events generated by analyzing the audio data Mode does nat apply to this event Referenced Mode Refer to 3 4 Referenced Mode Testin Processes on page 48 and 2 1 2 Referenced Mode Event Reporting on page 7 Non Referenced Mode Referto 1 1 Non Referenced Mode on page 1 and 2 1 1 Non Referenced Mode Event Reporting on page 3 The system generated identification for a specific frame Details and explanation about this event A system generated time stamp for each frame Sorting Event table entries are sortable by column Left click on the column heading to sort Event Table Pop Up Menu Right clicking with the cursor over the Event Table will open a menu of additional options For more on this option see Wave Panel amp Event Table Pop up Menu on page 44 43 Audio Expert System Reference User Guide Chapter 3 Audio Expert System User Guide Lock Event Table lock bad Take The Lock Event Table checkbox is available in live mode only Clicking to check the box will prevent the Event Table from scrolling during live capture Un checking the box will r
39. eger Application Maximum Minimum form decibel level in dBFS and Channel Gain in dB 23196 5826 Example 1 For the case where the Info1 parameter is i converted to Channel Gain if the audio is speech i e mea aa transported via a SCO channel then a value of 11 9 dB is 3dBFS 15dBFS acceptable and a value of 12 1 dB is not 3dBFS 20dBFS Example 2 For the case where the Info1 parameter is Ta converted to Channel Gain if the audio is music i e kababasa 12 dB transported via an A2DP connection then a value of 16 9 17 dB dB is acceptable and a value of 17 1 dB is not For both cases at the high volume end a value of 0 1 dB is acceptable a value of 0 1 dB is not 2 6 3 Adjusting for Optimal Volume Levels The exact steps that need to be taken depend on the exact devices being used and their device specific setup requirements and the speech or audio configuration under test For the simplest case where for example a music audio file is to be played by a smartphone to a set of Bluetooth speakers the typical steps would include the following 1 Choose an audio reference file to be played at DUT1 appropriate for the configuration to be tested The test files are stored on the users computer In the directory Frontline ComProbe Protocol Analysis System Development Tools Audio Expert Test Files For example Test 1 03 48kHz 16Bit 3Loops 2Ch wav Note Reference test files are perio
40. egrees Info2 Not used Msg Not used 2 1 2 10 Clipping Event The purpose of the Clipping event is to report the detection of suspected distortion that occurs when the amplitude of a signal exceeds a digital systems ability to represent it accurately Clipping is a type of amplitude distortion The system reports a Clipping event when consecutive samples at the maximum value that can be represented by the digital system have been detected Note that the maximum value that can be represented is different 11 Audio Expert System Reference User Guide Chapter 2 Audio Expert System Reference Guide depending on the number of bits per sample i e bits of resolution of the audio stream The system limits the number of reported Clipping events to typically 10 to 20 per sec The number of consecutive samples needed to qualify as a clipping event depends on both sample rate and number of bits per sample Table 2 2 specifies the number of consecutive samples at the maximum value level that will generate a Clipping event Table 2 2 Clipping Event Threshold The info reported with the Clipping event is listed below e Info1 Contains an integer value representing the RMS average audio level when the clipping event was detected The value can be converted to a decibel level via the relationship 20 LOG 10 info1 32767 e Info2 Not used e Msg Not used 2 1 2 11 Excess Noise Event In Referenced Mode the system know
41. erenced Mode Testing Processes z cee cee cee cee cee cece eee cece cece ee eeeeeeeeeeees 48 3 5 Air Sniffing Positioning Devices 50 3 1 Starting the AudioExpert System Sodera and BPA 600 only To use the Audio Expert System the user must have e Current Premium Maintenance purchased from Frontline e ComProbe hardware with Audio Expert System license installed connected to the PC This is a requirement for both live capture and when viewing a saved capture file For live capture set up the ComProbe Sodera or ComProbe BPA 600 datasource and begin capturing data Note Proper positioning of the ComProbe hardware relative to the devices under test DUT 1 Si source DUT2 sink will contribute to effective data capture Air Sniffing Positioning Devices on page 50 For viewing a capture file load the saved file from the ComProbe Control window File menu When an audio stream is available the open the Audio Expert System Window by clicking on the Control window Audio Expert System button ry If the ComProbe analyzer is not licensed for Audio Expert System the button will not be present 231 2 Audio Expert System Reference User Guide Chapter 3 Audio Expert System User Guide 3 2 Audio Expert System Window This window is the working space for the Audio Expert System Upon opening Audio Expert System the window shown below will open with four main areas displayed e Global Toolbar Provides play cursor co
42. ering Figure 3 20 Example Poor Capture Environment 54 Chapter 3 Contacting Technical Support Technical support is available in several ways The online help system provides answers to many user related questions Frontline s website has documentation on common problems as well as software upgrades and utilities to use with our products On the Web http fte com support supportrequest aspx Email tech support fte com If you need to talk to a technical support representative about your Audio Expert System product support is available between 9 am and 5 pm U S Eastern Time zone Monday through Friday Technical support is not available on U S national holidays Phone 1 434 984 4500 Fax 1 434 984 4505 Instructional Videos Frontline provides a series of videos to assist the user and may answer your questions These videos can be accessed at fte com support videos aspx On this web page use the Video Filters sidebar to select instructional videos for your product 55
43. ession ended coenen loo bitstream type has been set For Blueooth it should be LATM Warning Single frame error During decoding a single frame error was concealment triggered detected which triggered built in concealment processing Error Codec setting change The codec has been re initialized due to a setting change 25 lt Audio Expert System Reference User Guide Chapter 2 Audio Expert System Reference Guide Table 2 4 Event Type Codec continued Unframed stream error A frame error was detected for an unframed stream The codec is being reset in order to continue processing Error Transport not initialized The codec cannot be initialized for the given transport Transport not supported The selected transport is not supported This could occur when an out of band LATM is selected opposed to in band Transport failure General failure in the transport Error Transport error This typically occurs when there isn t any configuration information available Codec initialization Codec session started Codec session started started Information Codec teardown tear down Codec session ended Codec session ended ended Data Non stereo data has been detected for incoming data stream 2 1 3 3 Event Type Audio Table 2 5 Event tee A Audio Warning Clipping See 2 1 1 2 Clipping Event on page 4 See 2 1 1 4 Dropout Event on page 6 Warning see 2 1 1 5 Glitch Event on page 7 TestID Found See 2 1 2 1 Test
44. esume scrolling of events as they are detected When analyzing a capture file the checkbox has no effect 3 2 4 Wave Panel 8 Event Table Pop up Menu Additional Wave Panel and Event Table options are available by right clicking the mouse with the cursor anywhere in the Wave Panel or in the Event Table Wave Panel Pop up Menu Actions Right clicking anywhere in the Wave Panel will provide you with a Clear Selection selection of the following actions Copy Selection Export Audio Data Loop Zoom to Selection Select rea Select All Table 3 7 Wave Panel Pop up Menu Selections Clear Selection Clears the current selection in the viewer Copy Selection Saves a copy of the selection to the computer clipboard The clipboard can be pasted into a Word document an e mail or other Windows clipboard compatible application Export Audio Data Opens the Export pop up menu with options to export the waveform as a raw wav or Event Data For additional details on exporting refer to Waveform Display Export Loops through the audio selected on the Wave Panel Zoom to Selection Expands or compresses the selection to fill the Wave Panel view 44 Chapter 3 Audio Expert System User Guide Audio Expert System Reference User Guide Table 3 7 Wave Panel Pop up Menu Selections continued Select Area When the mouse cursor is positioned over data not fill pause or gaps in the Wave Panel and selecting this o
45. hat this event will result in an attempt to resynchronize see Synchronization Lost Event on page 9 and Synchronization on page 19 if the measured duration is greater than expected The info reported with the Unexpected Duration event is listed below e Info1 The measured duration us e Info2 The expected duration us e Msg UD LvI SeglD N ixSeg l LastFreq F where N is the Segment ID is an index into the segment 10 Chapter 2 Audio Expert System Reference Guide Audio Expert System Reference User Guide description in a Segment Array and F is the measured frequency Note The above Msg example is one of several unexpected duration cases that can occur H The exact syntax will vary 2 1 2 8 Amplitude Fluctuation Event The Amplitude Fluctuation event i e amplitude modulation is reported if while in Referenced Mode the system detects unexpected amplitude changes over a given interval The test tones in Frontline s Reference Audio files have a fixed amplitude level over their duration Therefore if the corresponding audio levels received over the air by the system fluctuates more than a specified level this level is based on the received audio stream parameters then the system generates an Amplitude Fluctuations event The system calculates amplitude fluctuations as Max Level Min Level x100 Max Level Min Level The information reported with the Amplitude Fluctuations even
46. ice s local HFP memory able z e In AOLE e testngt E Hala Two DUTs Streaming Play the test in a browser onthe Make a call to 434 964 1407 or audio overa audio source device 434 964 1304 through a cellular cellular one network The phone number network https youtu be rmirD bikriM receiving the call playbacks recorded test signal Streaming Play the test in a browser onthe Make a call to 434 964 1407 or audio overa audio source device 434 964 1304 through a VoIP Wi Fi network ae provider such as Skype The https youtu be rmirD bikrtM phone number receiving the call playbacks recorded test signal Potential problem The VoIP provider might use custom codecs and cause undesirable behavior A2DP Playing the test file locally The simplest way to perform music data testing is to directly play the reference file from DUT1 to DUT2 To do that save the reference file provided with the ComProbe software on the Source device Then connect the Bluetooth enabled devices and play the music file from one device to the other The software will automatically detect the mode and present analysis for the user 48 Chapter 3 Audio Expert System User Guide Audio Expert System Reference User Guide Playing the test file via Internet If the user is testing a scenario where they need to analyze audio played through the internet either using Wi Fi or cellular data plan they may access the reference file on YouTube
47. in 11 8 dB TermFreq 400 0 y Note This is an example The display may vary with the reference file version The ComProbe analyzer has successfully detected the reference test signal and the system is locked into reference mode DUT 1 DUT 2 Audio Source Audio Sink Cellular network Wi Fi network Audio source Figure 3 16 Test Cases for Referenced Mode Testing 3 5 Air Sniffing Positioning Devices When capturing over the air packets proper positioning of the ComProbe hardware and the Devices Under Test DUTs will result in the best possible captures and will mitigate sources of path loss and interference The following procedures will help optimize the capture process especially if you are have problems obtaining reliable captures Problems with indoor radio propagation Even in free space it is well understood that radio frequencies attenuate over distance The free space rule of thumb dictates that radio energy decreases in strength by 20 dB by each 10 to 1 increase in range In the real world the effects of objects in an outdoor environment cause reflection diffraction and scattering resulting in greater signal losses Indoors the situation can be worse Reflections occur from walls and other large flat surfaces Diffraction occurs from objects with sharp edges Scattering is produced from objects with rough surfaces and from small objects Also any object directly in the path of
48. indicator the audio volume is at its highest level As the volume decreases the bars will move to the right linearly with no visible green bar indicating no audio The volume indicator will continue to operate if the audio stream has been muted Mute Mute Checking the Mute check box will silence the Wave Panel s audio output The volume indicator will respond to the audio volume but nothing will be heard All panels can be simultaneously muted using the Audio Expert System Global Toolbar The Wave Panel mute is a local control only However the Global Toolbar mute control will set the Stream Panel s Local Controls mute Vertical Zoom Each Wave Panel contains local Vertical Zoom controls that expands or reduces the waveform display vertically The waveform amplitude is always visible and the Vertical Zoom controls increases or decreases the entire vertical size of the display The vertical zoom buttons will turn gray and become inactive when the maximum and minimum values are reached Collapse Expand Control Collapse Expand button toggles between two views The top image indicates that the Wave Panel is expanded When the bottom image is visible it indicates that the Wave Panel is collapsed T When the top image is visible clicking on it will collapse the Wave Panel to the minimum size that shows only the Stream Info and the Local Controls When the bottom image is visible clicking on it expands the Wave Panel to full size
49. interest Release the mouse key The selection is surrounded by a blue border For either of the procuedures described in the table above once the selection is made details of the segment appear below and to the left of the waveform These details include selection start and stop range TO and T1 the time difference dT samples selected frequency and Bluetooth Frames selected Right clicking in the Waveform panel will open a pop up menu see Wave Panel amp Event Table Pop up Menu on page 44 Selecting Zoom to Selection will expand the selection to the full width of the Wave Panel Other selection option in the pop up are Select Area Clear Selection and Copy Selection 39 Audio Expert System Reference User Guide Chapter 3 Audio Expert System User Guide Bitrate Overlay Display 01 31 37 257 PM 01 31 44 990 PM 01 31 52 723 PM 01 32 00 457 PM 01 32 08 190 PM Figure 3 9 Actual Bitrate Overlay The Average and Actual audio stream bitrate graphs can be displayed over the audio waveform using the Global Toolbar Average Bitrate Overlay and Actual Bitrate Overlay buttons respectively These are presented as overlays onto the main Wave Panel so the user can correlate audio issues with bitrate changes and the like The scale is in kbps kilo bits per second Hovering over the bitrate scale will display a pop up showing the bitrate at the play cursor position Actual Bitrate is based on the throughput at the Codec leve
50. io Expert System User Guide Audio Expert System Reference User Guide Play Cursor The Play Cursor is identified by a white vertical line on the Wave Panel The Play Cursor appears when user clicks on any point in the waveform or if the cursor is already present it can be dragged to another position To drag the Play Cursor hover the mouse cursor over the Play Cursor until the mouse cursor changes to a pointing hand click and drag the cursor to a new position Waveform Segment Selection 03 02 10 159 PM 03 02 10 214 PM M dT 00 00 00 010 Samples Selected 458 Bluetooth Frames 42721 CLP IF FL Figure 3 8 Selection in the Audio Waveform A waveform segment selection is identified by a blue border surrounding the selection Procedures for selecting a segment depend on the desired actions Table 3 5 Segment Selection Procedures Loop play back 1 Zoom in to the waveform segment of interest 2 Click in the approximate center of the proposed selection This will place the Play Cursor in the area to be selected 3 Move the mouse cursor to the right or left of the Play Cursor click and hold then drag over the waveform segment of interest Release the mouse key The selection is surrounded by a blue border View waveform 1 Zoom in to the segment of interest details 2 Move the mouse cursor to the right or left limit of the waveform segment of interest click and hold then drag over the waveform segment of
51. ion to an in car speakerphone appliance C PSTN A Cum s BT T L T ie i Figure 2 5 Typical Topology for Voice Scenarios Figure 2 6 shows another example for music sourced from companies like Pandora and Spotify A media audio server renders the desired audio which then traverses the public internet then onto the cellular network of the mobile provider and is delivered to the smartphone which sets up the Bluetooth connection to the playback device a q a internet PLMN bra kb a Media Server ps BT T I Figure 2 6 Typical Topology for Music Scenarios What is important to understand is that the audio can pass through multiple devices and networks As a result it can go through multiple conversions from the original analog source to an initial digital format and then through multiple additional conversions as it passes through various networks or subsystems These conversions are referred to as coding and decoding operations or simply recoding operations and are performed by various codec implementations which generally perform audio compression An analog voice call for example may be encoded as uLaw PCM by the Public Switched Telephone Network Even worse when IP Telephony is involved the audio will be compressed before it ever leaves the premises The audio will then be recoded again for transmission through the Public Switched Mobile Network using the 24 Chapter 2 Audio Ex
52. is dependent on sample rate and bits per sample but it generally is 3 dB for speech and 11 dB for music Since the exact transmitted amplitude levels from the Source DUT depend on a number automatic and manual analog and digital gain settings the measured levels are normalized before comparing against limits See 2 6 Audio Volume Level Calibration on page 25 The information reported with the Unexpected Level event is listed below e Info1 The measured normalized level dB e Info2 The expected level dB e Msg A string containing the Normalization Gain factor dB 2 1 2 7 Unexpected Duration Event The Unexpected Duration UD event is reported when a tone segment of the Reference Audio file is shorter or longer than expected In Referenced Mode the system knows the Reference Audio file that is being played on the Source DUT and therefore knows how long a specific tone segment should last If either a change of amplitude or frequency arrives either before or after that programmed duration then the change is by definition unexpected Lost or corrupted data repeated data faulty packet loss concealment algorithms etc can cause this type of audio impairment The amount that a measured duration must deviate from the programmed duration of a tone segment before the system declares this event varies depending upon the negotiated over the air audio stream specific parameters but it is generally in the range of 5 to 10 Note t
53. l The Average Bitrate is the moving average over 0 1 sliding second window 12 52 40 365 PM 12 52 40 490 PM 12 52 40 615 PM 12 52 40 740 PM___ 12 52 40 865 PM Ah a E AE Figure 3 10 Average Bitrate Overlay All of the information for calculating the Acutal and Average Bitrate is in the codec data frame header 3 2 2 4 Event Timeline The Event Timeline in the Wave Panel shows the Bluetooth ES Codec Cra and Audio events related to the waveform being viewed The events are synchronized in time to the waveform displayed in the Wave Panel The event severity is displayed as Information A Warning A and Error F Figure 3 11 Event Timeline Shown with Wave Panel 40 Chapter 3 Audio Expert System User Guide Audio Expert System Reference User Guide Clicking on an event in the Event Timeline shows a relevant selection in the Audio Waveform Panel The size of the selection depends on the number of frames associated with the selected event This selection will appear in all Wave Panels however the event severity icon will only appear in the Wave Panel associated with the event To assist the user with viewing events in detail the Event Timeline will zoom in and out in sync with the Wave Panel Event Timeline Example This example shows that event 159 was selected in the Event Table resulting in the severity icon being enlarged in the Event Timeline The system automatically selected the surrounding area the blue ou
54. l Adjustment 0 coo 19 2 2 6 Amplitude Normalization Aa 20 2 3 Basic Measurement Operations 2222 cece ee ec cece eee ence cee eeeeeeceeeeeeee 20 23 L Peak DEC CHOM A E wouae coins eanbecusn cau aes S 20 2 322 Frequency DELECHON AA Ee aire EE Eeer aira 20 2 3 3 Amplitude Levels nananana ccc cece eee cece cece cece eee e cece eeceeeeeeeeeeeeeeees 21 o AA APA 21 2 4 Audio Analysis Concepts aaa 22 De TINO AA AP 22 pa WAAAH 22 2 4 3 Precision and Dynamic Range e cece cece c cece cece cece cece cece ee ceeeeeeeeeeeeees 22 2 4 4 Measurement Accuracy 2 a 23 2 5 Audio Chain Network Technologies cece cece cocoa 24 2 6 Audio Volume Level Calibration c eee cece cece ee eee eee ee cee cece nc nena 25 AAA 25 2062 The General PROCESS mensuracinatic irradia tro daa obs 26 2 6 3 Adjusting for Optimal Volume Levels _ _ 2 22 eee ee ee cece cece cee eee e cece eeeeeeeee 28 Chapter 3 Audio Expert System User Guide _ 2 22 2 lee cece naana cece ee ceeeeeeeeees 31 3 1 Starting the AudioExpert System Sodera and BPA 600 only e eee eee ee eee cece eeee 31 3 2 Audio Expert System Window 2 2 22 eee cece cece cece cece cece cece cence eeceeceeceeeeeeceeeees 32 3 2 1 Global Toolbar cx aciced occ sedconcteeodeecctantwawesabnceue oia 33 32 2 Wave Panel Aa rios dpto coastal 35 AE SMe
55. logy for Music Scenarios cnc 24 Heure 27 Example L ieee hese o ir to 25 Na o a ada ma AA AA AA cane NAGA DAAA baa UNO NA haka aaa Aka 26 Figure 2 9 Cu 26 Figure 2 10 Example Reference File Waveform 44 1kHz 16 bit 26 Figure 2 11 Test 1 02 Test ID Segment ec ce cece cece eee c nec ececeneceneceees 27 Figure 3 1 Audio Expert System Window 2 2 2 2 e cece cece ccc cece cece cece ee ceeeeeeeceeeees 32 ASUS 352 AA 35 Figure 3 3 Audio Stream Info in the Wave Panel 22 22 2ec eee cece ee cee cece cece eeceeeeeeee 36 Figure 3 4 SBC Codec Information Pop Up on Cursor Hover Over u 2 eee ee eee cece ee eeeeeees 36 Figure 3 5 Wave Panel Local Controls 0 22 Aa 37 Figure 3 6 Collapsed Wave Panel 0 a 38 Figure 3 7 Audio Waveform Panel in the Wave Panel 22 2 e eee eee eee ee eee eee eee 38 Figure 3 8 Selection in the Audio Waveform 2 cece eee cece cece cnn 39 Fig re 3 9 Actual Bitrate OVeNAY ccccccscceccevls cirio hal 40 Figure 3 10 Average Bitrate Overlay ee cee ccc ccc rr 40 Audio Expert System Reference User Guide Figure 3 11 Event Timeline Shown with Wave Panel 2 22 eee eee eee eee eee cece ee eeeeees 40 Figure 3 12 Example Event Table Selection Shown in Event Timeline 0 0 0 2 cece eee eee 41 Figure 3 13 Event Timeline Selected Event Pop U
56. nd delimiter requirements If any of these parameters do not match the process is terminated and is reset to the initial conditions Referenced Mode therefore no events related to false starts are reported This is because for arbitrary audio there is no expectation of any Test ID Note Until a Test ID is successfully recognized the system will continue to operate in Non The info reported with the Test ID Found event is listed below e Info1 The Test ID e Info2 The measured RMS average audio of the 400 Hz terminator tone This level can be converted to dBFS by the relationship Chapter 2 Audio Expert System Reference Guide Audio Expert System Reference User Guide Optionally the value can be converted to Channel Gain via the relationship dB 20109 19 22 23196 e Msg The measured frequency Hz of the terminator tone 2 1 2 2 Test Script Not Found Event This event occurs if a valid Test ID was found see Transition to Referenced Mode on page 18 but the script for that Test ID was not found The system reverts to Non Referenced Mode if this happens This event should not occur if valid Reference Audio files provided by Frontline are being used The information reported with the Test Script Not Found event is listed below e Info1 The Test ID e Info2 The peak level amplitude of the recognized Test ID terminator tone dB e Msg A message indicating the exact measure frequency of Test ID Terminator
57. nly reported in Non Referenced Mode A High Volume Alarm event is generated when the detected audio volume is continuously above the high volume threshold see Figure 2 2 for 10 consecutive 0 5 sec measurement intervals i e 5 sec total The event will not be repeated again until the detected volume level drops below the high volume threshold The High Volume Alarm event will then be reported again if the high volume condition returns for 10 more consecutive 0 5 sec measurement intervals Listed below are the high volume threshold levels Note that the thresholds are different for music versus speech since the crest factor difference between peak and average levels see Crest Factor on page 21 is different for those types of audio Note that speech is typically transported in A2DP connections and music is typically transported in SCO connections e Speech greater than 6 dBFS e Music greater than 12 dBFS The info reported with the High Volume Alarm event is listed below Audio Expert System Reference User Guide Chapter 2 Audio Expert System Reference Guide e Info1 Average RMS level at the start of the 5 sec interval e Info2 The time at when the system began detecting high volume i e 5 sec before the event was reported e Msg Not used 0 dBFS Full Scale ald for Musie _ Figure 2 2 High Volume Threshold Example 2 1 1 4 Dropout Event The purpose of the Dropout event is to report the detection of an unusu
58. nt Type Codec Table 2 4 Event Type Codec 14 Chapter 2 Audio Expert System Reference Guide Audio Expert System Reference User Guide Table 2 4 Event Type Codec continued Error Incorrect Configuration SBC Codec detected a change in audio Detected parameters Error Lost Sync SBC Codec expected to find synch word Ox9C instead found Ox typically due to corrupted data Error Bad Header SBC Codec detected corrupted header typically due to corrupted data Error CRC Failure SBC Codec detected bad CRC typically due to corrupted data Error No output SBC Codec generated no output due to corrupted data Codec tear down Codec Session Ended Stream Re configuration Stream Re configuration Warning Packet Loss Concealment mSBC Codec detected a bad frame and generated substitute data to compensate for it Error Incorrect Configuration mSBC Codec detected a change in audio Detected parameters Error Lost Sync mSBC Codec expected to find synch word OxAD instead found Ox typically due to corrupted data Error Bad Header mSBC Codec detected corrupted header typically due to corrupted data Error CRC Failure mSBC Codec detected bad CRC typically due to corrupted data Error No output mSBC Codec generated no output due to corrupted data when PLC not configured Information Codec initialization initialization Codec session started Codec session started started mee eee tear down Codec s
59. ntrols waveform viewing controls and volume controls that affect all Wave Panels e Wave Panel Displays the waveforms for each captured audio stream There is a separate Wave Panel for each stream Each panel contains local information controls and an event timeline specific to the displayed audio stream being shown Other Wave Panels that may be off scream may be viewed using the vertical scroll control or by collapsing other Wave Panels e Event Timeline The Event Timeline shows Bluetooth events Codec events and Audio events synchronized to the displayed waveform There is an Event Timeline in each Wave Panel e Event Table A tabular listing of Bluetooth codec and audio events with information on event severity related Bluetooth frame timestamp and event information Global Toolbar Wave Panel pare Audio Expert Susie aa nia proa 03 39 35 343 PM 03 39 35 436 PM 03 39 35 528 PM 03 39 35 621 PM 03 39 35 713 PM 10 Lock Event Table i Event Severity Stream Id Event Type Mode Frame Number Description Timestamp 1 E N A 2822 A2DP connection is starting between devices 00 18 6B 35 A2 86 and A0 82 1F F5 00 62 for stream 1 Dec 30 2014 03 39 35 371056 PM Mo N A 2832 A2DP connection has started and data should follow between devices 00 18 6B 35 A2 86 and A0 82 1F F5 00 6 Dec 30 2014 03 39 35 386681 PM 1 N A 2839 Unable to process AptX data as extracted It appears that SBC encoded data is being sen
60. o Expert System in either live capture or captured file mode Chapter 2 Audio Expert System Reference Guide 2 1 Audio Expert System Overview The primary goal of the Audio Expert System is to expedite the detection and resolution of Bluetooth protocol related audio impairments To achieve this the system automatically identifies audio impairments and reports them to a user as events It also correlates the audio events with any detected codec or Bluetooth protocol anomalies events The system allows a user to view the audio waveform audio events codec events and Bluetooth protocol events on a time aligned display The system works in two modes of operation e Non Referenced Mode The system is processing audio of completely unknown program content see Non Referenced Mode Event Reporting on page 3 e Referenced Mode The system is operating in a pseudo closed loop test scenario where the user plays specific Reference Audio files pre recorded audio test files provided by Frontline on the Source DUT device under test that is sent via Bluetooth to the Sink DUT Frontline s Audio Expert System analyzes resulting over the air audio see Referenced Mode Event Reporting on page 7 The Audio Event Processor is a subsystem of Frontline s Audio Expert System and it monitors and analyzes Bluetooth audio streams with the purpose of detecting and reporting audio impairments 2 1 1 Non Referenced Mode Event Reporting In Non
61. oding done inside the Sink DUT Frontline s Audio Expert System automatically detects that a Reference Audio file is being received and then analyzes the resulting audio for deviations from expected parameters The Reference Audio files are specific audio files that exercise the system so that audio impairments can more efficiently and accurately be identified and reported The Reference Audio files are composed of a series of back to back and relatively short duration tones of changing amplitude frequency and duration The analysis of the received audio results in a series of Audio Events being reported by comparing changes in the received audio to expected changes of the Reference Audio and reporting deviation events when they occur 2 2 1 Transition to Referenced Mode The system starts up in Non Referenced Mode and is continuously looking for a valid Reference Audio file by measuring frequency and amplitude of the received over the air audio Transitioning to Referenced Mode requires the successful detection of a TestID tone sequence of proper frequencies and durations Once the Referenced Mode state is achieved the expectation is that all tones encountered will conform to the script identified by the collected Test ID digits The system remains in the Referenced Mode state until either the end of test is reached or a loss of synchronization occurs see Synchronization Lost Event on page 9 and Synchronization on page 19 18
62. of time equal to 1 Sample Frequency 2 3 2 Frequency Detection The instantaneous frequency is computed from the received audio based on the discovery of each peak This is done by maintaining a last peak sample number subtracting that value from the sample number of the current peak and multiplying by the sample time For example a 1 kHz tone sampled at 8 kHz precisely at its peaks will result in 8 samples between each peak So if a current peak is at sample number 8088 and the previous peak was sampled at sample number 8080 and since the sample time of 8000 Hz is 0 000125 sec then e 8088 8080 0 000125 0 001 and e 1 0 001 1000 Hz The system maintains a Frequency Found state and Last Average Frequency variable to aid in the determination of frequency changes To improve accuracy the system averages frequency measurements over 20 Chapter 2 Audio Expert System Reference Guide Audio Expert System Reference User Guide several samples This becomes especially important at higher frequencies where there are fewer samples per cycle and the detected peak may actually be 1 or more samples different than the location of the actual peak On each peak after initialization or reset the measured peak to peak frequency is added to an accumulator and when the configured number of measurements is reached the average is computed If the percent change from the Last Average Frequency exceeds an audio
63. one or more events in the table will highlight the associated frames in the standard ComProbe software windows such as Frame Display Coexistence View Bluetooth Timeline etc Lock Event Table F Seventy Stream ld Event Type Mode Frame Number Description Timestamp N A Packet retransmission Mar 31 2014 12 52 38 080991 PM N A A2DP paused between devices 98 0D 2E 23 B6 2E and 00 07 62 0F 00 00 for stream 1 using codec SBC Mar 31 2014 12 52 45 553569 PM N A A2DP paused between devices 98 0D 2E 23 B6 2E and 00 07 62 0F 00 00 for stream 1 using codec SBC Mar 31 2014 12 52 45 617944 PM N A SCO connection request Mar 31 2014 12 52 46 151071 PM N A SCO connection established between devices 98 0D 2E 23 B6 2E and 00 07 62 0F 00 00 for stream 2 using cod Mar 31 2014 12 52 46 504191 PM N A SCO connection established between devices 00 07 62 0F 00 00 and 98 0D 2E 23 B6 2E for stream 3 using cod Mar 31 2014 12 52 46 504191 PM N A Codec CVSD Frequency 64000 Bits Per Sample 16 Channels 1 Mar 31 2014 12 52 46 806067 PM N A Codec CVSD Frequency 64000 Bits Per Sample 16 Channels 1 Mar 31 2014 12 52 47 357946 PM N A SCO disconnected between devices 98 0D 2E 23 B6 2E and 00 07 62 0F 00 00 for stream 2 using codec CVSD Mar 31 2014 12 53 04 151789 PM N A SCO disconnected between devices 00 07 62 0F 00 00 and 98 0D 2E 23 B6 2E for stream 3 using codec CVSD Mar 31 2014 12 53 04 151789 PM N A A2DP resumed between devices 98 0D 2E
64. orted when a measured frequency deviates from an expected frequency by a specific percentage 6 Unexpected Level Event Error Reported when the measured level at the start of a tone segment is not within tolerance 7 Unexpected Duration Event Error Reported when a tone segment of the Reference Audio file is shorter or longer than expected 8 Amplitude Fluctuations Event Warning Reported if the system detects unexpected amplitude changes over a given interval 9 Clipping Event Error Reported when consecutive samples at the maximum value that can be represented by the digital system have been detected 10 Unexpected Phase Change Event Error Reported if the system detects a phase change during programmed intervals of constant frequency 11 Excess Noise Event Error Reported when energy sufficiently above the Silence Threshold is detected during programmed segments of silence 12 CVSD HF Level Too High Event Error Reported when a CVSD encoded audio stream is detected and there is high frequency energy above 4 kHz that is greater than 20 dBFS 13 Endof Test Event Information Reported to indicate that the system has completed processing a test script for a Reference Audio file and that the system has exited Reference Mode 2 1 2 1 Test ID Found Event The Test ID Found event see Test ID on page 19 occurs when a valid Test ID has been recognized A valid Test ID must meet the level frequency duration a
65. p _ 2 22 ieee eee eee cnn 42 Figure 3 14 Event Table 0 20 2 ec ccc cee cece cece cece cece cee ceeceececeeceececeeceeceeeees 42 Figure 3 15 Export Audio Data dialog 2 occ cece cece cece cence ec cececcececceceecs 46 Figure 3 16 Test Cases for Referenced Mode Testing 0 ccc cece cece ecccecceececeeeees 50 Figure 3 17 Devices Equally Spaced in the Same Horizontal Plane 2222 22 eee eee eee eee 52 Figure 3 18 Wideband Capture Devices Equally Spaced in the Same Horizontal Plane 53 Figure 3 19 For Audio A2DP Position Closer to SINK DUT HA 54 Figure 3 20 Example Poor Capture Environment 2 eee cece cece cece ccceccceccecccceeceeeees 54 V Chapter 1 About this Guide This guideis divided into two parts e Frontline Audio Expert System Reference guide and e Audio Expert System User Guide The Reference Guide Chapters 2 is a thorough description of Frontline s Bluetooth audio analysis features terminology event types and audio measurements The User Guide Chapter 3 provides specific instructions for using the Audio Expert System as an integral component of the ComProbe Protocol Analysis System Users new to Bluetooth audio analysis will find the Reference Guide chapter useful reading prior to using the Audio Expert System for troubleshooting Bluetooth audio Experienced users may want to skip to the Use
66. pert System Reference Guide Audio Expert System Reference User Guide provider s internal Codec technology of choice and then decoded back to analog by a smartphone When a Bluetooth appliance is added at the end of the chain yet another recoding operation is inserted For music similar recoding operations take place as for voice The original audio is digitally compressed into a format such as MP3 AAC etc traverses the internet and PLMN and is then recoded across the Bluetooth channel The important point is that Bluetooth rendered audio typically passes through multiple coding and recoding operations The resultant audio is often characterized as having been subjected to the copy of a copy syndrome Also at various points along the path digital or analog controls over volume level mayexist Interactions between multiple serial concatenated digital processes and improper level settings across the audio chain are among the root causes for poor audio at the final listening point 2 6 Audio Volume Level Calibration 2 6 1 Overview As mentioned in section Audio Chain Network Technologies on page 24 the audio that is ultimately delivered to a pair of speakers or headphones passes through a number of processing functions and or appliances Some of processing functions include a volume control and Bluetooth playback devices typically have a volume control A volume control on the Bluetooth source device DUT1 such
67. ping is a type of amplitude distortion The system reports a Clipping event when consecutive samples at the maximum value that can be represented by the digital system have been detected Note that the maximum value that can be represented is different Chapter 2 Audio Expert System Reference Guide Audio Expert System Reference User Guide depending on the number of bits per sample i e bits of resolution of the audio stream The system limits the number of reported clipping events to typically 10 to 20 per second The number of consecutive samples needed to qualify as a clipping event depends on both sample rate and number of bits per sample Table 2 1 Clipping Event Threshold specifies the number of consecutive samples at the maximum value level that will generate a Clipping event Table 2 1 Clipping Event Threshold The information reported with the Clipping Event is listed below e Info1 contains an integer value representing the RMS average audio level when the clipping event was detected The value can be converted to a decibel level via the relationship Info1 20Log Mp e Info2 Not used e Msg Not used 2 1 1 3 High Volume Alarm Event The purpose of the High Volume Alarm event is to warn the user that the volume level of the detected audio is above the maximum acceptable level for performing meaningful audio analysis i e approaching a level where the audio will likely become distorted This event is o
68. problems with capture methods including positioning and environment because it will point out missing frames For hands free profile data captures both DUTs send and receive data Therefore position the devices following the equilateral triangle arrangement as mentioned above However in A2DP data capture scenario the equilateral positioning of devices is not optimum because normally only one device is sending data to the other It is recommended that the ComProbe hardware be positioned closer to the device receiving data so that ComProbe better mimics the receiving DUT Position the DUTs 1 2 meters apart for Class 1 and 2 transmitters and 1 2 meter apart for Class 3 transmitters Ds Audio Expert System Reference User Guide Chapter 3 Audio Expert System User Guide Figure 3 19 For Audio A2DP Position Closer to SINK DUT Poor Placement A poor test configuration for the analyzer is placing the DUTs very close to each other and the analyzer far away The DUTs being in close proximity to each other reduce their transmission power and thus make it hard for the analyzer to hear the conversation If the analyzer is far away from DUTs there are chances that the analyzer may miss those frames which could lead to failure in decryption of the data Obstacles in close proximity to or in between the analyzer and the DUTs can interfere and cause reduction in signal strength or interference Even small objects can cause signal scatt
69. provided by Frontline https youtu be rmirDbikrtM Note that the software is only analyzing the Bluetooth link between the two DUTs Any abnormalities at the Wi Fi and cellular network level will affect the audio quality that may not be Bluetooth protocol related and the software will not be able to detect that HFP Playing the test file by calling a phone number Frontline provides the following phone numbers 434 964 1407 and 434 964 1304 that users can call to conduct speech audio data analysis over Bluetooth The calls can be made using the cellular network most common method or VoIP Again the VoIP provider might use custom codecs and cause undesirable behavior which cannot be detected by Audio Expert System software Playing the test file using Third party Apps Bluetooth Audio Expert System Reference mode testing can be accomplished using third party apps on Android OS and Windows phones The following apps are available from their respective App stores e BTmono Android e Blue2Car lOS e Windows Headset player lite Note When selecting and using these apps thoroughly review all the vendor documentation While Frontline has conducted testing of these apps Frontline has not completed full Si interoperability testing with our library of Bluetooth devices and does not warrant the use of these apps with every device when using the following procedures Frontline does not provide support or maintenance for third party apps
70. ption will select all the data between and fills pauses or gaps Select All Selects the entire waveform Event Table Pop up Menu Actions Right clicking in the Event Table will provide you with a selection of the following actions b Clear Selection d Export Event Table Go Loop 38 A Zoom to Selection pr f Select Area pr w Select All 30 Table 3 8 Event Table Pop up Menu Selectioin Copies the selected events to Windows clipboard as text Clear Selection Clears the current event selection in the table Export Event Table Copies the current event selection and saves it as a csv file For additional details on exporting refer to Event Table Export Loops through the audio selected on the Wave Panel Zoom to Selection Expands the Event Table selection to fill the Wave Panel view Select Area Expands the selection Select All Selects all events 3 2 4 1 Export Audio Data There are two ways to export audio data 1 Clicking the Audio Expert System window Global Toolbar Export button En 2 Right click in a Stream Panel Wave Panel and a pop up menu will appear Select Export 45 Audio Expert System Reference User Guide Chapter 3 Audio Expert System User Guide Two windows will appear 1 The standard Windows Save As 2 The Export Audio Data dialog In the Windows Save As window enter a File name and directory location Click on Save ad Export Audio D
71. r Guide Chapter but they will find the Reference Guide chapter a useful refresher or a quick reference when troubleshooting Bluetooth audio problems All users will find the User Guide beneficial as it describes all of the Audio Expert System functions buttons controls views and options in detail The User Guide information is also found in the ComProbe software help in the ComProbe Sodera User Manual PDF format ComProbe BPA 600 User Manual PDF format the ComProbe Sodera On Line Help and ComProbe BPA 600 On Line Help These last two documents are available at FTE com 1 1 What You Will Need The Audio Expert System operates in conjunction with the ComProbe Protocol Analysis System installed on your computer and with a ComProbe BPA 600 Dual Mode Bluetooth Protocol Analyzer or ComProbe Sodera Wideband Bluetooth Protocol Analyzer The ComProbe analyzer must be licensed for operation with the Audio Expert System To use the Audio Expert System you will need e Audio Expert System license The license is purchased from Frontline Sales that can be contacted at sales fte com or 1 800 359 8570 or 1 434 984 4500 e ComProbe Protocol Analysis System software installed on your computer The software release must be after 1 August 2014 Audio Expert System Reference User Guide Chapter 1 About this Guide e ComProbe BPA 600 or ComProbe Sodera with an active license The ComProbe analyzer must be connected to your computer to use the Audi
72. rage Bit Rate Overlay Displays an overlay graph of the average bit rate for the audio stream in each Wave Panel The average is based on a 0 10 second moving window Actual Bit Rate Overlay Displays an overlay graph of the instantaneous bit rate for the audio stream in each Wave Panel Export Data Exports audio data in raw and or wav format for selected Wave Panels or all the Wave Panels This button also lets user export Event Table data in csv format Refer to Waveform Export Audio Data for more details Help Opens ComProbe software help 34 Chapter 3 Audio Expert System User Guide Audio Expert System Reference User Guide Table 3 2 Global Toolbar Controls continued AENA A IA Collapse Expand Toggles between collapsing and expanding all Wave Panels Note that the Wave Panel Local Controls Collapse Expand control will locally override the Global Toolbar Collapse Expand control 3 2 2 Wave Panel The Stream Panel is where the details of the captured audio stream are presented The Stream Panel displays the captured audio waveform along with an event timeline that displays discrete Bluetooth Codec and Audio events synchronized to the captured waveform Audio Stream Info Local Controls Go ala 03 39 35 564 PM 03 39 35 671 PM 03 39 35 778 PM 03 39 35 351 PM 03 39 35 457 PM Linear Audio Waveform Event Timeline Figure 3 2 Wave Panel The Wave Panel contains
73. s the Reference Audio file that is being played on the Source DUT therefore the system knows when to expect a period of silence and can measure the noise level during this period The Excess Noise event is reported when energy sufficiently above the Silence Threshold is detected during programmed segments of silence Excess noise can indicate a poor analog audio chain with an inherently poor noise floor glitches occurring during silence intervals or codecs that do not transition to silence instantaneously The info reported with the Excess Noise event is listed below Both info1 and info2 values are integers indicating sample level and can be converted to dBFS via the relationship ABFS 20Log uf e Info1 The peak noise amplitude dB e Info2 The RMS Average of the noise level dB e Msg Unused 2 1 2 12 CVSD HF Level Too High Event The CVSD High Frequency Level Too High event is reported when a CVSD encoded audio stream is detected and there is high frequency energy above 4 kHz that is greater than 20 dBFS The information reported with the CVSD High Frequency Level Too High event is listed below 12 Chapter 2 Audio Expert System Reference Guide Audio Expert System Reference User Guide e Info1 The measured sample level It can be converted to dBFS via the relationship ABFS 20Log f gt e Info2 Not used e Msg Not used 2 1 2 13 End of Test Event The End of Test event is reported to indic
74. st Factor is 18 to 20 dB The significance of clipping events reported for arbitrary audio in Non Referenced mode can be determined by assessing the Crest Factor of the original audio and verifying the volume settings at the source to ensure sufficient headroom is provided to accommodate that ratio 21 Audio Expert System Reference User Guide Chapter 2 Audio Expert System Reference Guide 2 4 Audio Analysis Concepts 2 4 1 The Decibel It is useful to understand the use of the decibel It is simply a convenient way of expressing differences in signal levels In terms of power if a signal is reduced to half power we say it is 3 dB If a signal is reduced to half voltage we say it is 6 dB A decibel is also convenient for expressing ratios such as signal to noise ratio A 20 dB signal to noise ratio sounds like listening to music on a car radio speaker with the windows open and a jack hammer operating close by or that a 90 dB ratio allows us to listen to even the quietest passages of a symphony with little disturbance 2 4 2 Units of Measure Digital representations of analog audio often referred to as PCM Data Pulse Code Modulation represent a series of value snapshots or samples of the analog signal s instantaneous amplitude equivalent to voltage over time These samples are recorded at a rate of speed referred to as the Sample Rate For example 8000 samples sec is typically the minimum for
75. stream specific limit an internal frequency change event is declared and the Frequency Found state is set to true On all subsequent peaks the instantaneous peak to peak frequency continues to be computed If the instantaneous frequency differs from the current average frequency by a sufficient amount the Frequency Found state is set to false and the algorithm returns to its initial conditions searching for a stable frequency Such resets will ultimately result in a reported event of some type If a new frequency is established an internal frequency change event occurs lf the same frequency is found as the previous frequency an Unexpected Phase Change event occurs If the measurement is not successful a Synchronization Lost event will ultimately occur 2 3 3 Amplitude Levels Amplitude level measurements include peak and average RMS sample values encountered over the measurement interval For some of the lower bit rate and bits of resolution codecs peaks are averaged and reported on a change of the average For the highest quality settings a change may be reported on the first occurrence The measurement intervals are typically 0 05 to 0 1 sec representing 10 to 20 reports per sec The peak values are commonly used to drive user interface volume level indicators The minimum change that will trigger events is configured according to the audio stream specific parameters in effect but it is generally between 2 to 4 dB 2
76. t is listed below e Info1 Percent Amplitude Modulation expressed as an integer e g 40 40 e Info2 Not used e Msg Not used 2 1 2 9 Unexpected Phase Change Event The Unexpected Phase Change event provides a fine grained indication of lost or repeated energy In Referenced Mode the system knows the Reference Audio file that is being played on the Source DUT therefore the system knows when a specific tone should be expected During this interval the system checks that the measured average frequency is the same as the expected frequency If this is correct the system will continue to monitor the instantaneous frequency lf the instantaneous frequency deviates sufficiently from the current average frequency the frequency measurement state machine will reset and begin re measuring Typically the outcome is the discovery of the next scripted expected frequency see Unexpected Frequency Event on page 10 and Frequency Detection on page 20 However another outcome can be that the same frequency as the previous average frequency is rediscovered and this is reported as an Unexpected Phase Change event Such phase changes are an indicator of losses of signal that do not result in amplitude dropouts or signal substitution repetition of previous audio energy due to things such as packet loss concealment tactics The information reported with the Unexpected Phase Change event is listed below Info1 The phase jump expressed in d
77. t over this stream Dec 30 2014 03 39 35 418557 PM A 1 2850 A 21 25 millisecond delay has been detected between this frame and the previous frame compared to the overall Dec 30 2014 03 39 35 422307 PM a pn ALO Processing II Analysis os Event Table Event Timeline Figure 3 1 Audio Expert System Window Color Codes and Icons The Audio Expert System uses standard color codes and icons to assist the user in focusing on specific issues e Chapter 3 Audio Expert System User Guide Audio Expert System Reference User Guide Table 3 1 Audio Expert System Color Codes and Icons Note If an Event Severity icon is surrounded by a dark line the event is a global event and not H applying to a particular captured waveform The event is assigned to Stream 0 in the Event Table The following topics describe the Global Toolbar Wave Panel Event Timeline and Event Table in more detail 3 2 1 Global Toolbar The global toolbar provides audio play controls audio play cursor positioning controls waveform viewing controls and volume controls Global toolbar controls apply simultaneously to all waveform panels n n a a a Table 3 2 Global Toolbar Controls Home Moves play cursor to beginning of the waveform Play Start playing the audio from the current play cursor position Toggles to Pause when clicked Pause Stops audio play back at its current position toggles to Play when clicked End
78. t resolution When a Test ID is validated and the system transitions into Referenced Mode the measured amplitude of the Test ID Terminator Tone is used to calculate a normalization gain factor This is the amount of gain required to adjust the measured level of the received audio to 3 dBFS This normalization gain is then applied to measured levels before comparing them to expected levels As an example if the Test ID Terminator is received at 5 dBFS 2 dB below the expected value then the Normalization Gain is 2 dB A subsequent tone scripted at 12 dB but actually received at 14 dB would be increased by 2 dB to 12 dB before comparing it to the 12 dB value in the script and would therefore be accepted as a correct level 2 3 Basic Measurement Operations 2 3 1 Peak Detection Detecting peaks is a basic function that underlies all operations because it is common to both amplitude and frequency metrics and calculations of amplitude and frequency are made continuously by the system Sample to sample changes in amplitude are tracked to determine transitions from a positive to negative first derivative slope When a peak is detected its real time level is immediately available for amplitude related measurements Also its sample index is saved and when the next peak is detected the real time frequency is computed based on the time between peaks Frequency F is the reciprocal of time t F 1 t and each sample represents a unit
79. the radiation can present a hard or soft partition depending on the partition s material properties Path losses from partitions are difficult to estimate 50 Chapter 3 Audio Expert System User Guide Audio Expert System Reference User Guide Estimating indoor propagation loss 1 One estimate of indoor path loss based on path loss data from a typical building provides a range power rule At 2 4 GHz the following relationship provides an approximate estimate of indoor path loss Indoor Path Loss in dB 40 35Log range in meters This approximation is expected to have a variance of 13 dB Mitigating path loss and interference Bluetooth device design contributes to mitigating environmental effects on propagation through spread spectrum radio design for example However careful planning of the testing environment can also contribute to reliable data capture process The first step to ensuring reliable air sniffing data capture is to understand the RF characteristics of the Devices Under Test DUTs The Bluetooth Class antenna types and radiation patterns are all important factors that can affect the placement of the DUTs and the ComProbe analyzer Radiation patterns are rarely spherical so understanding your device s radiation patterns can greatly enhance successful data capture Position devices to avoid radiation attenuation by the surroundings This step is optional Consider conductive testing to establish
80. tline 01 32 47 320 PM 01 32 47 339 PM 01 32 47 357 PM 1 32 47581PM dT 00 00 00 500 Samples Selected S000 Frequency 2 Blueto Event Saray Stream ld Event raan Mode Frame Number Description 159 39561 Unexpected Level Measured 6 0 dB Expected 3 0d6 o MP O fe FOQS Report FOGS Repot 1 83213 Fchg Fret 0 000 Fr Figure 3 12 Example Event Table Selection Shown in Event Timeline Event Pop Up When the cursor hovers over a selected event severity icon in the Event Timeline a pop up will display the event class severity and associated Bluetooth frame 41 Audio Expert System Reference User Guide Chapter 3 Audio Expert System User Guide 01 32 47 320 PM 01 32 47 339 PM T0 01 32 47 331 PM 11 01 32 47 331 PM dT 00 00 A y Me Class of event Audio Seventy Error Bluetooth Frame 39561 her 3300 REF 139561 Figure 3 13 Event Timeline Selected Event Pop Up 3 2 3 Event Table The Event Table lists all audio stream events Clicking on an event will select that event in the Event Timeline in the Wave Panel If the selected event is outside the visible area of the waveform the waveform will move and bring the selected event to the center of the display The event icon in the Event Timeline is also centered and the selected icon will be larger than the non selected event icons Selecting
81. tone Hz 2 1 2 3 Invalid Test Script Event This event is generated when an error occurs while accessing information in a script This event should not occur if using a valid reference audio file provided by Frontline The information reported with the Invalid Test Script event is listed below e Info1 The Test ID e Info2 For most cases this is the audio segment number at which the error occurred For the special case that initialization fails Info2 holds the number of segments found e Msg Contains a case specific message indicating the nature of the message 2 1 2 4 Synchronization Lost Event This event is generated when the system is in Referenced Mode after a successful initial Test ID recognition see Synchronization on page 19 and encounters unexpected frequencies or durations of audio segments while analyzing a received Reference Audio file If this situation occurs the internal segment tracking logic attempts to look forward and or backward in the test script to determine if the currently measured characteristics are consistent with the previous or next segment of the script If there is a match the internal segment pointer is advanced or retarded appropriately the Synchronization Lost event is not generated and the audio analysis continues However if a match cannot be found the system declares itself out of sync and generates the Synchronization Lost Event terminates any active test script and reverts to Non Referenced Mo
82. ude level is checked to ensure that it is above a minimum Measurement Threshold required for accurate operation To ensure adequate margin the user must adjust the system under test such that the over the air Test ID tone amplitude level is within the acceptable ranges listed below Note that the system reports the Test ID tone level in the Info1 parameter of the Test ID Found event 19 Audio Expert System Reference User Guide Chapter 2 Audio Expert System Reference Guide e Speech 15 dBFS to 3 dBFS e Music 20 dBFS to 3 dBFS In addition to the acceptable amplitude ranges listed above see Audio Volume Level Calibration on page 25 and Adjusting for Optimal Volume Levels on page 28 for the optimal over the air amplitude levels and description of how properly to adjust the levels 2 2 6 Amplitude Normalization Because the exact transmitted over the air level from the Source DUT devices depends on a number automatic and manual analog and digital gain settings and because the system needs to determine deviations from expected levels Frontline s Audio Expert System provides a mechanism for automatic level normalization Normalization is the process of adjusting measured level to account for known gain or loss within the system The Test ID Terminator Tone plays a key role in this normalization process All Test ID tones in Frontline s Reference Audio files are generated at 3 dBFS a maximum sample value of 23196 for 16 bi

Audio Expert System Reference & User Guide

Contents

Download Pdf Manuals

Related Search

Related Contents

Audio Expert System Reference &amp; User Guide

Contents

Download Pdf Manuals

Related Search

Related Contents

Audio Expert System Reference & User Guide