No, integration (-6 dB/octave) of white noise is brown/Brownian/red
noise. Pink noise is -3dB/octave from white noise.
This is correct.
Both are incorrect. Brown/Brownian/red is -3 dB/octave from pink,
and pink is -3dB/octave from white. Brown/Brownian/red noise is
-6dB/octave from white noise, as in integrated / antidifferentiated
white noise.
-3 dB is half power, and -6 dB is half voltage. dB and dBV are the
same changes if impedance is constant or considered irrelevant.
If it's -3dBV, then such a
That is a white-to-pink filter, which does not exist exactly. However,
there are useful approximations, especially for the audio frequency range.
True, at least both largely and generally.
Pink noise is only approximated by feeding white noise into a filter
approximating having a -3dB/octave slope. There are circuits easy to
find by web searching that achieve such useful approximations for the
20-20,000 Hz range or so, to +/- 1/2 or 1/4 dB or something like that.
The Wikipedia article on "colors of noise" is correct, except for two
minor issues, and then it's minor:
1. Analogy of white noise to white light: Light with a spectral power
distribution of "pink noise" is very slightly more purplish than blackbody
having color temperature of 7000 K.
Light with a spectral power distribution of "white noise" is very
slightly more purplish than blackbody having color temperature of 9000 K.
+6 dB/octave from white noise is the characteristic of blackbody
radiation having infinite color temperature. Such color temperature
has color being close to that of a deep side of sky blue, very slightly
more purplish and less greenish, and not "infinitely deep blue" or violet.
(However, ideal isotropic radio antennas have a -6 dB/octave
characteristic, due to capture area being a constant times square of
wavelength - their output from reception of infinite temperature thermal
radiation is white noise.)
Light having spectral power distribution of brown/Brownian/red noise
is "equal energy per unit wavelength" "white light", very slightly more
purplish than blackbody having color temp. of 5400 K.
2. That article's plots of spectral power distribution:
That has "flat" being equal energy/power per unit frequency bandwidth,
which makes white noise "flat". However, pink noise has equal power
per unit log of frequency, as in equal power per octave. That makes
pink noise "flat" for most audio considerations. And for visible light,
brown/Brownian/red noise (equal energy / power per unit inverse frequency
bandwidth, which is equal power / energy per unitwavelength bandwidth)
is considered "flat" due to being generally not yellowish / orangish
since it has color close to that of light radiated by a 5400 K blackbody.
<SNIP stuff on name-calling and on someone who is quick to do such>
- Don Klipstein (
[email protected])
