Voice equalization:

Often misunderstood by Hams, equalizers (EQ) should only be used judiciously to achieve improved transmit audio reinforcement. As such, knowing how to properly EQ your transmit audio is one of the most critical tasks to master.   From correcting problems and enhancing your sound to adding cohesion to your voice, there’s a lot you can accomplish with proper equalization.  Excessive EQing should be avoided!   Don't ruin a great sounding voice and microphone with heavy-handed EQing!  Concentrate on achieving great sounding audio at the source, and you will achieve far better results. This means choosing a high-quality microphone, minimizing shack noise, and employing the proper microphone technique.  As such, understanding how to use it to your advantage can greatly enhance your sound.  Learn to "work your mic by adhering to these techniques:

1. Placement of the microphone, relative to your mouth, plays a large role in the clarity and character of your voice.  Experiment with mic placement.  A good starting point is 3 - 5 inches.

2. Avoid lateral movements to either side of the microphone. Generally, it is necessary to remain "on-axis" (in front of the microphone) to ensure a clear tone. 

3. It is preferable to remain the same distance from the microphone to ensure a consistent volume.

4. Consider proximity effect whereby base sounding tones are enhanced by "close talking" a directional microphone, the type most hams use.  Be careful doing this as it may make you more prone to "popping your Ps" when a burst of air from your mouth overloads and distorts the microphone. Popping occurs mostly on "plosives" (words that begin with "p," "b," and "t.") A windscreen or pop filter is a useful deterrent.

Follow these techniques, and you will sound better and appear more experienced. While equalization can do wonders, it’s important to consider the bigger picture every time you reach for the EQ. 
An equalizer (EQ) is a filter that allows you to adjust the level of a frequency, or range of frequencies, of a human voice audio signal. In its simplest form, an EQ will let you turn the treble and bass up or down, allowing you to adjust the coloration of your transmit or receive audio. Equalization is a sophisticated art. Good equalization is something to strive for. 
Parametric EQ 
The parametric EQ is the most common equalizer found because it offers continuous control over all parameters. A parametric EQ offers continuous control over the audio signal’s frequency content, which is divided into several bands of frequencies (most commonly three to seven bands).  A fully parametric EQ offers control over the bandwidth (basically, the range of frequencies affected), the center frequency of the band, and the level (boost/cut) of the designated frequency band. It also offers separate control over the Q, which is the ratio of the center frequency to the bandwidth. A semi-parametric EQ provides control over most of these parameters but the Q is fixed.
Q is the ratio of center frequency to bandwidth, and if the center frequency is fixed, then bandwidth is inversely proportional to Q—meaning that as you raise the Q, you narrow the bandwidth. In fully parametric EQs, you have continuous bandwidth control and/or continuous Q control, which allows you to attenuate or boost a very narrow or wide range of frequencies.
A narrow bandwidth (higher Q) has obvious benefits for removing unpleasant tones. Let’s say you have a particularly annoying nasal quality to your audio.  With a very narrow bandwidth, you can isolate this one frequency (usually around 650) and remove, or reject, it. This type of narrowband-reject filter is also known as a notch filter. By notching out the offending frequency, you can remove the problem without removing the instrument from the mix. A narrow bandwidth is also useful in boosting pleasant tones as well.
A broad bandwidth accentuates or attenuates a larger band of frequencies. The broad and narrow bandwidths (high and low Q) are usually used in conjunction with one another to achieve the desired effect.
A shelving EQ attenuates or boosts frequencies above or below a specified cutoff point. Shelving equalizers come in two different varieties: high-pass and low-pass.
Low-pass shelving filters pass all frequencies below the specified cutoff frequency while attenuating all the frequencies above it. A high-pass filter does the opposite: passing all frequencies above the specified cut-off frequency while attenuating everything below.

Note:  Once popular graphic EQs use sliders to adjust the amplitude for each frequency band.  K4QKY does not recommend their use in audio processing. 

One of the easiest ways you can clear up your mix and reclaim a large amount of wasted headroom is by applying a high pass (low-cut) filter since extremely low voice frequencies do not contribute to effective, clean and pleasant sounding transmit audio. 
Cut First, Boost Second
Before you boost what you want to hear, cut out what you don’t want to hear. The best reason for doing this is to remove problem frequencies from your particular voice profile. Once you achieve this goal, you’ll find you often don’t need to boost much else.  Many hams will resist doing this for the commonly held belief that "more is better".  In short, boost only as necessary and always with care.   As you adjust the EQ, you’ll notice that frequency boosts are significantly easier to hear than cuts. This phenomenon causes many hams to boost frequencies they want to bring out, rather than to cut problem frequencies. There are two major issues associated with doing this. First, if you boost at 3kHz to achieve greater presence, your audio will likely become harsh and cutting. The other problem is that if you boost all of the frequencies around a problem frequency rather than simply cutting the problem frequency (like boosting the extreme lows, upper midrange, and high end instead of just cutting the lower-mid which is really the issue), you can easily overload the EQ gain stage and introduce distortion that you may not initially notice.
Understand the frequency spectrum of your voice

Your vocal tone needs to be as perfect as possible. That’s because as humans, we can’t help but scrutinize what we hear in an extremely critical manner. Trouble is that hams often disagree with what constitutes ideal sounding transmit audio.  So, develop your own style as you see fit but try to avoid unpolished or harsh sounding audio that will likely annoy and distract your listeners.  Here are a few suggestions for properly EQing your voice:

  • Body (200–500Hz)

    This frequency range is where muddiness lives, but it’s also where the warmth of your voice comes from. If your vocals sound mushy, try cutting low frequencies in this range. If your vocals are clear but lacking warmth, try boosting in this range.

  • Nasal (1-1.5kHz)

    Almost universally, 1-3kHz is where the nasal frequencies lie. Try cutting somewhere within this frequency range. Don’t go overboard though.

  • Presence (1.5 to 3kHz)

    When it comes to intelligibility, presence is absolutely critical but be careful boosting too much as this can render your vocals harsh and jarring.

EQ with Your Ears

It is important to point out that the best tools you have for EQing are your ears. You can memorize tables of important frequencies for all kinds of instruments and applications, but the most important thing is that your voice sounds great. The best way to evaluate your sound is to listen to yourself from your rig's built-in monitor or, better yet, on air from a separate receiver, perhaps an SDR.

The bottom line for EQing your voice is to find the biggest offender and fix that first. Then, boost sparingly to polish the results.  If this technique fails, then consider reevaluating the quality of your microphone and the correctness of audio level settings in the audio chain.   Remember that an EQ can't fix poor unprocessed input.   So before you resort to the EQ, listen closely to that input to avoid falling victim to the "Garbage in... garbage out" syndrome.

Dynamic range is the ratio between the loudest possible audio level and the lowest possible level.  For example, if an audio processor states that the maximum input level before distortion is +24 dBu, and the output noise floor is -92 dBu, then the processor has a total dynamic range of 24 + 92 = 116 dB.  The average dynamic range of an orchestral performance can range from -50 dBu to +10 dBu, on average. This equates to a 60 dB dynamic range.   Although 60 dB may not appear to be a large dynamic range, you’ll discover that +10 dBu is 1,000 times louder than -50 dBu!

Equalizers can't fix the character of your voice transmission but they can help emphasize the good stuff and help minimize the bad stuff... but only if it’s of the highest quality in the first place.


Do we need compression for voice processing?  

The average dynamic range of an uncompressed vocal is around 40 dB.   In other words, a vocal can go from -30 dBu to +10 dBu.  The passages that are +10 dBu and higher will be heard over prevailing noise.  However, the passages that are at -30 dBu and below will never be heard over the roar of the noise. A compressor can be beneficial in this situation to reduce (compress) the dynamic range of the vocal to around 10 dB. The vocal can now be placed at around +5 dBu.  At this level, the dynamic range of the vocal is from 0 dBu to +10 dBu. The lower level phrases will now be well above the lower level of noise, and louder phrases will not overpower the noise, allowing the vocal to “sit above the noise.”

Can we have too much compression? 

Over-compression often sounds horrible.  That statement can be qualified by defining over compression. The term itself is derived from the fact that you can hear the compressor working. Therefore, the over compressed sound is likely to have been caused an improper adjustment of the compressor.  Most importantly, a well-designed and properly adjusted compressor should never be audible to other hams!  

So, what constitutes compression/limiting?

Punch, apparent loudness, presence—these are just three of the many terms used to describe the effects of compression/limiting.  Compression and limiting are forms of dynamic-range (gain) control. Audio signals have very wide peak-to-average signal-level ratios (sometimes called dynamic range, which is the difference between the loudest level and the softest level). The peak signal can cause overload in the audio-recording or sound-reinforcement chain, resulting in signal distortion.

A compressor/limiter is a type of amplifier in which gain is dependent on the signal level passing through it. You can set the maximum level a compressor/limiter allows to pass through, thereby causing automatic gain reduction above some predetermined signal level, or threshold. Compression refers, basically, to the ability to reduce, by a fixed ratio, the amount by which a signal’s output level can increase relative to the input level. It is useful for lowering the dynamic range of a vocal, making it easier to be heard over the air without distortion.  

Take, for example, a ham who moves around in front of the microphone during a QSO, making the output level vary up and down unnaturally. A compressor can be applied to the signal to help correct this phenomenon by reducing the louder passages enough to be compatible with the overall signal.

How severely the compressor reduces the signal is determined by the compression ratio and compression threshold. A ratio of 2:1 or less is considered mild compression, reducing the output by a factor of two for signals that exceed the compression threshold. Ratios above 10:1 are considered hard limiting.

As the compression threshold is lowered, more of the input signal is compressed (assuming a nominal input-signal level). Care must be taken not to over compress a signal, as too much compression destroys the acoustic dynamic response.

Limiting refers to the processing that prevents the signal from getting any louder (that is, it prevents an increase in the signal’s amplitude) at the output.

Vocals usually have a wide dynamic range. Transients (normally the loudest portions of the signal) can be far outside the average level of the vocal signal. Because the level can change continuously and dramatically, it is extremely difficult to ride the level with a console fader. A compressor/limiter automatically controls gain without altering the subtleties of the transmission.

Compressor terminology

Threshold. The compressor threshold sets the level at which compression begins. When the signal is above the threshold setting, it becomes eligible for compression. Basically, as you turn the threshold knob counterclockwise, more of the input signal becomes compressed (assuming you have a ratio setting greater than 1:1).

Ratio. The ratio is the relationship between the output level and the input level. In other words, the ratio sets the compression slope. For example, if you have the ratio set to 2:1, any signal levels above the threshold setting will be compressed such that for every 1 dB of level increase into the compressor, the output will only increase 0.5 dB. As you increase the ratio, the compressor gradually becomes a limiter.

Limiter. A limiter is a compressor that is set to prevent any increase in the level of a signal above the threshold. For example, if you have the threshold knob set at 0 dB, and the ratio turned fully clockwise, the compressor becomes a limiter at 0 dB, so that the output signal cannot exceed 0 dB regardless of the level of the input signal.

Attack. Attack sets the speed at which the compressor acts on the input signal. A slow attack time allows the beginning envelope of a signal (commonly referred to as the initial transient) to pass through the compressor unprocessed, whereas a fast attack time immediately subjects the signal to the ratio and threshold settings of the compressor.

Release. Release sets the length of time the compressor takes to return the gain reduction back to zero (no gain reduction) after the signal level drops below the compression threshold. Very short release times can produce a very choppy or “jittery” sound, especially in low-frequency instruments such as a bass guitar. Very long release times can result in an overcompressed sound; this is sometimes referred to as “squashing” the sound. All ranges of release can be useful at different times, however, and you should experiment to become familiar with the different sonic possibilities.

Hard/Soft Knee. With hard-knee compression, the gain reduction applied to the signal occurs as soon as the signal exceeds the level set by the threshold. With soft-knee compression, the onset of gain reduction occurs gradually after the signal has exceeded the threshold, producing a more musical response (to some folks).

Auto. Places a compressor in automatic attack and release mode. The attack and release knobs become inoperative and a preprogrammed attack and release curve are used.

Makeup Gain. When compressing a signal, gain reduction usually results in an overall reduction of the level. The gain control allows you to restore the loss in level due to compression (like readjusting the volume).


Noise Gate

Do we also need a noise gate?

Problems sometimes arise when shack background noise (air conditioner, linear amp fan, etc.) become more audible after the lower end of the dynamic range is raised.  This calls for the use of a noise gate. The noise-gate threshold could be set at the bottom of the dynamic range of the vocal, say -10 dBu, such that the gate would shut out the unwanted signals between the phrases.

What is noise gating?

Noise gating is the process of removing unwanted sounds from a signal by attenuating all signals below a set threshold. As described, the gate works independently of the audio signal after being “triggered” by the signal crossing the gate threshold. The gate will remain open as long as the signal is above the threshold. How fast the gate opens to let the “good” signal through is determined by the attack time. How long the gate stays open after the signal has gone below the threshold is determined by the hold time. How fast the gate closes is determined by the release. How much the gate attenuates the unwanted signal while closed is determined by the range.

Noise gates were originally designed to help eliminate extraneous noise and unwanted artifacts from a recording, such as hiss, rumble, or transients from other instruments in the room. Since hiss and noise are not as loud as the instrument being recorded, a properly set gate will only allow the intended sound to pass through; the volume of everything else is lowered. Not only will this strip away unwanted artifacts like hiss, but it will also add definition and clarity to the desired sound. This is a very popular application for noise gates, especially with percussion instruments, as it will add punch or “tighten” the percussive sound and make it more pronounced.

Noise gate terminology:

Threshold. The gate threshold sets the level at which the gate opens. Essentially, all signals above the threshold setting are passed through unaffected, whereas signals below the threshold setting are reduced in level by the amount set by the range control. If the threshold is set fully counterclockwise, the gate is turned off (always open), allowing all signals to pass through unaffected.

Attack. The gate attack time sets the rate at which the gate opens. A fast attack rate is crucial for percussive instruments, whereas signals such as vocals and bass guitar require a slower attack. Too fast of an attack can, on these slow-rising signals, cause an artifact in the signal, which is heard as a click. All gates have the ability to click when opening but a properly set gate will never click.

Hold. Hold time is used to keep the gate open for a fixed period after the signal drops below the gate threshold. This can be very useful for effects such as gated snare, where the gate remains open after the snare hit for the duration of the hold time, then abruptly closes.

Release. The gate-release time determines the rate at which the gate closes. Release times should typically be set so that the natural decay of the instrument or vocal being gated is not affected. Shorter release times help to clean up the noise in a signal but may cause “chattering” in percussive instruments. Longer release times usually eliminate “chattering” and should be set by listening carefully for the most natural release of the signal.

Range. The gate range is the amount of gain reduction that the gate produces. Therefore, if the range is set at 0 dB, there will be no change in the signal as it crosses the threshold. If the range is set to -60 dB, the signal will be gated (reduced) by 60 dB, etc.