Avoiding Clipping

[randythom]

Sorry if this topic has already been addressed, but I thought it might be a good idea to talk about distortion.  There are a few production sound guys and gals out there who don't know enough about avoiding clipping when using microphones.  One of the most common false assumptions is that the mic itself is the first thing in the circuit to clip... wrong.  The mic preamp in the mixer or recorder is almost always where the distortion happens.  And if your knob or fader is downstream from that preamp then turning it down is going to do nothing to fix the distortion.  All you'll be doing is lowering the level of the signal and the distortion in equal amounts.  To avoid distortion in this kind of situation you either need to attenuate the signal before it gets to the preamp (by using a lower output mic, moving the mic farther away, or inserting an inline pad between the mic and preamp) or lower the internal gain of the preamp itself, which some (but not all) mixer "gain knobs" let you do.

Which of the popular location mixers allow you to adjust internal mic preamp gain?  (as opposed to simply attenuating the output of the preamp, or padding the input)

By the way, "distortion amelioration" software is beginning to arrive in the marketplace these days.  It's far from perfect, but it can sometimes take a very badly clipped piece of dialog and improve it enough to make it usable.  It isn't just a low pass filter.  The software apparently uses pretty sophisticated algorithms that allow you to reconstruct the distorted regions to some degree.  As I say, it's far from a panacea, but is a useful last resort.

This is one such piece of software:  http://www.cedar-audio.com/products/cambridge/camdeclip.html

Randy

[Jeff Wexler]

There are a few production sound guys and gals out there who don't know enough about avoiding clipping when using microphones.  One of the most common false assumptions is that the mic itself is the first thing in the circuit to clip... wrong.  The mic preamp in the mixer or recorder is almost always where the distortion happens.

Which of the popular location mixers allow you to adjust internal mic preamp gain?  (as opposed to simply attenuating the output of the preamp, or padding the input)

By the way, "distortion amelioration" software is beginning to arrive in the marketplace these days.  It's far from perfect, but it can sometimes take a very badly clipped piece of dialog and improve it enough to make it usable.

Randy

I learned early on that distortion is one of the nastiest things to have to deal with, and this was before there was any software to help out on this. I also learned  fairly early on to try and understand GAIN and where distortion is happening. Randy is right that it is rarely the fault of the microphone --- most modern microphones that have been in use for as many years as I can remember, are good to well over 130 db SPL (there can, of course, be a defective mic which could distort even at low level).

THE most important knob on the mixing panel is the MIC TRIM pot which must be adjusted properly from the beginning. I am not sure exactly what Randy is referring to when making the assumption that the mic preamp gain is not adjustable on some or many mixers. We are talking about the gain trim, right? This is where the level of output signal from the microphone is adjusted on its way to the mic preamp and almost all mixers have this. It is also so true that turning down the channel fader (gain) will NOT get rid of the distortion once it's there. The channel fader controls the gain to the OUTPUT bus --- so, if the sound is distorted you are just changing the level of a bad signal, you are not getting rid of any distortion.

This is all quite elementary but as Randy points out there are quite a few out there who really don't quite understand these things.

One more thing I will add which is that when dealing with the human voice, there is the phenomenon of what we basically call "vocal tearing" which is sort of real world human distortion caused by the operation of the vocal chords while yelling. There have been times when it is difficult to differentiate this true acoustic distortion from distortion that can be present from improper level/gain adjustments. If I get the chance, I usually try to listen, without headphones or any other device, to what the actor's voice sounds like when yelling. This helps me know whether I am faithfully reproducing this sound.

Regards,  Jeff Wexler

[Philip Perkins]

I'm curious--why did this come up now?  Is this referring to another thread? 

Re distortion fixes, in docs we've used a number of apps for many years now, some of which were originally developed for restoring audio from vinyl records.  There is also lots than can be done by "cut/replace" software like reNOVAtor and ReTouch.  There are a few "clipped peak" restoration plugs around, which mostly work when the plugs sees a "flat-topped" waveform--they are kind of a crapshoot.  The Big Daddy of this work, in my experience, is the Cedar Cambridge stand-alone computer system, which can do several operations (NR, de-clip, de-buzz, etc) at once across a whole 5.1 mix if wanted.  But the cheaper apps are very effective too.  Luck seems to play a large role in whether a specific app "likes" a specific piece of distorted audio--one ends up with several possibilities on hand.

Philip Perkins

[randythom]

As elementary a concept as this is, Jeff, I do encounter working production mixers who don't understand it. Obviously there are lots of people in post who don't understand it either.  And there are some relative newbies who monitor this forum who aren't aware of it.  I bring it up here because even with these new pieces of software in our post toolbox, distortion, as you say, is one of the nastiest problems we have to deal with.

Anybody, no matter how experienced, can and will be surprised by an actor suddenly screaming, but in post it isn't as uncommon as you might think for us to receive ten takes of a line, all equally distorted.  The thing that tells us the production mixer may not understand optimizing gain is that each  take will be recorded at a lower level than the one before it, but all of them are equally distorted.  The mixer is attempting to fix the problem, but doesn't know how.

You're right, of course, that a knob labeled "gain trim" is probably exactly that.  It's the one to grab when you hear what sounds like distortion, not the main fader.

Randy

[randythom]

Hi, Phil!

I started this topic because I thought it would be useful information for a few of the less experienced people who monitor the forum.... and because I aspire to someday be a "hero" level poster on Jeff's group, so I have to keep submitting things!  : )

Randy

[old school]

I love this topic. Gain structure is the big difference between mixing and booming. If I don't set it up right on my end, I could have Don Coufal, Nelson Stoll, Randy Johnson, or me back in the day booming and the whole thing can still go to hell if we distort or clip the recording. Hard to go back when you do this in the digital world. Much of what the Vets here post about is obvious to us, but there are many who read this and are young and learning. The reason I enjoy this forum besides talking with friends, is passing a little info to the next generation. Also Glen, on another thread, brings up a valid point about using the tools in front of you. Thats why we/they call it mixing. BTW, R T is a hero to many.

CrewC

[curleysound]

Gain structure is the big difference between mixing and booming. If I don't set it up right on my end...

CrewC

This is the part that gets me. The getting it set up "Right" part. As far as I can understand, the gain allows voltage from a source (the microphone) to enter the system at a set signal level from 0 to 100 percent of what the source produces. If the signal level sent into the system exceeds its specs, then distortion occurs, so anything less than that would produce a usable signal. When setting up, I usually set the faders at unity or 0 and then adjust the gain knob until I get an acceptable signal level, peaking between -12 and -4 on my meters. Is there any more finesse to it than that, or is this about "right"? I guess when limiters and EQ comes into play, it can get more complicated, but isn't this how it's supposed to be done?

[Jeff Wexler]

I started this topic because I thought it would be useful information for a few of the less experienced people who monitor the forum.... and because I aspire to someday be a "hero" level poster on Jeff's group, so I have to keep submitting things!   : )

Randy

Okay ... You are a HERO to us all (but please, don't let it go to your head and do keep on posting). Remember, it's not the number of posts but the quality ... and you have certainly shown this.

Best regards,  Jeff Wexler

[Eric Toline]

This is the part that gets me. The getting it set up "Right" part. As far as I can understand, the gain allows voltage from a source (the microphone) to enter the system at a set signal level from 0 to 100 percent of what the source produces. If the signal level sent into the system exceeds its specs, then distortion occurs, so anything less than that would produce a usable signal. When setting up, I usually set the faders at unity or 0 and then adjust the gain knob until I get an acceptable signal level, peaking between -12 and -4 on my meters. Is there any more finesse to it than that, or is this about "right"? I guess when limiters and EQ comes into play, it can get more complicated, but isn't this how it's supposed to be done?

The way that I set up gain staging is to keep the fader closed and set the level via the trim knob while listening thru the PFL (pre fade listen) position. After I've got the level I want (say -12dbfs) I then open the channel fader to give me the same level on the output meter. On some mixers like the SD 442 there is a "unity" mark on the fader scale, setting the fader to the "U" and adjusting the trim to your preferred level is best way to avoid any distortion problems.

Eric

[greg sextro]

Here's a followup question for you the board:

What if you are dealing with a particular scene where an actor is delivering lines that goes from really quiet or whispering, to screaming at the top of his or her lungs? 

do you ride the input trim during the take?  bring the trim down before the next take to level it would be set at during the yelling and ride the output fader?  (and therefor, maybe have something closer to a working mix, but a low level iso track till the yelling occurs)?  and/or, set it somewhere in between and hope the input limiter on your mixer can handle most of the level without distorting?

-greg-

[Daniel McIntosh]

I'll bet you are probably also experiencing the newer phenomenon of wireless boom transmitter distortion.  A good unity gain for dialogue on a Lectro or Audio Ltd. transmitter in conjunction with a Schoeps/Sennheiser offers essentially no protection against the shout or door slam. 

An otherwise excellent boom operator should not be expected to ride the trim on the transmitter, and giving the mic extra air for the shout doesn't always work.

[Jeff Wexler]

do you ride the input trim during the take?  bring the trim down before the next take to level it would be set at during the yelling and ride the output fader?  (and therefor, maybe have something closer to a working mix, but a low level iso track till the yelling occurs)?  and/or, set it somewhere in between and hope the input limiter on your mixer can handle most of the level without distorting?

-greg-

All of the above at times. So-called "two handed mixing" where you are riding mic preamp gain trim AND the channel fader is quite common --- and used to be necessary with some mixers that had really poor gain structure and little headroom (room for error). Since most of us now are able to record iso pre-fader signals (which by definition are subject only to the mic preamp gain trim) a common practice in a scene with  huge dynamics is to set the trim low enough that loud stuff does not clip on the iso track. The low stuff will be uncharacteristically low on this track but can be raised later. On the mix track, the track that is subject to both the mic preamp gain AND the channel fader (to the output bus) often you do a lot of wild gain riding to bring up low stuff. This technique avoids two-handed mixing but it may be impossible to get everything into an optimum range without making mic preamp gain trims. Your mention of input limiter is also significant. A limiter, on the INPUT, can help massage the signal into a workable range.

-  Jeff Wexler

[David Waelder]

I'll bet you are probably also experiencing the newer phenomenon of wireless boom transmitter distortion. 

Daniel raises an interesting point.

Most mixers I know are now working with radio booms. The relentlessly fast pace of most TV work more or less demands it. Also, boom operators who experience the freedom of the cable free boom favor it and come to expect it. I know only a very few mixers (Jeff W among them) who are still working with a direct connection to the mixer.

With a cordless rig it is still possible to have an error in the gain structure and bring about overload at the mixer preamp. But the most likely source is at the gain setting on that little belt pack transmitter. And, there is, of course, no possibility of riding that during the take.

I've taken to using proper preamps in my wireless rig to have better control over the issue. Instead of a phantom power box, I use Sound Devices MM-1 and MP-1 preamps. They supply the phantom power but also are the first stage of amplification, feeding a line level signal to the Lectro transmitter via a padded cable. If carefully set up, one can calibrate the system so that the limiter on the MM-1 is activated before the limiter on the Lectro. It doesn't seem possible to align everything so that the limiter on the Lectro never comes into play but one can configure so the MM-1 limiter has the primary action. I find that this permits better control of the signal and I am less often surprised by sudden overloads with this rig. My mobile boom ops do find that the bigger MM-1 (& MP-1) preamps are a more clumsy weight on the belt but they accept it with minimal grumbling.

Even with this rig I can occasionally be surprised. I was playing back some files recently and encountered a take where a word delivered more vigorously than normal came out sounding distorted on the second boom. Since I could monitor levels on playback, I can confirm that it was not overloaded at the recorder and I think it most unlikely that it saturated the preamp in the Cooper. Probably a momentary overload of the transmitter.

David Waelder

[Eric Toline]

Here's a followup question for you the board:

What if you are dealing with a particular scene where an actor is delivering lines that goes from really quiet or whispering, to screaming at the top of his or her lungs? 

do you ride the input trim during the take?  bring the trim down before the next take to level it would be set at during the yelling and ride the output fader?  (and therefor, maybe have something closer to a working mix, but a low level iso track till the yelling occurs)?  and/or, set it somewhere in between and hope the input limiter on your mixer can handle most of the level without distorting?

-greg-

If the scene is being boomed then move the mic further away when you know the loud part is going to happen. This way the background ambience never changes level. If it's a wireless mic then ride the trimmer after you've adjusted the input level on the tx to handle the loud peaks.

Eric

[Philip Perkins]

Much as I love wireless boom, this is one thing I miss about being cabled--it is much faster/easier to ride input trims on the mixer than try to guess what they should be on a remote preamp and TX.  The idea of using an MM1 etc is attractive (more stuff for the boomist to carry though)--I'm going to see about that. 

Philip Perkins

[randythom]

The term "gain" in audio engineering refers to how much a signal is amplified or attenuated in a particular part of a signal path.  So, there is positive gain and negative gain.  Positive gain happens in an amplifier or preamplifier.  A mic produces a relatively tiny voltage.  The main job of the preamp is to boost that voltage.  Most high quality professional mic preamps have a control (the "gain trim") that allows you to adjust how much amplification happens in the preamp.  If your preamp is set to maximum gain (maximum amplification) and it receives a signal from a high output condenser mic three feet from a screaming person, the preamp will produce a signal with a voltage too "loud" for it's own internal circuits to handle, and the signal will be "clipped," which means that the peaks of the waveform will be "sheared off" (clipped) into "square waves"... in other words:   distortion.

So, as Jeff indicates, the main tool you have to avoid preamp distortion, which is by far the most common place for distortion to happen when recording with mics, is the preamp gain trim knob, not the knob or fader which comes after the preamp in the signal path.

Randy

In response[ftp][/ft

As far as I can understand, the gain allows voltage from a source (the microphone) to enter the system at a set signal level from 0 to 100 percent of what the source produces. If the signal level sent into the system exceeds its specs, then distortion occurs, so anything less than that would produce a usable signal.

[nawvernmonkee]

FWIW I am finding your area of work massively more dynamic than live work, obviously you might think for rock n roll, but even more than theatre, classical  or conferences, and as such I am learning to take more care of the source material and have been inclined lately to employ both methods of moving the boom mic and where not possible (low budgets = low ceilings quite a bit, it seems) adjusting the (what i would call) the input gain. Not something I would ever consider in my usual line of work.

It's jolly good fun though !

D.

[larry long]

Great topic Randy.

I've had some output clipping issues occasionally, not really preamp specific just overall gain stages to the recorder especially in noisy enviroments or crazy overlapping frenetic scenes.

Ditto RVD,

I need a solid line from the mic, I only use the plug on when I need to. I also enjoy private talkback from my Boom op.

My problem is while riding the trim I'm changing my iso's and I've had a complaint from post about that in the past.

Along the same lines I'm curious, I'm still using the Cooper 106 till I find an upgrade that I like, with the Deva 4.  Does anyone use the input limiters on the Deva? I sent tones to a channel from the Cooper with the limiters on and got different levels with 100hz ,1k and 10k. So I didn't use them thinking that they were changing the EQ. Now I'm thinking about using them on my ISO's only so I can safely raise my ISO levels.

Great thread,

LL

[Noah Timan]

Daniel raises an interesting point.

Most mixers I know are now working with radio booms. The relentlessly fast pace of most TV work more or less demands it. Also, boom operators who experience the freedom of the cable free boom favor it and come to expect it. I know only a very few mixers (Jeff W among them) who are still working with a direct connection to the mixer.

I'm one of those, too, for all of the reasons discussed in this thread.  It's not that I insist that the boom op be on a cable for every single shot -- sometimes we'll get better sound if the operator has more freedom of movement, and sometimes we'll get better sound if the mixer has a preamp with better headroom (ie one on a good mixing panel as opposed to one in a wireless transmitter).  I find the many varied situations on set tend to define what's the best approach, rather than an arbitary "it's always got to be done this way" procedure.

Having the preamp on my Cooper be the first stage instead of the preamp in a transmitter allows for a lot of minute control and adjustment that can be the difference between hitting the bulls-eye and just landing somewhere on the dartboard.  It can also be the difference as to whether something is going to overload or not.  (I still lug around a bunch of old heavy Audio Limited wireless for this reason because I find their preamps simply have more headroom -- I've found the Lectrosonics ones to be frustratingly shallow comparatively and find myself constantly having to ask my operator to make gain adjustments on the transmitter when we use these).  But the older Audios have their own set of annoyances and problems, so despite all of this wonderful new technology available with transmitters that fit under your fingernail, tune themselves, record themselves and do your laundry for you on Saturday, I find that the most reliable way to get a signal into the mixer is STILL on a piece of copper wire between the mixing panel and the boom pole.

I do work with some operators who insist on being wireless all the time, but not that many, and even then they will generally acquiesce for a scene that's heavy on dynamics, so long as it's practical to cable (ie not a shot where the boom op is running backwards next to dolly track or following a steadicam through a crowd of extras stepping all over the cable). 

.02 nvt

[Noah Timan]

Here's a followup question for you the board:

What if you are dealing with a particular scene where an actor is delivering lines that goes from really quiet or whispering, to screaming at the top of his or her lungs? 

Duck, you sucker!  <g>

Being able to make those adjustments imperceptibly -- and keep the background consistent -- is one of the real challenges of our craft.  Sometimes using an open mic on set can help stabilize the background when you have to make these kind of audible gain/fader adjustments on the primary microphone (so long as the open mic is far enough away from the screamer that it won't itself distort).  Sometimes it's not possible and in order to record both the whisper and the scream properly, it has to be left to post to add room tone to hide the change in level.

(and therefor, maybe have something closer to a working mix, but a low level iso track till the yelling occurs)?

This is one I haven't really found a great solution to yet -- it still seems like sometimes I've got to sacrifice the ideal level of the iso track in order to make the mix work.  Still looking for a method to have my cake and eat it too in that regard, although I'm usually pretty confident that if I nail it in the mix, no one will ever need the iso.

[FSBELLA]

Great topic Randy.

I've had some output clipping issues occasionally, not really preamp specific just overall gain stages to the recorder especially in noisy enviroments or crazy overlapping frenetic scenes.

Ditto RVD,

I need a solid line from the mic, I only use the plug on when I need to. I also enjoy private talkback from my Boom op.

My problem is while riding the trim I'm changing my iso's and I've had a complaint from post about that in the past.

Along the same lines I'm curious, I'm still using the Cooper 106 till I find an upgrade that I like, with the Deva 4.  Does anyone use the input limiters on the Deva? I sent tones to a channel from the Cooper with the limiters on and got different levels with 100hz ,1k and 10k. So I didn't use them thinking that they were changing the EQ. Now I'm thinking about using them on my ISO's only so I can safely raise my ISO levels.

Great thread,

LL

LL,RVD,  or any other mixers who use the wireless boom system.

when you guys do go with the "wireless boom setup" are you going with a Lectro plug on TX ?  or the UM400/denecke 48v ps w/mic level cable?

thanks,

[Izen Ears]

  Love this topic and it's timely since I just vowed to avoid distortion in the isos by having them peak at -20 (-15 at the highest).  So far it's worked, I just had to get used to the isos being quieter, the general level being between -35 and -20!  I ride the trim all the time and that's been working pretty well to avoid peaks in dynamic dialog.

  My problem was I was trying to get the isos to peak up in the -10 to -5 range and unexpected yells would occasionally peak on the isos.  It was almost always because the trim set too high, though the Zax txs max out easily and the limiter is ridiculous in how it changes the sound when it kicks in.  I have the boom person adjust if necessary.  I'd love to use a corded boom for the gain control reason, plus the clarity it gives the boom op compared to anything wireless, but I'm out on location outside a structure they're shooting in, they're adding lights and cables by the second and a boom cable would get mangles and lost in the mess in an instant.  If I were on a stage I'd try it.

  Is it true that 24-bit sound can be boosted without adding noise?  Like even 20db?  I don't do any post and don't really have the tools to test this, but it would ease my mind to know this.  10 years ago I remember 16-bit would get noisy if boosted more that about 10 db.

  Keeping the background consistent is huge to me, but if someone only yells for one line in a 4-pager, that background is going to get quiet for that yell.  I love that idea of putting another mic in the room to ride the background level though...

  Dan Izen

[soundtrane]

when you guys do go with the "wireless boom setup" are you going with a Lectro plug on TX ?  or the UM400/denecke 48v ps w/mic level cable?

I initially used a UM400 -> Denecke ph48 system, now i have a Lectro buttplug.

-vin

[Guest Eric Lamontagne]

For my crew, the boomers always lean towards wireless. I don't think this is only because they don't like wrapping cable. I feel that they can get better sound, sometimes, being wireless. I, being the idealist, lean towards a hardline cable whenever possible. So, it ends up to be a compromise with certain benifits between us. Sometimes cable and sometimes wireless with the deciding factor often our content and situation.

I have moved into the 700 series Micron and love its flexibility and audio quality. Paired with a SD MM1, its an awesome system with two compressor/limiters before transmission. These can be set generally or specifically according to content and drastically reduce 'distortion' possibility.

Still, whispers always sound better on copper.

My Audio Developments 149 has great headroom and so I will often let the track ride out to the iso's while mixing. To the mix track.

Digital recorders have changed the way I do things on set. I will often do a quick playback and isolate a specific track to check it's level, make adjustments and reset before the next take. Often this perfects the mix.

Good luck'

Eric

[Guest Eric Lamontagne]

Correction AD147 is my mixer.