Old School: multiple path processing-how did you do it?

[PirateJohnny]

I've seen some photos and read descriptions of radio processing set-ups from the 70s and 80s with multiple path processing. It starts with a crossover and the different frequency bands are processed through different equipment and then re-combined. I've seen crossovers for this purpose (not for speakers) and of course EQs and compressors. My question is:

How did the signals get re-combined into one? Was it with a simple resistor bridge or something more complicated?

 

[RadioMan203]

I heard of one station back in the 70's (or 80's) that was using Crown crossovers and custom processing. I don't recall in detail the combining network but it may have been as simple as a resistive network as I do recall the final mix was determined by the output level of the three processors.

 

[Goat Rodeo Cowboy]

I've seen some photos and read descriptions of radio processing set-ups from the 70s and 80s with multiple path processing. It starts with a crossover and the different frequency bands are processed through different equipment and then re-combined.

Does that same description not also fit our modern, state-of-the-art audio processors?

 

[PirateJohnny]

I've seen some photos and read descriptions of radio processing set-ups from the 70s and 80s with multiple path processing. It starts with a crossover and the different frequency bands are processed through different equipment and then re-combined.

Does that same description not also fit our modern, state-of-the-art audio processors?

All the "modern processing" I'm reading about is software. I'm talking about analog processing where the signal is split, by a crossover, into different frequency bands and each band is sent through different processors and then re-combined into the final signal.

I read about WHBQ-AM in it's heyday where all the music went through the program side of the board and the mics were sent through audition for different processing and then re-combined.

Is the re-combing as simple as a resistive network? Anybody have a schematic for that?

 

[Goat Rodeo Cowboy]

I read about WHBQ-AM in it's heyday where all the music went through the program side of the board and the mics were sent through audition for different processing and then re-combined.

Is the re-combing as simple as a resistive network? Anybody have a schematic for that?

Well, let's just probe around and see "how pure your religion is" on this project.

Are you going "all the way" and replicate the WHBQ system and have vacuum tubes and big honking transformers for the goze-intas and the cumz-outaz? Simple resistive mixing pads might work well.

Are you going to design an "illegitimate child" chain where you use solid-state analog amps? With discrete transistors? Or are you going to use analog amps with those little black-box integrated circuit chips? At this point you may have opened Pandora's fabled box because what you choose to hang onto the non-transformer output may affect response and gain and noise. The same would be true going into the first amplifier following the "mixer mechanism". If it also is a solid state amp with no input transformer, it is possible the mixing pad could adversely affect noise and gain and distortion and frequency response.

Just do what engineers and techs used to do back in the day. Hook something together and see if it works. If not, hood something else together and see if it works. Continue until "the roast appears done, and taste's good." I can tell you the name of the young engineer, fresh out of college, who was the very first person to tell me that the time had arrived where we designed things via computer, and build little integrated chips and they worked. No more "skunkworks" where you solder resistors together and hope for the best.

If you will search the Internet for "resistor audio pads" you may find some diagrams that would be a good starting point for the "skunkworks" methodology of cut-and-try, cut-again-and-try etc.

For a guy who did not have "engineer" in his title, I've built my share of little pads. Some worked. Some didn't. Have fun.

 

[jboyd]

In 1960, while in radio school in Memphis, I had an opportunity to visit the WHBQ transmitter site and see their "Echo chamber", a long box with a speaker at one end, and a mike at the other, with an assortment of tube amps, equalizers and patchwork circuit boards. On the air it souned great on the 60's music and promos...but not a beautiful sight to see. Wish I had taken pictures....Memories...JBI

 

[PirateJohnny]

Well, let's just probe around and see "how pure your religion is" on this project.

I don't know if I'll ever actually do this. I was intrigued when I read about it and saw pictures of racks full of equipment with little explanation of how it all came apart and went back together. My thoughts were a simple stereo mixer or resistor pads.

Thanks for all the tips so far, everybody!

 

[PTBoardOp94]

I can tell you the name of the young engineer, fresh out of college, who was the very first person to tell me that the time had arrived where we designed things via computer, and build little integrated chips and they worked. No more "skunkworks" where you solder resistors together and hope for the best.

Well, designing things on a computer without breadboarding every little thing is a good goal. That doesn't mean that the "green" engineer is never testing ideas before drawing them up for production.

 

[Goat Rodeo Cowboy]

Well, designing things on a computer without breadboarding every little thing is a good goal. That doesn't mean that the "green" engineer is never testing ideas before drawing them up for production.

The conversation took place circa 1970. A well seasoned engineer who was head of design for black-and-white TV for a major TV manufacturer was a mentor to me and taught me how to become acquainted with solid state by bread-boarding some transistor circuits. The freshly minted enginer in the first story worked at the same company and was on the leading edge of "we don't need to bread-board anymore".

About ten years ago I was working for a company where ALL the young engineers seemed to come from the "we let the computer design it" school of thinking. Fresh graduates. Everything was digital. They had little understanding or patience with analog issues. I took a transmitter final tube (4-400) and "potted it" in a flower pot with some greenery to decorate my cubicle. All the young engineers who wandered by stopped to ask "What is it?" It's a 4-400. "No, no, no. What is it.. a rectifier or something?"

No, it's a vacuum tube. "Really! I've never seen one. What's it good for?" Thank goodness he had a computer on his desk or he would have been pretty useless to the company. (P. S. I some grizzly old-timers in purchasing and production come walking by, and calling the tube by model number as they kept on walking. Very popular with ham radio operators today.)

 

[michael hagerty]

What the original poster is asking about was the forerunner to the Durrough and Orban processors.

Credit is usually given to KFRC, San Francisco chief engineer Bob Kanner, who split the audio coming out of the console into 10 separate bands (9 frequency bands of music and one mic channel), then compressed and limited each individually before combining it back into one mono signal. It avoided the common "pumping" sound caused by too-aggressive processing to that point.

Kanner was sent around to the other RKO AM Top 40 stations (KHJ, Los Angeles; WHBQ, Memphis; WRKO, Boston) the following year (1976) to install the same system.

By that time, though, Durrough and Orban were on the trail and were releasing all-in-one-box multiband processors.

Not sure about the other RKO stations after 1980, but KFRC continued to tweak and refine its homegrown solution at least until 1985.

 

[Goat Rodeo Cowboy]

I've poked a bit of fun at the OP on this one but yes, back in the day, we stood in awe of what they were doing at RKO and some other innovators. And yes, if you have the time and a bit of spare cash, recreating something along that line would be a fun project.

I've never gotten to tinker with an Orban or one of the other players in TODAY'S market, but I have a feeling that if someone went back and recreated that 9 channel beast we might be disappointed with the results when put in an a - b test with the modern boxes. I hear some stations that to my taste, my ears, are asking Orban and the others to do way too much. Just back off a bit and give it room to breathe now and then. (And room to let ME breathe now and then!)

And you guys who are drooling over the idea of recreating the 1960s.... I am taking the time and a bit of cash and experimenting around with a tube-valve condenser mic. I know how to have a little fun too. Who knows. Maybe we can find something that makes even me sound decent. ;D

(Remember- I'm the guy who said: Why would I pay $450 for a mic when I've got an $89.95 voice? )

 

[PTBoardOp94]

About ten years ago I was working for a company where ALL the young engineers seemed to come from the "we let the computer design it" school of thinking. Fresh graduates. Everything was digital. They had little understanding or patience with analog issues. I took a transmitter final tube (4-400) and "potted it" in a flower pot with some greenery to decorate my cubicle. All the young engineers who wandered by stopped to ask "What is it?" It's a 4-400. "No, no, no. What is it.. a rectifier or something?"

No, it's a vacuum tube. "Really! I've never seen one. What's it good for?" Thank goodness he had a computer on his desk or he would have been pretty useless to the company. (P. S. I some grizzly old-timers in purchasing and production come walking by, and calling the tube by model number as they kept on walking. Very popular with ham radio operators today.)

LOL. I have some vintage textbooks from circa 1960 where the engineers of that era were being taught about the power of transistors. I'm looking at a table of appropriate transistors to use for a certain type of NOR gate (unnamed but uses fewer transistors than CMOS or TTL), including such obsolete parts as 2N167 that commanded the princely sum of $6.47 in 1960! ($49.79 in modern dollars) Its no wonder they wanted to do the logic with minimal transistors!

 

[Goat Rodeo Cowboy]

LOL. I have some vintage textbooks from circa 1960 where the engineers of that era were being taught about the power of transistors. I'm looking at a table of appropriate transistors to use for a certain type of NOR gate (unnamed but uses fewer transistors than CMOS or TTL), including such obsolete parts as 2N167 that commanded the princely sum of $6.47 in 1960! ($49.79 in modern dollars) Its no wonder they wanted to do the logic with minimal transistors!

When you look at the economics of the 1950s and the 1960s and try to translate into today, you realize that this can be a nutty world we live in. My first job in radio paid the princely sum of $40.00 per week and they worked out a side deal that I would get another $10 per week for carrying out the trash and doing some clean-up.

Today, I draw Social Security and the formula for what I draw reaches all the way back there and takes that into consideration. (I'm not complaining... some of the folks today who are so upset and claiming Social Security can't possibly survive just need to chill-out and do what I did about 20 years ago. Build a spread sheet and build in what SSA will do for you, and what your own retirement (401k and whatever else) will be doing. It will scare you, but it IS helpful.)

Sorry for the sidetrack. Anyway, if I had been a good boy and maybe put 5% of that 40 per week into savings, what would that 5% of $40 buy for me at today's prices.

Magnecorders were $600 to $900 in that era. Today I own a little pocket recorder that I bought for $90 in todays money that blows away the old "Maggies". In college I got to tote around an RCA 44-BX to make recordings. I think they sold for $300 or so. In today's world mics can be purchased for $100 - $200 that pretty well match their performance.

It was a violation of patent law, but some engineers used to buy ONE of an item from Gates or RCA and then take the schematics in the manual and run down to the parts place and buy components to build duplicates for maybe 1/4th what the manufacturer was charging.

For the most part... that won't work today! You can't even buy the parts over the counter that are used today. And if you can... the prices are amazing. I recently wanted to buy what we used to call a "key switch". I think the word key came from the phone industry. The kind of lever switch that just above the rotary pots of mixer boards of that era... the switch used to select Program or Audition. If you can find such a switch today.... what is the price?

Back to the thought that got us here: The academics of electronic design have changed. The economics of electronics design (and build) are from another planet.

 

[robgrayson]

I actually worked with the Kanner system at WHBQ in the mid 70's, and somewhere I have a picture of the installation. It took a whole rack of stuff. There was a splitter/combiner, and component compressors in sets of three. And indeed, you ran the mike into audition and the music in program for separate processing. We joked that if you paused while talking over dead air, you would be sucked into the microphone and spit out the antenna.

 

[radiorob2.0]

The WHBQ rack of processing. http://www.w3am.com/audiocha.html

Johnny Randolph reflected on the processing chain on WAKY Louisville. By memory it began at the console that had essentially a bass and treble control on the master output. The bass was rolled off a click and the treble was boosted a click. The program output went through a Marti compressor (optical if I remember) that squashed it 30 db except morning drive where it ran 10db. Then it went though a dynamic enhancer that ran gain reduction/enhancement around 5k. Then it ran through a EMT reverb. From there it was sent to the transmitter down a designated pair, even though they had a 15k line the 8k sounded better. Then at the transmitter site the audio was fed into a final processor of whatever was hip at the time (CBS and beyond). Randolph did say he loaded the meter on the transmitter processor so the engineers believed it wasn't squashing so much.

 

[engdept]

This sounds like a fun project. Anyone who has the time to setup an all analog air chain will find an infinite number of possibilities. I strongly suggest attempting this with a patchbay. When setup correctly you can quickly move your signal processors around in the chain.

Even if you don't use vacuum tubes or transformers you can still get ear pleasing results. Go on ebay, there are many different types of crossovers, compressors, limiters, aural exciters, etc for cheap. A simple multiband compression program bus could go like this:

Stereo output from console --->3-way cross over-->[Low band]--[Med band]--[High band]

[Low band]-->Compressor1-->to sum mix
[Med band]-->Compressor2-->to sum mix
[High band]-->Compressor3-->to sum mix

Now you can buy a small 3 channel stereo mixer, or create a passive summing mixer. Take all the compressor outputs and mix them together. Spend the next 8 hours tweaking compressor settings and making small adjustments to mix levels for desired sound ;D

Or you could keep going some more. Maybe put a limiter after each band compressor then sum it together. Or put another crossover at the beginning and split into a low and high band, send that to two compressors with slow attack and release, sum that together and go into the 3 band crossover. ....or throw in a bass only processor, create a special bass processing chain..or stereo expander but only being signaled by a certain frequency range.....

Look at today's Omnia and Orban, they take these large chains and put them into a box. They are great! But they are limited as the entire operation is basically being done in a "clean room". When you start stacking analog processors together interesting things happen! This is also a great way to learn the fundamentals of audio processing. If you have a good audio workstation, record your results and watch your frequency analysis, stereo phasing, and average db or loudness compared to the original input.

 

[PirateJohnny]

I actually worked with the Kanner system at WHBQ in the mid 70's, and somewhere I have a picture of the installation. It took a whole rack of stuff. There was a splitter/combiner, and component compressors in sets of three. And indeed, you ran the mike into audition and the music in program for separate processing. We joked that if you paused while talking over dead air, you would be sucked into the microphone and spit out the antenna.

Rob, it was your WHBQ pictures and comments I was referring to.

What the original poster is asking about was the forerunner to the Durrough and Orban processors.

Exactly, as described by robgrayson, above.

I've poked a bit of fun at the OP on this one but yes, back in the day, we stood in awe of what they were doing at RKO and some other innovators. And yes, if you have the time and a bit of spare cash, recreating something along that line would be a fun project.

(Remember- I'm the guy who said: Why would I pay $450 for a mic when I've got an $89.95 voice? )

I expected the full spectrum of comments. (See what I did there?) I remember reading that comment, GRC.

For my own personal entertainment I have installed automation and streaming programs on my Mac. (RadiologiK and Nicecast.) Don't even start ragging the Mac. I found a processing plug-in in Nicecast with four separate bands of processing (compression) and a simple 10 band EQ plug-in. After 3 weeks of tweaking I like what I hear. I would like to duplicate it with analog gear when I get my BE 8S250 up and running with my Russco Studio Pro turntables, ITC Delta cart machines and Otari MX-505 reel to reels. I have three Aphex 320A compellors, aDBX 266XS, a DBX 266XL and an AirCorp 500PH to start with.

Man cave.

 

[RadioMan203]

Something else to keep in mind with the audio processing experimentation is what are you going to be playing through it? If you're going to subject these processing creations to modern music it almost seems wasteful as much of the current pop - Top 40 - A/C releases are already so processed that a typical broadcast chain doesn't do much of the heavy lifting it once did.

If you're playing programming with some dynamic range still in it then you'll likely see (and hear) what you can really do with the processing. If you're taking "vintage" audio or some very mildly processed audio like Jazz, Folk and the like you should be able to perform some very interesting experimentation.

Somebody mentioned about people making today's processors perform too much. I think it's pretty evident when you get burned out by listener fatigue that a station's processing is simply too much. A little dynamic range won't kill anybody.

@PirateJohnny: Keep us updated on your trip through the wonders of audio processing. Perhaps once you get something going you can post a mock aircheck of your audio.

 

[PirateJohnny]

I'm mostly playing the "Top 100" from 1954 to 1987, MP3s ripped from CDs I have purchased. I have acquired all the jingles from the Top 40 station I listened to and later worked at. My "station" sounds like the radio I grew up with. My automation program has dayparts, etc. programmed in, so the music mix changes through the day. I have streaming capability so I could post the link here when I'm up.

 

[michael hagerty]

I'm mostly playing the "Top 100" from 1954 to 1987, MP3s ripped from CDs I have purchased. I have acquired all the jingles from the Top 40 station I listened to and later worked at. My "station" sounds like the radio I grew up with. My automation program has dayparts, etc. programmed in, so the music mix changes through the day. I have streaming capability so I could post the link here when I'm up.

I'm not an engineer, but I'd be concerned about 70s-era aggressive processing emphasizing quality deficiencies and digital artifacts in mp3s. Real engineers, am I right?