Free online grey noise generator

What is grey noise?

Grey noise, also spelled gray noise, is the only color here defined by how hearing works rather than by a mathematical slope. White noise is flat when you measure it with an instrument, but human hearing is not flat: your ears are far more sensitive around 3 kHz than they are in the bass, so flat-by-measurement sounds bright and top-heavy to a person.

Grey noise corrects for that. It is shaped along the inverse of an equal-loudness contour, scooped where your hearing is most sensitive and lifted where it is least. On a spectrum analyzer it looks distinctly uneven. To you it should sound level at every pitch, with no part of it standing out.

When grey or gray noise may fit

The practical case for grey is narrow and worth stating honestly: it is the color to try after the obvious ones have not worked. If white noise reads as piercing and brown noise reads as muffled, the problem is usually not the amount of masking but the balance, and grey is the one shaped to fix balance specifically.

That evenness is also why grey is the color most often raised in connection with tinnitus masking, since a noise with no prominent peak of its own is less likely to compete with a tone you are already hearing. That is a reason to try it, not a medical claim, and tinnitus is worth taking to an audiologist rather than to a web page.

Grey corrects, it does not upgrade. If pink or brown is already working for you, there is nothing here to gain.

How loud should you set it?

Grey noise is shaped so that no part of it stands out, which makes it unusually easy to misjudge: it sounds quieter than it measures, because nothing in it is poking at the sensitive part of your hearing. Set it lower than instinct suggests, and judge it purely on whether the disturbance has stopped registering.

Set it, then leave it alone for a full night before adjusting. Grey rewards that more than the other colors, because what it changes is balance rather than loudness, and balance is not something a ten second sample can tell you about. The full three-step method is here.

How this generator actually works

Grey noise is the hardest of the six to serve as a file, because the shaping is the whole point of it. Nothing here is recorded: the audio is generated in your browser by a port of the generator that runs inside the Veil app.

It starts from raw white noise, then applies two peaking filters: a broad 13 decibel scoop at 3 kHz, where hearing is most sensitive, and a 9 decibel lift at 110 Hz, where it is least. A low-pass at 7 kHz finishes it.

One honest caveat about the shaping. It is fitted along one equal-loudness curve rather than derived from a definitive one, because those curves shift with playback level and differ between listeners, so there is no single correct answer to fit. We also roll off above 7 kHz instead of lifting the top end the way a strict inversion would, which keeps the most intense hiss out of it at sleeping volume.

Grey carries real low-end energy after that bass lift, so an untreated loop boundary clicks audibly. Each of the two buffers, about 30 and 37 seconds with coprime sample counts, is sealed by crossfading its head against real continuation samples, which removes the seam. Mixed together, they take roughly 1.55 years at 44.1 kHz to realign.

Try the other colors.

Which color works is personal, and a minute of listening settles it faster than any chart. Each of these is a live generator like this one: White noise Pink noise Brown noise Green noise Blue noise. Or hear all six side by side. If noise is not the right texture, there are also seven fan sounds.

Grey noise for sleep, in depth