your low end is mono before anyone hears it

it’s 2am and the low end finally sits. at some point you widened it, because wide felt bigger, and it was. it filled the headphones.

then the track goes out into the world and something quietly takes the weight back out of it. the club. a phone speaker. the festival you were picturing while you made it.

nothing broke. the low end was mono before anyone heard it, and the part you widened is the part that got subtracted.

the frame

every large sound system feeds its subwoofers one shared signal. not as a taste, as a requirement: the trick used to aim bass at the audience and away from the neighbours only works when every subwoofer receives the same waveform. so anything that exists only in the difference between your left and right channels isn’t quiet down there. it’s absent.

i just spent six weeks building the film version of this, following one track from a bedroom to the biggest speakers on earth. this is the part i couldn’t fit into it properly.

why anyone aims bass in the first place

a subwoofer is close to omnidirectional. at 40 Hz it doesn’t have a front. it loads the whole space, which includes backwards, onto the stage, and outwards, over the fence, into a town that has opinions about it.

so the people who design these systems stop treating subs as individual boxes and start treating them as one instrument. several subs, arranged in a line or stacked front-to-back, each fed the same signal but offset slightly in time, sometimes with one polarity flipped. forwards, those offsets line up and the waves add. backwards, they arrive out of step and cancel. you get a loud field in front and a comparatively dead zone behind. cardioid and end-fire arrays, and d&b quote 15 to 20 dB of rear rejection for a three-box cardioid setup.

that’s interference used deliberately, and it has a condition attached. interference needs the waves to be related. d&b’s own manual for the arrangement is blunt about it: all the amplifiers “have to be set to the same input gain and fed with the same input signal”. hand those boxes two different signals and there’s no stable relationship between them, so there’s no dependable place where they add and none where they cancel. the steering collapses.

which is why the feed is mono. not because anyone dislikes stereo bass. because the aiming doesn’t function without one shared signal.

your head is too small for the job

sound travels about 343 metres per second in ordinary air, the figure the ISO propagation standards run on. divide by frequency and you get wavelength. at 40 Hz that’s 8.6 metres. at 80 Hz, 4.3 metres.

now measure your head. about 18 centimetres across, which at 40 Hz is a fiftieth of a wavelength. it casts no acoustic shadow at all down there, so there’s no level difference between your two ears to read. below roughly 1.5 kHz your hearing has already switched to comparing arrival time between the ears rather than loudness (Mills, J. Acoust. Soc. Am. 30, 237, 1958), and at 40 Hz that timing cue is tiny too.

i want to be careful here, because the confident version of this claim is everywhere and it isn’t supported. people will tell you flatly that bass has no direction, usually with a number attached, and that number is a loudspeaker crossover convention wearing a lab coat. a 2012 review of fourteen studies on low-frequency localization found six concluding you can, six concluding you can’t, and two undecided (Hill, Lewis and Hawksford, Reproduced Sound, 2012). recent anechoic work has people locating broadband bass down to 31.5 Hz.

so the honest statement is narrower: the system never hands you the cue.

what the system actually does with it

subwoofers run off a single auxiliary send: one signal, every box. the AES technical document on outdoor events describes the main PA as behaving “multiple-mono”, and notes the stereo effect is “lost on everyone other than a relatively narrow area of audience central to the main PA” (AESTD1007.1.20-05, 2020).

sit with that second part, because it’s bigger than the bass question. almost nobody standing in that field is hearing your stereo image at any frequency. the low end is just where it’s most total.

and if a system does run its subs in left and right, what happens off-centre still isn’t your stereo picture. two spaced sources carrying the same signal build a loud corridor straight down the centreline and cancellation either side, the power alley that same document names and tells you to avoid. the standard fix is to deliberately decorrelate the two sides, which is the detail that gives the game away: down there, the relationship between your channels is something engineers manage as an interference problem, not a picture anyone is trying to preserve.

either way it doesn’t arrive. mono-summed, the difference is discarded at the console. not mono-summed, the difference is overwritten by a position-dependent interference pattern that has nothing to do with what you authored.

what actually gets subtracted

think of your stereo mix as two other things: the part common to both channels, and the part that differs. mono summing keeps the common part and cancels the difference.

so it matters where you put the weight. if the low end is mostly common to both channels, summing leaves it standing. if you widened it, and especially if you widened it with anything that pushed energy into the difference, summing removes that. not softens. removes.

this is the uncomfortable bit: the wider it sounded in your headphones, the more of it is stored in the channel that doesn’t survive the trip. the sensation of size and the fragility are the same fact viewed from two sides.

and it isn’t only festivals. a club rig folds the same way. a phone speaker is one driver. most bluetooth speakers are one driver. a laptop’s two speakers are close enough together to be one for anything low. the majority of listening your track will ever get is closer to mono at the bottom than the thing you mixed on.

what to do about it

keep the bottom common to both channels. the handover to subs usually sits around 100 Hz, and the coherent-interference problem the AES document describes runs up to about 200, so somewhere in that band is where stereo stops being decoration and starts being risk. treat it as a judgment rather than a law: pick the line by where your track actually keeps its weight, not by a number you read.

then check the sum instead of assuming it. this is the whole move. play the track, fold it to mono, and listen to whether the weight stays.

the test is blunt and it’s reliable: if the low end gets smaller when you sum it, that size was never in the track. it was in the difference between two channels, and almost nobody is going to hear that difference.

do it early, while it’s still a mixing decision. once the arrangement is leaning on a bass sound that only exists in stereo, fixing it stops being an EQ move and starts being a rewrite.

the same shape outside the studio

the parts of anything that survive contact with other people are the parts held in common. the private part, the bit that only exists in the gap between two views, tends not to make the journey. the work wasn’t worthless. it was just never shared to begin with.

your mix has a mid and a side. so does most of what you believe.

the tools

this is the entire reason CHECK is free and always will be.

it splits the signal into 40 bands spaced by how your hearing actually resolves frequency, and scores each one for what survives a mono fold, rather than handing you a single correlation number that averages a dead 60 Hz together with a fine 3 kHz and reads reassuringly in the middle. and it weights the bottom deliberately: below 200 Hz counts three times a midrange band. that is a weighted measurement rather than a neutral one, and it doesn’t pretend otherwise: the weighting is aimed at the thing that actually costs you the mix.

it costs nothing because a mono problem you can’t see isn’t a fair fight, and charging for the ability to see it felt like the wrong business to be in.

the film

the piece above is one stage of a longer trip. i built the whole thing as a film: what happens to your track at tomorrowland, about twelve minutes, following one bedroom mix from a 2am session to the biggest speakers on earth and through every bit of physics waiting for it on the way.

it covers why a stack of speakers loses half as much level over distance as a single one, why the air itself is an equalizer nobody set the curve on, why a festival sounds louder at night, and why your ears lie to you about bass at low volume. it’s the first proper film i’ve made rather than a short, and it took six weeks.

there’s nothing to buy in it. i’d just like it watched.

the channel is new enough that a like or a subscribe actually moves it. i’m not going to pretend that’s a small ask on a twelve minute video, so if you only do one thing, send it to the one person you know who would find it useful. that’s worth more than either.

from the studio

tell me about the bass that vanished on someone else’s system, or the widener you undid a week later. i read every reply, and they turn into the next of these.

jonas