Follow the chain
When a sound is missing or wrong, I start at the source and work forwards, one stage at a time. Most faults turn out to be at the first two stages: the performer, the mic or the cable.
Live sound and mixing
Live sound, from the mic to the back row
Sound engineering is my hobby, and it uses the same discipline as network design: a clear signal path, good cables and a plan for failure. This page covers the signal chain, balanced lines and levels, Allen & Heath Qu mixers and digital stageboxes, and mixing and monitors. It then covers rigs for music and conferences, the physics of the public address (PA) system, safety, recording and a fault finder.
Signal chain
From the stage to the back row, one signal chain
Every sound system, from a pub gig to a conference hall, follows the same chain. A source makes sound, a microphone or a DI (direct injection) box turns it into a signal, the mixer shapes and routes it, and amplifiers drive the loudspeakers. When something goes wrong, I follow this chain in order.
Swipe sideways to see the whole diagram.
Live view: the main mix runs to the speakers, while an auxiliary mix feeds a wedge on stage and a separate mix feeds the recording and the stream.
Level changes on the way
The signal starts as sound, becomes a tiny mic level voltage, is raised to line level, turns into numbers in a digital mixer, is amplified to speaker level and becomes sound again.
Two jobs for one mixer
Front of house mixes for the audience. Monitors mix for the performers. On smaller shows one digital mixer does both, using auxiliary mixes for the stage.
Inside the channel strip
Swipe sideways to see the whole diagram.
Gain. I set the preamp while the performer plays or speaks at full level, listening on PFL (pre fade listen). The aim is a strong signal with headroom left, well clear of the red. Every later stage depends on getting this right.
High pass filter (HPF). Removes low frequencies the source does not need, such as stage rumble and handling noise. I start around 80 to 100 Hz on vocals (typical) and lower on bass and kick. It also cleans up the low end of the whole mix.
Gate. Mutes the channel when the signal falls below a threshold, so tom mics stop picking up the rest of the kit. I set the threshold just under the softest real hit and leave a little hold and release so notes are not chopped.
Compressor. Turns down peaks above the threshold by a ratio. A 3:1 or 4:1 ratio with a medium attack (typical starting points) evens out a vocal. Gain added afterwards brings the level back up.
Equalisation (EQ). I cut before I boost: muddy build up around 200 to 400 Hz is the usual culprit, and a small lift around 2 to 4 kHz helps words cut through. Broad, gentle moves sound better than narrow, deep ones.
Fader, pan and sends. Monitor auxes are sent pre fade, so my front of house moves do not change what the band hears. Effects sends are post fade, so reverb follows the fader. The fader and pan place the channel in the main mix.
Cables
Balanced lines and the connectors that carry them
Most live sound faults are cable faults, so I start with the basics: which signal runs on which connector, why balanced lines reject noise, and where phantom power goes. Copper cabling in general, including Cat5e and coax, is on the copper cabling page.
Swipe sideways to see the whole diagram.
Balanced. The signal travels twice: normal on pin 2 (hot) and inverted on pin 3 (cold). Noise picked up on the way lands on both equally. The receiver subtracts one from the other, so the signal doubles and the noise cancels. That is why mic cables can run 50 m or more.
Unbalanced. One conductor and a shield, as on a guitar lead or a phono cable. Noise added on the way stays in the signal, so I keep these runs short and put a DI box at the stage end to turn them into balanced signals.
Phantom power. The mixer sends +48 V down pins 2 and 3 to power condenser mics and active DI boxes. Dynamic mics ignore it, but I switch it off before plugging anything in, and keep it away from older ribbon mics and unbalanced gear.
Speaker level. An amplifier's output is tens of volts and several amps, so passive speakers use Speakon connectors and heavy cable. Powered speakers take line level on XLR instead, and the amplifier is inside the box.
The connectors I meet on a job
| Connector | Where I use it | Signal | Watch out for |
|---|---|---|---|
| XLR, 3 pin | Microphones, DI outputs, line feeds between mixers and powered speakers | Balanced mic or line level; also AES3 (the Audio Engineering Society digital audio standard) on 110 ohm cable | Phantom power on every input; pin 2 is hot by modern convention (AES14) |
| TRS (tip, ring, sleeve), 6.35 mm (1/4 inch) | Balanced line outputs, headphones, insert points | Balanced mono line, stereo headphones, or send and return on inserts | A TRS plug and a TS (tip, sleeve) plug look alike; the wrong one unbalances the line |
| TS, 6.35 mm (1/4 inch) | Guitars, basses and keyboards | Unbalanced instrument or line level | Keep runs short; use a DI box at the stage |
| RCA (phono) and 3.5 mm jack | Laptops, phones and consumer players | Unbalanced consumer line level, about minus 10 dBV (decibels relative to 1 volt) | Use a stereo DI box for playback, never a long unbalanced run |
| Speakon (NL2 and NL4) | Passive loudspeakers and amplifiers | Speaker level, tens of volts | Locks in place; never use it for line level, never use a jack for speakers |
| RJ45 in an etherCON shell | Stagebox links, Dante and network control | Digital audio and control over shielded Cat5e | Shielded cable for stagebox links; protect the latch on stage |
| 5 pin and 3 pin XLR for DMX (Digital Multiplex) | Lighting control | Lighting data, not audio | DMX cable looks like mic cable but is 110 ohm data cable; keep them apart |
DI boxes
A DI box turns an unbalanced, high impedance source such as a bass guitar or a laptop into a balanced mic level signal that can travel down a long cable. Active DI boxes need phantom power; passive ones do not. Most have a ground lift switch for hum.
Cable length and tone
Long unbalanced instrument cables lose treble, because the cable's capacitance acts as a filter. Balanced mic cables can run tens of metres, and digital stagebox links take the long runs.
Testing and labelling
I test every cable with a cable tester before a job, label both ends of long runs, and coil them over and under so they lie flat and do not twist. Faulty cables get a tag and go in a separate bag.
Levels
Mic, line and speaker level, and the decibels in between
A decibel is a ratio, so it always needs a reference. Knowing which reference a meter uses, and which level each connector expects, explains most distortion and most hiss. Gain staging is how I keep every stage in its sweet spot.
| Level | Typical voltage | In decibels | Where it appears |
|---|---|---|---|
| Mic level | About 1 to 10 millivolts | About minus 60 to minus 40 dBu | Microphones and passive DI boxes |
| Instrument level | Tens of millivolts to about 1 V | Varies widely, high impedance | Guitar and bass pickups |
| Consumer line level | 0.316 V | Minus 10 dBV (about minus 7.8 dBu) | Laptops, phones, hi fi equipment |
| Professional line level | 1.228 V | +4 dBu | Mixer outputs, stagebox outputs, powered speakers |
| Speaker level | Tens of volts | Set by amplifier power and load | Amplifier outputs to passive speakers |
Typical values. Real levels depend on the source, the mic and how loud the performer is.
The decibels I use
- dBu
- Voltage against 0.775 V. Professional audio gear is specified in dBu.
- dBV
- Voltage against 1 V. Consumer gear is often specified in dBV.
- dBFS (decibels relative to full scale)
- Digital level against full scale. 0 dBFS is the top; anything above it clips.
- dB SPL (sound pressure level)
- Sound pressure against 20 micropascals, the threshold of hearing.
- dB(A)
- Sound level weighted like human hearing at moderate levels; used for noise limits.
- dB(C)
- A flatter weighting that keeps the bass; used for peak limits and loud music.
Swipe sideways to see the whole diagram.
A microphone gives a tiny voltage, often around minus 50 dBu for a voice (typical). It is far too small to mix, and it changes with how loud the performer is.
I set the preamp while the performer gives their loudest, listening on PFL. The meter should sit well up the scale with headroom left above. This is the most important control on the desk.
EQ and dynamics change the level a little. If a compressor or an EQ boost pushes a channel towards clipping, I take it back with the processing, not by changing the preamp.
With gain set well, faders sit around their 0 dB mark, where they have the finest control. A fader at the bottom with a channel still too loud means the gain is too high.
The mix bus adds every channel together, so its meter climbs as channels are added. I keep it out of the red and use the master fader as the last trim.
Amplifiers and powered speakers are set so the full mixer output gives the full safe level of the system, and no more. Then the mixer is the one place where level is controlled.
Try it: convert between dBu, dBV and volts
Type a level and choose its unit. The tool shows the same level in the other units, and where it sits between mic level and speaker level.
- dBu
- +4.0 dBu
- dBV
- +1.8 dBV
- Volts
- 1.228 V RMS
Professional line level. This is the level a mixer's outputs and a stagebox's line outputs work at.
Stageboxes
Digital stageboxes: one cable instead of forty
An analogue multicore carries one pair of wires per channel and weighs as much as a person. A digital stagebox puts the preamps on stage and sends every input to the mixer, and every output back, over one Ethernet style cable. On Allen & Heath mixers that link is dSnake, or SLink on newer models.
Swipe sideways to see the whole diagram.
Live view: inputs travel from the stagebox to the mixer, and monitor mixes travel back the other way on the same cable.
Two generations of Qu
| First generation Qu | Qu range launched in 2025 | |
|---|---|---|
| Models | Qu-16, Qu-24, Qu-32, Qu-Pac and Qu-SB | Qu-5, Qu-6 and Qu-7, each also as a D version with Dante |
| Local mic inputs | 16, 24 or 32, depending on the model | 16, 24 or 32 (Qu-5, Qu-6, Qu-7) |
| Sample rate | 48 kHz | 96 kHz processing and converters |
| Remote audio port | dSnake | SLink |
| Stageboxes | AR2412, AR84 and AB168 over dSnake | DX expanders such as the DX168 at 96 kHz |
| Recording | Qu-Drive: stereo or multitrack to a USB drive (up to 18 tracks on models such as the Qu-SB) | Multitrack to an SD card (32 tracks at 48 kHz, 16 at 96 kHz), stereo to USB A |
| Computer audio | USB B, 32 in and 32 out | USB C, 32 in and 32 out |
| Dante | Not built in | Built into the D models, 16 in and 16 out |
From Allen & Heath's product information and its SLink support article, checked in October 2026. Details vary by model; check the manual for the one in front of you.
What the SLink port can talk to
| SLink mode | Sample rate | Connects | Runs over |
|---|---|---|---|
| dSnake, with ME | 48 kHz | AR2412, AR84 and AB168 racks, and ME personal mixers | Fast Ethernet signalling on shielded Cat5e |
| DX | 96 kHz | DX168, DX164-W, DX012 and other DX expanders | Fast Ethernet signalling on shielded Cat5e |
| gigaACE | 96 kHz | Other Allen & Heath systems, and the GX4816 expander | Gigabit signalling on shielded Cat5e |
SLink modes as Allen & Heath documents them for its SLink systems. The port picks the mode from whatever is plugged in; I check the manual for which modes a given mixer supports.
My rules for stage links
- Shielded cable only. Allen & Heath asks for shielded (STP) Cat5e or better; unshielded cable is not supported on stage links.
- 100 m per run. Each cable can be up to 100 m. Two expanders can be cascaded for up to 200 m from the mixer.
- No joins in the middle. Passive couplers and tired connectors shorten the usable length. I use one continuous cable on a drum, with etherCON ends.
- Point to point, not a network. The link uses Ethernet cable and signalling, but it runs straight from mixer to stagebox. It does not go through a network switch.
- A spare in the van. One cable now carries every input, so a spare cable and a short analogue backup for the vocal mics are part of my kit.
Mixer families
Where the Qu sits among Allen & Heath mixers
Allen & Heath builds digital mixers from small app controlled units to large touring systems, and they share much of the same thinking. I mix on the Qu series; the ideas below carry across the range.
| Family | Scale | Processing | Where it fits |
|---|---|---|---|
| CQ | Compact, with app control and automatic gain and feedback assistants | Digital | Small venues, quick setups, rehearsal rooms |
| Qu | 16 to 32 local mic inputs, expandable with stageboxes | 48 kHz on the first generation, 96 kHz from 2025 | Bands, churches, schools and conferences |
| SQ | 48 input channels | 96 kHz | Larger bands, theatre and corporate events |
| Avantis | 64 input channels and 42 buses | 96 kHz | Touring and larger installations |
| dLive | Up to 128 input channels, with separate surface and MixRack | 96 kHz | Large touring, broadcast and big conference rooms |
Allen & Heath product information, October 2026. Channel counts are processing channels, not local sockets.
Scenes and safes
A scene stores the whole desk: gains, EQ, routing and mixes. For a conference I build one scene per session; safes stop a recall from changing what must stay put, such as the stream feed.
Layers and custom banks
Faders come in layers. I put the channels I need most, such as presenters' mics and the playback, on one custom layer, so nothing important is a page away.
Mixing from an app
Allen & Heath has tablet and phone apps (Qu-Pad and Qu-You on the first generation Qu). I walk the room with a tablet during soundcheck, and performers can set their own monitor mix.
Libraries and show files
Channel libraries hold my starting points for vocals, kick and lectern mics. The show file goes on a USB stick before and after every job, so a failed desk does not lose the show.
Mixing
Gain, EQ and dynamics, and keeping feedback away
Mixing live is mostly about clarity and control: each source in its own space, words that can be understood, and no feedback. These are the moves I start from, before the room and the performers tell me what to change.
Swipe sideways to see the whole diagram.
Mics in their show positions, wedges where the performers will stand, and the PA at show level. Ringing out an empty stage with the mics somewhere else proves nothing.
With one mic live, I raise the monitor mix slowly until a single tone starts to ring. That tone is the frequency where the loop between wedge and mic has the most gain.
I find it by ear, with a real time analyser (RTA) on the desk or a phone, or with the feedback assistant on newer Allen & Heath desks. Most rings sit in the mids and upper mids.
A narrow cut of 3 to 6 dB (typical) at that frequency on the monitor mix's EQ removes the ring with little change to the sound. Then I bring the gain up again.
The next ring appears a little higher in level, at a new frequency. I cut that one too, and the next, each time gaining a few more decibels of gain before feedback.
After two or three cuts, more notches start to thin the sound. Then the better fixes are mic technique, wedge position, a directional mic or in ear monitors.
EQ, by range
| Range | What lives there | My usual move |
|---|---|---|
| Below 80 Hz | Rumble, handling noise, kick and bass weight | High pass on everything that does not need it |
| 200 to 400 Hz | Warmth, then mud and boxiness when it builds up | Gentle cuts on vocals, guitars and toms |
| 500 Hz to 1 kHz | Honk and nasal tone | Small cuts if a source sounds like a telephone |
| 2 to 4 kHz | Presence and the consonants that make words clear | Careful small boosts for speech; cuts for harsh guitars |
| 5 to 8 kHz | Sibilance on S sounds, cymbal bite | A de esser or a narrow cut on bright vocals |
| Above 10 kHz | Air and sparkle | A shelf lift on dull sources, a cut on noisy ones |
Typical ranges. Every voice, instrument and room is different, so I listen first and use the analyser to confirm.
Dynamics starting points
| Source | Tool | Starting point (typical) |
|---|---|---|
| Lead vocal | Compressor | Ratio 3:1 to 4:1, medium attack, release about 100 ms, 3 to 6 dB of gain reduction on loud phrases |
| Speech at a lectern | Compressor and high pass filter | Ratio about 3:1, high pass around 100 Hz, gentle so pauses stay natural |
| Toms | Gate | Threshold just under the softest hit, fast attack, release set by the drum's ring |
| Kick drum | Gate and EQ | Gate to stop bleed, a cut around 300 to 400 Hz, a lift for the beater click |
| Bass guitar | Compressor | Ratio about 4:1, slower attack so the note's start comes through |
Habits that make a better mix
- Get it right at the source. Mic choice and position change the sound more than any EQ. I move the mic before I reach for a knob.
- Cut before boosting. Removing what is wrong keeps headroom and sounds more natural than adding what seems missing.
- Fewer mics, less feedback. Every open mic adds noise and feedback risk. I mute what is not in use and use scenes for changes.
- Effects on sends. Reverb and delay sit on post fade sends and return on their own channels, so one reverb serves many voices.
- Listen from the audience. The mix position can lie. I walk the room during soundcheck with a tablet.
Monitors
Monitor mixes the performers can trust
Performers play and speak better when they can hear themselves. Monitor mixes are separate mixes for the stage, sent from the same desk, and they are where most feedback starts.
Swipe sideways to see the whole diagram.
Wedges. A floor monitor fed by a mono auxiliary mix, sent pre fade so my front of house moves do not change it. Simple and loud, but every wedge is a feedback risk and adds sound to the stage.
In ear monitors (IEM). A stereo auxiliary mix (which uses two mixes on most desks) feeds a radio transmitter and a belt pack. Quieter stages and less feedback, but performers need good earphones and a limiter to protect their hearing.
Personal mixers. The desk sends many channels to a mixer at each performer, who sets their own balance. On Allen & Heath systems the ME range does this over the stage link. It frees the engineer from a stream of requests.
Talkback. A mic at front of house that I can route to the monitor mixes only, never the PA. I use it at soundcheck and between songs, and mute it before doors open.
| Wedges | In ear monitors | |
|---|---|---|
| Feedback risk | Higher, every wedge points at a mic | Very low |
| Stage volume | Adds to the sound in the room | Quiet stage, cleaner front of house mix |
| Mixes used | One mono mix each | Usually a stereo mix, two buses |
| Hearing | Loud stages tire ears | Lower levels, but a limiter is essential |
| Kit | Powered speakers and cable | Transmitters, belt packs, earphones, radio licences |
| Best for | Speech, simple gigs, short changeovers | Bands that play often, and in ear clicks and tracks |
Pre fade, always
Monitor sends are pre fade, so turning a channel down in the room does not take it out of a performer's mix. Effects sends are the exception: post fade, so reverb follows the voice.
One mix per need
I group performers who want the same mix. The lead vocal, the drummer and the keys usually need their own. A Qu has a set number of mixes, so I plan them in the input list stage.
Ask, then listen
At soundcheck I ask each performer what they need, set it, and then stand in their spot. "More me" is the most common request; it is often more of everything else that is missing.
Music rigs
Stage plots, input lists and a patch that makes sense
A live music show is won or lost in the paperwork. With a clear stage plot and input list, setup is a matter of following the numbers; without them, it is guesswork under time pressure.
Stage plot
A drawing of the stage from above: where each performer stands, their amps, wedges, power points and DI boxes. I ask the act for one weeks ahead, and draw it myself if they have none.
Input list
Every source, numbered in the order it appears on the desk: drums first, then bass, guitars, keys, vocals and playback. The same numbers go on the stagebox, the desk and the patch sheet.
Patch logic
Inputs on the stagebox follow the list, so input 14 on the desk is socket 14 on stage. Line inputs from radio mic receivers and playback go to the sockets nearest their kit.
Splits
When a band brings its own monitor desk, the inputs are split: through a transformer split for analogue, or by sharing the stagebox digitally when both desks support it. Only one desk sends phantom power.
Line check
Before the band arrives I speak into or tap every mic and check every DI with a phone or test tone, in list order, watching the meter. That finds wrong patches and dead cables in minutes.
Changeovers
For several acts I build a scene per act from their input list, label the faders, and keep the drum kit and shared mics on the same channels across acts.
Try it: build an input list and patch
Choose what is on stage. The tool numbers the inputs in the order I lay them out on the desk, suggests a mic or DI for each, flags phantom power, adds two spares and says how many sockets the show needs.
20 inputs: 4 need phantom power, 4 more if their DI boxes are active. More than 16 local sockets: a 24 input frame, or a 16 input frame with a stagebox, takes the rest.
| In | Source | Mic or DI | Phantom | Stand |
|---|---|---|---|---|
| 1 | Kick | Large dynamic kick mic | No | Short boom |
| 2 | Snare | Dynamic instrument mic | No | Short boom or clip |
| 3 | Hi hat | Small diaphragm condenser | Yes | Short boom |
| 4 | Tom 1 | Dynamic or clip on mic | No | Clip |
| 5 | Tom 2 | Dynamic or clip on mic | No | Clip |
| 6 | Floor tom | Dynamic or clip on mic | No | Clip |
| 7 | Overhead left | Small diaphragm condenser | Yes | Tall boom |
| 8 | Overhead right | Small diaphragm condenser | Yes | Tall boom |
| 9 | Bass | Active DI box | Yes | None |
| 10 | Guitar 1 | Dynamic mic on the amp | No | Short boom |
| 11 | Guitar 2 | Dynamic mic on the amp | No | Short boom |
| 12 | Keys left | DI box | If active | None |
| 13 | Keys right | DI box | If active | None |
| 14 | Vocal 1 | Dynamic handheld vocal mic | No | Tall boom |
| 15 | Vocal 2 | Dynamic handheld vocal mic | No | Tall boom |
| 16 | Vocal 3 | Dynamic handheld vocal mic | No | Tall boom |
| 17 | Playback left | Stereo DI box from laptop | If active | None |
| 18 | Playback right | Stereo DI box from laptop | If active | None |
| 19 | Spare 1 | Dynamic vocal mic | No | Tall boom |
| 20 | Spare 2 | Dynamic vocal mic | No | Tall boom |
The list without scripts shows a five piece band with playback. Mic types are generic suggestions, not brands.
Conference rigs
Speech every seat can understand, and a feed for everyone else
Conference sound is less about loudness and more about clarity, reliability and feeds: the room, the stream, the recording and the remote presenters all need a different mix of the same voices.
Swipe sideways to see the whole diagram.
Live view: speech goes to the PA and the delay speaker, a separate mix feeds the stream, and the remote presenter hears everything except their own voice.
The lectern
A gooseneck condenser mic about a hand's width from the mouth, angled at the chin. Presenters move, so I prefer a headset or lavalier radio mic for anyone who walks.
Speech first
Speech must be clear more than loud. I set a high pass filter, gentle compression and a little presence, and keep music beds under the voice.
Feeds for everyone else
The stream, the recording and any overflow room get their own mix from a matrix, so I can change the room mix without changing what remote viewers hear. More on the live streaming page.
Mix minus
A remote presenter on a video call hears the whole room except their own voice. Without that, they hear themselves late and stop talking.
Delay speakers
In a long room, speakers halfway back are delayed so their sound arrives just after the sound from the front, and the voice still seems to come from the stage. The calculator below gives the delay.
Hearing loops
Venues often need an induction loop for hearing aid users, fed from its own clean speech mix. The loop is covered on the audio visual (AV) over IP and meeting rooms page.
Radio mic frequencies in the UK
| Band | Licence | What I use it for |
|---|---|---|
| 863 to 865 MHz (channel 70) | Licence exempt | A few radio mics, typically up to four at once, at low power |
| 173.7 to 175.1 MHz (VHF) | Licence exempt | Budget systems; a small number of channels |
| About 606 to 614 MHz (channel 38) | Shared UK licence from Ofcom's Programme Making and Special Events (PMSE) team | Most professional systems; around ten channels with a good plan |
| Other UHF (ultra high frequency) TV channels | Coordinated licence for the site or event | Large events that need many channels, where local TV leaves gaps |
UK frequencies as summarised by Canford from Ofcom's rules, checked in October 2026. Frequencies change as spectrum is reallocated, so I check before buying or hiring radio mics.
The input list, cable schedule and run sheet I write before a conference are on my AV event workflows page.
Loudspeakers
Coverage, level and timing: the physics of the PA
A PA system has three jobs: reach every seat at about the same level, stay clear of the microphones, and keep sound arriving from the stage. Two simple rules, the inverse square law and the speed of sound (about 343 m per second at 20 °C), explain most of it.
Swipe sideways to see the whole diagram.
A loudspeaker's maximum level is quoted at 1 m. Here it is 120 dB SPL, an illustrative figure for a mid sized powered speaker.
At 2 m the same sound covers four times the area, so the level falls by 6 dB to 114 dB SPL. This is the inverse square law.
Each doubling of distance costs another 6 dB: 108 dB SPL at 4 m.
At 8 m the level is 102 dB SPL. A room 32 m deep loses 30 dB from the first metre to the back wall, which is why long rooms need delay speakers.
A tall line array behaves more like a line than a point up close, losing about 3 dB per doubling in its near field (typical). That keeps the front and back of a big room closer in level.
| System | How it spreads | Where I use it |
|---|---|---|
| Point source (a single box) | Like a sphere, minus 6 dB per doubling | Small rooms, bars, side fills, delays |
| Column speaker | Wide horizontally, narrow vertically | Conference rooms, churches, reverberant halls |
| Line array | Like a cylinder up close, about minus 3 dB per doubling | Large rooms and outdoor stages |
| Subwoofers | Almost everywhere at low frequencies | Music; often placed or steered to keep bass off the stage |
Try it: level at a seat, and the delay for a delay speaker
Enter a speaker's sensitivity and the amplifier power, the distance to a listener, and how far behind the main speakers a delay speaker sits. The results are indicative: real speakers compress near their limits, and rooms add reflected sound.
- Maximum at 1 m
- 125.0 dB SPL
- At the listener
- 100.9 dB SPL
- Sound travel time
- 58.3 ms
- Delay to set
- 68 ms
At 16 m this system gives about 101 dB SPL at full power: plenty for speech, and enough for moderate music with headroom to spare.
How I place speakers
- Aim at ears, not walls. Point the speakers at the audience and away from reflective walls and the stage mics. Coverage matters more than size.
- Height helps. Speakers above head height reach the back rows over the front rows and are kinder to the front seats.
- Delays for depth. When the back of the room is more than about 10 dB quieter than the front, I add delay speakers rather than turning up the front.
- Headroom, not volume. An amplifier and speaker that run at half their limit sound cleaner than a smaller system at full stretch.
Power and safety
Clean power, safe earths and ears that last
Sound systems meet mains power, heavy speakers, cables across floors and very high sound levels. The safe way is also the quiet way: most hum and buzz comes from poor power planning.
One clean power source
Audio kit shares one power distribution, away from lighting dimmers and catering. Different earths at the desk and the stage are the usual cause of hum.
Never lift a mains earth
Hum is fixed with the ground lift switch on a DI box or an isolating transformer in the signal path. Removing the earth from a mains plug is dangerous and is never the answer.
Inspected and tested kit
Portable equipment is visually inspected before every job and tested on a schedule, called portable appliance testing. Damaged cables are tagged and taken out of use.
Temporary power done properly
Event power from a generator or a distribution board follows BS 7909, the UK code of practice for temporary electrical systems in entertainment, with residual current device (RCD) protection where it is needed.
Cables and rigging
Cables cross walkways under ramps or mats, taped flat where people walk. Speakers on stands are secured and within the stand's rated load, and anything flown is rigged by a competent rigger.
Hearing
I wear hearing protection when levels are high, keep in ear mixes behind limiters, and log levels at the mix position against the venue's limit.
Noise limits for people at work
| Value under the Control of Noise at Work Regulations 2005 | Daily or weekly exposure | Peak sound pressure |
|---|---|---|
| Lower exposure action value | 80 dB(A) | 135 dB(C) |
| Upper exposure action value | 85 dB(A) | 137 dB(C) |
| Exposure limit value (allowing for hearing protection) | 87 dB(A) | 140 dB(C) |
These protect crew and performers at work. Audience levels are usually set by the venue's licence conditions. Guidance for music and entertainment from the Health and Safety Executive (HSE) is called Sound Advice.
Recording
Recording the show, and a soundcheck without the band
A multitrack recording gives the act a souvenir and the stream a backup. It also gives me the best rehearsal tool there is: the virtual soundcheck, where last night's show plays back through the desk as if the band were on stage.
Swipe sideways to see the whole diagram.
Each channel is recorded straight after its preamp, before EQ and dynamics, to a USB drive, an SD card or a laptop over USB.
Next time, before the band arrives, each channel's source is switched from its live input to the matching recorded track.
I adjust EQ, dynamics, effects and the PA with the band's real performance playing through the system, with no one waiting on stage.
Every channel goes back to its live source before doors. Saving that as part of the scene stops a recorded track being left in the mix.
| Option | What it records | When I use it |
|---|---|---|
| Stereo to a USB drive | The main mix, or any mix I choose | A quick reference copy for the act or the organiser |
| Multitrack to a USB drive (Qu-Drive, first generation Qu) | Up to 18 tracks on models such as the Qu-SB | Virtual soundcheck and simple mixes afterwards |
| Multitrack to an SD card (2025 Qu) | 32 tracks at 48 kHz, or 16 at 96 kHz | Full shows, virtual soundcheck |
| USB audio to a computer | 32 channels each way into a recording program (DAW) | Full multitrack with a laptop, and playback for virtual soundcheck |
| Dante (D models) | 16 channels each way on the network | Recording or sharing audio with other Dante kit |
From Allen & Heath product information. Track counts vary by model.
Editing and mixing the recordings afterwards, including loudness targets for online video, is on the editing and post production page.
A multitrack recording gets a second mix after the show, as on my post production page, and a separate matrix mix feeds the stream on my live streaming page.
Troubleshooting
Finding a fault along the chain
Live sound faults have to be found fast, often with an audience waiting. I follow the signal chain from the source forwards, and halve the problem with swaps. Choose a symptom to see where it usually sits and what I check first.
Swipe sideways to see the whole diagram.
No signal on one channel. Wrong socket, a dead cable, a muted channel or a condenser mic without phantom power. I follow the input list from the stage, swap the cable for a known good one, and check the channel's source, mute and phantom settings.
Hum or buzz. Usually an earth loop: two pieces of kit earthed at different points, such as a laptop on its charger feeding a DI. I press the DI's ground lift, run the laptop on battery, and keep audio kit on one power source.
A radio mic drops out. Flat batteries, antennas hidden behind metal, a body blocking the signal, or interference on that frequency. I start with fresh batteries, move the antennas into line of sight, and run a new frequency scan before doors.
All stage inputs are lost at once. The stage link is down: a damaged cable, a coupler in the run, unshielded cable or a run over 100 m. I reseat both ends, then swap to the spare cable, and keep a short analogue backup for the main vocal mics.
Feedback. Too much gain in the loop between a speaker and a mic. I move the wedge or the mic first, mute unused mics, then cut the ringing frequency on that mix. If it persists, the mix is asking for more than the room can give.
A distorted vocal. The preamp gain is too high for how loud the singer is tonight, or processing gain pushes the channel into clipping. I lower the preamp, not the fader, and check the compressor's added gain.
One side of the PA is silent. The amplifier is off or in protection, a Speakon is not locked, or the output patch is wrong. I check the amplifier's lights, reseat the speaker cable, and then the desk's output routing.
Voices sound doubled at the back. The delay speaker's delay is missing or wrong, so its sound arrives at a different time from the main PA. I measure the distance again, set the delay with the calculator above, and listen from the seats.
The tools I reach for
| Question | Tool | What it tells me |
|---|---|---|
| Is the signal reaching the desk? | PFL and the channel meter | Whether the problem is before or after the mixer |
| Is the cable good? | A cable tester | Open, shorted or crossed conductors on XLR, jack, Speakon and Cat5e |
| Is phantom power there? | A phantom power tester or a multimeter | About 48 V on pins 2 and 3, against pin 1 |
| Which frequency is ringing? | The desk's RTA, or an analyser app | The exact frequency to cut |
| How loud is it? | A sound level meter, ideally Class 1 or 2 | Levels against the venue's limit and for hearing safety |
| Which radio channels are clear? | The receiver's frequency scan | Free frequencies before doors open |
I change one thing at a time, and I keep a known good mic and cable at the desk for swapping. My general method is on the troubleshooting page.
Engagement
How I plan a live sound job, from first call to pack down
Whether it is a band night or a conference, I plan sound the same way. I agree who must hear what, find out what the room and the power can give, and draw the signal flow. Then I prove it at line check, run the show and pack down with nothing lost.
Swipe sideways to see the whole diagram.
I start with the brief: what kind of event it is, who performs or speaks, how many people will listen, and whether the show is streamed or recorded. I also ask what sound kit the venue already has, because it changes what I bring and how long setup takes.
I turn the brief into requirements: every input and monitor mix, the level the room needs and the limit the venue allows, coverage for every seat, and the feeds for the stream, the recording and any remote presenters. Each one has a number where a number can be agreed.
Before designing anything I visit or ask for photos and plans: the room's size and surfaces, where the power is and how it is protected, the route for load in, and the venue's noise limit and curfew. A hard, empty hall and a carpeted room need different systems.
The design is a set of drawings and lists. There is the input list and patch, the stage plot with monitors and power, and a signal flow diagram from every source to every output. A speaker plan sets heights, angles and delay times, and the stream team gets the feed they asked for.
Setup follows a fixed order: power and earths, speakers, cables, patching, then a line check of every input in list order. Then I ring out the monitors and the PA, soundcheck with the performers, and test the recording and stream feeds before doors.
During the show I mix, keep an eye on levels against the venue's limit, and keep spare mics and batteries within reach. The show file is saved before doors, and anything that goes wrong is noted at the time so it can be fixed for next time.
Pack down runs in reverse order, with every item counted against the kit list. Recordings are copied to two places before anything is deleted, faulty cables are tagged and repaired, and the notes from the show update my input list templates and checklists.
Functional requirements
What it must do.
- Give every performer and presenter a clear, feedback free sound in the room.
- Give each performer the monitor mix they need, on wedges, in ear monitors or personal mixers.
- Feed the stream, the recording and any overflow room with their own mix.
- Give remote presenters a mix minus, so they hear the room but not themselves.
- Record every channel for a backup and for a virtual soundcheck next time.
- Keep spare mics and a backup path for the main vocal or lectern mic.
Non functional requirements
How well it must do it, with a number wherever one can be agreed.
- Speech level
- Around 70 to 80 dB(A) at each seat, and 15 dB or more above the room's background noise (typical).
- Music level
- Set by the venue's licence and the Noise at Work Regulations; bands often run around 95 to 100 dB(A) at the mix position (typical).
- Coverage
- Level within about 6 dB from the front row to the back row (typical).
- Intelligibility
- A speech transmission index (STI) of 0.5 or more for conference speech (typical target).
- Inputs
- Every input on the list, plus at least two spare channels and two spare mics.
- Monitors
- Every performer hears what they asked for at line check, without feedback.
- Changeovers
- Between acts within the running order, typically 10 to 15 minutes, using one scene per act (typical).
- Resilience
- A spare vocal or lectern mic live on stage, and a recording running from the start.
Numbers are typical starting points. The venue's licence conditions and the Control of Noise at Work Regulations 2005 come first.
Who I talk to
- The organiser, who knows the running order, the audience and the budget
- The performers or presenters, who know their inputs and monitor needs
- The venue technician, who knows the room, the house kit and the rules
- The venue electrician or manager, who knows where safe power is
- The stream or video team, who need clean feeds and timings
How I find out
- A site visit, or photos and a floor plan with measurements
- The act's technical rider, stage plot and input list
- A walk of the room with a sound level meter to hear background noise
- A frequency scan for radio mics at the venue
- A call with the stream team to agree feeds, delays and talkback
Discovery tools
| Tool or source | What it tells me |
|---|---|
| Sound level meter (Class 1 or 2) | Background noise, show levels and the venue's limit |
| Measurement mic with Room EQ Wizard or Smaart | How the room responds, and where reflections and long reverb are |
| Laser distance meter | Distances for speaker coverage and delay times |
| Socket tester and the venue's electrical certificate | That earths and RCDs are present and correct |
| Cable tester | That every cable in the kit is good before it leaves |
| Radio mic receiver scan | Which frequencies are clear at the venue |
| Technical rider and stage plot | Inputs, monitors, power and backline the act expects |
Live sound faults must be found while people wait, so I follow the signal chain from the source forwards and halve the problem with swaps.
My approach
- Scope. Which channel, which output, which listeners, and since when? What changed: a cable, a scene, a performer?
- Follow the chain. Source, cable, stage link, mixer, amplifier, speaker, in that order. PFL tells me if the signal reaches the desk.
- Halve the problem. Swap the cable, swap the channel, swap the mic. A known good spare finds the fault in one move.
- Check the obvious. Mutes, faders, phantom power, the channel's source setting and the amplifier's power.
- Change one thing. One change at a time, then the same test again, so I know what fixed it.
- Write it down. Note the fault and the fix for the debrief, and tag any faulty kit.
Faults I see most
- A condenser mic with phantom power switched off
- Hum from a laptop on its charger feeding a DI box
- Feedback from a wedge pointed at a vocal mic
- A radio mic with a flat battery or hidden antennas
- A channel left on a playback source after a virtual soundcheck
- A delay speaker without its delay set
Layer by layer
| Layer | What I check |
|---|---|
| 1 · Source | The performer, the mic position, the radio mic battery and frequency. |
| 2 · Cable and DI | XLR and jack cables, DI boxes, ground lift and phantom power. |
| 3 · Stage link | The stagebox, its shielded Cat5e link and the etherCON ends. |
| 4 · Mixer | Gain, routing, mutes, scenes, sends and the output patch. |
| 5 · Amplifier and speaker | Power, protection, speaker cables, delays and positions. |
My general method, with worked network faults, is on the troubleshooting page.
A show runs smoothly when everyone knows their part and nothing depends on memory.
Show file
Saved to the desk and a USB stick before doors, with one scene per act or session.
Comms
Talkback to the stage, and a way to reach the stage manager and the stream producer at once.
Spares
A spare vocal mic, spare radio mic batteries, a spare stage link cable and a DI box at the desk.
Levels
Levels checked at the mix position against the venue's limit through the show.
Changeovers
Changeover plans from the input lists, with labelled faders for each act.
Debrief
A short note after every show: what went wrong, what worked, what to change.
Kit and habits age between shows, so I keep both current.
- Cables. Tested after every job; faulty ones tagged, repaired or replaced.
- Firmware. Desk and stagebox firmware updated between jobs, never on the day, with the show file checked afterwards.
- Templates. Input list and scene templates updated from each debrief.
- Batteries. Rechargeable radio mic batteries cycled and replaced on a schedule.
- Hearing. Hearing protection in the kit, and in ear mixes behind limiters.
- Kit list. The kit list, spares and cases checked against what the last job needed.
Projects
Where I have used it
Sound engineering is my hobby. I mix in my own time on an Allen & Heath Qu series mixer, and Kramer's audio training backs the AV side. My network and voice work explains the parts that run over cables and IP.
Live sound on an Allen & Heath Qu
I mix live sound in my own time on an Allen & Heath Qu series digital mixer: gain staging, monitor mixes, effects, scenes and recording the show.
Audio Specialist and AV Fundamentals
Kramer's Audio Specialist course and AV Fundamentals: training in professional audio for AV installations.
AV over IP
AV over IP Installer and AV over IP Sales Associate: audio and video carried over networks, covered on the AV over IP and meeting rooms page.
Voice across three lab sites
Nine IP phones across three sites joined by an MPLS (Multiprotocol Label Switching) core, with Cisco Unified Communications Manager (CUCM), Unity Connection, Presence and Contact Center Express. It shows in practice how codecs, delay and packet loss change what people hear. See the topology.
Voice over IP (VoIP) across the group
Local and wide area networks, VoIP, virtualisation and security across the group, including the warehouse and cold chain systems at the food depots.
CCVP for voice
CCVP (Cisco Certified Voice Professional), later renamed CCNP Voice: voice over IP, call control and quality of service for audio on networks.
Backed by Kramer's Audio Specialist and AV Fundamentals courses, and CCVP (Cisco Certified Voice Professional) for voice. See all credentials
Explore
More areas I design and deliver
Every page follows the same idea: how it works, the choices that matter, and where I have used it. The foundations pages explain the basics the others build on.