Your browser is out-of-date!

Update your browser to view this website correctly. Update my browser now



AM Preemphasis — Time for a New Curve?

Larry Langford insists that AM can sound better and without additional regulations or specifications

Larry Langford is owner of WGTO(AM) and W244ds in Cassopolis, Mich. He has been in radio since 1965. His commentaries on radio issues such as those facing AM owners are a recurring feature. Read his past articles by searching for “Langford.”

There have been numerous articles about reducing AM bandwidth and the overall AM listener experience. Both sides of the issue make some good arguments. Those opposed to reduction say it ruins what’s left of AM and 5 kHz cutoff filters cause excessive ringing where it can be heard most. Those in favor say almost all AM radios have little response above 4 kHz so why waste the juice transmitting a full 10 kHz signal that is pre-emphasized where few receivers can benefit.

Well, I am taking a different approach to this issue. I say let’s just make what we have work as well as it can. I will be blunt. The NRSC curve leaves a lot to be desired but it was well-intentioned. We all agree that despite the hype decades ago extended range radios were never built. So, the reality is the 10 dB boost at 10 kHz has little or no high-frequency impact on 99% of receivers because radios made from the most common chipset have an intermediate frequency (IF) amplifier bandpass extremely tight above 4 kHz. Efforts to ram more highs to the receiver by jacking up the HF EQ even more just results in added distortion and “highs” that are more ringing and smeared than musical. My experiment provides more highs on the typical radio, reduces dynamic bandwidth with no ringing and much less receiver distortion.

[Read: Let’s Toss Another Expensive and Useless Rule!]

The original Orban AM 9000 Optimod (1977) was a wonderful creation and had a boost of up to 22 dB peaked at 5 kHz then shelved out past that. For radios through the early ’80s it was great because the IF slope in the typical receiver was gentle enough that some more highs could be passed with a big enough push. But if you try and run the old Optimod curve now you will find that it sounds bad on most radios unless you follow it with a 4.5 kHz cutoff, but then there is that ringing problem again. This is again due to the newer much sharper receiver IF cutoff.

So how about a new curve that reflects the real world without heavy distortion, ringing and waste of sideband energy where it cannot be heard?

The advantage to owning the station and being the chief is that you can tinker at will and that’s what I started doing.

Since modern IF sections are so unforgiving of high-frequency shelf boost, let’s see just how we can get more highs without the shelf. The NRSC curve has a boost of 5 dB at 5 kHz so let’s improve on that with a boost of 9 dB at 4 kHz. But instead of the Orban shelving method let’s make it a bell curve boost and allow it to roll off back to 0 dB as it approaches 10 kHz. I did not alter the NRSC cutoff filter so as not to introduce any complications that would muddy the results.

I set the curve up and gave it some critical listening.

The new curve has an additional 4 dB boost over the NRSC curve but only at 4 kHz. The rolloff to zero past 4 kHz cleans up the ringing and smear that the IF produces when hit with excessive boost over 5 kHz. Since the curve rolls gently back to 0 dB additional gain at 10 kHz the processor does not need a brickwall at 5 kHz and that stops the ringing on the transmission side.

So how does it sound?

Amazingly clean and the consonants have the sharpness that was the sound of the first AM Optimod. On music it really brightens things up without the nasty artifacts of before. To make sure I was hearing the action of the curve and not so much the processor, I used a wideband limiter that would preserve the response curve shape while limiting the peak amplitude. I chose a vintage rebuilt CBS AM Volumax 4300 preceded by a UREI graphic equalizer. The Volumax was followed by an NRSC low-pass filter. I have been running it for a while and I am definitely not going back to the NRSC preemphasis curve. WGTO now sounds louder, brighter and much cleaner on a wide variety of car radios I have sampled. The increased sideband power near 4 kHz helps mask noise that the weaker, wider NRSC curve could not.

If you can try it, please do so I am curious as to what you find. When I did some research after making the design I found that noted audio engineer, the late Neville Thiele had written a paper back in the ’80s proposing a bell curve for AM preemphasis rather than the NRSC modified 75 microsecond boost. He argued; “The modified preemphasis standard proposed by the National Radio Systems Committee could prove deleterious to AM broadcasting, both in the quality received through the wanted channel and in the interference produced in the wanted channel by the adjacent channels. An alternative standard is proposed that could produce the improvements sought by the NRSC proposal with little or none of its ill effects…,” If you are a member of the AES the paper can be downloaded free. He makes some very interesting points that I have now verified with my own experiences. Never stop experimenting!

What are your own thoughts? Email with “Letter to the Editor” in the subject field.

[Want more information like this? Subscribe to our newsletter and get it delivered right to your inbox.]