Actions

Managing Audio Dynamics and Leveling

From The RadioReference Wiki

How not to blow someone's speakers

This article will describe an approach to managing the audio dynamics (loudness levels) of your feed to deal with varying output levels from the systems you are feeding.

Equipment

Concept

The difference between your softest signal and your loudest signal is generically referred to as dynamic range or headroom. In serving up an audio feed you have a few different sections of signal path, each with their own limitations on available dynamic range.

Components of the signal path in order of a typical feed connection
  • The scanner audio output driver itself
  • The amp that is in the input circuit (mic or line input) of the soundcard interface
  • The software encoder

To begin with this concept, the assumption will be that you have adjusted any or all of those so that your loudest anticipated signal from a feed will not clip or distort any of those stages in the feed configuration. Ideally you would want to drive your loudest signal to a level where it will be well clear of any clipping or distortion points, but still quite strong so that it is well above the noise floor of the next step in the signal path.

As a rule you never want a very quiet signal being 'pulled' up drastically by the next stage in the chain, nor would you want any point in the middle having to greatly attenuate (turn down) the level of the feed, only to boost it later in the chain. Both of these problems will introduce hiss, hum and possible distortion to the signal. In drastic cases of overdriving one input into another, you can actually damage your equipment.

You ideally want an even gain stage across all the inputs and until we introduce a compressor, it should be either even levels or a very gradual increase in levels at each stage until the final output.

All of this now assumed... Where are we going with this?

O.k. so you have configured for your loudest feed. When this system comes on line you monitor and set your listening volume to a comfortable level. Now...the quiet mumbly dispatcher or mobile unit comes on and you can't hear them, so you turn it up. Blam! The alert tones come blaring in from your loud system all the sudden. This is how we fix this.

How a Compressor Works

At a high level of explanation, a compressor/limiter/dynamics processor in our context will be serving as a gate keeper riding a volume knob automatically for you. Quiet signals will be audible and a loud signal will be 'grabbed' and attenuated to a level much closer to the quiet signal. The basic technical components of a compressor are as follows.

Input gain
This is the input drive level for the signal you are feeding into the compressor
Threshold
This is the level the signal has to reach before the compressor will 'grab' it
Your threshold level will vary depending on the signal levels you are trying to manage. You have to play with this one.
Attack and Release
This is how fast it will react, grab the signal, process and how fast it will then roll off and let it go again
For our needs this should be an immediate attack and a release about half a second to one second will do.
Ratio
This is how hard it will grab and apply limiting to the level of the signal
For our use, this should be quite high 10:1 or infinity:1 will work nicely.
Output level
As a result of compressing signals, there is a drop in the overall output
This has to be adjusted so that you have the maximum clean level for encoding to the stream and will vary based on your audio feed characteristics.

External Hardware Compressor Application

Connections
  • Scanner audio output cable to input of compressor unit
  • Compressor audio output cable to input on sound card device

You will have to read the manual on your compressor as these are very generic hints at how to configure the unit. The concepts are all the same however, so this should apply.

Set your scanner to an output level that is typically comfortable for decently loud, clear listening from its speaker.

Connect your cable to the input of the compressor. Use of properly matched balanced or unbalanced cable wiring and shielded cable is important for a clean signal.

Set the levels of the compressor
Input to zero
Threshold to zero
Attack to zero or as fast as it will go
Release to .5 to 1s roughly
Ratio to 10:1 or higher...up to infinity:1
Output to 0db (generally straight up and down
Configure your input on the computer so you can monitor it as you work on this. The level should be approximately what you would use now to listen to the feed source.

Now, very gradually start bringing up the input level on the compressor. Depending on the model, you will start to see output level indicated on its meter. The first light should be -24db or so and you should see it come up as you gradually raise the input level.

What you would ideally want is to find your local quiet system, one that represents the typical lowest level signal you hear. Adjust the input level on the compressor until this signal is showing just shy or barely blinking the 0db light on the compressors output meter reading. As you are doing this, adjust the computer input gain for mic or line in so that the level will work with a 0db output indicated from your compressor. This will likely mean your input slider on the computer is barely opened for the input source.

O.k. now we have the starting point that we can apply dynamics processing to.

Be careful...don't let it scan right now or something loud will be REALLY loud.

Now, take the threshold knob and start turning it during a transmission until you see the compression or 'reduction' meter just blink the first light. Back the threshold off until it no longer blinks on the level of this quiet system.

By doing this, you have set the quiet system level output as your baseline output that you basically don't want anything to exceed in any drastic way.

The next step...you may want to turn the compressor output gain down for this one. Fine your loudest local system, NOAA weather around here is super overdriven and is an easy quick reference, but this may vary by you. Once you get the loudest system going, bring the output level control of the compressor up so that it is reading 0db and maybe the +3db light occasionally blinking.

Compressor.jpg

In this image the left hand row of lights at the top is the indicated gain reduction of a loud signal. The compressor is reducing gain by about 6db in this case. The right hand row of output level lights is showing just shy of 0db. This is typical of what your gain reduction lights and output level should look like with a moderately loud input signal. On my loudest systems here, the left hand gain reduction lights will be pinned to -16db+ while the output remains right at 0db. This is the compressor at work.

The whole idea is, the quietest signal is as loud as you will allow the audio to get no matter how loud another system is, before it goes into the computer sound input. You have limited the maximum level. At this point, readjust the computer line input if needed and you should be on autopilot.

If you encounter distortion, you may have to tinker with the input level of the compressor vs the threshold and figure out how hard to drive into the compressor and where to make it start to 'grab' the signals.

Software Dynamics Processing

Connections
All of the concepts of above apply with one important caveat
Your audio routing is done internal to the software and you will need what is called the Virtual Audio Cable software to do this.

Picture it this way

Scanner audio out -> Computer line or mic in -> Processing Software -> Virtual audio cable software input -> VAC output -> scanner feed stream software input

Then you basically route that audio into the processing software, do all that we just did in the Hardware steps above and then route the output of the software to the VAC (virtual audio cable) and it will hand it off to your scanner feed software.

The benefit of this approach is that the processing software can be had for free and the VAC software license is cheaper than buying a compressor for sure.

IMPORTANT
This part is theoretical. I have used software processing for my transmit audio for ham radio many times and it worked. However, I do NOT know for sure if it will work properly with the scanner feed encoding software.
If someone can verify this...remove this comment block and add any further details applicable please.

Go to Integrating Feeds
Return to Live Audio