Broadcast compression/limiter settings

Posted: Mon Mar 24, 2008 9:22 pm

Any input on "typical" compressor/limiter settings for broadcast? I tend to use slow attack times, fast release times, and 4:1 to 6:1 ratios for voices. I've been experimenting with fast attacks and slow releases but it kinda weirds me out. It dulls everything and nothing sounds as close as the other way.

Also, I've seen an Aphex Dominator doing some SERIOUS limiting to air. I was shocked at how much was limited, but it sounded okay at the time to me. If I try that much limiting w/ my Dominator everything gets "sticky" sounding, like exceptionally "peaky". Maybe mine has an issue.

Thanx

Posted: Tue Mar 25, 2008 11:06 am

A few questions first.....what is the target here...are you producing music specifically for a broadcast client, or are you making music (CDs, etc.) that will end up in a lot of places, including FM radio broadcast?

If so, are you working with a specific station, or is it going national/syndicated? The genre will, unfortunately, dictate how much and what type of comp/limiting is required. In most cases, however, it's best to leave that up to the station and their processing, no matter how tempting it is to mess with it yourself ahead of time.

There's a myth out there that states it's better to do your processing ahead of time, before the radio station Optimods get a crack at it. According to the folks at Orban (and many others), this isn't necessarily so. Comp/limiting music twice (or more) sounds even worse.

If you're just making CDs for general consumption, make it your best, and hope for the best when it gets played on radio stations.

If you're making specific productions specifically for radio b'cast, you may want to check with the station's engineer and staff to see what they prefer. I routinely provide classical, operatic, jazz and world music shows to a local Public Radio station here in Phila.; www.WRTI.org. (They have one optimod setting for daytime/classical music, and it's programmed to flip automatically at 6 p.m. to a different, faster and harder setting (although it's not very noticable at all, overall) for Jazz in the evening. At 6 a.m., it goes back to the classical setting.

So, I make no major changes in my own work that I provide to them; I simply create the best mixes I can. In the old days of analog tape, Voice overs were somewhere around -10, with music peaks at +2, 3 and 4, letting the station do the limiting and compression.

These days, I create the mix & levels as if it were a CD produced for general public consumption. I do, of course, do some gentle peak limiting and overall level-smoothing, mostly to make it all consistent, but I don't change dynamics within a work or a production. I may remove some long empty/quiet passages (esp between movements) so as not to trigger the station's "Dead-air" alarm, but otherwise, that's about it.

Up until a few months ago, we provided broadcasts on CDs to the radio station for on-air use. They'd simply cue it up same as they would a "regular" CD - which it is, essentially - and play the show as-is, on the air. (Before that, it was a DAT tape, and before that it was a reel to reel...)

Now, we're uploading to their server instead. (saves me a trip to the station with a physical copy. Sweet.)

One caveat if you end up uploading files instead of sending CDs: You may want to convert it ahead of time to 48K sample rate, since that's what most digital broadcast facilities do anyway. Even with CDs - they convert (gearbox) the 44.1k sample rate right out of the players going into the digital console before it hits the air. You may be able to send them raw WAV (or broadcast wav) files this way instead of a "music" CD.

And it might interest you to know that most digital FM stations now convert everything to a high-res MP3 format as it goes down the wire, out of the station to the transmitter- sometimes miles away.

Posted: Tue Mar 25, 2008 8:16 pm

Thanks for the reply. I guess I should have been more clear in describing my situation. This is mostly a "talk show" type of environment. I may have up to 10 or so lavs on a set, some scripted, some not, followed by a live Q&A with phoners. There may be some music for the "open" and "close" and a bit under some voice-over.

This is internal corporate comms, so there are no Optimods (I am unfamiliar with that, honestly) after anything I do. What I do is what gets sent to the satellite and there is no further processing as far as I know. I work very closely w/ the engineers, so I know pretty much what's going on, although I have asked if there is more processing and haven't REALLY gotten a clear answer, but as far as I know, I'm it. So it seems to me that I should get it the way I want it.

So to be more specific about my question;

...regarding talking heads, what do most of you end up using for compressions settings. Like I said in my initial post, I am personally used to slowish attack times, fast release times and 4:1 to 6:1 ratios. I've been playing w/ fast attacks and slow releases and get level results that I'm looking for but not as pleased w/ the sonics of those settings.

Just looking for some ideas and "typical" practices. I know it always depends on the content, but,.......

Posted: Tue Mar 25, 2008 10:23 pm

Tracey, I used to do what you are trying to do for NBC-TV Washington DC for the Today Show, Meet The Press, The McLaughlin shows and others. For nearly 20 years I might add. I guess I added that?

Either way, lavaliers high on the lapel or tie. Remember to put it on the side of the lapel in the direction in which they're speaking most. If they are moving their heads back and forth, put it on their tie. Too often I see them much too low on the torso. It should be nearly under their chin. From the center of your chest (between your nipples) to the lowest part of your neck (but hopefully connects to the rest of your body unless you are a weightlifter or a football player that has no neck)

Don't even think about adding dynamics processing on each and every one of those microphones. Not wise. Won't work. Why won't it work?

I know you want to know this.

I'm just taking my time telling you.

Basically, unless you can electronically link each and every compressor/limiter/gate on each one of those microphones, level detector circuits together, good sounding levels will be nearly impossible to attain. While one person is speaking and gain reduction is occurring you'll have five other compressors balls to the wall open at full gain. Picking up this horrible sounding like hollow mush in an oil drum.

Conversely, you could just insert a compressor on the vocal subgroup bus so that your bumper music and/or commercials are not double compressed.

I just don't recommend any gating. A little downward expansion perhaps but NOT GATING. Utilizing downward expansion can do a lot to clean and tighten up any talking head show. The threshold level must be set at or beneath the point of the posts breathing. The amount of downward expansion should be limited to a realistic voiceover/pulldown. So downward expansion should not really exceed 6 to 15 DB maximum. This allows realistic ambience and no one will ever think anything is being censored. Downward expansion is one of the more difficult settings of a dynamics processing device as it cannot be generic as compressors/limiters can be.

Attack and release times:

This is a big discussion. The human speaking voice is one of the most dynamic instruments to record. Numerous compressors have been developed over the years that provide for pleasant compression of spoken word. DBX units with their RMS detection circuitry seems to do quite nice when in automatic for spoken word. If you want something of a more stylistic sound, I'd recommend a peak detector limiting amplifier such as a Universal Audio 1176. This has no automatic settings. But the benefits of being able to have a fast attack time with a fast release time will provide for a more in your face sound with greater apparent loudness. Apparent loudness is dependent upon release time. The faster the release the louder the signal will appear to be while not really being louder.

Common side effects are dulling of the high frequency response with high levels of limiting. For that, equalization post dynamics is suggested. This will still keep your dynamics in check while being able to bring back elements of the spectral content that have been affected by the gain reduction process.

Bottom line is, you are there to make sure that everything that everybody says is perfectly understood. This is where I feel that bandwidth limiting is a necessity. We do not need 20 to 20kHz response for talking head shows. Not needed. Not necessary. Greater intelligibility will be had by limiting the bandwidth from 80 to 15kHz. And even less depending upon the room. Less as in 100 to 13kHz.

You really won't need to use much if any equalization on each microphone in fact, best if you use none, (except for the bandwidth limiting which can be done on the bus or per channel whichever you desire). When you start adding presence & extra high frequency response you'll then be having problems with sibelance, which ain't nice to listen to. Then you'll have to start fumbling with a De-Esser, that ends up giving everybody a peculiar speech impediment if turned up too high. It's good for joke at least.

What you should do is properly set your gain trims for each of the microphones. You'll then slap a stereo compressor on the stereo output bus of the console. And then mix your ass off. YES YOU HAVE TO RIDE THE LEVELS THROUGHOUT THE SHOW. So look alive and don't let anything distract you. Just listen to the director and make the sound match the mouths.

Do 2 shows and call me in the morning
Ms. Remy Ann David

Posted: Wed Mar 26, 2008 6:56 am

RemyRAD and JoeH

I have been in this business we call audio for 39 years and this is some of the best advice I have seen on the topic of recording voices. You two should get together and write a book. I am sure it would be a million seller.

Have a GREAT DAY!

Posted: Wed Mar 26, 2008 9:52 am

Thanks RemyRAD, great advice.

I have been mixing my ass of, but now I have 16 channels of Dan Dugan Model E, so my ass stays put A LOT more! Those things are like magic boxes.

So, regarding lav placement, that's what I do, high up as close as possible to the chin. Ties are great, women's clothes, not so good. I see other shows on TV and the mic is practically on the talent's stomach and I'm wondering what they are thinking!

1176s on all my mics, that would be heaven but just not gonna happen. I convinced them to buy me an SSL C-100. I won't be buying any outboard comp/lims probably ever. Why should I, I have 64 channels of digital processing, but that's a whole other discussion.

Anyway, back to the Dugans. They are on the insert AFTER the fader AND after the EQ and comp. I know that's a weird way to work 'cause then the fader determines the input to the comp, but it works. Then the Dugans aren't trying to react to the compression or EQ.

Most of the EQing I do is pretty much what you said,....hi-pass at around 90Hz, lo-pass anywhere from 9K to on up with maybe a little boxiness dug out at about 250. So I don't think I'm over doing it.

I definitely like faster release times for the reasons you stated, but I think I get a better, more stable level to air with slow release times. I am the last link to air, there is no other processing after me, it goes straight to fiber or sat uplink. I've been mostly using slow attacks with fast releases OR fast attacks and slow releases. I think I'll try what you said and go for fast attacks and fast release. Don't know why I didn't try that already. I guess I just have to keep futzing with it until I get the happy medium I'm looking for.

Thanks for your input![/quote]

Posted: Wed Mar 26, 2008 11:30 am

Sounds like you're well on your way, Stacey! Congrats on the SSL, too.

With most wireless transciever systems; as long as the input trims (on the beltpack) are set properly, you're already a bit ahead of the game with the built-in companding going on with better wireless systems. I may be mistaken, but I've always felt that some (limiting) work is already being done FOR you this way.

I'm fairly certain there's at least a LITTLE bit of peak limit on most of these things before it goes to the transmitter, in order to keep things from peaking/clipping and overmodulating. (this may be why some artists still prefer hard-wired mics vs wireless. The cheaper wireless don't do it as well or as seamlessly as the truly pro models.) You can of course clip the front end with an overly sensitive mic input, but once levels are set, it's hard to get distortion out of them.

With any professional wireless LAV system, I've had very little used for comp/limiting on the receiver side, once the trims are all set properly. That will probably help keep things even on the LAV side for you.

As for mic placement, you'd think that would be second nature, but not always... I recently saw a CNN report with a well-known female anchor, who had her body turned to the camera just slightly, with her left shoulder more to the camera, and her head was turned that way as well. I have no idea why they then put the lav way down on her RIGHT lapel, making every other word muffled and disappear. (Maybe it was a rookie working that day.... Rolling Eyes

)

This topic sponsored by: