Rabu, 30 November 2011

Noise Gate

3.1. Introduction 
    The aim of a pure noise gate is to mute signals with low volume. Most gates are also able to close just partially, providing a set attenuation as opposed to muting the signal completely.
    There exist several reasons why a noise gate may be needed. For recording applications, gates are used to eliminate background noise when the instrument or voice is not generating a signal (during pauses, and just before the song starts and after the song has just finished). For sound reinforcement applications gates are not as necessary since background noise is not such a big problem once the show is taking place, so gates tend to be used more with the idea of avoiding other sounds from leaking into a microphone that is not being used. The most sophisticated drum micing might use gates to avoid picking up the sound of other drums with the microphone that is supposed to pick up just the sound of a specific drum. For instance, the bass drum mic would only be active during bass drum hits, while the snare drum gate would only be open for snare hits, such that the snare does not leak into the bass drum microphone, and vice versa.
    Typically, gating takes place in a slower way compared to compression (and modifies only signal RMS levels). Also, gates are often closed for a relatively long time. Like for compressors, the detection circuit is extremely important, and may vary widely from model to model.
    Also, it is important to understand that gates do not eliminate the noise. They just hide it by attenuating it when it is most obvious (during pauses), and let it through when it is mixed with the signal.
3.2. Controls
Gating is a difficult task that may require very different characteristics depending of the type of signal. Numerous controls are therefore needed. The drawing below shows a noise gate (downward expander) with the most common 

The most common controls provided on gates are given below. You may not always find all of them, or you may get additional ones: 
  • Threshold. When the signal falls below the threshold level the dynamic processing starts and the gate begins to close. In general, this control needs to be adjusted as low as possible (without false gate openings), since that way the original signal will be preserved.
  • Attack time. This is the time it takes for the gate to fully open once the signal exceeds the threshold level. Lowest available attack times would be in the 10-100 us (microsecond) range depending on the unit, while maximum attack times vary considerably depending on manufacturer and model, but would go anywhere from 200 ms (milliseconds) to 1 second. Fast times may create distortion artifacts, as they modify the waveform of low frequencies, which are slow. For example, a 100 Hz cycle lasts 10 ms, so that an attack time of 1 ms has the time to modify the wave form and generate distortion. On the other hand, a time that is too long will remove the percussive transient from a drum sound. In general, it should be as fast as possible, but not too fast to cause clicks or distort the attack of a signal.
  • Release time, sometimes referred to as decay time. It is the opposite of attack time, i.e. the time it takes to go from the unprocessed state (no attenuation) to the fully processed state where the signal is attenuated or muted. The release times available on a unit are a lot slower than attack times and may range from 2-10 ms to 3-5 seconds. The illustration shows the difference between gating with a slower (left) or faster (right) release time.

  • Hold time. This is the time that the gate is kept open after the signal level has fallen below the threshold, and is therefore the minimum time that the gate will remain open. It is used to prevent altering the decay of sounds by avoiding short low signal levels triggering the gate to close and entering the release phase. It can typically be adjusted between zero and several seconds. Sometimes an indicator is provided that lights during the hold phase. The graph below shows how the different times come into play on a noise gate.

  • Range or attenuation ratio. Simpler gates just shut down the output completely, though a dedicated gate would normally provide control over the amount of attenuation provided to signals below the threshold level, letting an amount of signal through. One advantage of just providing some degree of attenuation as opposed to muting the signal completely is that when the gate opens, it will do so more smoothly, since the signal does not have to rise from zero. The level reduction can be performed with two different families of devices:
    • Noise gate. It provides a pre-defined attenuation, be it total muting or an intermediate level, which is known as "range", or, sometimes depth. For instance, we could attenuate 20 dB or 40 dB, or just close the gate completely (-infinity attenuation).
    • Downward expander. This type of unit works like the opposite of a compressor, and provides a "ratio" control or attenuation slope. That way the signal is attenuated more highly the lower the signal level is. In fact, if the threshold is set high enough, we could "expand" the dynamics of the signal, increasing its dynamic range, though this is not normally done. The attenuation ratio works in an equivalent way to that of the compressor, defining the amount of a attenuation (compression) that is applied to the signal. These ratios are expressed in dB, so that, for example, 1:6, means a signal that is 1 dB below the threshold will get reduced to 6 dB below it, while a signal  3 dB below the threshold will get reduced to 18 dB below it. Likewise, a 1:3 (one to three) means a signal 1 dB below the threshold will be attenuated 2 dB (as the level will go from -1 dB to -3 dB; we use a negative sign as these levels are below the threshold, which is the 0 dB reference in this case). With a ratio of 1:10 and higher, the expander is considered to work as a pure noise gate, though an ideal gate would have a theoretical ratio of 1:infinity (any level below the threshold would be totally muted).
    In practice, expanders are used in a very similar way to pure noise gates, with the only difference that downward expansion is smoother and more gradual, and it is therefore more difficult to get the attack and release times wrong. Some units provide the possibility of choosing between gate or expander action. The illustration below shows the difference between a expander with a variety of ratios and a gate with different attenuation "ranges" (depths). If you are having a hard time understanding the input/output graph, do have a look at the compression tutorial.

Dynamic filter. Some gates are equipped with a dynamic filter whose job is to add more attenuation as the signal is lower in lower. This provides more natural gating action, as instrument and vocal sounds tend to show less harmonics as the instrument is played more softly or the voice is lowered.
  • Stereo link. When dynamics processors in general are used to process a stereo signal, it is necessary to synchronize the processing action on the channels so it happens at the same time on both. Otherwise the stereo image will be confusing and move from the center to one side or the other. Mono units often provide a connector that allows linking of two units.
  • Automatic. It is becoming more and more common for equipment to provide the possibility of controlling some of the parameters (normally attack and release times) automatically as a function of the signal characteristics. In general, the automatic mode works well when a transparent subtle effect is looked for.
  • Side chain listen. Noise gates that provide a side chain (explained later) often allow routing of the side chain signal to the main output so that it can be listened to, thus making the setting of the gate easier.
  • Bypass. For comparison of the original and processed signals.
3.3. Indicators
  • Gating indicator. There is normally a LED indicator that shows whether the gate is open or closed. Normally, the operator also gets some kind of indication of whether the threshold level has been crossed and the release phase is finished.

3.4. The side chain
    Normally, the detection circuit uses a copy of the signal being gated to check whether it exceeds the threshold level or not. However, many compressors allow using an external signal that is fed to the detector via the Side Chain (sometimes also called key) input. That way it is the external signal that triggers the compression, though it is the main signal that gets compressed. There may be a switch that toggles the detection signal between the main and the side chain signal, or sometimes, if the side chain input uses a 1/4" connector (often wrongly referred to as jack in many non-English speaking countries!), it is the connector that enables the function when the 1/4" plug is inserted. This 1/4" jack is an insert type connector that carries both a send and a return signal, the send carrying a copy of the main signal to facilitate its connection to a processor (e.g., an equalizer) and then feeding it back to the detector through the return part of the side chain connector.
    An equalizer is commonly used for the side chain; in fact, some noise gates come equipped with built-in EQ facilities. We could, for instance, attenuate the high frequencies on the detection circuit to avoid the cymbals triggering the noise gate for the bass drum.
3.5. Setting a gate depending on the application
    Before using a gate, we need to connect it in the right place. If we use it in combination with a mixer, we will connect it to an insert point, such that will send the original signal and return the gated signal. The insert outputs are always pre-fader, which means we do not have to change the gate's threshold every time the fader position is changed. If we do not have enough gates for every channel, we could also insert a gate on mixer subgroups carrying similar signals (for instance, a brass section). If we use compressors in the same channel that has a noise gate inserted, we will insert the gate first; otherwise the reduced dynamic range of the compressed signal will make adjusting the threshold on the noise gate very difficult.
    Another recommendation is to avoid gating a signal with reverb, as the effect is normally very obvious.
    In general, the criteria in this article are given as overall guidelines and starting points, but they will depend on the specific noise gate or expander model and they may have to be fiddled with by ear.
Words may start with a soft vowel sound or, alternatively, with a percussive consonant such as "p" or "t". The attack time should therefore be fast for vocals, but not too fast that produces an audible "click" when opening the gate. Slow attack times loose the attack of the starting consonants and reduce intelligibility, making it difficult to understand the first word after a pause. A time around 1 millisecond is a good compromise solution.
As far as the decay time, around 0.5 seconds works well keep the noise under control and yet not truncate the end of the words. Sometimes the gate will tend to open and close if there is a signal whose level happens to oscillate around the threshold level. In those cases we can use the "hold time" to avoid that.
As far as the threshold, use the lowest level possible, though with the risk of background noise opening the gate, experimenting may be needed to set the optimum threshold level. 
Electric guitar
Electric guitars in general, particularly Stratocaster type ones with single coil pickups that act as aerials for the interference that might exist in the room, produce a lot of background noise. Sometimes the position and orientation of the guitarist helps minimizing the noise, though this is not easy to do in a live concert. Also, distortion and other effects count on using a lot of gain before clipping the signal, which also amplifies any hiss and hum that might have been already present. All of which makes electric guitars very good candidates for gating.
There are many types of guitars and guitar sounds, as well as many ways of playing them, so attack and releases times will vary widely. A guitar with a lot of sustain may need a decay time as long as 1 seconds, while for a clean percussive sound 50 ms may be sufficient.
When a drum kit is closed miked using individual microphones for each instrument, every channel is often gated to improve instrument separation, since instruments leak into adjacent microphones. The threshold level may have to be quite high, and we will have to be careful to to miss the quieter hits. The task is simplified if we have a gate with filtering capabilities in the detection circuit. For instance, for the bass drum not to open the snare drum's gate, we will EQ the low frequencies out on the snare drum's noise gate.
A corrective possibility on lower quality toms with a decay that is too resonant (and too ugly) is to use a decay time that accelerates the natural decay of the tom's sound.
A way to use a gate with bass guitar is to synchronize it with the bass drum, which gets connected to the side chain input. If a bass note is played ahead of time, it will not be heard until the bass drum hit comes in, such that we will increase the feeling of "playing together", though the effect is quite radical and more appropriate for studio than for live sound.
Gated reverb
This is a classic setup, so much so that most reverb units feature it and it becomes not very practical to do it with a gate. The sound has a long reverb effect added to it; the reverb tail is then abruptly cut off with closing the gate very rapidly. For this effect a hold time of around half a second will be used. The effects is easier achieved by plugging the dry drum signal into the side chain input. Phil Collins made gated reverb very popular in its day.
Nowadays the range of sounds a keyboard or tone module can produce is enormous, so it is not possible to provide general settings. Use the general guidelines and criteria that have been explained on this tutorial.