The problem is that wires are also antennas. When you plug a mic into a camera, or a DAT player into your NLE, you don't just get the desired signal. Any nearby electric fields are also picked up on the wire, adding a slight voltage which the equipment can't distinguish from the desired audio.
You can't avoid these fields. They're created by any other wires that carry a current. This includes video, timecode, and data cables, which can add all sorts of high-frequency whines and whistles to your track.
Manufacturers reduce this antenna effect in a signal wire by wrapping a shield around it, usually a copper braid or metal foil. The shield is connected to ground and shorts out the interference before it can reach the signal wire in the center of the cable. That's why phono, BNC, and cable TV plugs have a center pin and outer metal shell: the pin is signal, and the shell carries the shielding that in the wire.
Furthermore, the same shielding that protects against high-frequency noise can contribute to hum.
That's fine for a single piece of equipment, but when you hook two devices together, both have to agree on the reference. Since the cable shield has to be grounded at least at one end, the usual scheme is to use it to connect the two devices' reference points together. It works in very simple systems.
But remember, that shield is picking up hum from the building wiring. And if the shield is carrying current -- something unavoidable if it's part of the audio path -- it has a slight voltage drop. Both these factors mean that the two devices are going to have slightly different references, and the difference is constantly varying. The input circuit can't tell that this variation isn't part of the signal, so it amplifies it. Again, in a simple NLE with short wires this interference may be tolerable. But in a complex room or studio shoot, it becomes hum and noise.
It's called "60 Hz hum", but it's not just 60 Hz
When power-line frequencies leak into an audio circuit, they generate harmonics. The 60 Hz base signal also hums at 120 Hz, 240 Hz, and up the band. That's why filters don't do a good job removing hum... you have to fix it at the source
Complex setups have other problems as well. If there are multiple ground paths, they combine to make a very efficient loop antenna for the 60 Hz noise. These "ground loops" are almost impossible to predict, since you don't know the internal details of your equipment, and can crop up even in very simple setups... particularly if both pieces of audio equipment also share a ground connection through their power plugs' grounding pins. In a practical video studio, the situation is apt to be far worse: the non-audio cables -- RS-232 and RS-422 control, video wires, and even cable TV -- all have their own grounds.
Balanced wires also reject noise that isn't coming from a ground loop. The two conductors are twisted closely together, so any interference radiated into the cable is picked up equally by both. But remember: the equipment is looking for a voltage difference between those wires. Noise is the same on both wires, so the equipment can't hear it.
Answer: it does. Noises from nearby video or computer cables are picked up on each conductor. But remember, a balanced audio input cares only about the voltage difference between the two wires. Interference is radiated equally into each wire. Since the interference is equal on each, there's no voltage difference from it! The balanced input can't even see that the noise is there.
Or to put it into a chart:
Conductor Audio Signal Noise Total on wire Black +1 v +1 v +2 v White -1 v +1 v 0 v Transmitted difference 2 v Received difference 2 vThis noise-immunity of balanced wiring is why it's also used for high-speed computer networks. Category-5 cable contains four tightly-balanced pairs of wires. In fact, if your balanced input and output circuits are good enough, you can use Cat-5 cable for professional audio wiring!
"Star Quad" is four-conductor shielded balanced cable. The four wires form a tighter, more consistent pack than two wires can and can resist even more noise. If you're using Star Quad, you must tie the two pairs of similarly-colored wires together at each end... reducing it effectively to two conductors. Don't try to use it as two balanced pairs for two different signals: this won't give you any noise-reduction benefits at all.
You can also get into trouble assuming that other kinds of connectors aren't balanced. Some manufacturers save money and space by putting balanced signals on three-conductor phone jacks. From the front panel, these look identical to two-conductor unbalanced jacks. But if you connect a three-conductor phone plug -- also known as Tip-Ring-Sleeve, or TRS -- it can carry a balanced signal. Again, check the specs.
And a balanced connection works only if both ends are balanced. If either of the two internal wires are shorted to ground at either end, the whole thing becomes unbalanced. This can happen easily if:
|Got balanced wiring on a TRS jack (found in many compact mixers)?
The worst thing you can do is split it to two channels with a stereo Y cable. You'll hear a signal on each end of the Y, but the two signals will be of opposite polarity. If you then listen in mono, they'll cancel each other and the sound will disappear!
If you want to connect a balanced mic to a prosumer camera, use one of the small transformer adapters from BeachTek or Studio 1, or put a mixer (with balanced inputs, of course) near the camera. This is the only way to run long mic cables to a boom or wired lav without risking noise pickup.
Connecting an unbalanced consumer output to a balanced professional input -- say, from a sound card to a professional VTR -- is slightly more complicated, because the voltage levels are different as well. For this situation, the best bet is to get a balanced interface for the computer. This is a small powered box with an unbalanced input at -10 dBV (usually on an RCA jack) and balanced output at +4 dBm (usually on an XLR male). Run a very short cable from the source to the RCA jack, and as long a balanced XLR cable as you need to the pro equipment. Similar interfaces are available to connect balanced sources to unbalanced equipment. They cost between $50 - $75 per channel
If you don't need balanced wiring and don't care about levels, start with the schemes below. They'll work with most modern equipment... but there are some oddballs.
* - Some balanced phone-plug inputs are designed so that you can plug an unbalanced two-conductor plug directly into them.
* - If the source has a transformer output (some high-end gear and most of the classic stuff), connect this to ground as well.
If you want to connect a +4 dBm professional balanced source to -10 dBV unbalanced prosumer equipment, and you don't care about balancing but you do want to adapt the level, get some resistors at Radio Shack:
* - You'll need this jumper if the source has a transformer output. It may cause problems with transformerless devices.
In 2010, reader Hugh Robjohns told me I'd made a common formula error in figuring the above resistors, and that the circuit actually has about 1 dB more attenuation than I describe. I apologize. The difference wouldn't be noticeable in most setups, but he is technically correct... the 3.9kΩ resistor should be 2.87kΩ. Readily available 3kΩ resistors will do the job. I thank him for the correction.
But those are with my assumptions about the internal impedances of whatever random devices you're connecting. This is a critical factor if you want the attenuation to be precise. Hugh assumes differently, and his numbers call for 2.7kΩ and 8.2kΩ respectively. I'd guess he deals with slightly different kinds of equipment.
Now while Hugh's numbers are very different, the ratio between them is about the same as mine. Believe it or not, both his version and my original one will give you about the same results in the field!
This assumes you do a little fine-tuning with the devices' volume controls, watching the input meter. If you want lab precision in your attenuator, you'll precise values for the device impedances. That's something manufacturers rarely put in equipment spec sheets or manuals. You'll have to examine the actual circuits, and ideally use an impedance bridge to measure it at critical frequencies.
To Tutorials Index
To DV Magazine Readers' Page
To top of site
Mail to Jay Rose
© Jay Rose 2001