Green vs Pink Noise
The only comparison here that is not a comparison of slopes. Green noise on this site is made from pink noise — a bell at 500 Hz, both ends cut away — so the question is what that recipe is worth.
One of These Is Literally Made From the Other
Every other comparison on this site puts two independent signals side by side. This one does not, because the green noise here is pink noise with a filter chain on top of it. Knowing the chain answers most of the questions people have about the pair.
The build, in order: a pink noise foundation at −3.07 dB per octave; a wide +8 dB bell centred at 500 Hz with a Q of 0.7, which is roughly two octaves across; two cascaded 100 Hz high-pass filters to pull back the rumble; and a 6 kHz low-pass to pull back the hiss. Measured on the output, the third-octave peak lands in the 400–500 Hz bands with gentle skirts of about 6 dB per octave either side — a broad tilt toward the middle rather than a resonance or a honk.
So "green versus pink" is really the question what does that chain buy, and what does it cost? — which is a far more answerable question than most colour comparisons, because both terms of it are known exactly.
One thing that cannot be done is to give green noise a dB-per-octave figure the way the other four have one. It is a shape rather than a tilt, and quoting a slope for it would be inventing a number. What can be stated precisely is the centre frequency, the bell width and the filter corners, which is what is above.
One Is Defined Tightly Enough to Be a Test Signal. The Other Is Not Defined at All.
Pink noise has a definition exact enough to do professional work with. Equal power in every octave band is a specification, not a description, and it is precise enough that pink noise is the standard signal for calibrating loudspeakers, aligning PA systems and measuring rooms: feed it in, and any bump the analyser shows belongs to the room rather than to the signal. Two implementations of pink noise from different vendors are, for practical purposes, the same sound.
Green noise has no standards-body definition in audio at all. No agreed slope, no agreed centre frequency, no agreed bandwidth. The familiar description — the mid-frequency component of white noise, somewhere around 500 Hz — traces back to Lau, Arce and Gallagher's Green-noise digital halftoning (Proceedings of the IEEE, 1998), which introduced green noise as the mid-frequency component of white noise for arranging dots on paper. The audio usage borrowed the idea from an image-processing paper and added a nature metaphor.
This has a direct and rarely stated consequence for anyone comparing the two: everything on this page is true of this generator's green noise and may not be true of anyone else's. A green noise centred an octave higher, or shaped with a narrow resonant peak instead of a broad bell, is a different sound with different strengths, and it is not wrong — there is nothing for it to be wrong about. Pink noise carries no such caveat, and that reliability is itself a reason some people stay with it.
What Gets Removed, and What the Room Gets Instead
Both colours are normalised to −20 LUFS under ITU-R BS.1770-4 K-weighting, so they arrive at the same perceived loudness. Since green noise is confined to a narrower range, that same loudness is concentrated into fewer octaves — which means that inside its range, green noise is genuinely more energetic than pink noise is. It is not a subset of pink noise at a lower level; it is the same loudness, packed tighter.
Below 100 Hz, green noise has given up. Two cascaded high-pass filters remove the region entirely. Pink noise does not have much down there either compared with brown, but it has some, and against a distant road or a boiler that difference is one-sided. If low-frequency intrusion is the problem, neither of these colours is the answer and green is emphatically not.
Above 6 kHz, green noise has also given up. Pink noise keeps audible content right to the top of the range. That upper content is what produces the impression of air, and it is also what does the small amount of consonant work pink noise manages. Removing it makes green noise sound closer and more enclosed — like a room rather than a space.
Between roughly 300 Hz and 1 kHz, green noise is meaningfully louder than pink. That is the +8 dB bell, and it is where the sound of the outdoors mostly lives: moving water at a distance, wind through foliage, the general hush of a large space. It is also, not coincidentally, where the human voice puts most of its body.
Voices: Body Against Intelligibility
This is the most useful practical difference between them, and it runs in an unexpected direction.
A voice through a wall has two components that matter. Its body — vowel fundamentals and general presence — sits around 300 to 800 Hz, and its intelligibility is carried by consonants around 2 to 4 kHz. Green noise's bell sits exactly on the body and its low-pass rolls away before the consonants. Pink noise covers both regions at a moderate level and neither of them strongly.
The result is that green noise is better at making a conversation stop feeling like a presence in your room, and no better — arguably slightly worse — at stopping you from following the words. Pink noise is the reverse in miniature. If you can still make out sentences through either of them, that is expected behaviour for both and the fix is white noise, which has full strength in the consonant band and is the only colour here that reliably takes speech from audible to unintelligible.
There is a related quirk with no equivalent on the pink side. Green noise emphasises the same frequencies your own voice occupies most strongly, so some people find it feels like it is competing with them when they are on a call, reading aloud or talking. Pink noise, being a straight line, never does this.
The Hardware Argument, Weighed Honestly
Green noise's most-repeated advantage is that it survives cheap speakers, and this is true. The regions it deliberately excludes — below 100 Hz and above 6 kHz — are precisely the two regions small drivers handle worst: they cannot reproduce the bottom at all, and many of them have a spiky, unpleasant top end. Green noise is the colour least damaged by a phone, a laptop or a small Bluetooth unit, and if sleeping in headphones is not an option and a real speaker is not available, that is a genuine reason to choose it.
The honest qualification is that the gap is smaller than the argument implies, because pink noise degrades gracefully too. A phone speaker takes away pink noise's bottom octave or two, and what remains still sounds like pink noise — lighter, but recognisably itself. This is not brown noise, which collapses entirely under the same treatment.
So the hardware case for green noise over pink is real but incremental: it is the difference between hearing all of a signal and hearing most of one. It becomes decisive mainly on genuinely poor speakers, or where the speaker's top end is harsh enough that removing everything above 6 kHz is a relief in itself.
The Evidence Gap Is Total and Almost Meaningless
Pink noise has appeared in published research. Green noise never has — not thin evidence, none — and while the term has no agreed definition it is hard to see how it could, since there would be no consistent stimulus to test. Any page claiming clinical support for green noise specifically is inventing it.
That sounds like a decisive point in pink noise's favour, and it is worth looking at what pink noise's record actually contains before treating it as one. Nigg and colleagues (Journal of the American Academy of Child and Adolescent Psychiatry, 2024) pooled white and pink noise together as a single exposure category and found a small benefit on laboratory attention tasks in young people with ADHD or elevated attention problems, and none outside that group — a result about a pair of colours, on lab tasks, not about pink noise for sleep. The famous 2017 slow-wave study used short pink noise pulses fired by a real-time algorithm at the predicted upstate of thirteen older adults' own slow oscillations, and found slow wave activity across the whole night similar between stimulation and sham; it is not a study of playing pink noise continuously. And the 2021 systematic review in Sleep Medicine Reviews rated the evidence for continuous noise improving sleep as very low quality across 38 articles.
So the accurate comparison is between a colour with no research and a colour whose research does not transfer to the way you would actually use it. Choose on acoustics and on what you will be playing it through. Nothing on this site is medical advice, and a sleep problem that persists is worth raising with a clinician.
Which One
Pink noise is the better default. It is fully defined, it is consistent between devices and apps, it keeps content across the entire spectrum, and it is the one to pick if you have no particular complaint and simply want something running.
Green noise is the better answer to two specific situations. The first is poor hardware — a phone or laptop speaker that mangles both ends of anything you give it. The second is a specific dislike: if you have concluded that brown noise is too muffled and white noise is too harsh, green does not merely split the difference the way pink does, it removes both of the things you objected to.
The comparison is easy to run because they are close relatives — start on pink, switch to green, and what you will hear is the bass and the air disappearing while the middle steps forward. If that trade appeals, green is your colour; if it feels like something has been taken away, it has been, and pink is where you belong.
Both are generated live in the browser rather than looped, so neither has a repeat point, and free sessions run 30 minutes. The free Android app carries both alongside the other three colours with no session cap, plus a sleep timer and offline playback — and green noise in particular pairs well with rain layered over the top, since it leaves the whole upper spectrum free for the rain to occupy.
Frequently Asked Questions
Is green noise just pink noise with a boost?
On this site, essentially yes, and the recipe is public: a pink noise base at -3.07 dB per octave, a wide +8 dB bell centred at 500 Hz with a Q of 0.7, two cascaded 100 Hz high-pass filters and a 6 kHz low-pass. The measured third-octave peak lands in the 400 to 500 Hz bands. Because both are normalised to the same perceived loudness, green noise is not a quieter subset of pink — it is the same loudness concentrated into a narrower range, so within the midrange it is genuinely more energetic.
Which is better for sleep, green or pink?
Neither is better in general and no research compares them, because no research on green noise exists at all. The useful distinction is practical: pink noise keeps content across the whole spectrum and is the safer default, while green noise removes everything below 100 Hz and above 6 kHz, which makes it the better performer on small speakers and the better fit for anyone who finds bass heavy and treble harsh. If what wakes you is low-frequency — traffic, machinery, footsteps — neither one is the right tool and brown noise is.
Why does green noise sound different in every app when pink noise does not?
Because pink noise has a specification and green noise does not. Equal power per octave is precise enough that any two correct implementations sound the same, which is why pink noise is used as a calibration signal in professional audio. Green noise is defined only by a loose verbal description — the mid-frequency component of white noise — which leaves the centre frequency, the bandwidth and the amount of boost entirely to whoever builds it. Comparing green noise across apps is comparing interpretations rather than implementations.
Which sounds better through a phone speaker?
Green noise, though by less than its reputation suggests. It was deliberately built to avoid the two regions small drivers handle worst — nothing below 100 Hz, which they cannot reproduce, and nothing above 6 kHz, where cheap tweeters often sound spiky — so what comes out of a phone is close to the whole signal. Pink noise loses its bottom octave or two and still sounds like itself, which is a much better outcome than brown noise manages. The gap is real but incremental: all of a signal against most of one.
Which is better for covering a conversation next door?
They do different halves of the job, and neither does it completely. Green noise emphasises 300 to 800 Hz, which is where the body and presence of a voice sit, so it is better at making a conversation stop feeling like something happening in your room. Pink noise spreads its energy across the consonant band as well, so it blunts intelligibility slightly more. If you can still follow the words through either, that is expected for both dark colours, and white noise is the one that reliably removes intelligibility.
Is there any research comparing green and pink noise?
None, and there is no research on green noise at all — not weak evidence, none. It is difficult to see how there could be while the term lacks an agreed definition, since there would be no consistent stimulus to test. Pink noise has appeared in studies, but its record is narrower than it sounds: the 2024 attention meta-analysis pooled pink together with white noise as one exposure category, and the well-known 2017 slow-wave study used short pink noise pulses timed to brain rhythms in thirteen people rather than continuous playback.