Line / Network Echo Canceller

VOCAL’s Line/Network Echo Canceller is used in telecommunication applications to remove audible echoes caused primarily by electric signal reflections from the telephone hybrid circuit (a.k.a. telephone line transformer).

Telephone hybrid circuits convert 4-wire interface into 2-wire interface in POTS/PSTN systems. Historically, the introduction of the 4-wire-to-2-wire interface was motivated by the need to reduce the cost of copper loops connecting the POTS telephone line facility (access network) to individual subscribers. Figure 1 illustrates a typical line echo canceller operating at the near-end of the voice connection.

Signals travelling from the far end to the near end are reflected from the hybrid circuit, as a result of inadequate impedance matching between the telephone line and the hybrid terminals facing the telephone line.  Although hybrid circuits are equipped with balance circuits that are selectable, ideal impedance matching can hardly be achieved due to line makeup/telephone set variations, which are not under telephone company control.  These variations may occur even during the telephone call. Therefore, one of the requirements for the Line/Network Echo Canceller is that the adaptive filter, whose role is to rapidly identify the echo path (and specifically, the echo path impulse response), enters in the state of adaptation as soon as the voice channel between the near end and far end is open.

Typically, the echo path impulse response of the hybrid circuit alone is relatively short and seldom exceeds 8-12 milliseconds. However, additional signal delays, due to signal paths linking the trunk facility and the line facility, have to be accommodated. This is why often Line/Network Echo Cancellers offer echo span coverage as large as 128 milliseconds.

Typical building-blocks of Line/Network Echo Canceller are illustrated in Figure 2. The Adaptive Filter is one of the most important blocks. The internal structure of the adaptive filter as well as the adaptation algorithm are critical from the viewpoint of the echo canceller adaptation speed and depth characteristics. Double-talk detector block controls the adaption mechanism: if the near-end signal  (Sgen) is present, the double talk halts the filter adaptation and thus prevents the filter from divergence. The Nonlinear processor removes traces of residual echo from the error signal. The delay block (which often is optional) shifts in time the input signal to the adaptive filter in order to compensate signal delays between the echo canceller ports (Rout and Sin) and the ports of the line hybrid circuit.

VOCAL’s Line/Network Echo Canceller offers reconfigurable echo path coverage of up to 128 ms.

Features

• Fully compliant with ITU G.168-2015 Recommendations
• Rapid convergence and re-convergence ( for echo path changes)
• Double talk detection
• No divergence during double talk
• Adjustable echo tail length to 32 msec for full compliance
• Adjustable echo tail length to 128 msec with restricted compliance
• Proper operation during facsimile and low speed (<9.6 kbit/s) voice-band data transmissions
• Disables automatically in the presence of ANSam
• Excellent tone rejection and narrow band stability
• Non-linear processing (NLP) with Comfort Noise Generator (CNG)
• Tone detector and hold release logic
• DC Offset Correction
• Bulk Echo Delay Estimator
• Utilizes Normalized LMS (NLMS)
• Supports multiple channels
• Supports user callable functions
• Easily portable, re-entrant and re-locatable code

Operational Characteristics

Exceeds those set forth by the ITU-G.168 Recommendations

• Fast Convergence with NLP Disabled: L_Rin – L_RES > 20 dB in 1 second
• Steady State Convergence with NLP Disabled: L_Rin – L_RES > 30 dB
• Fast Convergence with NLP Enabled: L_Rin – L_RET > 20 dB less than 1 second
• Steady State Convergence with NLP Enabled: L_Rin – L_RET > 55 dB
• Fast Re-Convergence for Echo Path Changes: L_Rin – L_RET > 20 dB less than 1 second
• Steady-State Leak Rate: Increase Residual Echo < 10dB
• No generation of echo of when echo path is open, ∞ ERL
• Comfort Noise Generator tracks background noise level changes
• DTMF Transparency

Design Constraints

• The standard VOCAL echo canceller operates on 16-bit signed input values, and returns 16-bit signed values.
• A variable internal excitation delay is available to provide full echo canceller tail length in the presence of known system latencies.

Supported Processors

• Texas Instruments – C6000 (TMS320C62x, TMS320C64x, TMS320C645x, TMS320C66x, TMS320C67x, DaVinci), OMAP, C5000 (TMS320C54x, TMS320C55x)
• Analog Devices – Blackfin, TigerSHARC, SHARC
• PowerPC
• MIPS – MIPS32, MIPS64, MIPS4Kc
• ARM Architectures: ARMv5TE(J), ARMv6, ARMv7-M, ARMv7E-M, ARMv7-A, ARMv8-M, ARMv8-A
• Intel / AMD – x86, x64 (both 32- and 64-bit modes)

Supported Operating Systems

• Google Android
• Apple iOS / iPhone / iPad & MacOS
• Unix, Linux, µClinux, BSD
• Microsoft Windows ACM / RTC / CE / Mobile
• Symbian
• eCOS / eCOSPro
• Wind River VxWorks
• Green Hills Integrity
• VOCAL LANsEND
• Micrium µCOS

Standards Compliance

VOCAL’s LEC/NEC software solutions are designed to pass G.168-2015 requirements using any echo path models defined by G.168-2015 as well as customer models, provide they fit to the echo path coverage of the echo canceller using G.168-2015 testing.

VOCAL’s line echo cancellation modules pass the standard ITU G.168 Test Suite:

• Test No. 2 – Convergence and Steady state residual and returned echo level tests
• Test No. 2A – Convergence and Re-convergence tests with NLP enabled
• Test No. 2B – Convergence and Re-convergence tests with NLP disabled
• Test No. 2C – Convergence test in the presence of background noise
• Test No. 3 – Performance under conditions of double talk
• Test No. 3A – Double talk test with low cancelled-end levels
• Test No. 3B – Double talk test with high cancelled-end levels
• Test No. 3C – Double talk test under simulated conversation
• Test No. 4 – Leak rate test
• Test No. 5A – Infinite return loss convergence test
• Test No. 5B – Re-convergence during call transfer test
• Test No. 6 – Non-divergence on narrow-band signals
• Test No. 7 – Stability test
• Test No. 9A – Comfort noise test with step change of background noise level
• Test No. 9B – Comfort noise test with dynamic change of background noise level
• Test No. 10 – Facsimile test during call establishment phase
• Test No. 10A – Canceller operation on the calling station side
• Test No. 10B – Canceller operation on the called station side
• Test No. 14 – Performance with V.Series Low-speed Data Modems
• Test No. 15 – PCM offset test
• Test No. 16 – DTMF Transparency Test

Proprietary G.168 Line Echo Canceller Test Compliance

As an experienced leader in this field, VOCAL has identified shortcomings in the G.168 standard and has implemented several additional test sets to guarantee world-class performance of our G.168 line echo canceller software in the harshest acoustic environments. All of our line echo cancellation modules must pass these VOCAL specific tests in addition to the standard ITU G.168 Test Suite, as well as undergo voice testing in our labs.

• VOCAL Test 2D – improved echo power test
• VOCAL Test 3D – double talk in the presence of tones
• VOCAL Test 5D – improved line change detection
• VOCAL Test 7D – improved narrow band rejection
• VOCAL Test 9D – improved comfort noise tracking
• VOCAL Test 9E – comfort noise tracking during singletalk
• VOCAL Test 9F – comfort noise tracking during doubletalk
• VOCAL Test 9G – comfort noise transitions
• VOCAL Test 21 – ANSam detection

More Information

LEC Subsystems

Double-Talk
Non-Linear Processing in Echo Cancellation
Tone Detectors in Echo Cancellers
High Pass Filters in Echo Cancellers

More Information on VOCAL’s categories of Echo Cancellation Solutions