Pink Noise White Noise

[firubbi]

Can anyone explane this pink noise and white noise.
Thanks
**it that true that the human ear responds to sounds in the range of 16Hz to 16kHz.

<font size=-1>[ This Message was edited by: firubbi on 2004-08-31 09:57 ]</font>

[BingoTheClowno]

On 2004-08-31 09:55, firubbi wrote:

**it that true that the human ear responds to sounds in the range of 16Hz to 16kHz.

<font size=-1>[ This Message was edited by: firubbi on 2004-08-31 09:57 ]</font>

It depends from human to human, some can hear up 20 KHz. I think Celmo had a plugin utility that had a frequency generator which you can use to test your speakers. But, be aware that if you cannot hear a high pitched sound it does not necessarily mean that your ears are not able to hear it, but maybe your speakers or headphones are not able to render it.

Regarding pink/white noise you'll have to use Google, there's pages of information on this. Small example:

http://www.firstpr.com.au/dsp/pink-noise/

The characteristics of pink noise
For the purposes of this discussion, "power" means the average power or energy contained in a signal over a long period of time.
White noise has the same distribution of power for all frequencies, so there is the same amount of power between 0 and 500 Hz, 500 and 1,000 Hz or 20,000 and 20,500 Hz.

Pink noise has the same distribution of power for each octave, so the power between 0.5 Hz and 1 Hz is the same as between 5,000 Hz and 10,000 Hz.

Since power is proportional to amplitude squared, the energy per Hz will decline at higher frequencies at the rate of about -3dB per octave. To be absolutely precise, the rolloff should be -10dB/decade, which is about 3.0102999 dB/octave.


The uses of pink noise
The most obvious use of pink noise is as an audio signal, to be used directly, to be filtered or to be used to modulate something.
I am also interested in pink noise way below 1 Hz as a control signal for simulating randomly fluctuating aspects of music. For instance, I might want some aspect of a piece to fluctuate on a minute-by-minute basis, so I need random numbers with energy at 0.01 Hz and below.

My particular interest is being able to sculpt such control signals from a pink noise source purely by the use of filters. The idea would be that I could have a band-pass filter with a certain bandwidth in octaves (or fractions of an octave) and that I could choose to set its frequency as I liked, without affecting the RMS level of its output. Without a pink noise source – for instance by using a white noise source – it becomes very difficult to adjust the piece by changing the frequency of the sculpting filter, because this also affects the resultant signal level,

Ideally, for control purposes (eg. "k" rate noise in Csound) I would like to be able to specify noise with:

1 - A certain lower bounding frequency. Eg. 0.1 Hz. Below that, there would either be little energy, or the energy would remain flat per Hz, rather than rising per Hz to give the 3dB/octave characteristic of pink noise. So it would be "white" below 0.1 Hz.

2 - A certain "RMS level per octave. For instance 5.0 RMS per octave. Therefore a perfect filter which excluded everything but an octave – no matter which octave above the lower bounding frequency – would average an RMS level of 5.0. Noise being noise, it would take a long time to average out the fluctuations to measure this accurately.

3 - It might also be desirable to specify an upper bounding frequency, to reduce computational load where high frequencies were not required.

In the future I intend to write a Csound / Quasimodo unit generator for "a" or "k" rate output, of an "industrial" quality, rather than "analytical" grade. Industrial or technical grade nitric acid specifies it being of a particular minimum and approximate strength. Analytical grade specifies exactly its strength and the tolerance for that specification, as well as noting the maximum permissible levels of the most important contaminants. Such parameters would be "i" rate: set at the start of the ugen's instantiate, not changeable over time. So specifying a lower limit frequency limit of the pink quality of the noise would result in that boundary being set to the nearest octave, not precisely.

It is possible to conceive of a pink noise generator with "k" rate control of level, upper and lower boundary frequencies and the slopes of those boundaries. In this way, precise control over the noise frequency distribution could be achieved without changing the RMS level of the noise. I will leave this idea for now, but it would be mighty handy!

<font size=-1>[ This Message was edited by: BingoTheClowno on 2004-08-31 10:59 ]</font>

[firubbi]

Thanks

[BingoTheClowno]

You're welcome.