Measurements From pink noise to white noise, don't understand !

[Romain74]

Hi everyone !

Today, having a discussion with an engineer, the guy told me that for measurement you have to spread a pink noise with your monitors, and the best curve you should have with your mic would be flat (as white noise).

He told me about energy separation (60% low, 30%mid, 10% high or something like that), and i've to admit that i'm completly lost !

i tought that if you spread white noise the best is to have white noise in return, its seens logic to me > you play a noise, you heard the same noise

I said that the sound in my room is really -boomy-, he said that's normal because i've made test/correction wrong.

So, Have I made a HUGE mistake ? could someone explain me ?

(havent seen post with this subject)

Thanks !! Rom

[Ethan Winer]

White noise has equal energy per Hz bandwidth, and pink noise has equal energy per octave bandwidth. So white noise has the same amount of energy in the range of 100 Hz to 200 Hz as the range 5,100 Hz to 5,200 Hz. Pink noise has the same energy in 100 to 200 as 5,000 to 10,000 Hz. Technically, pink noise is created by running white noise through a low-pass filter having a roll-off of 3 dB per octave.

Okay, that's the theoretical explanation. The practical explanation is a little simpler. Audio response measurements are usually made with a logarithmic scale, where each octave is the same width in the graph. So it makes sense to use pink noise which has the same logarithmic relation.

--Ethan

[Bubbagump]

White noise on a graph compensated for pink noise will look like your high end is way out of proportion with the lows as things will climb at 3bd/oct assuming everything is flat.

[DanDan]

Rom, Pink noise was often used by live sound engineers. In conjunction with a third octave Real Time Analyser and third octave graphic eq's.
We would play the Pink Noise before the SoundCheck, move the Graphic faders up and down (these of course have corresponding third octave level bars in the RTA display) until we got a reasonably flat response.
It was and is useful, and if you want a quick and easy look at your room, then play some pink noise, and view the result in an RTA . Both Noise generation and RTA are often available as DAW plug-ins. Even for live work but certainly in studio and more serious measurement, things have moved on, there are a variety of methods in use these days. Take a look at FuzzMeasure Pro, or REW. They are more complete measurement systems.

DD

[Romain74]

The fact that i dont understand is that pink noise value around 20khz is near from -inf dB.

So how can we get a white noise in return (flat) if there is almost no dB in the noise that we send at the begining ?

[DanDan]

Rom, with respect, it seems like you are asking questions about a whole topic with little understanding of the basics. What do you want to achieve here?
If you are worried about your room, your listening environment, why don't you ask someone you trust to check it out. You would have to learn quite a bit before you could do that yourself.
DD

[andrebrito]

white noise noise is constant in frequency domain

pink noise diminushes 3 db per octave, 10 db per decade