The passion of processing...

[jrobert]

Being an old AM guy from the 60's, and hearing the PD's scream about how our station is too soft in the AGC area, I am surprised, with all the modern processors out there, that AM radio has actually turned worse. With corporate DE's dictating narrower bandwidths and higher limiting and clipping levels, the audio portion of programming is almost un-listenable. With the narrowed bandwidth cutting off even more of the high end and the clipping shaving off every 'S' and soft 'C', it's nothing more than high volume sibilance noise.

I was listening (trying to, anyway) to Monday Night Football from Westwood One on our local, Houston, sports station. 790 KBME. All I heard was noise. With Westwood adding some compression at the ballpark and 790 limiting/clipping the crap out of an already narrow audio stream, I couldn't make out a single word spoken from the announcers.

I dearly miss the good old days of the old tried and true CBS Audimax and Volumax. Set right, they gave us the constant volume and gated AGC that sounded natural. I understand "why" we do things the way we do now. 150% modulation certainly is louder than the 125% required by the FCC. Narrow casting allows this without the adjacent channel splatter. But if your going to do high ambient noise programming, then you have to lower the gain reduction a bit.

Am I too old school?

 

[cgould]

I was listening (trying to, anyway) to Monday Night Football from Westwood One on our local, Houston, sports station. 790 KBME. All I heard was noise. With Westwood adding some compression at the ballpark and 790 limiting/clipping the crap out of an already narrow audio stream, I couldn't make out a single word spoken from the announcers.

I dearly miss the good old days of the old tried and true CBS Audimax and Volumax. Set right, they gave us the constant volume and gated AGC that sounded natural. I understand "why" we do things the way we do now. 150% modulation certainly is louder than the 125% required by the FCC. Narrow casting allows this without the adjacent channel splatter. But if your going to do high ambient noise programming, then you have to lower the gain reduction a bit.

Am I too old school?

I am aware of some operations that do operate this way. Their audio processing is based around the ball games as that's where the bulk of the audience comes. As a result, their normal talk programming is MUCH easier to listen to...as long as you agree with the political slant of the programs!

;-)

I don't listen to AM talk much either as most stations sound like high powered CB transmissions anymore. Heck, the CB transmissions sound better than some of the AM stations I hear out there! Now that I think about it, there's an FM in town (Cleveland) that sounds worse than a CB. They're LOUD, but painful to listen to!

I wouldn't say you are old school -- just being logical. We could use more of that in the biz these days!

-C

 

[oldguy]

I seems that almost everyone posting about this topic (except r. orban) is missing a VERY important point--the quality of the music being inserting into the processors (or, lack of quality). No names will be used (to protect the guilty)--very large corporate entities are limiting, clipping, and squishing their audio products. This includes networks, syndicators, music distributors, and the music industry itself. Compare an old CD (from the late 80's to mid 90's) to one from today. The older ones had dynamic range. Today the CD's are recorded to be as loud as possible. I have examined audio pieces that are coming into our stations. They sound like they were shoved through a pipe, and the waveforms look that way on an analyzer. To make it worse, everyone's ripping audio. Ripping audio is the fastest way, but not the best. Hyperloud audio from CD's tends to get louder, and sometimes clipped, during the ripping process. I have done tests, and I have demonstrated this. I can, and I have, recorded better sounding and more dynamic audio in my home's basement from CD's and, yes(!), good quality cassette tapes. If done properly you will get better results recording a CD in analog into your on air computer (with proper equipment and human attention). You will preserve most of the audio response and preserve the dynamic range of the material. The ripping process, by its very definition, is a distortion of the audio. You can have the best audio processor on the market in your radio station audio chain and not be able to correct the trashed audio you are given. You cannot "redynamic" a hypercompressed song. Once the audio is squished it will stay squished. Or, in other words, once you send your T-Bone steak through the meat grinder it is always hamburger.

 

[cgould]

I seems that almost everyone posting about this topic (except r. orban) is missing a VERY important point--the quality of the music being inserting into the processors (or, lack of quality). No names will be used (to protect the guilty)--very large corporate entities are limiting, clipping, and squishing their audio products. This includes networks, syndicators, music distributors, and the music industry itself. Compare an old CD (from the late 80's to mid 90's) to one from today. The older ones had dynamic range. Today the CD's are recorded to be as loud as possible. I have examined audio pieces that are coming into our stations. They sound like they were shoved through a pipe, and the waveforms look that way on an analyzer. To make it worse, everyone's ripping audio. Ripping audio is the fastest way, but not the best. Hyperloud audio from CD's tends to get louder, and sometimes clipped, during the ripping process. I have done tests, and I have demonstrated this. I can, and I have, recorded better sounding and more dynamic audio in my home's basement from CD's and, yes(!), good quality cassette tapes. If done properly you will get better results recording a CD in analog into your on air computer (with proper equipment and human attention). You will preserve most of the audio response and preserve the dynamic range of the material. The ripping process, by its very definition, is a distortion of the audio. You can have the best audio processor on the market in your radio station audio chain and not be able to correct the trashed audio you are given. You cannot "redynamic" a hypercompressed song. Once the audio is squished it will stay squished. Or, in other words, once you send your T-Bone steak through the meat grinder it is always hamburger.

Yes, that should be noted, I think that this topic is certainly in all of our heads as we post. The source material has been a problem for quite some time, and I know it is a driving force behind David Reaves' work, and the stuff I am working on as well. This is something that's kind of hard for us designers to ignore ;-)

All valid points, "oldguy"



-C

 

[W1DAN]

Hi all:

Interesting and long thread!

Ahh yes, one more opinion coming right up!

Having enjoyed studying and some building of dynamics processing technology for many years, I feel today's digital processors can sound better and louder than analog ones. I am frustrated that I cannot hobby-breadboard one of these things and hope to learn some DSP programming eventually.

That said, as a listener (and broadcaster) today's broadcast audio is too often too loud and undefined. As I tune into a station it may take a few seconds t determine what song is being played as the amplitude clues are now fully removed from the song. I call this the "noise" factor and believe this creates tuneout. When set well (in the Boston area WROR sounds good) and not too smushed, I find the cleanliness and lack of distortion in digital processor a breath of fresh air. Maybe today's boxes are allowed to become "too loud"?

One theory I have is that with analog processors the clipping distortion grew pretty linearly as one pushed the clippers harder. We learned to expect the sound signature. In well designed digital, this clue is gone today. It is good to see manufacturers working hard to help the customer out here.

Now to end the discussion, I'd like someone to donate an Onmia One and an 8100 with XT2. I'll give you my definitive answer as to which one is "best"!

Dan
W1DAN

 

[Sgeirk]

AFAIK, a H U G E broadcast chain used to...or still might real-time record their audio in at 48k...then compress it, of course.

But, I have yet to notice bricked limiting and clipping causing audible effects--- CD rip vs. one played in real time. For real ?!?

 

[RadeoEngineer]

Yes. I just worked a project to combine six stations in two locations into one. There were two automation systems so one had to be chosen. The music from stations "A" was ripped into the system and grunge and distortion in what was previously pristine files immediately appeared. We found out the system in use "processed" the files during ripping and could clearly see clipping on a scope.

 

[Sgeirk]

Okay, but if you're using Adobe or cdex or just a bare bones ripper that does nothing to the ripped files, there shouldn't be any issue...the copied file should be a relatively exact replica of the original file. Still makes no sense to me.

 

[RolfTaylor]

Yes. I just worked a project to combine six stations in two locations into one. There were two automation systems so one had to be chosen. The music from stations "A" was ripped into the system and grunge and distortion in what was previously pristine files immediately appeared. We found out the system in use "processed" the files during ripping and could clearly see clipping on a scope.

There may have been other factors involved as well. For example 16 vs 24 bits. Also, a popular function in rippers is "level normalization". If this was not turned off it could be a factor. Some of the more quality conscious broadcaster have all material dubbed in by a live person so that normalization can be done "by ear" rather than by algorithm.

Finally, if the ripped material was originally encoded with a perceptual codec, and is being ripped to another codec (or possibly even the same codec) you will have the "cascaded codec" problem.

In a merging library scenario like this, ideally you would find an automation system that could handle both types of files and both libraries could be handled without changes. Even then, if the two were normalized to different values and/or of the codecs sound too different this could still be a problem.

All good reasons to stick with Linear, or a gentle algorithm like Enhanced apt-X when creating a library in the first place.

I know of stations where the PD personally oversaw this who process and if a cut sounded bad would locate the original source and have it re-dubbed. Sadly thee are less of these folks every day.

Rolf Taylor
Applications/Support Engineer
Worldcast Systems: APT

 

[Goat Rodeo Cowboy]

To make it worse, everyone's ripping audio. Ripping audio is the fastest way, but not the best. Hyperloud audio from CD's tends to get louder, and sometimes clipped, during the ripping process. I have done tests, and I have demonstrated this. I can, and I have, recorded better sounding and more dynamic audio in my home's basement from CD's and, yes(!), good quality cassette tapes. If done properly you will get better results recording a CD in analog into your on air computer (with proper equipment and human attention). You will preserve most of the audio response and preserve the dynamic range of the material. The ripping process, by its very definition, is a distortion of the audio.

I am interested in some additional input and discussion on this topic. My experience does not exactly parallel the results described here. If I am operating in some fringe area... in danger of falling over the edge, I want to know about it.

OLDGUY: I would be interested in knowing what software(s) you have used for ripping that produced the bad results.

One of the things I do regularly is take CDs recorded during worship services at a church and RIP them, and then edit them. Take out non-pertinent content. Do some reduction of the dynamic range. I haven't had this experience of the ripping process doing harm to the audio material. And I don't want it to show up suddenly with a vengeance!

If anyone has information and advice to offer that doesn't fit within the boundaries of what we are discussing in this thread, use the following address for OFF LINE feedback. <[email protected]>

 

[OttoM]

The ripping process, by its very definition, is a distortion of the audio.

Whoa, nelly! Let's stop this right here. To be accurate replace "ripping" in your assertion with "converting analog audio to digital". That's the first place where distortion enters the chain through dithering. At that point it doesn't matter if the ripper adds more dither - even if the ripper doesn't try to re-process the file.

 

[HowardMBurgers]

Whereas I recognize processing is necessary, I disagree with the 1960's-1980's mentality (primarily by the programmer's), that in the digital era, loudness makes a station somehow more popular or listenable. The fact that radio competes now with streaming, MP3 players, CD's etc., why would radio purposely heavily distort the audio of their station?

A about three years ago, a Program Director in Seattle and I had a bet, that if he allowed me to reduce the processing of his AC station to no more than 10dB of total compression or clipping, (including the sum of all the multi-bands), that the stations TSL would double in three or less books. Other radio stations were laughing behind the scenes. They couldn't believe that the station was so quiet, and compared the sound to a raw CD. (I took that as a compliment actually). Not only did the TSL double, it did so in the first book, then quadrupled over four books. Even with PPM, the station ranks #1 in Women 25-54, and has been top 3 25-54 persons since. Of course as a PD, this person is probably one of the best out there. Since he knows what music to play and how to position the station, the audience isn't being driven away by poor psychoacoustics.

Coincidence? I don't think so.. Here's why.. Heavy processing (clipping) and fast release time creates amplitude phase distortion which can appear to the listener as multipath. I received numerous comments and E-mails that the stations signal seemed to be clearer on their radio than the competition. Also listening to heavily processed audio (minus PD's) is wearing on the average listener. Any more the listener probably has the song on a CD or IPod. When heavy processing trashes the music trust me, they notice the difference.

An example of the wrong way of doing things is DC101 here in the Washington DC area. I enjoy the music, but the agressive processing and lack of dynamic range is completely out of step with todays listener.

 

[rrsounds]

Ripping can only become problematic if there are changes made to the file as it is being ripped. Strictly speaking, by definition ripping can be a bit-for-bit duplication process which is completely lossless and transparent.

It is only when the data is modified that problems arise, and modification is part of the ripping process only if you want it to be.

Keep in mind that some forms of modification which seem benign at first listen will then rear their pimple-faced heads further down the chain. LOL!

Kind regards,
David

 

[RadeoEngineer]

What happened in this case that I refer to, is that this particular system has normalization built into the ripper program, and it can be on in more than one place. Like I said, it was picked as the system to use when combining the stations into one facility and there was some unfamiliarity involved with what it was doing. It took a little time to figure out, but we know that's what happened. I'm obviously choosing my words carefully as I don't want to point fingers.

 

[Research_Weenie]

I'm not an engineer. But I know that songs at my station when ripped were done with a 5 to 1 data compression ratio (so the file size was a fifth of full 16bit 44.1Khz), and that audio compression was added in the process - so when they hit the on-air chain there was an added layer of compression on top of everything else. Possibly to make the levels during segues uniform for automation. It detracts from what would otherwise be cleaner sound.

Having said that - for fun I enjoy trying to match the 'sound' of big radio stations where i live. If the CHR is playing Britney Spears - I'll fire up the same song in iTunes - running it through several ProTools multiband compressor plug ins, including multiband "transient controllers", then out to a Finalizer Express. I'll A/B the on air audio with my own, tweaking to match the high hats, the lows, the mids.. once during a Lady Gaga song, i closed my eyes, switched from one to the other, and couldn't hear any difference. A moment of personal pride. (it didn't last - only worked on that one song..) Matching the classic rock chain, the Hot AC chain, all require retweaking.

I have a lot of respect for the Omnia family of processors I hear all across the dial. A rich sound, balanced, and when set right: not fatiguing. My amateurish attempt to recreate the sound with whatever resources i could muster reflects MY passion (and appreciation) for processing - despite being an outsider.

 

[rrsounds]

I'm not an engineer.
<snip>for fun I enjoy trying to match the 'sound' of big radio stations where i live.<snip>My amateurish attempt to recreate the sound with whatever resources i could muster reflects MY passion (and appreciation) for processing - despite being an outsider.

If you can maintain the enthusiasm while learning by paying close attention to the interaction the processing has with the music, you won't be an amateur for long.
We all started somewhere!

Kind Regards,
David

 

[cgould]

Having said that - for fun I enjoy trying to match the 'sound' of big radio stations where i live. If the CHR is playing Britney Spears - I'll fire up the same song in iTunes - running it through several ProTools multiband compressor plug ins, including multiband "transient controllers", then out to a Finalizer Express. I'll A/B the on air audio with my own, tweaking to match the high hats, the lows, the mids.. once during a Lady Gaga song, i closed my eyes, switched from one to the other, and couldn't hear any difference. A moment of personal pride.

...And that's how it all begins.

I was in a similar situation 25 years ago, except it was my ..umm..."neighborhood low power station" that I was trying to get to sound like the big stations. As that young high school freshmen named Cornelius got older, and tinkered more and more, he learned more and more...one thing led to another, and the rest is history, so to speak!

Keep on plugging on!

-C

 

[HadYourPhil]

I'm with you, Jrobert!

 

[Greg Strickland]

Interesting thread.

 

[dannyscott101]

Having said that - for fun I enjoy trying to match the 'sound' of big radio stations where i live. If the CHR is playing Britney Spears - I'll fire up the same song in iTunes - running it through several ProTools multiband compressor plug ins, including multiband "transient controllers", then out to a Finalizer Express. I'll A/B the on air audio with my own, tweaking to match the high hats, the lows, the mids.. once during a Lady Gaga song, i closed my eyes, switched from one to the other, and couldn't hear any difference. A moment of personal pride.

...And that's how it all begins.

I was in a similar situation 25 years ago, except it was my ..umm..."neighborhood low power station" that I was trying to get to sound like the big stations. As that young high school freshmen named Cornelius got older, and tinkered more and more, he learned more and more...one thing led to another, and the rest is history, so to speak!

Keep on plugging on!

-C

I have worked part-time in commercial radio for the last 16 years. Still, nothing beats the thrill of the "neighborhood low power station"! Mine will be running this Halloween night, chugging out CHR hits for the kiddies, complete with professional sweepers and weather forecasts. Semi-automation is a free download of Winamp, along with a basic single-playlist crossfade program. (Songs are trimmed and adjusted for tight segues) It all goes through an Aphex Compellor, refurbed Prisms, and a Bill & Kim Sacks rebuilt 8100A!

This hobby is already out of hand, so sadly, I can't afford the upgrade to an Omnia 6 or 8500...