Signal Flow — AV System Signal Chain

Signal flow is the path audio and video signals travel from source to destination, through every processing and distribution stage. Understanding signal flow is the foundation of AV system design, commissioning, and troubleshooting. A clear signal flow diagram, drawn before any equipment is specified or installed, defines every component's role and exposes gaps or conflicts before they become problems on-site.

Audio Signal Levels

Before following the signal path, understanding the level standards is essential:

Mic level — Output of a microphone: approximately -60 to -40 dBu (1 mV to 10 mV). Extremely low voltage; requires a preamp to boost to line level before any significant cable run or processing.

Line level (professional, +4 dBu) — Standard operating level for professional audio equipment (DSPs, console inputs/outputs, amplifier inputs, recording interfaces). 0 dBu = 0.775 VRMS. Professional equipment uses +4 dBu nominal level (approximately 1.23 V) with headroom to +24 dBu before clipping.

Line level (consumer, -10 dBV) — Consumer and prosumer equipment (audio interfaces, some AV receivers, older multimedia hardware) operates at -10 dBV nominal (316 mV). The 14 dBV difference between professional and consumer levels causes distortion if a consumer output drives a professional input without level matching.

Instrument level — Guitar and bass pickups: similar to mic level but high impedance (typically 1 MΩ input required). Not commonly encountered in AV systems.

Speaker level (amplified) — Post-amplifier signal driving loudspeakers: typically 1–100 V RMS depending on power output and impedance. Never connect a speaker output to any line-level input.

Digital audio levels — In digital systems, levels are expressed in dBFS (decibels full scale), where 0 dBFS is the maximum digital value. Professional convention: nominal operating level is -18 dBFS (leaving 18 dB of headroom). Consumer convention: -12 to -14 dBFS nominal.

Audio Signal Chain — Microphone to Output

A complete audio signal chain for a conference room microphone:

1. Transducer → Microphone capsule converts sound pressure to analog voltage (-60 to -40 dBu).

2. Mic preamp → Boost mic level to line level. Phantom power (48V) supplied for condenser mics. Gain typically +40 to +60 dB. In professional AV: built into DSP analog inputs (QSC Q-SYS, Biamp Tesira), mixing console, or dedicated mic preamp. Set input gain so nominal talker level reads -18 dBFS at the ADC output.

3. ADC → Analog-to-digital conversion (if not already in digital domain from a digital mic). 48 kHz, 24-bit standard for AV systems.

4. Digital processing in DSP → Signal enters DSP domain:

• High-pass filter (80–120 Hz) to remove low-frequency noise
• Noise gate to suppress background noise between speech
• Automatic mic mixer (AMM) to manage multi-mic systems
• AEC (with reference from speaker output — see below)
• Parametric EQ for tonal correction
• Compressor/limiter for level management
• Digital output to next stage

5. DAC or digital output → Conversion back to analog (for amplifier input) or remaining in digital domain for Dante, AES3, or USB output.

6. Power amplifier → Line-level signal amplified to speaker-driving voltage. Amplifier sensitivity setting must match DSP output level.

7. Loudspeaker → Transducer converts amplified electrical signal back to sound pressure.

The AEC Reference Loop

In a conferencing system, signal flow is bidirectional: near-end microphones capture and transmit local speech while the loudspeaker plays far-end speech. The AEC algorithm requires a reference signal from the loudspeaker output to subtract echo from the microphone signal.

The reference loop within the DSP signal flow:

Far-end audio IN (from codec) 
    → DSP mixing → Loudspeaker output OUT → Amplifier → Speaker
                         ↓
                   AEC Reference Input (within DSP)
                         ↓
Near-end mic IN → AEC (subtracts loudspeaker echo) → Conferencing codec OUT

Critical: The AEC reference must tap the signal at exactly the point where it leaves the DSP for the amplifier, post-mixing. If the reference is taken pre-mix (before combining far-end audio with local audio sources), the AEC model will not include all of what plays through the speaker, leaving residual echo. See aec.

Video Signal Chain

A video signal chain for a presentation system:

1. Source → Laptop HDMI output, media player, camera. Carries video, audio, and control (CEC) over HDMI. HDMI supports resolutions through 4K60 (4:2:0 at 18 Gbps bandwidth for HDMI 2.0, 4:4:4 at 48 Gbps for HDMI 2.1).

2. EDID management → EDID (Extended Display Identification Data) is a data structure the display sends to the source describing its resolution and timing capabilities. The source uses EDID to select its output resolution. In systems with matrix switchers, extenders, or AV-over-IP between source and display, EDID must be managed: the switcher or extender must present appropriate EDID to the source so it knows what to output. EDID errors cause sources to output at wrong resolutions or to not output at all.

3. Signal distribution — Options:

• Direct HDMI — Source to display cable. Simple; 15' maximum without active cable or extender.
• HDBaseT extension (Crestron DTP, Extron DTP/XTP, standalone extenders) — HDMI + Ethernet + power over single Cat cable to 100m. Requires 10.2 Gbps or 18 Gbps HDBaseT for 4K/HDR.
• AV-over-IP (Crestron NVX, Extron NAV, QSC NV-32-H) — HDMI encoded over 1GbE or 10GbE. Flexible routing; requires managed switching. See notes on each platform.

4. HDCP handshake → HDCP (High-bandwidth Digital Content Protection) authenticates the source, distribution chain, and display before allowing protected content to display. Any device in the chain that does not support the correct HDCP version causes a blank screen. HDCP 1.4 is required for 1080p content from streaming sources; HDCP 2.2 is required for 4K content from streaming and Ultra HD Blu-ray. All distribution equipment must support the same or higher HDCP version as the source content.

5. Display → Final destination. Receives HDMI, processes the signal, drives the panel. Display input must match source output: resolution, color space, HDR format, refresh rate.

Conferencing System Bidirectional Signal Flow

A full conferencing system has two simultaneous signal paths:

Near-end send path (local to remote):

Room mic → Preamp → ADC → HPF → Gate → AMM → AEC → EQ → Comp → DAC/Digital → Codec → Network → Far-end

Far-end receive path (remote to local):

Network → Codec → DAC/Digital → DSP mix bus → (AEC Reference tap) → Amp → Speaker

These two paths share the DSP and must be configured as separate routing paths. Mixing them incorrectly is a common commissioning error: if far-end audio accidentally routes back through the microphone send path (without AEC), remote participants hear their own voice echoed back. This is distinct from the AEC reference loop — it is a routing error, not an AEC failure.

Gain Structure

Proper gain structure ensures the signal is at the correct level at every stage, maximizing headroom and minimizing noise.

Target levels at each stage:

Common gain staging error: Setting the DSP output level to "maximum" and using the amplifier gain control to reduce volume. This maximizes noise generated by the amplifier's gain stage and creates clipping risk if the DSP ever sends a peak above 0 dBFS. Correct approach: set DSP output to nominal level, set amplifier sensitivity to match, control volume at the DSP output fader.

Digital Audio Clock Synchronization

In a system with multiple digital audio devices (DSP, Dante network, AES3 outputs, USB audio), all devices must run at the same sample rate — and for AES3, the same word clock phase.

Dante handles clock synchronization automatically via IEEE 1588 PTP. The network elects a PTP grandmaster clock; all Dante devices lock to it. No manual configuration needed for same-domain Dante systems.

AES3 — Devices receiving AES3 extract the clock from the incoming signal (self-clocking). In a chain of AES3 devices, the first device in the chain (or a dedicated word clock generator) is the master. All other devices sync to the incoming AES3 stream. If the AES3 clock disappears (cable cut, device power off), downstream devices lose lock and produce audio glitches.

USB audio — A computer's USB audio class driver runs on the computer's internal clock. A DSP receiving USB audio from a laptop must either resample the incoming USB audio or provide an external clock reference to the computer (not commonly supported). Most AV DSPs perform transparent sample rate conversion on USB inputs.

Troubleshooting by Signal Flow

When audio fails in a system, the signal flow framework provides a systematic diagnostic path:

1. Identify where signal is present and where it is absent — Use metering in the DSP software to identify the last stage with signal. "Mic level OK, no signal at DSP output" → problem in DSP routing or output assignment.

2. Test each section independently — Replace the source with a known-good test signal. Bypass processing blocks. Output to a known-good device.

3. Check clock synchronization — Digital clicks, glitches, or silence in an otherwise correct signal flow often indicate clock mismatch. Check Dante Controller, AES3 lock indicators, or USB sample rate settings.

4. Verify routing — In the DSP software, trace the signal path explicitly from input to output. Misrouted signals are the most common DSP commissioning error.

5. Check levels — Use the DSP metering to verify signal at each processing stage is at expected level. A -60 dBFS signal at the DSP output when the expected level is -18 dBFS indicates an upstream issue (gain not set, wrong input selected, source not active).

Common Pitfalls

• AEC reference not tapped at the right point — See aec. The loudspeaker output feed to the AEC reference input must post-mix. Pre-mix tapping leaves uncanceled echo from sources added to the mix after the tap point.
• EDID mismatch causing source to output wrong resolution — A 4K source receiving 1080p EDID from the switcher outputs 1080p even to a 4K display. Always configure EDID management in the switcher or extender. Set EDID to match the highest capability the display and distribution chain can both support.
• Consumer/professional level mismatch — Connecting a -10 dBV consumer output directly to a professional +4 dBu input (or vice versa) causes level and gain staging errors. Use a level-matching transformer or select the appropriate input sensitivity on the DSP.
• Amplifier gain set too high — Amplifiers with gain cranked to maximum amplify any noise in the signal chain, including hiss from preamps and DSP. Set amplifier sensitivity to match DSP output level for proper gain structure.
• Dante sample rate mismatch — All Dante devices must be set to the same sample rate (48 kHz standard). A device at 44.1 kHz appears in Dante Controller but routing to it produces silence or noise. Check sample rate settings in Dante Controller.