Home Gym Flooring for Apartments: Density, Thickness, Load

A floor covering in a flat is asked to do four separate jobs and is sold on a single number. Thickness answers two of them well, one partly, and the last one not at all. The last one is usually the reason for the purchase: what the household below ends up hearing. So this guide takes the jobs apart, then takes apart the four quantities the trade quotes when it discusses the fourth, because those quantities describe different objects and are taken by different methods, and once they appear together the reader draws a conclusion that none of them supports.

The four jobs a covering is doing

Surface protection. Keeping a metal foot, a plate edge or a set-down dumbbell off a finish. Mechanical, visible, and settled by the covering alone.

Stability. A bench or rack that rocks a few millimetres under load is a training problem before it is anything else. This one runs against cushioning: the softer the layer, the less steady the equipment standing on it.

Load spreading. Equipment concentrates its weight onto small contact patches. A covering can widen that patch, and the amount it widens it is a matter of stiffness rather than thickness.

Interrupting structure-borne impact. Contact with the floor sends vibration into the slab, which carries it and radiates it back out as sound in the rooms below. This is the job the marketing is written about, and it is the only one of the four that depends on a building you cannot inspect from inside your own flat.

The first three you can settle in your own room with a tape measure and a hand. How much of the fourth a covering does in your particular building is not generally knowable from a product page, and the rest of this guide is largely about why not.

Four numbers, and what each one is measured on

Four different measurements. No shared scale.

1. x body weight

   Ground-reaction force

   Measures the load a body drives into a floor, as a multiple of body weight, on a force plate.

   Treadmill running peaks near 2.37 to 2.62 times body weight at 9 to 13 km/h.

   Not a sound level. A force ratio says nothing about decibels anywhere.

2. L'_n,w in dB

   Building performance requirement

   Rates a separating floor as a piece of construction, using a standardised tapping machine, normalised to a reference room.

   Germany's DIN 4109 caps a dwelling-separating floor at 53 dB (1989 version) and 50 dB (2018 version).

   Says nothing about any individual exercise, and it rates the building rather than the occupant.

3. dB(A), L_eq / L_max

   Neighbour-noise legal limit

   A regulatory threshold for what a resident may transmit, measured as sound pressure in the affected dwelling by a prescribed procedure.

   South Korea sets a direct-impact limit of a one-minute 39 dB(A) by day and 34 dB(A) at night, with peaks capped at 57 and 52 dB(A).

   One country's rule, measured in the flat below. It cannot be derived from a building rating or from a covering's data sheet.

4. ΔL_w in dB

   Covering improvement rating

   The improvement a floor covering adds in a laboratory test on a standardised bare reference floor, per EN ISO 10140.

   A 20 mm dense rubber gym plate is rated at about 25 dB.

   A difference measured on a test floor, not a level left in your neighbour's ceiling, and it cannot be subtracted from either of the two figures above.

Ground-reaction force is a load, expressed as a multiple of body weight. A force-plate study of 100 recreational runners recorded peak vertical ground-reaction force at 2.37 times body weight at 9 km/h, 2.51 at 11 km/h and 2.62 at 13 km/h. That is what arrives at the floor. It is a force ratio and carries no acoustic information whatsoever.

The building requirement describes the construction. Germany’s DIN 4109 sets a maximum weighted normalised impact sound level for a dwelling-separating floor of 53 dB in the 1989 version that courts have applied to existing buildings, tightened to 50 dB in the 2018 revision. The rating is produced by dropping a standardised tapping machine on the floor above and normalising the result to a reference room, so it grades the floor as a piece of building fabric and has nothing to say about which exercise you chose. One practical consequence does reach a resident: a German court drew a line between replacing only the floor covering, which does not oblige you to meet the current standard, and an intervention in the floor build-up itself, which does.

The regulatory limit describes what a resident may transmit. South Korea caps direct-impact floor noise at a one-minute equivalent of 39 dB(A) by day and 34 dB(A) at night, with peaks limited to 57 and 52 dB(A). That is a sound pressure level, measured in the affected dwelling by a prescribed procedure, in one country.

The covering rating describes a product in a laboratory. A 20 mm dense rubber gym plate with a 1000 kg/m3 covering layer is rated at delta-Lw 25 dB under EN 10140-1. The test lays the covering on a standardised bare reference floor and reports the difference the covering makes to that floor.

Those are four different measurements of four different things, and the arithmetic that people want to perform on them does not exist. A covering’s rating cannot be subtracted from a national dB(A) limit, because the two are not comparable: one is a difference measured on a laboratory floor, the other a level measured in an occupied room. A building rating cannot be converted into a permitted level for a given exercise, because it grades a slab rather than an occupant. And no force ratio becomes a decibel by any route. What a covering rating is genuinely good for is comparing one covering against another, which is a real and useful thing to be able to do.

Density is the spec that is not printed

The category runs on percentages. “85 to 90 percent impact absorption” and “up to 80 percent noise reduction” are the two figures most often quoted at the point of sale, and across the pages that quote them no test method is named, which leaves them uncheckable rather than wrong.

Density is checkable and it is what separates two tiles with the same headline thickness. German flooring practice puts fitness covering at 850 to 950 kg/m3, with premium plates reaching about 1000. Both figures describe trade practice rather than a standard, so they are usefully read as the band a serious product tends to sit in, not as a threshold anything passes or fails. A foam-filled 20 mm tile and a rubber-granulate 20 mm tile are different products, and the number that says which is which is rarely on the box.

Two of those rows come straight from the trade sources: Warco gives 18 mm as a minimum, 30 mm as better, and 40 mm where shock absorption matters most, and Vibraplast puts a 140 kg/m3 composite heavy foam of 30 to 50 mm under machines. What the trade sources do not give is a pairing between a thickness and a weight, so the pairing below is ours and is offered as an editorial cutoff rather than a tested limit.

Dense rubber over foam for anything past stretching, then. Foam compresses permanently under a point load and dents under a set-down dumbbell, which earns it a place only as a soft top layer over a rubber base. Between a thicker tile of unstated density and a thinner one that publishes a figure, I would take the thinner one, on the grounds that a stated number can be checked and a headline millimetre cannot.

Matching the covering to what is under it

The same tile behaves differently depending on the floor beneath it, and most guidance in this category quietly assumes a garage slab. Four cases cover almost every flat.

• Hardwood. A sliding bench foot marks oak in one session, so a rubber sub-layer earns its place even for bodyweight work. A 10 mm rubber tile under a thin top mat covers the contact points, which is where the risk actually sits.
• Floating laminate. The panels move under a small heavy footprint. Spread the load across a wide rubber tile under any rack or bench foot instead of a small puck.
• Concrete slab. The most forgiving subfloor for weight, and the one where a sealed mat can trap condensation over a cold slab in winter. Breathable interlocking tiles, or a moisture barrier under a rolled mat.
• Carpet over slab. Kind to joints, poor for stability. A rigid board under each load-bearing foot, topped with a rubber tile, steadies a bench without lifting the carpet.

Those are the common cases rather than a complete set. A floating screed, a raised access floor, or a timber joist floor in an older building each behave differently again, and none of them can be identified reliably from the surface, so an old building generally means asking rather than assuming.

Floor load is a design reference, not a reading of your building

Weight is the question that arrives after the rack is already standing, and it gets answered with a single number that was never a single number. Under Japan’s Building Standard Law Enforcement Order Article 85, a residential room carries three live-load values, not one: 1800 N/m2 when calculating the floor structure, 1300 N/m2 for beams, columns and foundations, and 600 N/m2 for seismic force. The same room, three figures, because each supports a different calculation.

All three are loads per unit floor area, which is to say distributed loads. Equipment does the opposite.

So a wide rigid board under each load-bearing leg, laid before the tile goes down, is worth doing because it lowers contact pressure. It is not a structural assessment and it certifies nothing about your slab. For a loaded rack that stays standing, stacked plates in one corner, or anything water-filled, the honest next step is the building documents, the landlord or a structural professional, in that order. For bodyweight work and a pair of dumbbells this rarely binds at all.

A worked corner

For the portable mat that goes over the top of a build like this, the exercise mat guide applies the same density reasoning to yoga, strength and cardio mats.

What a covering does not change

A treadmill is the case where expectations run furthest ahead of the material. It is worth noting how the one regulator with a published limit handles it: South Korea’s floor-noise centre excludes the friction, impact and striking sounds arising from the use of exercise equipment from the floor noise it mediates, alongside washing machines and refrigerators. So in the jurisdiction that writes the strictest number in this article, a treadmill complaint is not a floor-noise complaint at all. That does not mean the neighbour hears nothing. It means the number is not the tool for that argument, and a machine isolation pad under the frame is a better one. On a machine that runs in a flat above someone, I would put the isolation pad ahead of another 10 mm of tile across the room, because the pad addresses the frame that is actually feeding the slab.

Beyond that, a covering works on one path and one path only. Airborne sound, the motor hum and the belt, moves through air and leaks past doors and gaps; a covering does nothing for it. And no thickness makes a dropped weight, a jumping session or a jogging pace appropriate above a sleeping neighbour. Those are scheduling and exercise-selection decisions, and they sit outside what any material can be asked to fix. The routine side of that is in quiet apartment workouts and in the apartment-friendly routine; the equipment side is in the small home gym setup guide and in adjustable dumbbells versus regular.

Because the covering is one input among several, the planner below weighs it against your exercise, your hour and your footwear, and names the change that shifts the ranking most.

Apartment impact-risk planner

Pick your situation. The result updates as you change any answer and names the one fix that helps most.

Your floor Ground floor / no neighbor below Upper floor, neighbor below

Time of day Daytime Evening, before ~10pm After ~10pm (quiet hours)

What you are doing Static / mobility (yoga, stretching, holds) Standing band work / set-down weights Dynamic floor (lunges, step-ups, mountain climbers) Jumping, skipping, or running

Mat under you No mat Thin foam (<10mm) Dense rubber 10-20mm Dense rubber 30mm+

Footwear Barefoot / socks Cushioned trainers

Impact-transmission risk: -

FitVilo editorial planning model. It ranks risk from weights FitVilo chose; it does not measure or predict a noise level. See methodology.

The same planner sits on the tools page, beside a clear-zone check for whether the equipment you are building this floor for fits the room at all.

Comparing one covering against another is the limit

It claims that density and stiffness are the properties worth shopping on, that a covering rating compares coverings, that a code figure is a design reference and not a reading of your building, and that the four quantities above measure four different things.

It does not claim that any thickness meets a limit, satisfies a standard, or produces a stated reduction in the flat below. FitVilo runs no laboratory and measures nothing. If you need an actual level or a compliance answer for your building, that takes a measurement in the affected apartment by someone qualified to take it, and there is no page, calculator or data sheet that substitutes for one. The rest of the home gym hub covers the equipment that then has to stand on the floor you built.