Last time we left off, we had our audio stream ready for signal processing. Processing involves not only regulating the volume and tone quality of the signal; it also prepares the audio for FM broadcast.
A quick explanation of FM is in order. FM stands for Frequency Modulation. The program audio actually instantaneously changes the frequency of the carrier that the transmitter is putting out. KPBX only transmits at exactly 91.1 MHz when there’s no audio. The louder the signal is, the farther we move away from 91.1 MHz. The FCC has strict guidelines as to how far we can deviate from the center frequency. With complex audio, especially music with loud high frequencies, we need to be very careful about the harmonics produced. High energy, high frequency sound can very easily drive our modulation past the legal limits.
Much of the enjoyment of music is due to the difference between the soft parts and the loud parts. This is referred to as dynamic range. At Spokane Public Radio, we strive to provide a good balance between not-too-quiet and not-loud-all-the-time. If we didn’t compress the dynamic range somewhat, it would be nearly impossible to hear softer passages while in your car. This also means we need to put the brakes on the really loud passages in order to stay legal. But that’s only part of the job of processing; it does much more.
Until very recently, the FCC had mandated that enhancements to broadcast technologies had to be backward compatible with existing equipment. The effect of this is that a person listening to a stereo program on a monaural radio must be able to hear the entire program. This is all done by the audio processor.
The processor receives the Left Channel and Right Channel audio signals from the mixing board. After adjusting the volume levels and the tone color, etc, the processor combines the two signals together to produce both a Left+Right signal and a Left–Right signal. The L+R part contains all the audio information and thereby becomes the monaural signal which occupies the center of the spectrum from –15 KHz to +15 kHz around the center frequency. That’s the only part of the signal that a monaural radio needs to receive.
When the station is in stereo mode, the processor produces a very precise 19,000 Hz sine wave (+/– 1 Hz !) This stereo pilot tells your radio that it’s receiving a stereo program and turns on the “Stereo” light. Your radio has a filter to keep this pilot tone out of the audio you hear (15 kHz “brick wall”.) The processor then uses this pilot to produce a second signal at twice that frequency (38 kHz) which is combined with the L-R information. This resultant carrier is parked out at +/–38 kHz from the center frequency.
Out beyond the 38 kHz L–R information, there are SCA (Subsidiary Communications Authority) channels. The Radio Data Service (RDS) at 57 kHz provides station information to radios that are equipped to use it. KPBX uses a 67 kHz SCA to transmit the Evergreen Radio Reading Service for the Blind. It takes a special radio to be able to listen to this service. There’s even a 92 kHz SCA that we can turn on as a backup return telemetry path for the remote control equipment at the transmitter site.
In a previous installment, I promised a discussion of digital audio, so let’s go back to some sound fundamentals When we hear sound, it is because all the objects producing sound are vibrating and producing tiny variations in the air pressure right around them. These pressure variations travel away from the source and combine with all the other sound producers. All of their pressure waves mix, adding here and subtracting there, which results in a single (but very complex) pressure wave that arrives at your eardrum (or a microphone) and causes it to vibrate in accordance.
We talked earlier about protecting this signal from outside interference with shielding and balanced wire techniques. An alternate way of reducing noise is to convert the sound information into code. After the voltage signal is received, it goes through an analog-to-digital (AD) conversion. This consists of taking thousands of measurements each second and assigning a number to the instantaneous voltage of each measurement. The beauty of this is that these numbers are very interference-resistant. They’re all encoded in binary form with each bit being either a “one” or a “zero”. These numbers can then be transported between computers and “reassembled” at the final destination using a digital-to-analog (DA) conversion. As you might surmise, the more measurements (samples) you take per second and the finer degree of voltage variation you can detect, the closer you can get back to the original sound, which is analog. As an example, the standard digital coding used for CDs is 44,100 samples per second. Each sample can have a value of between 0 and 65536 (16 bits).
Okay. So what’s the difference between AM, FM, and HD® radio?
Watch for Part 4.