Tape Guide (Mar. 1980)

Pink Noise Revisited

Editor's Note: In the October, 1979 issue's. "Tape Guide" column, there appeared an inaccurate answer to two questions on pink noise. The answer which appeared was not from the column's author, Herman Burstein, and this month's more accurate and complete answer is from Contributing Editor Howard Roberson, who first noticed the problem.

Q. What is pink noise? How is 1/3 octave pink noise used in testing equipment?

A. Let's first say that white noise, sometimes called flat noise, is characterized by having constant energy per Hz-bandwidth. In other words, white noise has the same energy in a one-Hz band at 100 Hz as it has in a one-Hz band at 1,000 Hz. Pink noise cannot be called "flat" noise: The energy in a one-Hz bandwidth falls off at the rate of 3 dB every time the frequency is doubled (up one octave). White noise has very much of an s-s-s sound, but pink noise with the high-frequency energy reduced compared to the low is much more sh-sh-sh in nature.

Pink noise is usually generated by filtering the output of a white-noise generator. The output from the filter is broadband, however, usually covering 20 Hz to 20 kHz or more.

For most tests, the pink noise is not restricted to a 1/3-octave bandwidth.

(Reverberation decay measurements are an exception.) For audio evaluation, the pink noise has two major advantages over white noise. First of all, its decreasing energy with higher frequencies is much more like music and better for testing some equipment such as loudspeaker systems, where the extra high-frequency energy in white noise could burn out tweeters.

The second advantage of pink noise is that it can be used with what are known as constant-percentage-bandwidth filters. The most popular of these are the octave and '/3-octave filters and analyzers of various types.

So, the pink noise is actually broadband: It's the analyzer filters which split it up into a series of 1/3-octave bands. If the equipment frequency response is flat, the output in each 1/3 octave band will be the same, and a scope display of all filter outputs would be a straight line. The basis for the results is this: The actual Hz bandwidth of the filters is increasing at the same rate as the noise power in each Hz bandwidth is decreasing. For example, the filter at 500 Hz will have twice the bandwidth as the filter at 250 Hz, but the pink noise power at 500 Hz will be half as much per Hz; and the result is the same noise power in the two bands.

Dubbing Dolbyized Cassettes

Q. I am planning to dub some of my friend's Dolbyized cassettes onto open-reel tapes. Will the recordings that I make from his cassettes be recorded in Dolbyized form or will they be just plain recordings?

-Scott MacGregor, Atlanta, Ga.

A. If your friend's Dolbyized recordings are played back on his cassette deck with the Dolby switch "on," the playback signal will be flat (not Dolbyized). Therefore, your dubbings will also be flat.

Variable Speed Ahead

Q. I would like to be able to adjust the speed of my tape recorder. I have done some experimenting with masking tape on the capstan, but this is a lot of trouble and requires quite a bit of trial and error to find the right variation in speed. How would you suggest that 1 go about achieving the control I desire?

-Terry Black, Springfield, Ill.

A. Your objective is a difficult one for the user to achieve, although it seems not a difficult one for the manufacturer; after all, many turntables have adjustable speed, and so do a few tape decks. The most feasible course for you would appear to be to acquire a power supply with variable frequency and use this to power the motor. I doubt that such power supplies are available at your audio dealers, although you might inquire. Variable-frequency power supplies have been described in the popular electronic literature, and you can research this at your library.

(Adapted from: Audio magazine, Mar. 1980; Herman Burstein )

= = = =