When you put it that way then definitely yes, the two are different and pink is preferred. As with many things, the application matters.
I'm not sure if you still have any confusion on this, so sorry if you already got this.
Pink noise does have a downward slope, depending on how you display it. That's why I included the two different plots.
If the plot is dB vs actual hertz (narrow band), pink noise will have a downward slope (like that in the wikipedia image above, though the color choices require a careful eye - white noise being gray on that one, and pink noise being the lower pink color).
If the plot is dB vs aggregated octave bands (or some other relatively wide bin based on a portion of an octave), pink noise will measure flat (like the dot-based plot above from softdb).
Also from the softdb link:
"First off, the spectra of white and pink noise can be a source of confusion. The confusion is primarily due to the fact that the graphical representation of white and pink noise will be completely different depending on the the type of analyzer being used. For instance, viewed in a spectrogram set in narrow bands, white noise looks just like the image below. Notice how flat the spectrum is? That’s because all frequency bands (ascending left to right on the x-axis) have more or less the same energy level.
And now here’s where things get a little confusing…
The very same white noise signal, but viewed in third-octave bands instead of in narrow bands, looks completely different from the image in figure 1 above. Notice how the spectrum isn’t flat anymore but sloping upward as we go from lower to higher frequencies?"
@mattstat these visualizations are all coalescing in my mind now.
So my observations were correct(ish) just formed on the basis of improper logic, hahah. In the case of of the REW RTA dB vs log freq, white noise is applicable to see a straight flat line (ignoring slope for desired taste for now).
So my observations were correct(ish) just formed on the basis of improper logic, hahah. In the case of of the REW RTA dB vs log freq, white noise is applicable to see a straight flat line (ignoring slope for desired taste for now).
Garboui,
Appears you may have linear and logarithmic scales mixed up.
On REW's RTA log (logarithmic) frequency scale pink noise being equal energy per octave produces a flat line, while white noise rises at 10dB per decade.
On a linear scale, pink noise decreases at 10dB per decade, white noise produces a flat line.
Pink noise energy level is the same in each and every octave.
White noise energy is equal per frequency, so progressively increases in the higher octaves.
Pink noise is white noise filtered at -10 dB/decade.
White noise is not recommended for testing speakers, both from the aspect of potentially burning out tweeters and it does not correspond to our logarithmic hearing response.
Art
Appears you may have linear and logarithmic scales mixed up.
On REW's RTA log (logarithmic) frequency scale pink noise being equal energy per octave produces a flat line, while white noise rises at 10dB per decade.
On a linear scale, pink noise decreases at 10dB per decade, white noise produces a flat line.
Pink noise energy level is the same in each and every octave.
White noise energy is equal per frequency, so progressively increases in the higher octaves.
Pink noise is white noise filtered at -10 dB/decade.
White noise is not recommended for testing speakers, both from the aspect of potentially burning out tweeters and it does not correspond to our logarithmic hearing response.
Art
I dont mean to drag this on as there are facts here. However I am trying to get aligned with my observation experience.
According to this (previously referenced) graph, pink noise is -10dB/decade on a log scale (4th colour band down).
Comparing RTA with White Noise and a measurement sweep in REW both response are flat. The pink noise REW RTA of the same measurement is sloped down. By these comparisons, tuning to a flat response with pink in RTA would be an overly bright system (confirmed by doing so).
There is definitely magnitudes more power going to the tweeters with white noise. I can see this being a real hazard with some lower efficiency/power rated drivers and will heed for future ventures. With 109dB/1W sensitivity CD's I imagine I have the safety overhead here that I would be deaf before I start doing damage to the driver.
NOte on the MMM, the <200Hx is much much better represented in the response.
I am not very familiar with REW, as I normally use SMAART for testing.
Here is what I should have wrote regarding REW, noise and the display mode:
On REW's RTA plot, pink noise being equal energy per octave produces a flat line, while white noise rises at 10dB per decade.
On REW's spectrum plot, pink noise decreases at 10dB per decade, white noise produces a flat line.
From REW Help Index, Page 209:
"Pink noise has energy that falls 3 dB with each doubling of frequency. On a spectrum plot it is a line that falls at that 3 dB per octave rate, on an RTA plot it is a horizontal line as the energy in the signal is falling at the same rate as the bins are widening. We perceive pink noise as having a uniform distribution of energy with frequency.
Single tones are a special case, they will appear at the same level on either style of plot as their energy is all at one frequency, so on a spectrum plot they show as a vertical line, on an RTA plot they show (typically) as a bar of the width of the bin at their frequency, but the height of the bar is the same as the height of the line on the spectrum as all the energy is at that one frequency."
A measurement sweep, being equal amplitude at all frequencies, will appear as flat in either mode.
A quick and dirty pink noise test on my iMac right speaker using the internal mic, using a "spectrum" plot for displaying pink noise response shows the response dropping about -10dB per decade more than the "RTA" plot.
At any rate, in REW if you are using pink noise to analyze or adjust your system in real time, you must use the "RTA" (Real Time Analyser) mode to see pink noise appear flat on the dB scale, assuming the device under test has flat response.
Cheers,
Art
Here is what I should have wrote regarding REW, noise and the display mode:
On REW's RTA plot, pink noise being equal energy per octave produces a flat line, while white noise rises at 10dB per decade.
On REW's spectrum plot, pink noise decreases at 10dB per decade, white noise produces a flat line.
From REW Help Index, Page 209:
"Pink noise has energy that falls 3 dB with each doubling of frequency. On a spectrum plot it is a line that falls at that 3 dB per octave rate, on an RTA plot it is a horizontal line as the energy in the signal is falling at the same rate as the bins are widening. We perceive pink noise as having a uniform distribution of energy with frequency.
Single tones are a special case, they will appear at the same level on either style of plot as their energy is all at one frequency, so on a spectrum plot they show as a vertical line, on an RTA plot they show (typically) as a bar of the width of the bin at their frequency, but the height of the bar is the same as the height of the line on the spectrum as all the energy is at that one frequency."
A measurement sweep, being equal amplitude at all frequencies, will appear as flat in either mode.
A quick and dirty pink noise test on my iMac right speaker using the internal mic, using a "spectrum" plot for displaying pink noise response shows the response dropping about -10dB per decade more than the "RTA" plot.
At any rate, in REW if you are using pink noise to analyze or adjust your system in real time, you must use the "RTA" (Real Time Analyser) mode to see pink noise appear flat on the dB scale, assuming the device under test has flat response.
Cheers,
Art
Last edited:
I believe we are both arguing the same thing now but with a different frame of reference. I have been commenting from the standpoint of using a spectrum plot in RTA (what I have been using and same as sweeps) and not the octave binned plot display.
To finally put this to rest, pink noise and RTA 1/4 octave is the best practice for using RTA for MMM tuning.
I use white noise because it is flat. No corrections are needed in analyzing white noise.
The downside is that white noise contains much more high-frequency energy than natural sound. It should not be played loudly - it is bad for the ears and (if loud enough) tweeters.
Ed
The downside is that white noise contains much more high-frequency energy than natural sound. It should not be played loudly - it is bad for the ears and (if loud enough) tweeters.
Ed
Why?
I am quite confused about the confusion on various noise colors. The right explanation has already been given. Of course you can EQ using white noise, as long as you realize what your measurement system is measuring…
Yes, you could.
About sweeps vs noise, you can gate sweeps to exclude your room. Noise can be used for steady state measurements.
About sweeps vs noise, you can gate sweeps to exclude your room. Noise can be used for steady state measurements.
I make periodic pink noise measurements gated most of the time, actually (ARTA). Gating is not the issue. Started measuring on Neutrik 3300 with warble sine sweep tho, IIRC that had a following filter to reduce room influence.
As Markbakk says, it is possible to extract a gated impulse response from ARTA using the periodic pink noise function. In fact, the preferred method is to use the PN noise signal combined with averaging to get a repeatable measurement with a low noise floor.