Radio SBN runs three dedicated feeds directly from its offices in Temple City, CA. Each feed is on a dedicated server, but route through a 14-channel Behringer MX1804X studio mixer. We are often asked how we get such great audio - this is how. In the process of getting clean audio we have discovered many ways to get clean audio to the servers.

The Behringer MX1804X mixer is used because of its ability to control the tonal quality of the scanner audio, as well as routing. Each scanner is assigned to a mixer channel and each mixer channel has a Send / Return loop. This allows the audio from the individual channel to be routed directly to a dedicated client server, AND, to be mixed simultaneously on one Teamspeak channel during mutual aid events. Additionally, the MX1804X has a mute button on each channel that when depressed the audio is removed from the master buss and routed to what is known as ALT 3/4, a secondary buss. This allows us to monitor the primary, or Program, feed at one volume level, and the secondary, or Monitor, feed at another volume level. Regardless of the mixer configuration the audio from each channel is uninterrupted and streams to its respective Teamspeak channel. During a multi-agency event, however, you may be listening to CHP on the CHP channel, and LAPD on the LAPD channel while BOTH are being mixed on the In-Progress channel.

You don't need to get this fancy with your feed at all, but an external mixer is highly recommended if you want to have a very clean stream. This may become important with P25 digital, as in the case of LAPD. Helicopter audio always seems to come over very bassy and by using an external mixer we are able to clean up their muffled voices. In fact, radio audio is best confined to 300-3000 hertz. Not only does this help remove the effects of static on the radio signal, but it cleans up the voices and helps remove 60 cycle hum. Depending on the mixer, you may also be able to employ mid-boost that helps to pull the human voice above the rest of the audio. Lastly, an external mixer helps in matching line levels. A suitable mixer is the Behringer UB502 or the Behringer UB1002.

Clean, hum-free audio is easy to obtain by watching the stage gain, line level matching, and balanced audio lines.

Gain Structure is one of the things, which can improve or spoil audio systems totally. Every piece of audio equipment has a certain operation range, determined by noise floor and clipping point. Equipment can only work properly with enough footroom and enough headroom. Different pieces of analog audio equipment need different input levels and supply different output levels.

The questionable theory of unity gain says that the same level should travel through every unit. Every unit of the signal chain will be set upstream, beginning with the final device , to unity gain. The system level for actual working conditions is then set at the first unit in the signal chain (mixer, preamp, etc.). This approach has one big disadvantage: The maximum gain in the chain is limited by the weakest unit and the signal/noise ration is mostly bad. The maximum level reaching the amplifier is much too much for real live sound reinforcement and on the first level stage (the mixer) the level will be reduced dramatically. The result is that most signal LEDs in all devices don't even show a a single time. Analog devices work much too close to the noise floor and some (PCM) digital equipment works (in worst case) like 8bit-equipment.

The better approach is to look for proper headroom in every single piece of equipment. Every unit should be driven to have the optimum signal/noise ratio with the proper headroom. This level range will be determined by the output level of the previous device in the chain and by setting the input level controls of the unit itself. The system will be set downstream, beginning with the first unit and ending at the computer.

Setting up the system:

An reliable steady input source must be available. A pink noise generator is the best choice. The gain structure will be set up from the first input to the last output. The last unit in the chain is the computer. It will be turned down first. The pink noise input signal will be fed into the first unit in the chain (a mixer, preamp, DSP unit, etc.). This level should represent the real world level that will be present at this input under normal circumstances. Now this level should be set to the maximum possible level under real world conditions (peak). With this maximum input level, fed to the first unit in the chain into the regular system signal input, the input level control of this unit should be set to operate the unit just before clipping with clean signal and just enough headroom for maximum signal. Now signal LEDs should indicate perfect working conditions.

If this unit works fine, we move on to the next stage. Now level will be set on the output level control of the first unit and the input level control of the second device. This device should work under similar conditions as the first one. Levels within the unit should be set just under clipping point with sufficient headroom.

As the next step the output control and/or the input control of the third device will be set with the same procedure. This will be continued all the way through the signal chain to the last stage, the computer's input stage. If the computer has no input level control, the level must be set on the last level control in the signal chain. Now the computer will be set to the proper level to ensure the appropriate record level.

A possible high dynamic range of the entire system could be reduced to an unacceptable small amount … Audible hiss in installed loudspeakers is always a clear indicator for such a mismatching gain structure (even if it is UNITY GAIN) !

Excerpt from H. Ott 'Noise Reduction Techniques in Electronic Systems': '... a balanced circuit is a 2-conductor circuit in which both conductors and all circuits connected to them have the same impedance with respect to ground and to all other conductors. The purpose of balancing is to make the noise pickup equal in both conductors, in which case it will be a common-mode signal that can be made to cancel out in the load...'

The two conductors of a balanced line operate in a push-pull manner. As the voltage in one conductor becomes positive, the voltage in the other conductor becomes negative (both by the same amount and at the same time). Both conductors are equal in voltage but opposite in polarity. The receiving circuit for a balanced signal is called a differential amplifier. The wanted audio signal itself is opposite in polarity in both conductors, the effect in the differential amplifier is that this signal is processed by the differential amplifier. Any unwanted electrical interference will interact with both conductors the same and also with the same polarity. The same polarity voltages aren't processed in the differential amplifier and the unwanted signal, the interference, is cancelled out. The unwanted voltage is a common-mode voltage, it is common to both inputs of the differential amplifier. Noise immunity on balanced audio lines is completely dependant on how WELL the input and output impedances are balanced and on the common mode rejection ratio of the audio input.

The shield on regular balanced lines is only an additional protection against interference, but it is actually not a substantial part of the balanced line. Balanced lines without shield work very well too. Especially the number of twists between the two conductors are more important. A very well twisted cable like unshielded CAT5 cable is perfectly usable for balanced audio lines.

The definition which of the two conductors is 'positive' and which is 'negative' is a pure definition to ensure the proper polarity of the signal throughout a device or a system, it is not a part of the definition of a balanced line. Using a device from input to output in the 'opposite' polarity as labeled (+ for - and - for +), wouldn't degrade the signal in any way. The definition of the polarity of connections and connectors (like the XLR connector) is necessary to create systems with a defined polarity but is not essential to the characteristics of a balanced line! The only needed definition for a balanced line is to have input and output in the SAME polarity.

Okay. You have your scanner connected and levels adjusted, but you still have hum. The section above discusses using balanced, three-wire lines, but yours only have two lines. How then do you convert to a balanced line? By using a Direct Box (also known as a DI.)

A Direct Box converts an unbalanced audio line to a balanced line using a special transformer and associated circuitry. Most DI's also have the ability to add attenuation in various ranges. Some DI's are passive and require no power, while others are active and run on a battery or phantom power supplied by the mixer. DI's are also very helpful in matching the output line to the input circuit. Most DI's also have a ground lift switch that is often helpful when dealing with hum.

An easy to use DI is the Rolls DB25. It sells for about $29 and accepts a 1/4" mono input and converts it to a balanced XLR output. This device is passive and requires no battery or phantom power.

Another method of using a balanced line is the use of two matching transformers with 1/4" plugs. You will need to use an adapter on the scanner end to match the output jack.

If you are using the Mic In on the mixer end you will simply plug the XLR cable into the mixer. If however, you are using the 1/4" Line In on the mixer you will need a second matching transformer unless the mixer has a "Smart Jack" that can detect a balanced line. There are several possible configurations. Your goal should be to keep things simple.

Building a solid audio streaming platform may be as simple or as complex as you choose to make it. We recommend the simpler approach of plugging into your scanner and then into the computer. If it works, STOP! If it ain't broke, why try to fix it? The only real requirement regardless of your connection method is to look at your signal in an audio application, such as Cool Edit Pro, and verify that you have no or little hum on your signal, and, that your levels fall below 0dB. Computers are digital and digital audio will not tolerate a 0dB signal. Too high of an input will result in distortion and the introduction of hum. When the input of the computer is too high you begin working within the noise floor of the computer and it only gets worse. So the general rule is to use as little signal as possible to achieve a -6dB output.