Pink Noise and the Behringer deq2496
I have a question based on an observation made while using a deq2496 equalizer to output noise.
For those that aren't familiar with this unit, it has an option to generate 'pink' noise. I'm my case, I'm using that noise to speed up the break-in of some new earphones.
When I select pink noise as a source on the deq2496, and remove all other processing options from the signal path to output (equalization etc), and then activate the RTA, I see frequencies of equal amplitude across the entire RTA spectrum. Even after zooming in on the RTA level (y axis), I see no reduction per octave in amplitude.
Having read the following Wiki article on 'colours of noise', I am lead to believe that white noise has equal power in any band (flat RTA response), while pink noise power decreases by 3dB per octave.
The RTA response I'm seeing on my deq2496 looks more like the white noise depiction on the Wiki page than pink noise. It isn't rolling off at 3dB per octave.
I've obviously miss-understood something, as I don't imagine Behringer could get their noise types mixed up.....
Anyone care to explain?
Colors of noise - Wikipedia, the free encyclopedia
Received a reply from Behringer:
Yes, this is correct - and yes, the DEQ2496 displays this information the way it does deliberately.
Pink noise - Wikipedia, the free encyclopedia
White noise is "equal power per Hz" while pink noise is "equal power per octave". Pink noise is used more often than white noise as a reference / test signal to EQ the response of professional musical applications, such as measuring the response of a PA system, studio monitor system, or acoustic response of a room.
On the DEQ2496 (as well as on many, if not most, other analyzer/EQ systems targeted for concert sound rather than science labs) the graphic readout is compensated so that pink noise looks "flat" on the display (i.e. a "flat response" will look like a horizontal line, rather than a diagonal line tilted toward the right) largely because it's easier to visualize on a small screen than a diagonal line would be.
You could also think of it as "if you measure a flat response at the test microphone input, you're going to see a flat line on the display", which tends to be more intuitive for users than the "diagonal" alternative.
So, if I've understood correctly, the unit is outputting pink noise, but the real time analyzer (RTA) is showing white noise instead because it's easier on the eye (?) Which would mean the dB scale on the y axis is useless as an absolute reference.
This also means you can't compare RTA measurements from the deq2496 to other branded RTAs, unless they share the same 'slope normalizing' algorithm.
I realise most people are going to use the same unit to generate pink noise and measure received audio (from a Behringer microphone), so relative differences on the dB axis should be correct.
I have to say, I'm a little sceptical about the whole RTA normalization reply. How does the unit know it should compensate, so that pink noise looks flat?
I hooked up some XLR cables from the analogue-out to the analogue-in, and set the RTA to measure analogue-in. Flat response (same as the analogue-out). I could understand a flat response from the a-out, but the a-in? The unit doesn't know where the other end of the XLR cable is plugged (so it can't assume noise). If the normalization answer is correct then the unit is appplying scale compensation to everything (not just the noise feature). The above test (to me) indicates that the RTA function is not to be trusted as an absolute measure. In fact, other than relative measurements, it's just a pretty light show.
Jump in if you think I'm wrong....
Bit of a late reply, but..
Pink noise is equal power per octave band, so the 1/3 octave display of the Behringer shows equal power per bar.
white noise would tilt upwards on the display as the number of Hz per bar increased.
20Hz to 40Hz = 1 octave, but =20Hz
2kHz to 4kHz = 1 octave, but =2kHz
so the higher band has more frequencies (Hz) in it, but the same overall power level.
|All times are GMT. The time now is 11:15 AM.|
vBulletin Optimisation provided by vB Optimise (Pro) - vBulletin Mods & Addons Copyright © 2014 DragonByte Technologies Ltd.
Copyright ©1999-2014 diyAudio