Hacker News Reader

Comment by tverbeure | Hacker News Reader

Very first sentence: "This is pink noise... if you measure it, you can make it look like a straight line."

He then shows a horizontal straight line: that's white noise?

Pink noise distributes energy logarithmically in a way that matches the the sensitivities of the human ear. The graph which he shows on screen is ALSO in that same logarithmic space, hence why it shows pink noise as flat: it's flat as far as how humans hear it. White noise is not perceptually flat to humans. Thus when testing audio, it's important to use pink noise instead of white noise.

https://en.wikipedia.org/wiki/Colors_of_noise#Pink_noise

> The graph which he shows on screen is ALSO in that same logarithmic space, hence why it shows pink noise as flat

Does this also mean that audio equipment in general will display a spectrum with this kind of logarithmic offset adjustment?

More often than not, yeah. On the music production (rather than listening) side the two EQs and spectrum analyzers I use most frequently, FabFilter Pro-Q and Voxengo SPAN, both use a 4.5dB/octave tilt by default. You can adjust it, but I've never done so.

Practically all of audio land is logarithmic.

The orders of magnitude in power/loudness is pretty astonishing.

When introducing decibels to new audio engineers, we generally introduce it as a 3db increase is a doubling in power, a 10db increase is a 10x increase in power.

It gets silly when you start talking sound pressure level, because how people perceive a 10x increase in the output power, is about a doubling of the perceived loudness.

Well, yes, of course it’s logarithmic.

The question was whether or not the audio displays have a correction to make pink noise render as a flat horizontal line, whereas general test and measurement spectrum analyzer would show a tilted line.

The main difference between white and pink noise, is white noise has the same level of power across all frequencies. Whilst pink noise has the same level of power across an octave.

There are more frequencies in the high-range, so it naturally sounds considerably higher pitched.

i.e. a C6 on a keyboard sits at around 1Khz (1046.502 to be exact) and a C#6 is 1108.731 KHz (a difference of 62.229 Hz)

While a C3 is 130.8128, and C#3 is 138.5913 (a difference of only 7.7785 Hz).

Then using an entire octave as an example, C3 (130.8128) to C4 (261.6256) = 130.8128 (0Hz up to C3)

vs C6 (1046.502) to C7 (2093.005) = 1046.503 (0hz up to c6 +- 0.001hz)

Reply on news.ycombinator.com