A café fits out its room with a pair of bookshelf speakers behind the counter and a second pair at the far wall, all four fed from one stereo amplifier. Near the counter it is loud. At the far table the music is present but thin, and turning the volume up makes the counter unbearable long before the far end becomes pleasant. Nobody has wired anything incorrectly. The far pair is being fed through a voltage divider that nobody drew on the plan.
The divider you did not install
A speaker cable has resistance. So does a speaker. Put them in series — which is exactly what a cable and a speaker are — and the amplifier's voltage splits between them in proportion to those resistances. Everything that lands on the cable is heat in the copper rather than sound in the room.
The numbers that matter are small enough to look harmless. Copper of 1.5 mm² has a resistance of about 0.0125 ohms per metre; 2.5 mm² comes in near 0.0069; thin 0.75 mm² bell wire is around 0.023. Then remember that current has to come back: a run of n metres puts 2n metres of copper in the circuit.
So a 30-metre run in 1.5 mm² is 60 metres of copper — 0.75 ohms in series with the speaker. Against a nominal 8-ohm cabinet, the speaker keeps 8 of every 8.75 volts, which costs about 0.8 dB. Barely audible, and this is the case people cite when they say cable resistance is a myth.
Now change two things a real installation changes. Make the cable the thin stuff that was already in the wall, and make the load 4 ohms because the installer wired two cabinets in parallel on that leg. Forty metres of 0.75 mm² is 1.84 ohms, and against 4 ohms the speaker now receives 4 of every 5.84 volts. In power terms that is −3.3 dB: less than half of what the amplifier sent arrives as sound. The counter pair, five metres away on the same amplifier, loses 0.2 dB. The far end is three decibels down on the near end before anybody touches a control, and no amount of volume closes a gap that scales with the volume.
The rule of thumb, and the distance it implies
Installation practice has a rule for this: cable resistance should stay under 5 per cent of the loudspeaker's nominal impedance. For an 8-ohm speaker that is 0.4 ohms of cable, total, there and back.
Turn that into a distance and the advice that circulates as folklore becomes arithmetic. In 1.5 mm², the loop costs 0.025 ohms per metre of run, so 0.4 ohms is reached at 16 metres. In 2.5 mm² it is about 29 metres. Drop the load to 4 ohms and the budget halves: 8 metres and 14 metres respectively.
This is why commercial audio texts say low-impedance distribution suits a small number of speakers within roughly 20 metres of the amplifier. That figure is not a tradition. It is the 5 per cent rule, evaluated for the cable people actually pull.
Why parallel cabinets make it worse twice over
The obvious way to feed four speakers from a two-channel amplifier is to parallel two per channel. That halves the impedance the amplifier sees — two 8-ohm cabinets in parallel present 4 ohms — and everything above gets worse in two ways at once.
The first is the divider. A 4-ohm load makes the same cable twice as costly in decibels, because the cable's share of the total resistance has doubled.
The second is the amplifier itself. A stereo amplifier rated into 8 ohms is not obliged to be happy at 4, and is certainly not obliged at 2, which is where a third pair would take it. What follows is heat, then protection, then a shop that has learned to keep the volume at a level where nothing cuts out — which is the level at which the far tables cannot hear the music.
There is a third effect that is less often noticed and easy to hear. Damping factor — the amplifier's ability to control the cone's own momentum — is the ratio of the load impedance to the total source impedance, and the cable is part of that source. Put 0.75 ohms of copper in front of an 8-ohm speaker and the damping factor at the cabinet cannot exceed about 10, no matter what the amplifier's specification claims. Bass loses its edges. The far pair does not just play quieter; it plays woollier.
The impedance on the box is one number for a curve
Everything above used "8 ohms" as though a loudspeaker had a resistance. It does not. A cabinet's impedance varies across the band: it rises to a peak at the bass driver's resonance, often 30 to 40 ohms, falls to a minimum somewhere above that, and climbs again at the top as the voice coil's inductance takes over. The nominal figure on the back describes roughly the low part of that curve, and the true minimum is allowed to sit somewhat below it.
So the divider is frequency-dependent. Cable resistance is constant, which means it claims a large share where the impedance dips and a negligible share where the impedance peaks. A run costing 3 dB at the minimum costs a fraction of that at resonance.
What you hear is therefore not a volume control. It is a tilt. The far speaker loses most where its impedance is lowest — typically the upper bass and lower midrange, where most of a room's musical energy sits — and loses least at resonance, which is the one region where the cone is already least controlled. That is why a long thin run sounds simultaneously thin and boomy rather than simply quiet, and why turning the amplifier up makes the imbalance louder instead of correcting it.
What commercial installations do instead
The distributed-audio answer is to stop sending large currents down long wires. A 100-volt line system puts a step-up transformer at the amplifier and a step-down transformer at every speaker, so the wire between them carries a high voltage and a small current.
The arithmetic is worth doing once, because it is the whole argument. A speaker set to a 10-watt tap on a 100-volt line presents, by definition, an impedance of 100² ÷ 10 — 1 000 ohms. Put our 30-metre run of 1.5 mm² in series with that: 0.75 ohms against 1 000 ohms is a loss of roughly 0.006 dB. The same copper that cost 3.3 dB in the low-impedance system has become unmeasurable.
The current tells the same story from the other side. Ten watts into 8 ohms is 1.12 amps; ten watts into a 1 000-ohm tap is 0.1 amps. Cable loss goes as the square of current, and the current has dropped by a factor of eleven.
That is why a 100-volt line lets an installer hang a dozen ceiling speakers off one pair of thin wires running the length of a building, and set each one's level by choosing a tap — 5 watts near the counter, 10 in the dining room — with a screwdriver rather than a mixer.
What the high-voltage system costs
None of this is free, and the trade is specific rather than vague.
- Insertion loss. Every transformer takes its cut. Quality line transformers are specified at 0.5 to 1.0 dB at mid frequencies, which is real power turned into heat in the core and windings — around 16 per cent of the amplifier's output at 0.75 dB.
- Low-frequency limits. A transformer that reproduces deep bass at full power has to be large and expensive, so small taps roll off early. Background music and speech survive this comfortably; a system meant for music with weight does not.
- Voltage drop budgets. Practice allows about 10 per cent voltage drop on a background-music run, and holds emergency and life-safety systems under 5 per cent. Long runs and many taps still need calculating; the high voltage widens the margin, it does not abolish it.
- Cost per speaker. Each cabinet needs its transformer, which is why a 100-volt ceiling speaker costs more than a passive one of similar quality.
Choosing between them without guessing
Three cases cover nearly every room, and the distances above decide which one you are in.
One room, one pair, under about fifteen metres. Low impedance, 1.5 mm² or better, and the losses stay under a decibel. This is a living room, a small studio, a shop counter. Nothing here justifies transformers.
One long run, still one or two speakers. Stay low-impedance and buy copper. Going from 1.5 mm² to 2.5 mm² buys you nearly double the distance for a fraction of the cost of converting the system, and the calculation above tells you exactly where the limit sits.
Several speakers, spread out, different levels in different areas. This is what 100-volt line was built for, and the case where trying to save money with low impedance produces the café at the top of this article. The moment you want four or more speakers at different distances and different volumes, the transformer path is cheaper than the copper path and far more predictable.
The check that takes a minute
Disconnect the far speaker and short its two cable ends together at the speaker end. Walk back to the amplifier and measure the resistance across that pair with a multimeter. What you read is the entire round-trip resistance of the run — no calculation, no guessing at what is inside the wall.
Compare it with 5 per cent of the speaker's nominal impedance: 0.4 ohms for an 8-ohm cabinet, 0.2 for a 4-ohm one. If you read more than that, you have found where your volume went, and you now know whether the answer is thicker cable or a different system entirely.
One purchase this does not justify: exotic speaker cable. What matters here is the cross-sectional area of the copper and the length of the run, both of which are printed on ordinary installation cable sold by the metre. Paying a premium for cable that claims better sound while leaving four cabinets in parallel at the end of a forty-metre run is treating the invoice rather than the divider.
How this was put together
Five independent sources sit under the figures above: published resistance figures for copper conductors of 0.75, 1.5 and 2.5 mm² cross-section; the installation rule that cable resistance should stay below 5 per cent of nominal loudspeaker impedance, and the corresponding 0.4-ohm budget for an 8-ohm speaker; commercial-audio practice limiting low-impedance distribution to a small number of speakers within roughly 20 metres; manufacturer and application-note figures for line-transformer insertion loss of 0.5 to 1.0 dB, and the 84 per cent power efficiency that a 0.75 dB loss implies; and the voltage-drop budgets used in distributed systems — about 10 per cent for background music, under 5 per cent for life-safety.
The derived figures are ours: the 0.8 dB loss of a 30-metre 1.5 mm² run into 8 ohms and the 3.3 dB loss of a 40-metre 0.75 mm² run into 4 ohms; the three-decibel gap between near and far pairs on one amplifier; the distance limits of 16 and 29 metres that the 5 per cent rule implies for 1.5 and 2.5 mm² cable; the 1 000-ohm impedance of a 10-watt tap and the 0.006 dB the same copper costs it; and the eleven-to-one current ratio between the two systems.
