[Elecraft] Noise gen / Spectrogram--who invented?

[Jim Brown]

On Sun, 16 Jan 2005 14:27:51 +0000, Steven Gibbs wrote:

>I recall using Rhode & Schwarz test-sets for filter
>alignment, also Tektronix tracking-generator and
>spectrum analyzer combos.  Further back I recall
>similar devices called wobbulators used for IFs.
>
>- when was the technique of using wideband noise
>   followed by audio or video spectrum analysis first
>   described?
>
>- ditto, for amateur radio applications?

For roughly 35 years, pink noise and real time analyers have been used to 
evaluate the amplitude response of audio systems and and equipment, adjust 
system equaliztion for a desired response. 

In the late 60's, the late Dick Heyser developed and patented a powerful 
analysis technique the he called "Time Delay Spectrometry," whereby a 
generator and narrowband detector were swept at the same rate, but with 
time offset that could be set by the user. This allowed coherent 
measurement of two-port systems -- that is, magnitude and phase -- and it 
also allowed the direct sound of a loudspeaker to be separated from 
reverberation in a room. Heyser demonstrated the technique to the pro audio 
community using a collection of HP and GR gear that filled a van. It took 
about ten years before Bruel and Kjaer built a dedicated product to do it. 
Crown Int'l introduced a dedicated product in 1982. Although it wasn't 
widely known, that box included FFT capability, and many of us, including 
myself, used the FFT capability for various functions. 

About that time, a number of folks began using single channel and dual 
channel FFT analyzers to to analyze and tune sound systems. With the single 
ended technique, the source was usually pink noise and you got only an 
amplitude response. With dual channel analysis, one could compare any 
excitation signal (including program material) with the output. These two 
forms now dominate the landscape in pro audio. A program called Smaart, at 
least some of which was written by a programmer who had been associated 
with the Crown box, was introduced at least ten years ago. It runs on any 
Windoze box, and uses the sound card for I/O. It is QUITE widely used to 
tune live systems. 

Nearly 25 years ago, I used my portable real time analyzer to determine the 
response of systems where I had access to both input and output.  The 
analyer could store and display two measurements, and I  subtracted them 
point by point to get the system response. I did this with the telco links 
between the Chicago FM Club's 146.76 repeater and two remote receivers in 
the late-70's -- the input was my radio listening to a station on the 
input, the output was the same radio listening to the repeater output. We 
then adjusted an equalizer at the repeater site to optimize (or at least 
improve) system response. I also used the technique professionally to 
analyze the response of the duplication systems used for TV spots. 

There are now dozens (hundreds?) of FFT programs written to run on PC's, 
and for use in many disciplines. Spectrogram was written to facilitate the 
analysis of bird calls! A very powerful new one called EASERA, for which I 
have been a beta tester, was written  by some good friends in Berlin. It is 
primarily a tool for acoustic testing, but from a single log sine sweep, it 
can quickly computer and display harmonic distortion vs. frequency, BY 
HARMONIC ORDER! 

Jim Brown  K9YC