I'm currently experimenting with an older FM transmitter that I ran across, and it has, I believe, 50us pre-emphasis. I note that when I listen to this transmitter with a receiver that uses 75us de-emphasis, I get enhanced bass and reduced higher frequencies. From my somewhat limited knowledge on the subject, that appears to be as it should, but I just wanted to verify those facts with people more in the know than I.
50us preemphasis gives less of a high frequency boost than 75us, so you would notice less highs and, therefore, more noticable low end.
But you can make up the difference by making a treble boost in your audio feed to the transmitter until the radio "sounds right."
Pre-emphasis is usually expressed in microseconds which may sound strange but it comes from a technical term known as the time constant. This is equal to R*C where R and C are in this case connected to form a high pass network which rolls up at 6 dB per octave. To calculate the "corner frequency", that is the frequency at which the pre-emphasis begins to accentuate the highs we use the equation f = 1/(2*pi*R*C) or 1/(2*pi*T) where T is the time constant. For 75us the f = 1/(2*pi*75E-6) = 2132 Hz. Similarly, the f for 50 us is 3183 Hz.
At the receiver, a low pass filter should be used with the same time constant. But if a transmitter uses 50 usec it begins boosting at 3183 Hz and a 75 usec receiver will begin to cut at 2122 Hz so in this case the receiver cut is happening too low in frequency and will reduce the audio high frequencies relative to the low frequencies which is what you reported.
This probably over answers your query but for those interested you can tuck this away in your technical notes.
Neil
Neil i am taking a class on broadcast engineering from the cleveland institute of electroincs and would like to understand that eqution a little better could you type it out in long format?
i am looking at it and beleiving it is f = 1/(2*pi* x T) in it's long form where t is 75 or is T 0.75?
i have been scoring an average of 95 (Scored 70 on one and 90-100 on all the others) on the 14 tests i've taken so far. there is 95 ten question exams and lessons in all. it appears i get about 19-20 lessons at a time by mail and complete the exams online.
All right so far, but now I need to know what to do about it.
I use VSTHost for my audio processing, and up to this point, a compressor plug-in. I've found an equalizer plug-in (IIEQ) which uses hardly any CPU and has 13 bands (the exact frequencies for each are adjustable).
My question now is - do I boost the upper frequencies to compensate for this incorrect (at least for the transmitter I'm experimenting with) pre-emphasis? Which frequencies, and by how much?
I did an initial experiment where I boosted the top 4 bands (from about 2Khz) by 12dB (probably overkill) and got a lot more treble in the transmitted signal, but I want the end result to be fairly accurate to the original music.
Below...
I had a very similar situation where I needed to build a pre-emphasis circuit but didn't have technical facts.
Analog television uses frequency modulation (FM) for its audio transmitter and also pre-emphasizes the signal at the transmitter and the TV set has a built in de-emphasis so that the final sound should match the original source.
Using a small transmitter called TV Genie, later found to be illegal in the U.S., I was able to send a video tape from one room to another on Channel 14. The picture quality was razor-sharp-perfect, but the audio did not have a pre-emphasis and sounded very muffled at the receiver.
So I setup an A/B listening arrangement where, when my monitor amp/speaker was set to "A" I could hear the source video by direct wire connection.
By flipping to "B" I could hear the same sound coming through the TV receiver, for instant comparison.
Working by trial and error I tried various combinations of resistor and capacitor for an "R-C" circuit, until B was indistinquishable from A.
I don't know what the pre-emphasis curve is for TV, but one of our posters will know.
Later I will dig out that circuit and add it here.
I think all that is needed is a resistor/capacitor in parallel with each other and placed in series at the input to the transmitter.
Here is a link to an NRSC document showing the pre-emphasis used by (some) AM broadcast stations.
See page 8.
http://www.nrscstandards.org/standards%20accept/standards-download%20NRSC-1-b.asp
So let's see if I understand the table correctly.
The pre-emphasis at 15Khz is 13.656 dB for 50us. If I use 75us de-emphasis, I would have to boost the treble at 15Khz by 17.073-13.656=3.417 dB. And so on.
I could, I know, build something, but I'm not really set up for that right now, and it appears that the equalizer method would allow me to get close enough to the original sound.
Good for you that you are taking class.
The "formal" equation is written fc = 1/(2πRC) and the time constant equation is just the RC part of this which is T = RC. R is in ohms, C is in farads, T is in seconds, and fc is the "corner" or "cutoff" frequency in Hertz. Using the time constant T in seconds this becomes fc = 1/(2πT).
Neil
If I use 75us de-emphasis, I would have to boost the treble at 15Khz by 17.073-13.656=3.417 dB.
Maybe its the way I read your post, but just wanted to be sure that it was not thought that 3.417 dB was the total treble boost when using 75 µs pre-emphasis.
The total boost at 15 kHz for transmit systems using 75 µs PRE-emphasis is 17 dB (close enough).
What I meant was - if I was using a transmitter with 50us pre-emphasis, I would have to boost the treble at 15Khz by 3.4 dB to make it sound like the original when listening to it on a receiver with 75us de-emphasis. Hope that makes sense.
And thanks, Rich.
If I follow this correctly about boosting at 15 kHz it is not this simple. The reason is that the pre/de-emphasis curves follow a specific amount of the rate of boost above the corner frequency. This boost is 6 dB/octave. For a 75us system this boost starts at 2132 Hz. For a 50us system it starts at 3183 Hz. The only way to change the 50us transmitter and keep the 6 dB/octave boost rate to track a 75us receiver is to change the corner frequency to 2132 Hz. Any other means of boosting the treble will not track properly since either the corner frequency or the boost rate will not be correct.
This could be a matter of changing one capacitor and maybe a resistor per channel in the transmitter.
Neil
Thanks again, Rich.
My intent is to approximate the original sound as closely as possible. The software equalizer I'm using has 13 bands. I did start off with the difference in dB, using the center point of each band and the values in the table you provided. I know that it's not 100% accurate, but it's a start. I also noted that the db difference is smaller in the higher frequencies, and increases rapidly the lower you go, so it stands to reason that the most playing around I'll have to do is there.
Any other means of boosting the treble will not track properly since either the corner frequency or the boost rate will not be correct. This could be a matter of changing one capacitor and maybe a resistor per channel in the transmitter.
Neil correctly points out that the audio response curve for 75 µs pre-emphasis needs to follow certain values throughout the audio bandwidth, not just the value at 15 kHz.
The "classic" way to do that would be to modify the pre-emphasis network(s) in the transmitter for a time constant of 75 µs.
But if that isn't desired, then external means used with a 50 µs transmitter might be able to approximate the 75 µs curve for that system. Such a system might be set up fairly accurately by A-B comparison of the audio output of an FM receiver having 75 µs de-emphasis with the audio from the original source. This might even be done "by ear."
As a starting point, the decibels of gain that the external device/process needs to add at the various audio frequencies can be determined from the table posted here earlier. Artisan already gave an example of this difference, for 15 kHz.
It is also good practice NOT to transmit audio frequencies much above 15 kHz, especially by an analog FM stereo transmitter. Two reasons:
1. They can interfere with the 19 kHz pilot carrier of FM stereo.
2. They can over-deviate the transmitter, and increase its occupied r-f bandwidth well beyond the 200 kHz FM channel -- which generates distortion products in the receiver, and adjacent-channel interference.