• Get involved.
    We want your input!
    Apply for Membership and join the conversations about everything related to broadcasting.

    After we receive your registration, a moderator will review it. After your registration is approved, you will be permitted to post.
    If you use a disposable or false email address, your registration will be rejected.

    After your membership is approved, please take a minute to tell us a little bit about yourself.
    https://www.radiodiscussions.com/forums/introduce-yourself.1088/

    Thanks in advance and have fun!
    RadioDiscussions Administrators

Where Does the FM Audio Processor Go -- Before or After the STL?

E

everydayguy

I'm learning about broadcast engineering so please excuse what may sound like a dumb question - here goes.

A friend who owns a small market class A FM has his audio processor at his studio feeding his Marti composite STL transmitter. The hop to the transmitter is about 10 miles.

1. Is there any sound quality advantage to having the audio processor moved to the transmitter site to feed the exciter directly?

2. If there is an advantage, what would have to be done techincally to make this happen -- Installing a different STL transmit/receive? Putting a limiter at the studio to feed the STL, then moving the processor to the transmitter site?

3. Or is it ok just the way it is? The audio sounds pretty good to me, but he says he wants it ..."cleaner and louder".

Thanks in advance for your answers.
 
Analog STLs come in two flavors: discrete and composite. Composite STLs are a widebanded system designed to have the stereo generator (and processing) at the studio. If the path is good, then these generally work well, but the problem with this system is that composite STLs exhibit something caled "bounce". If three successive peaks of 100% modulation are present, one might show up at 97%, one at 100% and one at 104%. These overshoots rob the station of small amounts of modulation that they would otherwise be entitled to. The result is that a station running a composite STL might be slightly less loud on the dial than one running a discrete STL where the processing is at the transmitter site.

The cure for this condition is an extra piece of gear called a composite clipper, or composite processor, placed at the transmitter site. In its simplest form, the composite clipper is a brick-wall that prevents any overshoots from reaching the transmitter, thus restoring the maximum loudness of the station without having to buy new STLs. When used for that purpose, the comosite clipper is highly effective. Two eary clippers were the CRL CC-300 and the Modulation Sciences CP-803. Of the two, the CP-803 was more sophisticated and enjoyed much more commercial sucess. Not long after their introduction, stations discovered that they could use the composite clippers to enhance loudness, too. Their use as loudness tools in the processing wars remains controversial because they add a rather unique form of distortion to the composite signal however, the discussion of that is beyond the scope of your question.

"Cleaner and louder" depends upon what he's using now. What processing is currently employed and is there already a composite clipper at the transmitter site?
 
If it is a composite STL, then it can only transmit composite or mono audio. You would need to change to a discrete analog STL system, or install a digital STL that can handle AES audio to the transmitter site. There are thousands of stations still using FM composite STL's the way you are doing it, and there are thousands of stations using L/R and digital STL's. Some composite STL brands offered a digital converter box that would allow discrete L/R, but at the expense of digital compression. If the composite STL is working correctly, there really should not be any difference with the quality or loudness.
 
Howdy!

There are no dumb questions. Well, okay, there are, but yours isn't one of them :).

everydayguy said:
1. Is there any sound quality advantage to having the audio processor moved to the transmitter site to feed the exciter directly?
Click to expand...

That depends on the performance of the STL. If the STL performs acceptably, it makes no difference. I define "acceptable" as "not audibly degrading performance".


everydayguy said:
2. If there is an advantage, what would have to be done techincally to make this happen -- Installing a different STL transmit/receive? Putting a limiter at the studio to feed the STL, then moving the processor to the transmitter site?
Click to expand...

It's pretty easy to verify performance of an STL. Do you own an oscilloscope? If not, you can pick one up for a hundred bucks on craigslist (not many people seem to know how to use them nowadays, so they sell them cheaply -- lucky for us!).

Hook the oscilloscope up to the composite output of the processor and look. Any FM processor made the last 20 years should have tight peak control, and if you're anywhere near competitive loudness, you'll see the waveform squaring off (clipped) at the edges. Assuming it's a stereo station, the pilot will be riding on top of the clipped audio, so you might see a nice "fat" squarewave any time the kick drum hits.

Then, take a 10 mile drive to the transmitter site, and hook up the oscilloscope there. Do you see the same flat edges and no peak overshoots? If so, the STL is performing acceptably. However, if those edges are tilting towards the center, or if there are peak overshoots, the STL is robbing you of loudness, and needs to be either serviced or replaced. Greg Ogonowski has written some excellent papers about this subject: http://www.indexcom.com/paper.html
They're a few years old, but although processors have gotten better since he wrote them, FM fundamentally hasn't changed, so his words are as valid as ever.


everydayguy said:
3. Or is it ok just the way it is? The audio sounds pretty good to me, but he says he wants it ..."cleaner and louder".
Click to expand...

If the STL causes overshoots, you could get it cleaner and louder by fixing (or replacing) the STL, adjusting for correct deviation (louder), and then backing off the processor (cleaner).
If the STL is performing as it should, the only way to get it cleaner AND louder is to replace the processor with a better one. The two are normally conflicting goals!

As a previous poster mentioned, a composite clipper like CP-803 will make it louder. However, it absolutely does not make it cleaner, so it wouldn't satisfy your friend's wishes. I personally would stay away from band-aids like this and get to the root-cause of the problem instead. :)

Best regards,
///Leif
 
I try to keep as little equipment at the transmitter site as necessary. I find that between lightning, bugs, mice, dirt and air conditioning failures, I have more equipment failures at the transmitter site than I do at the studio. So, as long as the STL sounds good, I say keep the processor at the studio.
 
konbaasiang said:
As a previous poster mentioned, a composite clipper like CP-803 will make it louder. However, it absolutely does not make it cleaner, so it wouldn't satisfy your friend's wishes. I personally would stay away from band-aids like this and get to the root-cause of the problem instead. :)
Click to expand...

Sorry, but I have to disagree with the last part of Leif's comments. Although some are worse than others, all composite STLs exhibit overshoot. It's unavoidable. When a composite clipper is used strictly to trim overshoots, it does not add any measurable distortion to the signal and I would not call that a band-aid solution. Instead, I would suggest that a composite processor is essential for any composite-based STL system that is intereseted in achieving maximum modulation. When used in this manner, even the simplest of composite clippers works well.

I wasn't going to get into this because it's not that relavent to your question, but the distortion issue occurs when the composite clipper is used more agressively as a loudness augmentation tool beyond mere overshoot control. There's a big difference in the effect, but even when used for loudeness augmentation, the effect of the distortion isn't what you might think. Artifacts of composite clipping mainly appear at upper baseband frequencies and affect the L-R more than the L+R. The result is manifest mainly as increased noise (splatter) that degrades the stereo image, but this tends to be somewhat unnoticable because it only occurs during loud passages where the clipper is working hardest. At those moments, the noise is well masked by the main audio anyway, somewhat along the sames concept as companding. So, the decision about how much composite clipping is right is subjective. You have to play with it and see what works. Incidently, there are more sophisticated composite processor designs out there than the CP-803. Some claim to enhance loudness without adding clipping artifacts, so do some homework and decide which is right for your station and budget.

I would avoid using a digital converter for your analog STL if the digital converter compresses the audio using a lossy algorithm. These systems introduce distortion that may be unacceptable, especially if your station is already using a lossy audio format for music storage like mp3, mp2 or wma. There's an old saying "There's no reason the throw the baby out with the bath water". Making wholesale changes to your system may be expensive and, if done incorrectly, can make things worse. You have a working setup, so the best solution may be to tweak it for best results.
 
Kmagrill, I will agree with you as long as the composite processor is used to clip tiny high-frequency overshoots. If the problem is low-frequency related (either tilt or bounce), "fixing" those with a composite clipper is absolutely audible to me, and makes the audio less clean. A CP-803 does not contain any of the advanced distortion masking algorithms of recent FM processors, so having it in-line clipping low-frequency overshoots will absolutely audibly (to me) degrade performance.

///Leif
 
Using a composite link is a classic way of transmitting your signal to the transmitter. Generally, there's nothing wrong with it and many stations are using composite links. However, the link will generate some overshoots - whether they are significant or not, largely depends on the manufacturer of the link, especially the modulator. You can leave it like that or (as has been suggested) you might put a composite clipper at the TX site to "shave off" the overshoots. Composite clipper adds distortion and to avoid degrading the quality of your audio, you want to set it up only to remove the overshoots generated by the link, and not composite clip the bulk of the MPX signal.

But if you were setting up a new radio station and asked what's the best way, I would tell you the ideal connection between the on-air processor and the transmitter is - a short piece of coax cable. So the ideal place for the processor is at the TX site. Now, the best way to get audio from the studio to the processor, is to use stereo uncompressed digital link with 24-bit depth. This configuration gives perfectly accurate peak modulation your processor produces and avoids any degradations of your signal such as overshoots, noise, changes in the sound (all modulators (and so the one in the composite link) affect the sound to a certain degree), etc. The uncompressed (important!) link with 24-bit depth, allows you to go straight from your console into the link without any pre-processing and worrying about clipping the link or the noise floor. You can leave 30 dB of headroom above your nominal operating level (the console will clip before your link does) and even if your jocks play something at -20 dB below the nominal level, you will still get over 90 dB of S/N ratio - more than enough for FM radio.

As there aren't many digital links with 24-bit depth, you second best alternative is to use the more common uncompressed digital links (again, make sure they are uncompressed) but precede it with the audio leveller that will make sure the level going into the link is more/less consistent and that the audio won't clip in the link. This leveller will have to be compatible with your on-air processor and/or achieve the desired effect in terms of sound texture. As the leveller interacts with your processor, it will affect the sound and you want this interaction to be synergistic or provide you with the colouration you prefer. Typically, you would turn off the leveller (AGC) in your processor in this scenario.


Regards,
Goran Tomas
 
It might help to see if you've got some kind of Internet access available at your transmitter site. Most, if not all, current processors can be controlled through a Web interface, making it less of a hassle to drive to your transmitter site if you want to make tweaks.
 
Goran Said it well. You are tied into a Composite STL meaning your Processing must be at the studio end, with the exeption of a composite clipper. A wise engineer taught me many years ago that a composite clipper is an absolute last resort after everything else is done to resolve issues. It can make your station louder, but the temptation to push it will almost certainly result in greater distortion, listener Fatigue, even if they don't know why, and lower books. Stick to your guns, process carefully making small changes and then listen over a range of songs for a reasonable long time before deciding. Get multiple opinions, including women and more from your target audience. THe listening of young folks who have had excessive loud music, older guys who have been around loud noise, i.e. guns, women who have better high frequency sense than men and many others has a big and different effect. As a 67 yr old veteran and ex hard rock jock, I would not use my hearing as the sole judge of our hard rock station's sound. If you can go to a dual system or high quality digital link system, by all means put your processing at the transmitter where artifacts won't have additonal place to enter the system. Also it keeps "magic fingers" DJ's away from your settings.
 
You'll find that there are many opinions about processing, in general, when you talk to various engineers and programmers. All of them are valid and some consideration must be made for the format. A high energy dance format can get away with more aggressive processing than soft rock or light jazz.

As for composite clipping, I'd bet that there are a dozen other processing artifacts that would be worse for listener fatigue than modest composite clipping. You certainly won't get into any trouble by employing composite clipping strictly as a means of overshoot control. That means no more than one to two dB at most. In other words, the little red LED on the CP-803 should not come on.
 
As the original poster for this thread, let me say that you guys are the BEST! Thanks to everyone who responded to my question -- I learned a lot and now have a much better idea about what can and can't be done at the station. You guys ROCK!
 
I don't mean to be too much of a heretic here on the engr topic, but 95%+ of your listeners will never tell a difference in audio quality if it's good now. Loudness--they can tell that. The time will be better spent on programming or promotion. But for the stereo light, probably 90% would not know if you if you flipped over to mono.
 
NE Miss Radio said:
I don't mean to be too much of a heretic here on the engr topic, but 95%+ of your listeners will never tell a difference in audio quality if it's good now. Loudness--they can tell that. The time will be better spent on programming or promotion. But for the stereo light, probably 90% would not know if you if you flipped over to mono.
Click to expand...

Very, very true.
 
I would respectfully disagree.

The listeners don't know anything about the radio, audio processing or much about audio at all. Expecting that a listener will call and say, look your audio doesn't sound good is ridiculous. Yet I can't count how many times people have told me "but nobody called and complained - so it must be good"... ...

But listeners do register - subconsciously. They may not be aware, certainly not able to articulate it and tell you. But if the quality is bad, it will bother them and unless the content is very interesting and compelling, they will eventually switch. Maybe not to the other station (which is probably just as bad, exceptions excluded) but to CD or iPod or something else.

I think it's exactly the "good enough" attitude and "who really hears the difference" attitude, a reason why radio has gone downhill in the last decade. And I don't mean just in technical aspects (from bad MP3s played on air to squashed audio processing), but in terms of content, talent and everything else.

Which is why radio feels threatened by devices such as iPods and other sources of juke-box music. Had radio taken care of it's product (including content AND audio quality) it would have been on a different level. Today, this is true only for some radio stations. The bulk are guided by "people can't tell the difference if it's stereo or mono" (you can replace this with WAV and MP3 or live and automated) and when you establish this as your norm, well... Listeners loose, we loose, everybody looses.

Don't underestimate your listeners!

Rant mode off.


Regards,
Goran Tomas
 
When I was a kid visiting a station with my older brother who was an engineer there, I happened to overhear a terse exchange between the group's VP of engineering, Vern Milton, and the programming consultant named Jerry who was having some success in New York with a new format that eventually became Urban. At one time, this station had been a dominant number one with CHR, but now Jerry was loosing the ratings war and he blamed the loss on the audio processing (A Volumax/Audimax combo) that was about 5 years old at the time. It was true that the station across town sounded brighter and better. Vern listen to Jerry's rant for about 10 minutes before he'd had enough. He spun on his heal and said, "The number one station in our radio group is a 1000W AM station in St. Louis. It wins in ratings and billing. It beats all the other stations, AM and FM, because they play what people want to hear. Period. Don't tell me that you cannot compete in Miami with a 100,000W FM station." Jerry had no answer. A few months later, the station recieved one of the first Optimod 8000s ever made. It was amazing to hear, not only compared to the CBS Labs stuff, but it also beat the competition across town in both loudness and in clarity. Did it help the ratings? Well, no, not really. The other station still dominated. Why did that happen? Well, they played more of what the audience wanted to hear. While the competition's sound wasn't great compared to the Optimod, it was good enough.

So, with respect, I would counter with the suggestion that the main purpose of audio processing is to keep the audio sounding consistent and loud enough to compete. While it's never a bad idea to have the best sound in town, you do not have to be the processing leader to get or retain listeners. As a matter of professional pride, always try to have the best sound possible. After all, the quality of the sound is your engineering product. Processing also creates a unique sound for your station and that's okay. While you can certainly foul up your audio enough to drive listers away, especially if the audio gets distorted, what you play is (usually) a lot more important than how you process it. Try to be roughly as loud as the competition, but don't resort to making the audio sound too busy by overcompressing it by turning up the density to maximum. If you just do that and play what your audience likes, you'll be okay.
 
I agree. Just loud enough so when you channel surf you don't get lost. And enough dynamics left so when you hear you favorite song it makes you want to turn it up loud. Win.
 
In the AM tube transmitter days, we were concerned about IM distortion. We knew way back then that there was a phenomenon known as "listener fatigue". It wasn't something you could put your finger on, but it was something real, causing listeners to move on to another station.

Today's processed music is enough to drive some away. Add to that additional lossy compression and digital artifacts, and you have a recipe for listener fatigue.

So while many, maybe even most, listeners won't notice anything from short-term listening, there will be significant numbers who will subconsciously find somthing else to listen to at some point. Why take that risk?
 
Status
This thread has been closed due to inactivity. You can create a new thread to discuss this topic.


Back
Top Bottom