2
\$\begingroup\$

A DSP guy like me should know this, and I think that I do, but I can't seem to find a definitive answer in my Carlson Communications Systems textbook.

We know that FM is like phase-modulation (PM) but with an integrator applied to the modulating signal before the PM modulator. And we know that, like AM, PM has a sensitivity to noise and interference that is pretty much flat throughout the channel bandwidth. So, even if FM demodulation is done by other means (like PLL), the implied differentiation to get the baseband signal back will boost noise and interference at higher frequencies more than the lower frequencies. To reduce that, the FM receiver has a de-emphasis filter, a 1st-order LPF, with a corner frequency at 2122 Hz and -6 dB/octave at frequencies higher than that and the broadcast FM transmitter has a complementary pre-emphasis filter. The pre-emphasis is applied to the (mono) audio before the FM modulator and the de-emphasis to the audio coming out of the FM receiver demodulator.

Now, for stereo in broadcast FM, they create out of the left (L) and right (R) channels, a composite L+R that remains at the baseband audio and another L-R that is bumped up to 38 kHz and added to the L+R baseband before going to the FM modulator. The stereo FM receiver undoes all that after the FM demodulation.

It seems to me that the only sensible way to apply the pre-emphasis would be to the composite audio signal (with ultra-sonic channels L-R and whatever else they're putting up there) just before the FM modulator, not to the L and R signals separately. But, although both topics are covered in my old, vintage 1980 Carlson text, it is not clear the order of operation. Can someone authoritatively settle that ambiguity?

If that is the case, that pre-emphasis applied in the stereo FM transmitter will really be boosting the L-R subcarrier (and the RDS and "DirectBand" subcarrier channels) by dozens of dB before going into the FM modulator. Is that correct? That seems a little bit problematic, in terms of the full-scale modulation signal.

\$\endgroup\$
4
  • 1
    \$\begingroup\$ On the base-band signal prior to modulation. \$\endgroup\$
    – Andy aka
    Commented Jun 16, 2023 at 21:35
  • \$\begingroup\$ Okay @Andyaka , I'm getting, what I think is, a different answer from you than from Justme. You're saying that the pre-emphasis is applied to the composite baseband signal before the FM modulator, right? This would be after L-R is bumped up to 38 kHz and added to the L+R at the bottom and to the RDS is DirectBand above 53 kHz, right? \$\endgroup\$ Commented Jun 17, 2023 at 2:40
  • \$\begingroup\$ Try reading his answer again. \$\endgroup\$
    – Andy aka
    Commented Jun 17, 2023 at 9:09
  • \$\begingroup\$ I did, @Andyaka . I read the standard. It's curious because the noise and interference sensitivity that is meant to be compensated with pre/de-emphasis applies to the whole "audio" baseband that contains the L-R all of the other subcarriers. It's not just the upper end of the audio that gets noisy and needs to be de-emphasized. \$\endgroup\$ Commented Jun 17, 2023 at 16:42

2 Answers 2

6
\$\begingroup\$

The recommendation how it should be done is stated in ITU-R BS.450-4, "Transmission standards for FM sound broadcasting at VHF". I know I've seen a better document with block diagrams but it will have to do.

It says that for stereo broadcast you first calculate the signal M which is a mono sum of 0.5*(L+R) signal and then pre-emphasis is applied. This way it will equal and be compatible with the monophonic pre-emphasized broadcast signal.

The difference signal S is calculated as the difference of 0.5*(L-R), and then the signal S is pre-emphasized.

These are then encoded as the RF signal.

However, as these are linear and time-invariable operations, the document has a note, search for "NOTE 4".

Same effect is obtained by pre-emphasizing the left and right signals before encoding, and is sometimes preferred to do this for technical reasons.

So in essence, it's the baseband signals M and S that are pre-emphasized and then modulated to the RF, in addition to the stereo pilot tone, possible RDS and other signals.

\$\endgroup\$
1
  • 1
    \$\begingroup\$ Okay, I'm getting, what I think, is an answer conflicting with the comment from @Andyaka, You're saying it s'posed to be done to the original audio before any of it is bumped up to 38 kHz, right? Not to the composite baseband signal, right? \$\endgroup\$ Commented Jun 17, 2023 at 2:36
0
\$\begingroup\$

Okay, I got it. The answer is spelled out in that standard ITU-R BS.450-4 at section 2.2.2.2:

A signal S is produced equal to one half of the difference between signals A and B mentioned above. This signal, S, is pre-emphasized in the same way as signal M. The pre-emphasized signal, S,is used for the suppressed-carrier amplitude modulation of a sub-carrier at 38 kHz ±4 Hz.

It is done on the original audio not on the composite signal. So it seems to me that the L-R and RDS and whatever subcarriers bumped up there are gonna have a worse S/N, because it's noisier up there above 53 kHz.

\$\endgroup\$
2
  • \$\begingroup\$ If you'd do pre-emphasis on the composite signal [instead], that would essentially apply AM to it, wouldn't it? And that would require more power at the transmitter etc. \$\endgroup\$ Commented Jun 1 at 11:56
  • \$\begingroup\$ Power in an FM transmitter is essentially constant. Only in music FM synthesizers, where you tune the modulator specifically/harmonically, can you get some persistent power boost. \$\endgroup\$ Commented Jun 1 at 12:05

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.