In case of relatively weak reception, but not so weak that click noise dominates, the noise spectrum at the output of a frequency demodulator increases with 20 dB per decade. Meanwhile, most audio signals have most of their spectral content at low frequencies.
You can therefore gain a lot in signal-to-noise ratio by emphasizing the high audio frequencies in the transmitter (pre-emphasis) and correcting for that in the receiver (de-emphasis). Unfortunately, different parts of the world use different standards for the pre- and de-emphasis: 75 us or 2122 Hz in the USA, 50 us or 3183 Hz in Europe.
While pre-emphasis/de-emphasis works well for FM mono, it is far less effective on the left minus right signal of FM stereo. Hence the huge noise increase when the signal is just strong enough to switch a radio to stereo (assuming an old-fashioned FM stereo radio, nowadays radios usually gradually fade from mono to stereo to make it less objectionable).
You can therefore gain a lot in signal-to-noise ratio by emphasizing the high audio frequencies in the transmitter (pre-emphasis) and correcting for that in the receiver (de-emphasis). Unfortunately, different parts of the world use different standards for the pre- and de-emphasis: 75 us or 2122 Hz in the USA, 50 us or 3183 Hz in Europe.
While pre-emphasis/de-emphasis works well for FM mono, it is far less effective on the left minus right signal of FM stereo. Hence the huge noise increase when the signal is just strong enough to switch a radio to stereo (assuming an old-fashioned FM stereo radio, nowadays radios usually gradually fade from mono to stereo to make it less objectionable).
And this is done (in real life) in a manner that gets correctly compensated in the receiver by a simple RC filter?
Yes.
That is, in real life, they put a multiband compressor/clipper before the transmitter that ensures the sound coming out of the radio has little resemblance to the original, but the correction in the transmitter itself is indeed an accurate complement of the roll-off of the first-order de-emphasis filter in the receiver up to 15 kHz (the normal audio bandwidth of FM broadcasting).
That is, in real life, they put a multiband compressor/clipper before the transmitter that ensures the sound coming out of the radio has little resemblance to the original, but the correction in the transmitter itself is indeed an accurate complement of the roll-off of the first-order de-emphasis filter in the receiver up to 15 kHz (the normal audio bandwidth of FM broadcasting).
That would be interesting to verify on some sunny day - only problem (mission impossible probably) is to find a radio station transmitting good quality noise.
I measured the response of a local broadcast station here all the way from the mixing console to the audio outputs of a Sony ST-SB920 tuner on several occasions. The compressor the station then had left small signals unharmed, so I measured with small sine waves, -40 dB or so, and of course in the middle of the night, assuming there was no one listening anyway. It turned out that a friend of a colleague at the radio station had actually listened to it once.
My wife did present a radio program on one of the local fm transmitters. We streamed the audio at 192kbit through an english server. that stream was picked up by the local studio (30sec delay), processed by the orban and the stereo mux signal was sent to the local mountain transmit station using small transmitter.
The mountain did not have telephone cables or fiber, not even an electricity meter as the electricity company did not want to climb the mountain for meter reading. Radio as a pure art....
The mountain did not have telephone cables or fiber, not even an electricity meter as the electricity company did not want to climb the mountain for meter reading. Radio as a pure art....
Well, since my goal is to use the FM tuner... I could take the Internet Stream via an Android device, hopefully synthesizing a 24/192 signal, send that to the DAC, then drive that to an RF transmitter of sorts -no antenna, just the signal, drive that to the input of the FM tuner and then use the expander.
Voila!
Wasn't that the point of FM Dolby?
I got a Marantz receiver with an FM Dolby deemphasis switch.
We were close to France, well in reach of the AM stations.
The French radio stations in Le Roussillon would broadcast some shows in Catalan and played the La Marseillaise in lieu of La Santa Espina. Everybody understand that, but La Guardia Civil couldn't just cross the border.
Somethings Franco just did not allow.
I hear the difference by tuning to different stations... some sound quite realistic, some sound "flat". So, I honestly think the tuner can sound quite good. I do plan on having a guy go through it and make sure it's all up to spec.... although the guy I bought it from did a pretty good job on maintaining it.
I have a KT-8300 which is the same design as the 9900. I'll knock it because ...
I found mine to not be anywhere selective enough to tune into my favourite FM station which is sandwiched between two more powerful adjacent carriers.
KT-8300 sits on the shelf, needing a fine filter upgrade. I also found it to overload, have images. It has no RF FE AGC. MOSFET-G2 hard-wired for max gain and there are two gain stages. KT-917 is a better design as it uses a balanced mixer. I do not know why Kenwood EE's did not use AGC whereas the Yamaha EE's managed to get it right.
I am using some Yamaha T-85 tuners which perform much better for my locale. I have a large FM yagi on the roof with a rotor.
The KT-8300/9900 tuner has a deviation meter on it, very few do, scan your dial and see where they are. Some actually over modulate.
The stations that sound a bit lower in volume are probably not as compressed or over modulated.
I offered my services a couple of times to repair our local LPFM stations transmitter (CFGI) when it was down. Its a Nuatel VS1.
These devices are quite sophisticated, they have built-in spectrum analyzers, to monitor modulation, power SWR ect and a web page for remote control. Lots of ways to configure them.
I know from my experience the CFGI radio station has very little knowledge on VS1 configuration. It was configured once by a tech/EE at first install and never touched again.
Crap when I designed a Si4735 radio and wrote the PI decoding code, I found out that many well known stations in our area, owned by Corus, etc did not even have they RBDS setup correctly. One sation was putting out "FFFF" pi code which is reserved for Mexico.
So based on my experience the stations are all over the place in regard to proper configuration. Nautel have excellent product support, they walked me through a complete upgade to the VS1 software and a hardware upgarde as well.
Long live FM radio, for me, streaming sucks the big one. But CFGI does stream, it does not sound anywhere as good as the over the air transmission.
I found mine to not be anywhere selective enough to tune into my favourite FM station which is sandwiched between two more powerful adjacent carriers.
KT-8300 sits on the shelf, needing a fine filter upgrade. I also found it to overload, have images. It has no RF FE AGC. MOSFET-G2 hard-wired for max gain and there are two gain stages. KT-917 is a better design as it uses a balanced mixer. I do not know why Kenwood EE's did not use AGC whereas the Yamaha EE's managed to get it right.
I am using some Yamaha T-85 tuners which perform much better for my locale. I have a large FM yagi on the roof with a rotor.
The KT-8300/9900 tuner has a deviation meter on it, very few do, scan your dial and see where they are. Some actually over modulate.
The stations that sound a bit lower in volume are probably not as compressed or over modulated.
I offered my services a couple of times to repair our local LPFM stations transmitter (CFGI) when it was down. Its a Nuatel VS1.
These devices are quite sophisticated, they have built-in spectrum analyzers, to monitor modulation, power SWR ect and a web page for remote control. Lots of ways to configure them.
I know from my experience the CFGI radio station has very little knowledge on VS1 configuration. It was configured once by a tech/EE at first install and never touched again.
Crap when I designed a Si4735 radio and wrote the PI decoding code, I found out that many well known stations in our area, owned by Corus, etc did not even have they RBDS setup correctly. One sation was putting out "FFFF" pi code which is reserved for Mexico.
So based on my experience the stations are all over the place in regard to proper configuration. Nautel have excellent product support, they walked me through a complete upgade to the VS1 software and a hardware upgarde as well.Long live FM radio, for me, streaming sucks the big one. But CFGI does stream, it does not sound anywhere as good as the over the air transmission.
Last edited:
No. it's not a matter of signal strength. They're all pretty much up fully deflect the signal strength meter.
I guess I ought to look at other meter. It does multipath or deviation.
Comparing two stations, a good sounding one and flat sounding one... both show the same multipath, but the good sounding one shows more "dynamic range" for deviation.
I'm using a powered indoor antenna.
I guess I ought to look at other meter. It does multipath or deviation.
Comparing two stations, a good sounding one and flat sounding one... both show the same multipath, but the good sounding one shows more "dynamic range" for deviation.
I'm using a powered indoor antenna.