An alternative look at broadcast FM radio

Everyone is familiar with broadcast FM radio, right? As a ubiquitous source of entertainment, news, discussion, or just background sound, it is perhaps something we take for granted these days.

Turning music and voice into wireless signals

The radio stations are all corralled into a section of frequency spectrum extending from 87.5MHz to 108MHz in Ireland. Each licenced radio station is allocated a small part of this spectrum, just 200kHz wide.

Frequency modulation (FM)

FM works by using the amplitude of an information signal to vary the frequency of the carrier signal. The carrier signal is what you tune to e.g. 104.4MHz (FM104), 105.2MHz (Phantom FM). The carrier frequency can swing by ±75kHz either side of this centre frequency. A very quiet or silent audio signal results in little or no deviation while a loud audio signal causes the frequency to swing by the maximum ±75kHz allowed.

Receivers reverse this process to recover the information. This information signal comprises the mono and stereo audio. Extra services can also be included in this. Radio Data System (RDS), a low rate data service that can provide traffic alerts, station and song identification, and alternative tuning frequencies is a widely used example of this.

Some examples and colourful plots

To help visualise how radio stations use their allocated chunk of spectrum, we can take a look at how the signal power is distributed using power spectrum plots. In the following, the power of the FM signal is represented by the variation in colour from dark blue (low power) to dark red (high power) over 15-30 minute observation periods.

Music has a lot of rapidly changing audio components and causes the frequency of the carrier signal to swing wildly between ±75kHz about the centre frequency. A single voice results in a slower rate of change. The silences between words mean that no modulation takes place and therefore no swing in the carrier frequency.

The width of the red/orange bands in these plots indicates the deviation and therefore the complexity of the audio content e.g. music and jingles. The narrow red segments in the centre indicate the voice and silence portions of the FM signal e.g. DJ talk and the news.

The above plot is for 2FM during a morning talk show. The programme content was a mixture of voice and music. The songs and jingles are the wider bands of red and orange and the concentrated red bands right in the middle indicate periods where the DJ or other presenter was speaking.

Raidió na Gaeltachta - voice & music: In this case, the programme consisted of more spoken word content than music. During the midpoint of the plot timeline, it looks like the DJ played some music with a lot of complex components e.g. traditional Irish music perhaps.

FM 104: This is a mainly music-based station hence the almost constant use of the entire range of deviation. In a bid to sound as 'loud' as possible, radio stations compress the audio signal so that the normally quieter parts of songs and voice sound as powerful as the normally loud parts. In some cases, this can be pushed to the extreme where the dynamics of the music are heavily restricted.

98FM - music: As a predominately music-oriented station during the period of capture, the spectrum usage is quite similar to FM104. However, take a look at the concentration of power about the centre frequency. This was the news at 11am.

RTE Radio 1 - voice & music: The strong concentration of power close to the centre frequency indicates this was a period of time where the programme content consisted of mostly spoken word.

Phantom FM - music and voice. On 105.2MHz, Phantom FM's programme content comprised music and voice during the observation period. However, based on this plot, they don't appear to push the compression to the limit meaning that the music retains more dynamics than for the other examples above.

Sticking with Phantom, we had a look at spectrum usage over a two hour period. The plot below is an entire afternoon show called 'Phantom Daily' by Phantom FM DJ Simon Maher.

What are those extra signals at the band edge?

If you've read this far, perhaps you may be wondering what those extra signals near the edges of the band for some of the above plots are. We were wondering also so decided to ask RTE. Apparently they are trialling a new low-rate data service that will enable some satellite navigation systems in cars to get traffic alert information that will enable them to update their routes to avoid problem areas. This will eventually be incorporated into the RDS system and will cover the same area as the radio station. For the moment though, it's being tested using a dedicated frequency out of harms way on some radio stations.