Why does my sub disappear on phones?

What a phone speaker can't play

A phone speaker is a very small driver in a case with no room for a proper enclosure or port. Low frequencies need a lot of air moved per cycle, and a driver that small can't move it, so its output falls off steeply below a cutoff somewhere in the low hundreds of hertz. The exact cutoff differs between phones. No published, citable measurement of a specific phone's speaker response could be found, so this article does not quote one. Instead, the measurements below use a stated simulation: a 4th-order high-pass filter (24 dB per octave) with its −3 dB point at 200 Hz, and a milder version at 150 Hz. Treat the numbers as "what a steep roll-off at 150–200 Hz does", not as a measurement of any phone.

With that filter, the attenuation at the frequencies that matter for a 55 Hz sub is:

The fundamental is gone (−44.9 dB is a 175:1 reduction in amplitude). The 3rd harmonic and above pass almost untouched.

The missing fundamental: how your brain fills in bass

A periodic note at 55 Hz is a fundamental at 55 Hz plus harmonics at whole-number multiples: 110, 165, 220, 275, 330 Hz and so on. The spacing between the harmonics, and the repetition rate of the waveform, are both 55 Hz.

If you remove the 55 Hz component and play only the harmonics, listeners still report a pitch at 55 Hz. Oxenham (2023) puts it this way: "We perceive a pitch corresponding to the F0 of a harmonic complex tone, even when the component at F0 itself is missing (the so-called pitch of the missing fundamental)." The classic experiments on this are by Schouten, Ritsma and Cardozo (1962), who called the perceived pitch the "residue". The mechanism is still debated (place coding versus timing coding in the auditory nerve; see Plack et al., 2005), but the effect itself is robust and easy to hear.

This is the whole reason any bass is audible on a phone. The phone cannot play 55 Hz. If the sound contains harmonics of 55 Hz, you hear a note at 55 Hz anyway. If it doesn't, you hear nothing.

Audio to try on a phone (8 s each): a 55 Hz sine, the same sine saturated, and the saturated version with the fundamental removed. On monitors or headphones the third one sounds thinner but at the same pitch. On a phone the first one is close to silent and the other two sound the same.

Why a clean sine sub vanishes

A sine wave has one component and no harmonics. When the fundamental is filtered out, nothing is left to carry the pitch. In the test sine, the 120–400 Hz band holds −88 dB of the signal's power, which is numerical noise. On a phone it is simply absent.

Bass music is where this hurts most. A sine or near-sine sub is the standard way to get a clean, loud low end that sits under a kick, and much of the track's energy is in that one component. In the three producer tracks measured below, 26–43% of the total power is in the 30–120 Hz band. On a phone, all of that is removed and only what sits above the cutoff is heard.

Your ears are less sensitive down there anyway

Even on a full-range system, low frequencies need more sound pressure than mids to sound equally loud. ISO 226:2003 gives the equal-loudness contours. Tytonic's analysis code implements the standard's formula and table, and these are its outputs:

Two things follow. First, the gap grows as playback gets quieter: at 40 phon a 50 Hz tone needs 37.8 dB more than 1 kHz to sound as loud, against 21.7 dB at 80 phon. Phone listening is quiet listening, so even the little sub a phone leaks is perceptually smaller than it is in the studio. Second, the harmonics at 200 Hz and above sit in a range where the ear is far more sensitive, which is part of why a harmonic layer is so audible on small speakers.

What Tytonic measured

Tytonic is an AI mix engineer. Its analysis code is the same code that produces the free diagnosis on tytonic.org. For this article it was run on the signals below, plus the phone simulation above.

Method:

• Sample rate 44.1 kHz. Spectra are Welch power spectra (16384-point, 2.7 Hz bins), averaged over left and right.
• "Sub share" is the share of total signal power in 30–120 Hz, in dB (0 dB would be all of it).
• Loudness is LUFS-I (ITU-R BS.1770, via pyloudnorm). RMS is the plain level in dBFS.
• Phone simulation: 4th-order Butterworth high-pass, applied causally, cutoff 200 Hz (main) and 150 Hz (milder). It is a simulation, not a phone.
• Harmonic levels are the spectrum peak within ±4 Hz of each multiple of the fundamental, relative to the fundamental's peak.

Synthetic signals (8 s, stereo, 55 Hz)

Three signals: a 55 Hz sine at −6 dBFS peak; the same sine through tanh(4x) saturation, gain-matched to the sine's LUFS; and the saturated signal with the fundamental removed (8th-order high-pass at 80 Hz, zero-phase), i.e. only the overtones.

Notes on the table:

• Symmetric saturation (tanh) produces odd harmonics only, so this sub gets 165 Hz and 275 Hz but no 110 Hz. Asymmetric saturation would also add 110 and 220 Hz.
• The saturated sub and the harmonics-only signal come out of the 200 Hz simulation at the same loudness (−29.2 LUFS). That is the point: everything a phone plays of a saturated sub is its harmonics. The fundamental contributes nothing.
• The pure sine loses 44.7 dB. The saturated sine, at the same loudness and same pitch, loses 19.1 dB. Harmonics 15–27 dB below the fundamental turned a signal the phone cannot play into one it can.

Three producer tracks

Three unreleased bass-music tracks by Tytonic's founder (TIMEMACHINE, LOWS, CRIME VIP), analysed on the loudest 45 s window of each (the drop). No commercial or third-party tracks were used.

What this shows:

• The simulation removes about 30 dB of sub share from every track (from −4…−6 dB to −34…−38 dB). The sub band itself is gone, as with the sine.
• The tracks lose only 1.5–2.7 dB of RMS and under 1 LU of loudness, because 57–74% of their power was already above 120 Hz, and because the sub notes in all three have harmonics only 10–22 dB below the fundamental. Those harmonics are what you hear on a phone, and they carry the note.
• LUFS understates the loss on purpose: BS.1770's K-weighting already discounts low frequencies, so a metric designed around loudness barely moves when the sub is removed. RMS, and the sub share figure, show it.
• The 150 Hz version of the simulation gives the same picture with smaller numbers (sub share −24…−28 dB, RMS loss 1.4–2.6 dB).

Tytonic's free diagnosis reports the sub share, the sub fundamental and the kick/sub relationship for any uploaded drop. The per-track harmonic content of the sub is part of the paid fix.

What engineers do about it

The general principle follows from the measurements: give the sub something above the phone's cutoff that is harmonically locked to it, and check the result on a phone. Common approaches, in general terms:

• Saturate the sub, or a copy of it, so the fundamental gains harmonics. Symmetric saturation adds odd harmonics (3×, 5×); asymmetric adds even ones too (2×, 4×). The second harmonic of a 40–55 Hz sub still falls below a 200 Hz cutoff, so the 3rd harmonic and above do most of the work on a phone.
• Add a harmonic layer (a mid-bass or "top bass" sound) that plays the same notes an octave or two up, and keep it in tune and in time with the sub.
• Don't fix it by boosting the sub. A boosted fundamental is still filtered out. It only eats headroom on full-range systems.
• Check on a phone, and at low volume. The equal-loudness table above shows that quiet listening makes the low end relatively smaller again.

How much harmonic content, at which frequencies, and whether the kick leaves room for it depends on the track. That is the per-track work.

FAQ

Does mono matter for phone playback? Yes. Many phones have one speaker, or two so close together that the playback is close to mono, so anything that cancels between left and right is lost on top of the roll-off. Keep the sub and its harmonics centred. Tytonic measures low-end stereo width (side/mid ratio below 120 Hz) in the free diagnosis.

Should I just boost the sub? No. In the simulation, a 55 Hz component is attenuated 44.9 dB regardless of its level. A boost of a few dB does not change that, and on club systems the extra level costs headroom. Add harmonics instead.

Do earbuds have the same problem? Much less, in general. A sealed earbud drives the small closed volume of the ear canal rather than open air, which is a far easier job at low frequencies than what a phone speaker does. Earbuds were not measured for this article. The equal-loudness effect still applies at low volume, though.

Is 200 Hz the right cutoff for my phone? Unknown without measuring it. Phones differ, and no citable measurement could be found. This article uses 200 Hz as the main case and 150 Hz as a milder case; the conclusions are the same for both, only the numbers change.

Will the missing fundamental make a 55 Hz sine audible if I just play it loud? No. The effect needs harmonics. A sine has none, so there is nothing for the ear to infer a pitch from.

Sources

• ISO 226:2003, Acoustics — Normal equal-loudness-level contours. International Organization for Standardization. https://www.iso.org/standard/34222.html
• Schouten, J. F., Ritsma, R. J. and Lopes Cardozo, B. (1962). "Pitch of the Residue." Journal of the Acoustical Society of America 34(9B), 1418–1424. https://doi.org/10.1121/1.1918360
• Oxenham, A. J. (2023). "Questions and controversies surrounding the perception and neural coding of pitch." Frontiers in Neuroscience 16, 1074752. https://doi.org/10.3389/fnins.2022.1074752
• Plack, C. J., Oxenham, A. J., Fay, R. R. and Popper, A. N. (eds.) (2005). Pitch: Neural Coding and Perception. Springer Handbook of Auditory Research, vol. 24. Springer. https://doi.org/10.1007/0-387-28958-5
• Measurements: Tytonic's analysis code, run on 2026-10-08.