General advice on monitoring sound level - Page 2 at DVinfo.net
DV Info Net

Go Back   DV Info Net > The Tools of DV and HD Production > All Things Audio

All Things Audio
Everything Audio, from acquisition to postproduction.

Reply
 
Thread Tools Search this Thread
Old May 3rd, 2006, 06:14 AM   #16
Trustee
 
Join Date: Aug 2003
Posts: 1,483
A limiter is like someone standing over your
Marantz and when they hear a loud sound
that it seems like will cause distortion, this
person will quick turn down the level knob,
and then turn the knob back to where it
was once the loud sound has passed.
So, when that guy turns down the
knob to prevent distortion, the loud
sounds and the quiet sounds get turned
down the same amount, so it's proportional.
Dave Largent is offline   Reply With Quote
Old May 3rd, 2006, 07:11 AM   #17
Major Player
 
Join Date: Jun 2004
Location: McLean, VA United States
Posts: 749
A limiter is, as the name implies, a circuit which limits the signal passed on to the innards of the recorder. A "hard" limiter is simply a clipper. Any level above the maximum is set to the maximum. Put a nice sin wave of excessive amplitude into it and the top gets lopped off flat and badly distorted. "Soft" limiters don't chop the top of the waveform but rather push it down - still distorting but not so badly. A soft limiter will never put out more than a maximum level.

An AGC (automatic gain control) on the other hand does not distort but simply adjusts the volume automatically to try to keep the average level at a value which loads the A/D converter FM modulator properly. Whereas the penalty for using limiters is distortion on peaks (low level stuff sounds fine) the penalty for using AGC is "pumping", the phemomenon in which you hear the background noise level going up and down in response to the ambient signal level. This is often most noticeable after something like a rifle report (recorded as a thud because even with an AGC it's likely to get clipped). Immediately afterwards there is dead silence as the AGC has turned the gain all the way down (quick attack). Then over the next second or two (slow decay) one hears background noise slowly increase.

To avoid paying either penalty it is necessary to increase system dynamic range. More bits is a good way to go. 24 bit audio has 48 dB more dynamic range than 16 bit (caution: just because the marketing stuff says 24 bits doesn't mean that the performance is really 24 bit). Another scheme is to split the audio into 2 channels set 10 - 20 dB apart in gain. In post you then pick the high gain channel if it's not overloaded and the low gain one if it is.
A. J. deLange is offline   Reply With Quote
Old May 3rd, 2006, 07:42 AM   #18
Major Player
 
Join Date: Sep 2005
Location: NYC, weeee.
Posts: 417
A soft limiter will clip, it just trys to do it softly.


a compressor and limiter are basically the same thing used differently. A compressor has a ratio that you can set, it has an attack and release and threashold setting. A ratio at say 4:1 means that for every 4 db over the threashold, 1 db will come out.

A limiter has a threashold, it has a ratio generally from 20:1 or above that is fixed as well as the attack and release being fixed.

a compressor is designed (generally) for controlling the average level while a limiter is used for controlling the peak level.

Most of the limiters that are built into say minidisc players and camcorders are a simple limiter. they are brutal and hard. I would say almost as unusable as distortion.

It will take a bit of practice to effectively use a compressor or limiter without having the side effects become noticable.

Basically you want to use the limiter as a last resort and then only sparingly. A compressor can be much more gentle and invisible.
Michael Fossenkemper is offline   Reply With Quote
Old May 3rd, 2006, 12:29 PM   #19
Regular Crew
 
Join Date: Apr 2006
Location: Brooklyn, NY
Posts: 111
A limiter will turn down the sounds disproportionately - i.e. if you have audio peaking at -3dB, and all of a sudden crowd noise puts it up to +4dB, the limiter will stop it from peaking and put the audio through at the level you select.

I believe most limiters are set to 0dB by default.
Abe Dolinger is offline   Reply With Quote
Old May 3rd, 2006, 05:59 PM   #20
Major Player
 
Join Date: Feb 2006
Location: New York
Posts: 508
Wait, okay, so some of you are saying limiters turn down sounds proportionally, while others say it turns sounds down disporportionally. Which is it?

Also, how would I split into two channels with different gains?

And hard vs. soft limiters. How do I know which one I have? And is a soft limiter the same as a compressor? One of you said soft limiter does not clip, while another said soft limiters do clip, just gently - which is it?
Alex Thames is offline   Reply With Quote
Old May 4th, 2006, 09:40 AM   #21
Major Player
 
Join Date: Jun 2004
Location: McLean, VA United States
Posts: 749
Limiters don't turn down sound. They implement a non linear relationship between input and output. If you picture a graph of output voltage vs. input voltage as a straight line with 100 mV in giving 100 mV out, 200 mV in giving 200 mV out, -150 mV in giving -150 mV out you are describing a linear (because of the straight line relationship) system with unity (0dB gain). If the line is straight with a steeper slope (100 mV in gives 200 mV out, 200 mV in gives 400 mV out...) you are describing a linear system with a gain of 2 (6 dB). In manual gain control you set the slope of the line. In Automatic Gain Control (AGC) there is circuitry which measures the average output and adjusts the slope to hold the average to value appropriate to the application (e.g. proper loading factor for an A/D converter). Thus AGC may lower gain (compress), raise gain (expand) or do both (compand) but not at the same time. The important things are 1) the transfer curve is linear and 2) the slope of the curve varies with time. Linear systems do not distort.

Limiters, by contrast, have non linear transfer functions that do not vary with time. The transfer function of a limiter may have a linear portion at low inputs but the slope will decrease at higher input levels. If you adjust gamma in images to increase shadow contrast at the expense of highlight detail you will know what the positive voltage half of a limiter curve looks like. Thus in a limiter the gain depends on the amplitude. A change of 100 mV near 0 may result in a change of 100 mV at the output but a change of 100 mV near 1 volt in may result in a change of only 50 or even fewer mV. Because of this low level sounds are not distorted but high level ones are. The amount of distortion depends on the amount of non linearity encountered which depends on the design of the limiter and the total input voltage swing. The most extreme distortion occurs with "hard" limiting in which any positive voltage input is converted to a fixed positive voltage and any negative one to a fixed negative voltage of the same magnitude as the positive limit. More gentle limiters such as the µ-law curve create less dramatic distortions.

The obvious and simplest way to split into 2 channels is to use 2 microphones each going to its own channel on a mixer with the mixer channel gains set as far apart as experiment indicates is necessary. There is another thread here in which a poster suggests doing this by setting the gain on a stero channel pair as required and then panning heavily to one side. Pretty clever. One could split a single microphone's output into two paths but some care would be required with respect to keeping impedances matched, getting phantom power to the mic (if required) and so on.

All devices are ultimately hard limiters and will clip which simply means that no device has infinite input dynamic range.

You need to be aware that semantics gets into the stew here. For example µ-law limiting is often referred to as µ-law compression because the dynamic range of the incoming signal (speech) is effectively compressed into fewer bits than are required with straight PCM. On the other hand what I have described as compression, expansion and companding are in reality AGC functions. It's the same with clipping. Clipping usually means the peaks of the waveforms are clipped off as if with scissors i.e. the tops are flat. In this sense a soft limiter does not clip (unless you apply voltage outside the specified input range it which case it, as any device ultimately must, clips). From this perspective a soft limiter does not clip but rather lowers the height of and rounds peaks.

To determine what you have you could take a recording of a tone (often comes with the SMPTE color bar pattern from various devices) and run it through your DAW's limiters at various settings. Look at the wave form out of the device and familiarlize yourself with the distortions that the various control settings give. Do the same with a speech or music signal. With a little experience you'll be able to tell what has happened to a signal just by looking at it. Also, of course, listen.
A. J. deLange is offline   Reply With Quote
Old May 5th, 2006, 05:15 AM   #22
Inner Circle
 
Join Date: Nov 2004
Location: Baltimore, MD USA
Posts: 2,337
Whoa!

A limiter does NOT turn things down. It sets a threshold above which sound level is controlled. The degreee of control depends on the settings of the limiter.

Some limiters do really nasty things to the sound when it crosses the threshold. Others are quite good.

A good limiter allows you to record at higher levels without fear of going into the red. Most AGC circuits on cameras I've heard suck. Literally, sucking up the noise in quieter sections. AGC circuits act more like compressors. Compressors and limiters are different.

Ty Ford
Ty Ford is offline   Reply With Quote
Old May 5th, 2006, 05:41 PM   #23
Major Player
 
Join Date: Feb 2006
Location: New York
Posts: 508
Okay, all that technical stuff vaguely made sense to me. Could someone explain to me how practically I could record audio without distortion when there are definitely going to be softer/quiet/near silent parts as well as extremely loud cheering/shouting/clapping? Or even if a singer has a soft part and then pumps her volume up for dramatic effect during the song? I want the audio to capture all this without distortion and increases/decreases in volume to be proportional (linear?) with reality, except in the cases of cheering/etc. where it doesn't matter (I actually prefer to bring cheering down and not get distortion).
Alex Thames is offline   Reply With Quote
Old May 5th, 2006, 06:33 PM   #24
Inner Circle
 
Join Date: Nov 2004
Location: Baltimore, MD USA
Posts: 2,337
Sure, hire a sound guy/gal who know what to do. That's what the pros do.

Ty Ford
Ty Ford is offline   Reply With Quote
Old May 5th, 2006, 10:26 PM   #25
Major Player
 
Join Date: Feb 2006
Location: New York
Posts: 508
Uhhh...that doesn't help much. In fact, that is completely out of the question as first, it does not teach me anything, second, it costs a load of money that neither I nor the club have, and third, there is no time to get them to the event.
Alex Thames is offline   Reply With Quote
Old May 6th, 2006, 04:51 AM   #26
Inner Circle
 
Join Date: Nov 2004
Location: Baltimore, MD USA
Posts: 2,337
You've already let us know you don't really know what to do to fix the problem. You've already let us know you don't understand how a limiter works. In order to get what you want, you have learn it, get it or pay someone else to do it. That's pretty simple.

I don't know how to do Calculus or my taxes. Tried, didn't get it. I pay someone to do my taxes. If it needed Caluclus, I'd pay someone.

Executive producers, directors and (sometimes) shooters don't really understand audio. Audio is NOT trivial. You have discovered that. Take the time to learn, or get someone else to do it and hope you make the right choice in getting someone who actually DOES know what they are doing. Those are your two options.

Regards,

Ty Ford
Ty Ford is offline   Reply With Quote
Old May 6th, 2006, 06:05 AM   #27
Major Player
 
Join Date: Jun 2004
Location: McLean, VA United States
Posts: 749
Probably the simplest solution is to use the AGC in the camera. They generally have fast attack thus protecting you from sudden loud noise but slow decay which leads to the annoying pumping. The slowness of the decay may be on your side as it is slow enough that you should be able to comensate for it to some extent in post. Another simple approach might be to use a pair of cardioid mikes with one aimed towards the performers (or more pertinantly with the low gain part of the pattern aimed at the audience) and the other aimed into the crowd. Adjust the gain on the one so the performer can't overload and the gain on the other so the crowd can't overload.

With either approach you will have to experiment to get the settings right. This is where the learning comes in.
A. J. deLange is offline   Reply With Quote
Old May 6th, 2006, 06:14 AM   #28
Inner Circle
 
Join Date: Nov 2004
Location: Baltimore, MD USA
Posts: 2,337
As I have mentioned many times in the past. I have yet to hear a camera AGC circuit that does more good than harm. An AGC is not a replacement for a good limiter.


Ty Ford
Ty Ford is offline   Reply With Quote
Old May 6th, 2006, 06:21 AM   #29
Inner Circle
 
Join Date: Mar 2005
Location: Hamilton, Ontario, Canada
Posts: 5,742
Quote:
Originally Posted by Alex Thames
Uhhh...that doesn't help much. In fact, that is completely out of the question as first, it does not teach me anything, second, it costs a load of money that neither I nor the club have, and third, there is no time to get them to the event.
How about this? Bring along a second recorder and record some of the audience applause. Record the performance normally using whatever method you prefer and don't worry if the applause is distorted. In post, edit out the distorted audience reaction and replace it with the undistorted audience reaction from the separate recording.
__________________
Good news, Cousins! This week's chocolate ration is 15 grams!
Steve House is offline   Reply With Quote
Old May 6th, 2006, 07:29 AM   #30
Major Player
 
Join Date: Sep 2005
Location: NYC, weeee.
Posts: 417
Compressors and limiters are different only in the sense of how they react. They basically do the same thing. I can take a compressor and turn it into a limiter with a few adjustments. Where as dedicated limiters usually don't include the controls to use it as a compressor. expanders, gates, companders, agc, de'ssers and limiters are just different ways to impliment a compressor. They are all doing the same things. How a designer decides to impliment the parameters of a compressor will dictate how you can use it. cost is usually the deciding factor.

Hiring a sound guy is usually not as expensive as you would think. In NYC, there are probably thousands of fresh out of school audio students who would be more than happy to man the controls for next to nothing. You could even contact an audio school in your area and request an intern who is interested in audio for film. Heck, they may even have their own gear.
Michael Fossenkemper is offline   Reply
Reply

DV Info Net refers all where-to-buy and where-to-rent questions exclusively to these trusted full line dealers and rental houses...

B&H Photo Video
(866) 521-7381
New York, NY USA

Scan Computers Int. Ltd.
+44 0871-472-4747
Bolton, Lancashire UK


DV Info Net also encourages you to support local businesses and buy from an authorized dealer in your neighborhood.
  You are here: DV Info Net > The Tools of DV and HD Production > All Things Audio

Thread Tools Search this Thread
Search this Thread:

Advanced Search

 



All times are GMT -6. The time now is 05:23 AM.


DV Info Net -- Real Names, Real People, Real Info!
1998-2024 The Digital Video Information Network