Clipping on overheads

[reshp1]

Ducking is a GATE, and not a compression. Yes, the triggering is different, but a it is not program dependent, which is to say, there is no ratio of output voltage to trigger voltage. When the gate turns on to duck, the signal goes down by x (level of gain reduction) amount. When a compressor goes over a certain level, the output goes down by x times y, where x=level of gain reduction and y=level of input over threshold. There is NO setting on a compressor which will give you a Ducking effect. If your precious "Amateur Toys" program won't let you do that, then that is simply one more thing to complain to them about. But I would bet they have a way of using their gate for ducking.

What part of the difference are you having a problem with? Or have you just never ducked anything?


Light

"Cowards can never be moral."
M.K. Gandhi

Wow, thanks for turning a friendly debate ugly. I still think you're wrong, but don't take my word for it

http://www.computermusic.co.uk/pdf/tutorials/sidechain.pdf
(3rd page, 1st paragraph "Two of the most useful applications you'll find for compressor sidechains are ducking and de-essing.)



http://www.mikekonopka.com/page26.html
(DUCKING: Ducking can be defined as one signal, or set of signals controlling the limiting/compression of another signal, or set of signals. Suppose you are doing a mixdown in which the producer specifies that the almighty kick drum should take precedence over all other tracks. With ducking, this is easily accomplished. Here's how.....)

http://www.pcmus.com/compressors.htm
(AUDIO "DUCKING"
Ok....This is a cool one for all you Pirate radio owners or Bedroom DJ's.......You know when your listening to the radio, and the music's pumping' away....Then the DJ says something over the music,and as soon as he does, the backing drops in volume a little to make way for the voice......Then when they stop chatting, the music just comes up again to it's normal level.......Well this is known as ducking....and it works like this....

As I mentioned in the Compressor section, the unit listens to the incoming signal through a "Side chain" curcuit......The Side chain gives instructions to the actual compressor by the settings you make in the front panel controls......as soon as the Side chain hears the signal go over the Threshold level, it tells the Amplifier (VCA) at the input to turn down the level....

OK......lets suppose that the music backing is coming through the compressor, but.....the DJ's voice is being fed into the Sidechain........Ah...yes indeed, as soon as the compressor "hears" the DJ's/presenters voice, it turns down the incoming signal .......which is the music backing.......Here's how:

YOUR COMPRESSOR MUST HAVE SIDE-CHAIN SOCKETS ON THE BACK PANEL..... )


Must be a lot of clueless people out there doing a lot of rambling huh?

 

[Light]

YOUR COMPRESSOR MUST HAVE SIDE-CHAIN SOCKETS ON THE BACK PANEL..... )

Sure, but that doesn't mean it is DUCKING. Do you know what the most common use of ducking is? Radio DJs and voice-overs for TV. Now, tell, me, what ratio setting would you use on your compressor to make the music go down by a specific, set amount when the voice comes it? I mean, would you REALLY want the gain reduction of the music to be changing over time while the VO is going? I sure don't. As a matter of a fact, I would be really pissed off by that. And how, exactly, would you have me do that with a compressor? Hmmmm? I've used a lot of compressors. A lot. And I use the side chain all the time. Very useful when you want the kick and the bass to really lock together. And de-essing is damned good when the recording engineer was not as careful as he should have been. But ducking is something else. Ducking is a gating effect.

But prove me wrong. Just tell me what ratio setting will give a ducking effect? There isn't one, but feel free to prove me wrong.



Light

"Cowards can never be moral."
M.K. Gandhi

 

[SouthSIDE Glen]

Radio DJs and voice-overs for TV. Now, tell, me, what ratio setting would you use on your compressor to make the music go down by a specific, set amount when the voice comes it?

That's easy, you set the compressor to almost it's highest compression setting (not quite hard limiting) and then set the threshold to 20dB or so below RMS.

You have your terms a bit backwards; it's not a gate that performs the duck, it's a gate that *triggers* it. Ducking is triggered by a gate but it's not a gate that performs it.

And ducking is best performed by an attenuation pad.

G.

 

[Farview]

That's easy, you set the compressor to almost it's highest compression setting (not quite hard limiting) and then set the threshold to 20dB or so below RMS.

You have your terms a bit backwards; it's not a gate that performs the duck, it's a gate that *triggers* it. Ducking is triggered by a gate but it's not a gate that performs it.

And ducking is best performed by an attenuation pad.

G.

You also set the release relatively long so it doesn't pump. The Drawmer has a ducking feature but most gates do not.

 

[Light]

That's easy, you set the compressor to almost it's highest compression setting (not quite hard limiting) and then set the threshold to 20dB or so below RMS.

You have your terms a bit backwards; it's not a gate that performs the duck, it's a gate that *triggers* it. Ducking is triggered by a gate but it's not a gate that performs it.

And ducking is best performed by an attenuation pad.

G.

16:1 is STILL input dependent. Ducking, such as with the 201, is not.



Light

"Cowards can never be moral."
M.K. Gandhi

 

[Farview]

16:1 is STILL input dependent. Ducking, such as with the 201, is not.



Light

"Cowards can never be moral."
M.K. Gandhi

Having worked at more radio stations that I care to admit to, none of them used the drawmer 201. It was a compressor. The attack and realease was set so it didn't pump. You have to take into account the processing on the voice, the mic is squashed to death to make the DJ sound big and to keep him from over modulating. It is the post-squashed signal that is sent to the ducking compressor, that siganl doesn't really have a lot of dynamics. The release time is set around 1.5 to 2 seconds, which is why it doesn't pump wildly as the DJ speaks.

 

[Farview]

BTW The way it is done in really cheap stations is everything goes through a compressor with a short attack and really long decay. The DJ's mic is 6 to 12db louder than the music is, when he talks the compressor clamps down, bringing the music back behind.

 

[SouthSIDE Glen]

16:1 is STILL input dependent. Ducking, such as with the 201, is not.

*sigh*, it's like talking to a brick wall. And I don't mean a brick wall limiter.

Light, the more posts you make, the more you prove that you have the vocabulary without the knowledge to back it up.

In radio, to duck a music channel to make room for a voiceover, one needs to attenuate the music channel. Now, frankly and for the record, much of the time neither a gate or a compressor are used for this; there is often an attenuator pad (i.e. a simple volume controller) that is not triggered by voice, but rather by the announcer/DJ pushing a button on his console. This momentary button (similar in physical action to a "talkback" button on a studio mixer) activates a (usually, but not always, and sometimes settable) -20dB pad [EDIT: OK, Jason tells us that it might be as small as 6-12dB. I stand corrected there] on the main program channel and de-mutes the announcer's microphone channel. Release the button, and the pad is disengaged and the microphone muted (or released to the fader setting) again.

Now as far as the gate/compressor thing goes, first Farview is absolutely correct, the vox trigger has to be a "gate" with very fast attack and very slow release. But this gate is *not* doing any ducking in and of itself; all it is soing is acting as a vox relay to automatically perform the function of the manual "duck" switch described above. In other words, it is detecting when the guy is talking and when he isn't. When he is talking, an attenuator kicks in on the main progaram channel and the announcer's voice is passed through on the vox channel. When he's not talking, the attenuator is disengaged and the vox channel is gated off. But the gate is acting as a *switch* or *trigger* for the ducking, not as the actual ducking device itself.

Now, as I said, in radio is usually an attenuator that's doing the ducking. But a compressor can be used. You're right, a compressor is input-dependant. But so is an attenuator! Neither one is a leveler; they neither have to be, nor are supposed to be. All you have to do knock down the volume by a relative amount, there is nothing that says the volumes have to be knocked down to identical levels. 16:1 might be slightly weak, 20:1 would be better, but frankly, either one would probably do the trick just fine.

Now, taken away from radio and into a studio, it's more common to use compressors and not so much simple pads, but all the rest of the details stay the same. It's the compressor doing the ducking as triggered by the special purpose "ducking gate".

Finally, and to take it back to where this all started, ducking a clipped signal will do no good whatsoever. The signal is already clipped; all ducking would do is to reduce the amplitude of the clipped signal. It will not restore the original peak.

G.

 

[mixsit]

Nice side show. So CalariasDead. What's the deal? You never said if they actually sound bad, were they useable? Did you have to snip em' all down, re-record?
Come on back.
Later.
Wayne

 

[Farview]

*[EDIT: OK, Jason tells us that it might be as small as 6-12dB. I stand corrected there] .

It really depends on the station. You would be surprised at the dynamic range of a broadcast before it goes through the optimods. The voice over is always louder than the music in the first place, so the voice will win even without the ducking. Commercials are sometimes louder than the music as well, you can't notice it as much now that the volume wars have gotten so out of hand. Everything is so crushed that ducking almost becomes unnecessary.

 

[Light]

edited to keep the post from being so dang long

So what you're saying is that in order to duck, I need at least two compressors (one for the input to the side chain, and one for the actual ducking), right? Or, I could use one gate.

I've done more than a bit of ducking, mostly for TV commercials, and I have ALWAYS used a gate for it. Usually a 201, because it is there, but most good gates can duck. Sure, I could go in and automate it, but why take the time to do all that when I can set up a gate in like 60 seconds, max. O.K., maybe two minutes. I mean, sure, the client might be able to afford it, but I don't want to listen to that crap a second longer than I have to, and the gate sounds at least as good. I actually like the consistency of the envelope a lot more than automating it.

I've never used a compressor for VO ducking, because that is not how it is done, at least not the way I learned. I don't want to have to compress the VO that much just so I can duck it. I suppose I could compress the VO only to the side chain of the other compressor, but why? The gate does a faster job.

Besides, the whole point of the exercise was to show that whoever said that Aardy didn't know how a gate works was talking out of their ass. My point was simply that gates can shut based on a side chain trigger.

Which brings me to the real problem. People have misrepresented what Aardy said (he NEVER said clipping was desirable, only that a few INAUDIBLE transient overs are far less of an issue that recording too low, which is demonstrably true), and have in the process apparently chased away a really fucking good engineer and guy. A guy who has a lot more to teach about engineering than most of the guys on this board. And for what? Because someone doesn't like that they might not be the big fish in the pond anymore? Give me a break. It's not like this is the first time, either. Look at the shit SOUNUSMAN gave to Mixie. He fucking banned from the place, and he STILL wants to be the only one around with any knowledge. People get so convinced that what they have read on the internet is some kind of biblical truth that when a real professional with DECADES more experience comes around with another, probably better, way, they can't fucking listen.

OK, you showed me a way to duck using a compressor. My way is faster and easier. Probably sounds better, but who knows. If I ever do another VO (chances are slim, I hate that shit), and if I have the time, I'll try it your way. But you try it mine. I bet you'll like it a lot.


Light

"Cowards can never be moral."
M.K. Gandhi

 

[Light]

Finally, and to take it back to where this all started, ducking a clipped signal will do no good whatsoever. The signal is already clipped; all ducking would do is to reduce the amplitude of the clipped signal. It will not restore the original peak.

G.

Who said anything about restoring anything. You've got a sound that is fucked. It is part of a track you need to use. Kill the fucked part, and move on with the mix. Sure, it's not ideal, but the whole question was, can anything be done. Sure it can. Would you RATHER not have to do it, sure. But if you are given the tracks as is, and the snare on one of the overheads is cliped, kill the offending sound (and I would probably kill it in the other overhead, just to keep some kind of balance). Yeah, it's gonna sound a little different, but if you can't retrack (and retracking drums is a PITA at best, and impossible if your drummer is not more than a little bit amazing), then maybe that is part of the sound of the mix. Using only one overhead, as someone suggested, MIGHT be possible, if you used a coincident pair for your overheads, but if they are a spaced pair the balence is gonna be pretty funky. So you kill the OH on the snare hits. What's gonna happen? Well, your snare is gonna be the only thing being heard, depending on how you mic'ed the kit. Your gonna loose the cymbals for a second. So what? Sure, it might suck, but it might really make that back beat stick out. On a lot of rock songs, that is far from being a bad thing.

No, it will not restore the original peak, but no one said it would. But hey, if putting words in other peoples mouths is the only way to make yourself feel big...


Light

"Cowards can never be moral."
M.K. Gandhi

 

[mshilarious]

Look at the shit SOUNUSMAN gave to Mixie. He fucking banned from the place, and he STILL wants to be the only one around with any knowledge. People get so convinced that what they have read on the internet is some kind of biblical truth that when a real professional with DECADES more experience comes around with another, probably better, way, they can't fucking listen.

So what are we supposed to do about that? This board has almost no moderation. Sonusman knows that twice a year all his posts will be deleted, but that actually makes it worse, because he will create new usernames, but be disinclined to contribute anything of value.

Mixerman I'm sure either knows the situation or figured it out pretty quick. Mixerman also must have realized that Sonusman was the only one giving him a hard time. If Mixerman is gone that's his prerogative, but there is nothing to be done.

Aardvark showed more attitude than he was given. In my view, he has not successfully defended his premise that overs are preferable to -12dBFS peaks, especially considering the typical conditions and ability of a newbie homereccer. Heck, post a clip of the two approaches on, say, classical piano. That ought to prove it one way or the other.

Anyway, if they have gone, it is no secret where they are: PSW. This board is what it is and so is PSW. There is no reason why they must be the same.

Or maybe they take weekends off . . .

 

[SouthSIDE Glen]

Well, light, I swear that either you must be from another planet whatsoever, or you're using one of those Internet language translators to translate plain English to your native tongue and back.

Nobody said anything about two compressors. What's more, the compression is not applied to the VO, it's applied to the program signal that's being ducked down. And third, gates do not attenuate volume, they set thresholds. They are only the front end of the ducking chain, they do not perform the ducking themselves. I can't say it any plainer than that.

As far as applying the ducking to the original problem, OK, I misunderstood your purpose there, my mistake. I didn't realize you were proposing nuking the offending clipping almost altogether and leaving no sound where there's supposed to be a snare hit. I don't even need to comment on that one.

And, one last time - maybe this will translate to your language properly this time - Aardy went MUCH further than to say that a slight over on a fast transient is inaudible. He did say that, and I agreed to it right off the bat. That is a NON-ISSUE. The issue was with his claiming that pro engineers push their stuff into clipping all the time in order to increase overall RMS volume. This is simply not the case, and is absolutely horseshit advice to try and pass along on a board like this.

If you can't understand this post, then I give up. I'm really only making this last one for the benefit of our readers, whom I sure will understand a whole lot better and hopefully learn that the information coming from that side of the classroom is a whole bunch of twisted nonsense.

This is not one of those "it's a matter of opinion" cases or even one of those "there's more than one way to skin a cat" cases. It is - plain and simple - that you are giving incorrect and misguided information because you are talking way over your head.

Folks, don't just take my word for it; listen to the half-dozen or so others who have come to this thread to try and correct Light's mistakes as well.

G.

 

[reshp1]

Besides, the whole point of the exercise was to show that whoever said that Aardy didn't know how a gate works was talking out of their ass. My point was simply that gates can shut based on a side chain trigger.

Just because a unit with the word "gate" in it's name has some bells and whistles that allow you to duck with it doesn't mean that it's the gate function that is doing the ducking. There are gates that can close based on a sidechain, but that is FAR from a standard feature. Your hero Aardy told someone he KNEW was using Protools to do something that wouldn't work with the protools gate, nor many, if not most, other gates - hardware or software. There's a lot of talking out of asses in this thread but it ain't being done by me.

 

[Cloneboy Studio]

Why do I get the feeling that Light and Aardy are sleeping together?

 

[mixsit]

... As far as applying the ducking to the original problem, OK, I misunderstood your purpose there, my mistake. I didn't realize you were proposing nuking the offending clipping almost altogether and leaving no sound where there's supposed to be a snare hit. I don't even need to comment on that one.

I assumed right off that was exactly what Ardy' was saying and all fairness I figure he most likely very well knows it's a comp-ducking application -and just made a boo boo.
Now it really comes down to the OP to say. If the peaks are brief we may indeed be able to get in and out fast enough to be done with it.

FWIW My Valley 610 has a duck function switch. It turns 'ratio to (depth) 'range in db. And it's on the compressor section, not the gate/expander for crying out loud.

Peace people...

 

[Farview]

So what you're saying is that in order to duck, I need at least two compressors (one for the input to the side chain, and one for the actual ducking), right? Or, I could use one gate.

That was me that said the thing about 2 compressors. I was talking about live on-air ducking. The DJs mic is limited beyond belief, that's how radio stations are set up. The mic triggers the ducking compressor, etc...

I believe your experience, but I have wired several radio stations, I know how this is done. About half do it my way and half use a pad connected to the mic switch.

 

[Reggie]

Geez guys, different ways to do similar things. Get off the semantics already.

Back on the thing with the stuff, as long as you are recording at 24 bit things should sound fine in the end even if your peaks don't hit -20. The only instance I can think of otherwise is if you are a freak purist and will do no further processing to the recorded signal. If it is already hitting at the max or desired level, no further processing will be needed to adjust volume and you won't have to worry about faulty math or added dither or whathaveyou in software.
I personally don't get too worked up anymore about recorded levels. I try to keep things at a reasonable level with enough headroom for most normal circumstances. But the unexpected almost always happens, and I might get the occasional over. Now I can't speak for everybody's setup, but I have no problem with the way my RME convertors clip a quick transient. One time I didn't have a mic pad handy and I had a whole lot of clipped samples on the snare track! Luckily it wasn't the end of the world and didn't even sound bad. But slower things like bass, distorted guitars (especially) ,and perhaps vocals can make for much uglier clipping-the crackling and popping type. THAT I stay far away from.
So back to the original poster - if it sounds OK, it is OK. If you hear the crackling, you are going to have to try the ducking/gating/compressing/whatever, or some serious editting.

 

[Psyko]

Why do you arguing all the time?!

I have the same problem that CalariasDead has.
But my both overheads are clliped. I don't know what to do except of recording just cymbals once again.

Help us and don't dissgusing about duckin cause it didn't made any good!