DSPRelated.com
Blogs

The History of CIC Filters: The Untold Story

Rick LyonsFebruary 20, 20126 comments

If you have ever studied or designed a cascaded integrator-comb (CIC) lowpass filter then surely you've read Eugene Hogenauer's seminal 1981 IEEE paper where he first introduced the CIC filter to the signal processing world [1]. As it turns out, Hogenauer's famous paper was not the first formal document describing and proposing CIC filters. Here's the story.

In the Fall of 1979 Eugene Hogenauer was finalizing his development of the CIC filter, the filter now used in so many multirate signal processing and communications systems. At that time Hogenauer worked for a premier signal processing company called ESL Inc. located in California's Silicon Valley. (ESL was subsequently bought out by TRW Inc., and then later acquired by Northrop Grumman Corp.)

Late in 1979 an engineering colleague of Hogenauer's at ESL, Richard Newbold, was working on a proposal for a spread spectrum system that had a super-high-order traditional FIR decimation filter that was slated for use with a zoom fast Fourier transform (zoom FFT). As Hogenauer was writing his initial IEEE paper, he told Newbold about his new, and unusual, CIC filter. (The filter was unusual because its block diagram, having feedback, was a recursive but its impulse response was finite in duration.) Newbold hoped to use this new-fangled filter in his complex-valued, decimate by 25, spectrum analysis system.

At first Newbold wasn't convinced that CIC decimation filters were practical. (As was well known at the time, implementing a filter containing feedback such that the filter has a z-domain pole located on the unit circle is a very risky proposition!) So, Newbold plotted a decimation CIC filter's frequency response by calculating each and every frequency response point with his trusty HP-35 pocket calculator.* That key-punching exercise took Newbold "all day" [2]. From his data Newbold created a frequency-domain plot that he inserted in his zoom FFT proposal and Hogenauer subsequently used that plot as his Figure 3 in his famous IEEE paper. That frequency magnitude response plot and much of Newbold's notation, Figure 1 below, is now the standard CIC filter magnitude response curve repeated in every description, and data sheet, of CIC filters in the literature of DSP. To this day, Newbold still owns his HP-35 calculator.

Figure 1

As Newbold recalls, "It is interesting to note that the proposal manager, and the systems engineering manager Jon Jenny, did not feel comfortable with the CIC filter because it has a pole at DC and they were convinced it would be unstable. It took an entire day of meetings and a technical demonstration to convince them that it not only worked but it was the only efficient solution available. Once everyone was on board, the filter was used in the proposal that's the way it was back on November 21, 1979."

At that time, according to Newbold, Hogenauer was somewhat unfamiliar with z-transforms and had not yet derived the algebraic expressions for the frequency response of the CIC filter. Newbold gave the transfer function equations he used in his proposal to Hogenauer some of which went into Hogenauer's IEEE paper. When I suggested to Newbold that I thought his name should have been mentioned in Hogenauer's famous paper, Newbold, with humility said, "I did contribute some information to Gene's paper but the situation was analogous to a person who works for Goodyear Tire claiming to have played a big part in winning the Indianapolis 500. That's OK with me. Gene was so far above me technically that I really didn't deserve mention especially since the whole concept was his development."

As it turns out, Newbold's name, and the name of another ESL signal processing expert Harry Glaze, were mentioned in the draft version of Hogenauer's IEEE paper. (When I worked at ESL I never met Hogenauer or Newbold, but I remember Harry Glaze. He was a highly-skilled and friendly DSP expert, always willing to take time out to help young engineers.) For some reason, Newbold's and Glaze's names were removed from the final published IEEE CIC filter paper. This didn't bother Newbold, because he said, "Gene was so far above me technically that I really didn't deserve mention especially since the whole concept was his development. I should not have taken any credit for his revolutionary idea. All I did was prove to myself that I could use it."

If you've reached this point of my blog, then there you have it. In his formal proposal to his Customer, Newbold described the structure, performance, and mathematics of CIC filters 16 months before the publication of Hogenauer's IEEE paper. Newbold gained the distinction of being the very first engineer to use a CIC decimation filter.

In March 1980 Hogenauer submitted his initial CIC filter manuscript to the IEEE for possible publication in their Transactions on Acoustics, Speech, and Signal Processing periodical. At that time it typically took 22 months from a paper's submission to its publication. However, Hogenauer's paper was published just 13 months after its submission.

In February of 2012 I had a question about the algebraic notation used in Hogenauer's paper, so I sent my question to Hogenauer by way of e-mail. In that e-mail, and a follow-on e-mail, I asked Hogenauer just what ideas germinated in his thinking in 1979 that led to the development of his powerful CIC filter. Hogenauer did answer, in some detail, my algebra notation question but chose not share any stories about how he invented the CIC filter. That's too bad. As for Newbold, he recently completed the manuscript for a DSP textbook which is currently under review by Prentice Hall Publishing.

* The Hewlett Packard model HP-35 calculator was the world's first 'pocket calculator', assuming that you had a pocket big enough to hold it. Forty years ago that calculator's model number was chosen by HP co-founder William Hewlett because the HP-35 had 35 keys. This ground-breaking device led the portable computing revolution that literally made the slide rule, used by generations of engineers and scientists, obsolete.

References
[1] Eugene Hogenauer, "An Economical Class of Digital Filters For Decimation and Interpolation," IEEE Trans. Acoust. Speech and Signal Proc., Vol. ASSP 29, April 1981, pp. 155-162.

[2] Richard Newbold, Private E-mail Communication, February, 2012.


[ - ]
Comment by Rick LyonsApril 12, 2015
Hi Steve, (I hope all is well with you.)
Yep, when it comes to DSP, Newbold is no "light weight."
[-Rick-]
[ - ]
Comment by ShepFebruary 20, 2012
An enjoyable and informative story. Thanks!
[ - ]
Comment by cfeltonFebruary 29, 2012
Rick, thanks for the great post. It is always great to hear some of the back story around development, especially around a design that is ubiquitous.
[ - ]
Comment by Steve RawlinsonApril 11, 2015
Hi Rick,

Nice little write up. I worked with Newbold on another project and thought that he was one of the most knowledgeable DSP guys I ever ran across. I still keep in touch with him.

Steve R
[ - ]
Comment by marianuxMay 3, 2024
Very interesting story Rick, thanks for sharing! I also share your curiosity about the fundamental ideas that led to the development of the CIC. Why do you think he didn't want to share that anecdote with you?
Reading your article, I find it surprising how Hogenauer was able to develop his idea without mastering the Z transform, given that stability in these filters is apparently so compromised.
[ - ]
Comment by Rick LyonsMay 6, 2024

Hello marianux.

It was over 20 years ago when I made single phone call to Eugene Hogenaurer. We only talked for a few minutes; maybe he was busy at that time. Anyway, yes, it's too bad he didn't tell me exactly how he arrived at the idea of CIC filters.

In 30 years of working in DSP I have "come up with" what I think are three novel ideas regarding DSP. But I don't exactly recall what caused me to think of the three unique ideas. Who knows how some interesting idea bubbles up into someone's brain. You think and think and think about some DSP topic and all of a sudden "Boom", an idea enters your brain.

So, for me I had a good idea once every 10 years.

To post reply to a comment, click on the 'reply' button attached to each comment. To post a new comment (not a reply to a comment) check out the 'Write a Comment' tab at the top of the comments.

Please login (on the right) if you already have an account on this platform.

Otherwise, please use this form to register (free) an join one of the largest online community for Electrical/Embedded/DSP/FPGA/ML engineers: