Do acoustic panels actually make your voice sound better?

Ten minutes into the Monday call, a colleague says it as kindly as she can: "Are you in a bathroom?" You are not. You are in a spare bedroom with a laminate floor, bare white walls, a desk pushed against one of them and a MacBook about 60 cm (two feet) from your face. Nothing in that room is loud. The fan is off, the window is shut. And still every word you say arrives with a faint metallic tail, as if you were talking into a bucket.

That evening you have three browser tabs open: a 24-pack of gray foam wedges, a set of oak slat panels on felt, and a video of someone building fabric panels in a garage. All three promise to fix exactly this.

The short answer: yes, acoustic panels can make your voice sound better on calls and recordings, but they fix one problem only, your own voice bouncing around the room. Thin panels mostly fix the upper part of the voice range, and even thick ones do less than moving the microphone closer, which costs nothing. They do nothing about sound coming in from outside. If the thing ruining your audio is a neighbor's drill or traffic, whether you can soundproof a room for calls is the page you want, because that is a construction problem and panels will not touch it. If it is a fan, a keyboard or a dog in the same room, how to stop your mic picking up background noise answers that faster. This article is for the reader whose room is quiet but sounds hollow.

How acoustic panels work, and what they cannot do

When you speak, sound leaves your mouth in every direction. A small part goes straight into the microphone. The rest hits the desk, the wall, the ceiling and the window, and some of it bounces back into the mic a few milliseconds later. Bare plaster, glass and laminate reflect almost all of that energy, so the reflections stack up into a short tail after each word. On a call this reads as distance and hollowness. In a recording it is what makes a voice sound like it was taped in a kitchen.

An acoustic panel is a slab of porous material: open-cell foam, mineral wool, fiberglass, or felt behind wooden slats. Sound pushes air into the pores, the air rubs against millions of fibers, and a little of the energy turns into heat. Less comes back out. That is the whole mechanism. Absorption, not blocking.

This is why the same panel can be excellent and useless at once:

The timing of a reflection is easy to work out. Say your mouth is 20 cm from the mic and the wall behind your head is 1 m away. The reflected sound travels to the wall and back past your head to the mic, roughly 2.2 m against 0.2 m for the direct path. At 343 m/s that extra 2 m arrives about 5.8 ms late. Too early to hear as a separate echo, just in time to smear the edges of consonants. That smear is what panels remove.

If people hear a true repeat of their own voice when they talk to you, that is a different fault: your speakers feeding back into your mic. Why people hear an echo from your mic covers that one, and no panel will fix it.

Do you need acoustic panels? A two-minute test

Before you buy anything, find out whether your room is the problem or your mic distance is. Both produce the same complaint.

1. Clap once, sharply, standing where you sit for calls. If you hear a bright ringing tail or a flutter (a fast "zzing" between two parallel walls), the room is live.
2. Open QuickTime Player, start a new audio recording, and read the same paragraph twice: once at your normal distance from the laptop, once with an external mic or headset mic about 15 to 20 cm (6 to 8 inches) from your mouth.
3. Play both back in headphones. Not on the laptop speakers, which hide exactly the detail you are listening for.

If the close take sounds clean and the far one sounds hollow, distance is the cheap fix and panels are optional. If the close take still has an obvious tail, the room is lively enough that treatment will be audible. If you want to hear yourself the way the other side of a Zoom call does, how to test your mic and hear what others hear walks through each app.

Acoustic foam panels vs dense panels: the numbers

Every absorber is rated with an absorption coefficient per frequency: 0 means everything bounces back, 1.0 means nothing does. (Lab values can read slightly above 1 because of edge effects in the test.) Here are two widely sold products, both 2 inches (about 5 cm) thick, tested by the same standard, ASTM C423, plus the 4-inch version of the second one:

Sources: the Riverbank lab report Auralex publishes for its 2-inch wedges and the Owens Corning 703/705 data sheet, both measured with the material flat against a solid wall.

Look at the 250 Hz column. That is where a speaking voice carries its body: a typical adult male voice has its fundamental between roughly 90 and 155 Hz, a female voice between 165 and 255 Hz, and the first harmonics sit just above. At 250 Hz the foam absorbs about a quarter of what hits it. The fiberglass board of the same thickness absorbs about 70 percent, and the 4-inch board more or less all of it.

So 2-inch foam is a high-frequency absorber. Cover a wall with it and the room stops ringing in the treble while the low-mid boom stays. NRC, the single number on most product pages, is an average of 250, 500, 1000 and 2000 Hz, so it hides exactly this. Always look for the 250 Hz figure.

There is a geometric reason thick and spaced panels do better. A porous absorber works where air is moving fastest, which for a given frequency is about a quarter of a wavelength out from the wall. At 343 m/s, a quarter wavelength is 34 cm at 250 Hz, 17 cm at 500 Hz and under 9 cm at 1000 Hz. A 5 cm slab glued flat to the wall sits where low-mid air barely moves. Panels mounted with an air gap behind them generally absorb more low frequency than the same panels flat on the wall, which is the cheapest upgrade in this whole article: hang the panel 5 to 10 cm off the wall instead of gluing it.

What about wooden slat panels?

Slat panels (thin wooden strips on a felt backing) are the version interior shops sell as acoustic wall panels. They do absorb, but how much depends on the mounting. Acupanel, one of the brands selling them, states that its class A absorption rating needs the panels screwed to battens with mineral wool in the gap, and that screwed straight onto the wall through the felt they are mainly decorative with some damping. If you buy slats for their looks, fine. If you buy them for your voice, budget for the battens and the wool behind them.

What 24 foam tiles do to a real room

Numbers on a spec sheet do not tell you what changes in your room, so here is a rough worked example using the Sabine formula for reverberation time (RT60, how long it takes a sound to fade by 60 dB). It is an approximation, and small rooms behave less neatly than the formula, but it shows proportions honestly.

Room: 3.0 m x 4.0 m, ceiling 2.6 m
Volume V = 31.2 m3, surface area = 60.4 m2
Assumed average absorption of a sparse room: 0.10
(lab coefficients above 1.0 counted as 1.0)
Total absorption A = 60.4 x 0.10 = 6.0 m2
RT60 = 0.161 x V / A = 0.83 s

Option 1: one 24-pack of 2" foam wedges (24 x 1 ft2 = 2.2 m2)
          250 Hz: 0.79 s   500 Hz: 0.70 s   1000 Hz: 0.63 s
          reverberant level drops 0.2 / 0.8 / 1.2 dB

Option 2: four 60 x 120 cm panels of 2" rigid fiberglass (2.9 m2)
          250 Hz: 0.64 s   500 Hz: 0.58 s   1000 Hz: 0.58 s
          reverberant level drops 1.1 / 1.6 / 1.6 dB

Option 3: move the mic from 60 cm to 20 cm, no panels
          direct voice rises about 9.5 dB, room stays the same

The decibel lines use the plain room-acoustics relation NR = 10 log(A after / A before). Doubling a room's total absorption lowers the reverberant sound by about 3 dB. In a bare room, doubling takes a few panels; in a room that already has a sofa, curtains and a rug, it takes a lot more material for the same 3 dB.

Two things fall out of this. First, the panels are not useless: going from 0.83 to about 0.6 seconds of decay is a shorter tail, and in a bare room you can hear that. Four thick panels beat 24 foam tiles at every frequency that matters for speech, especially the low-mids. Second, the mic distance line dwarfs both. Halving the distance between your mouth and the mic raises your direct voice by about 6 dB while the room's contribution stays where it was; going from 60 cm to 20 cm is three times closer, about 9.5 dB. To get the same improvement in voice-to-room ratio from absorption alone, you would need to multiply the room's total absorption roughly ninefold, which in practice means covering nearly every wall and the ceiling with thick absorber.

That is the turn most panel guides skip. In a call setup, panels are the second fix. The first one is getting the mic close to your mouth: a headset mic, a lavalier on your collar, a USB mic on an arm. Then panels clean up what remains.

Acoustic panel placement: where to place acoustic panels for your voice

Most placement guides are written for studio monitors and your ears: they tell you to find "first reflection points" on the side walls with a mirror. For a voice mic the geometry is different, because the source is your mouth and the receiver is the mic, usually within half a meter (20 inches) of each other.

The order that helps a talking voice most:

1. The wall behind your head. A mic with a cardioid pattern points at you, which means it also points at whatever is behind you. Sound bouncing off that wall goes straight into the most sensitive side of the mic. One or two panels at head height, centered behind your chair, are the highest-value spot in most rooms.
2. The surface right in front of you, if you face a wall. With a desk pushed against a wall, your voice hits that wall 40 to 70 cm away and comes straight back. A panel directly behind your monitor catches it.
3. The desk. A bare desk between mouth and mic is a reflector at very close range. A felt desk mat, a folded towel during recording, or simply raising the mic on an arm so it sits above the desk line all reduce it.
4. The ceiling above you. A panel hung flat above the desk takes out the floor to ceiling bounce. Worth it once the walls near you are done.
5. Corners, with thick panels only. Low-frequency energy collects in corners. Thin foam there does little; 10 cm of mineral wool across a corner does more.
6. Side walls at mouth height. Useful in narrow rooms where both side walls are within a meter (about three feet) of you.

An omnidirectional mic hears the whole room about equally, so for it panels behind you matter no more than panels to the side. If you are not sure which pattern your mic has or why it matters here, which mic pickup pattern is best for a noisy room explains the difference.

Placement beats panel count. Two panels in the right places do more than ten scattered across walls the mic does not face. A reasonable target for a small home office is four to six panels of 60 x 120 cm (2 x 4 ft), starting from the top of that list.

DIY acoustic panels: how to make acoustic panels that beat foam

Building your own is the cheapest way to get panels that work on the low-mids, because the material that actually absorbs low-mids (dense mineral wool or rigid fiberglass) is sold as building insulation. What you need for one panel:

Frame:     4 pieces of 1x4 lumber (19 x 89 mm), outer size 60 x 120 cm
Core:      rigid mineral wool or fiberglass board, 5 cm (good) or 10 cm (better)
           e.g. Owens Corning 703 in the US
Fabric:    any fabric you can easily breathe through when held to your mouth
Back:      optional thin fabric or gauze so fibers stay inside
Fixings:   staple gun, wood glue, two picture hooks
Mounting:  5 to 10 cm off the wall (spacers or hooks), not glued flat

The order of work: build the frame, drop the board in, stretch the fabric over the front and staple it on the back, then hang the panel like a heavy picture. The blow-through test on the fabric matters. A tightly woven or coated fabric reflects high frequencies before they ever reach the wool, which defeats the point.

Mineral wool and fiberglass shed fibers when cut. Wear gloves, long sleeves and a dust mask while cutting, cut outside or over a drop cloth, and keep every cut edge wrapped in fabric once the panel is done. That is also why the back layer is worth adding.

On cost, compare like with like: as of October 2026 Auralex lists its 24-pack of 2-inch Studiofoam Wedgies at $164.99 for 24 square feet of foam. Four DIY panels cover about 31 square feet with a core that, per the data above, absorbs roughly three times more at 250 Hz. Insulation board prices vary by country and store, so check locally before deciding.

If you would rather buy foam, buy the thickest you can and check the listing for a flammability standard. Auralex lists UL 94 HF-1 for its wedges. If a pack names no fire standard at all, treat that as the answer.

What you already own that works like a panel

Before buying anything, look at what absorbs in your house already. A bookcase full of books of uneven depth breaks up reflections. A closet full of hanging clothes is a thick porous absorber; recording voice-over inside one is an old trick for a reason. A heavy comforter hung on a stand behind you during a recording works on the same principle as a panel, with the same limits. A rug on a hard floor and heavy curtains over a window take out two large reflecting surfaces.

None of these look as tidy as panels, and a comforter on a drying rack is not a permanent setup. But if your test recording improves noticeably with a comforter behind your chair, you know panels in that spot will work, and you have found out for free.

For singing or voice-over work where you need a truly dry sound, a reflection filter or a small booth around the mic is a different and sometimes better approach than treating the whole room. Do you need a vocal booth or a reflection filter compares those options.

What software does about room echo

Panels change the room. Software changes the signal after the room has already done its work. Several tools now try to remove reverberation in real time:

• Krisp added room echo removal to its Mac app in version 1.17.13 in November 2020, alongside its noise cancellation.
• NVIDIA Broadcast has a Room Echo Removal effect, but it needs an NVIDIA RTX graphics card and Windows. There is no Mac version.
• Google Meet noise cancellation is documented as removing room echo along with typing and construction noise. It is a paid feature, available only on the Workspace and Google One plans Google lists, and you switch it on under More options > Settings > Audio.
• macOS has Voice Isolation, one of the Mic Mode options. On macOS 14 and later you pick Mic Mode from the Video or Audio menu that appears in the menu bar during a call; on macOS 12 and 13 it sits in Control Center. Apple describes Voice Isolation as minimizing background noise and does not promise to remove room reflections.
• TunedMic has a room echo control among its optional manual settings, and its one-minute Sound Check measures your voice, your mic and your room before building your profile.

There is a hard limit on all of them. Software has to separate your direct voice from its reflections, and the further the mic is from your mouth, the more those two overlap in time and level. A mic at 20 cm in a lively room gives processing a strong, clear voice to keep. A laptop mic at 70 cm in the same room hands it a voice already mixed with the room, and every tool has to guess harder. So the order stays the same: distance first, then the room, then processing. If you want a deeper comparison of these tools for noise as well as echo, which noise cancelling software actually works on calls goes through them one by one.

Where TunedMic fits, and where it does not

TunedMic is a Mac app that adds a virtual microphone called TunedMic, which you select in Zoom, Meet, Teams, Slack, Discord, OBS or your browser instead of your real mic. During setup you read a short passage for about a minute while it takes more than 30 measurements of your voice, your mic and your room. Processing happens on the Mac, adds less than a tenth of a second of delay, and the audio is never uploaded.

What that means for a lively room: the tuning starts from your real room rather than a stock preset, and the room echo control is there if you want to push further. What it does not mean: it cannot stop a neighbor's music crossing the wall, it cannot do to the room what thick panels in the corners do, and it does not turn a laptop mic across the room into a studio mic. The same rule as above applies. A closer mic gives it more to work with, and a wired or USB mic gives it more detail than a Bluetooth headset.

The real limits: it needs a Mac with Apple silicon (M1 or newer) and macOS 14 or later. There is no Windows or Intel version. It costs $19 once. Payments are final, with no refunds, which is why there is a free 7-day trial: it starts at first launch, includes every feature and needs no card or account.

Which fix for which room

Your next step

Tonight, before ordering anything, run the far-and-close recording test from earlier with a comforter hung behind your chair, and listen back in headphones. If the close take with the comforter sounds right, you know exactly where your first two panels go and that two is probably enough. If even the close take still rings, build or buy thick panels, not more thin foam, and hang them off the wall. And if the room is as good as you can make it but you want your voice on Zoom and Telegram to sound like you on a good day, watch the 38-second video on the home page with the sound on. It shows the one-minute read-aloud that tunes the mic to your voice and the Speech Report you get after a call. Then decide whether the free 7-day trial on your own mic is worth an evening. The price is on the pricing page.