Foundations of Audio: Compression and Dynamic Processing by Brian Lee White (Lynda.com Tutorial)
A) Compressors will express the threshold and ratio relationship using an X-Y graph called a transfer curve. A transfer curve graph plots the signal's input into the compressor on the X axis and the signal's output on the Y axis, the kink in the line, is the threshold point sometimes called the knee. B) Ratio of 2:1 ( for every 2dB in 1dB out)
A) Compressors will express the threshold and ratio relationship using an _/_ graph called a t_______ curve. A transfer curve graph plots the signal's input into the compressor on the X axis and the signal's output on the Y axis, the kink in the line, is the t________ point sometimes called the k___ B) R____ of 2:1 ( for every 2dB in 1dB out)
In the case of subtraction, I think about what part of the signal's frequency response I did not like or what I'm going to want to remove before it hits the mix bus. I generally like to get rid of any frequency material I don't want before I hit my compressors or limiters. In this case if I want to do additive or boost an EQ, I might consider saving that for after any compression, as a way of restoring some of the tonal response or shape to the signal--post-dynamics processing. Therefore compress (to tighten up the mix first) unless there is something specific that I want to cutout using EQ
Compression first or EQ first?
Dynamics control of a vocal is a two stage process of tucking in the louder words and phrases and turning up the result, allowing the softer notes to sit at or near the level of the louder ones. This is achieved by pulling down threshold until the louder words start to trigger compression, about 60 dbs on average of gain reduction in this case. And then making up the loss in signal level using the output gain. Notice that I am using a ratio of about six to one, an attack of 10 milliseconds in a release time that allows the compressor to recover between words.
Describe the basic process of compressing a vocal Dynamics control of a vocal is a t__ stage process of tucking in the louder words and phrases and turning up the result, allowing the softer notes to sit at or near the level of the louder ones. This is achieved by pulling down t________ until the louder words start to trigger compression, about 60 dbs on average of gain reduction in this case. And then making up the loss in signal level using the output gain.
One of the most common uses of side chains is ducking. Think about when a radio DJ talks over the music he's playing and how the music seems to automatically turn itself down, much quicker than a human can react with their hand on the volume fader. In post-production mixing or audio for video, it's fairly common to use side-chain compressors to duck music and effects under dialogue.
Ducking is an example of what technique? Give a 2 real-life examples.
Dynamic Processors are used to control the dynamic range of a recording to enhance the highs and/or lows of your tracks to make the recording more exciting or to emphasize a particular part of a song; or to make the amplitude more consistent; to eliminate unwanted noise; or, to adjust the transients (loud sounds at the start of a sound)
Dynamic Processors are used to c______ the d______ range of a recording to e______ the highs and/or lows of your tracks to make the recording more exciting or to emphasize a particular part of a song; or to make the a________ more consistent; to eliminate unwanted n_____; or, to adjust the t_________ (loud sounds at the start of a sound)
1) a value of 0 dBFS means that threshold is sitting at the digital clipping point therefore no effect 2) a threshold setting of -20 dBFS would cause the compressor to react to any signal whose amplitude is measured at over -20 dBFS up through 0 dBFS 3) the more negative the threshold the greater the effect
Explain how the compressor would react if the threshold was set at the follow 3 values 1) 0dBFS 2) -20dBFS 3) -40dBFS
16-bit recording systems offer 96 dBs of dynamic range. In comparison, 24-bit recording offers a much wider range of 144 dBs, exceeding the range of human perception, which is generally accepted to be around 120 dBs. Because of this, 24-bit recordings can accommodate a wider range of amplitude values before digital clipping.
Explain how this diagram relates to dynamic range. 16-bit recording systems offer __ dBs of dynamic range. In comparison, 24-bit recording offers a much wider range of ___ dBs. Because of this, 24-bit recordings can accommodate a w_____ range of amplitude values before digital clipping.
The second image displays the audio wave converted into an RMS curve. Here you can see that the (uncompressed) range is from -54.0 dB to +4.5 dB or a dynamic range of 58.5 dB.
Explain the diagram and how it helps us to measure dynamic range
A gate and its close sibling the expander are used to allow the desired clean signal to pass through while removing any unwanted low-level noise, like hum from a guitar amp or bleed from a track, like the sound of a headphone mix that bleeds into a vocal mic during a vocal recording session.
Explain the function of a noise gate
Headroom in the analog world allows the signal breathing room before the circuit starts to distort. Headroom can be thought of as a safety zone, allowing transient audio peaks to exceed this average working range without exceeding the signal capabilities of an audio system--digital clipping, for example. Another way to think about this is to imagine the ceilings in your house which are set at a height to allow everyone (except maybe NBA players!) to avoid bumping their heads.
Explain the term Headroom : Headroom in the a_____ world allows the signal breathing room before the circuit starts to distort. Headroom can be thought of as a s_____z___, allowing transient audio peaks to exceed this average working range without exceeding the signal capabilities of an audio system--digital clipping, for example. Another way to think about this is to imagine the ceilings in your house which are set at a height to allow everyone (except maybe NBA players!) to avoid bumping their heads.
A ratio of 1:1 means no compression takes place
Explain this graph
Softer audio signals below -30dB are not allowed through (muted out) and signals above -30bB pass through at a ration of 1:1
Explain this transfer curve of a gate
Dynamic range is the difference between the loudest and the quietest sounds. You could say that the dynamic range is represented by the red line shown above. However, if you tried to analyse the signal above by finding the average volume, you would end up with an average of 0 because the wave spends as much time about the 0 mark as below it. For this reason, dynamic processors often use the RMS (Root Mean Squared) measured in dB with 0 as one point of reference and -60 dB as another to determine the average loudness of a signal.
Explain what RMS means and how it relates to dynamic range. If you tried to analyse the signal above by finding the average volume, you would end up with an average of _ because the wave spends as much time about the _ mark as below it. For this reason, dynamic processors often use the RMS (R___ M___ S______) measured in dB with 0 as one point of reference and -60 dB as another to determine the average loudness of a signal.
Side-chaining is the technique of controlling one's signal dynamics using another signal's interaction with the compressor's threshold. Instead of the compressor listening and reacting to the input signal, the processor is told to listen to another signal. Ducking (DJ Effect), Pumping (EDM FX of creating a pumping synth part controlled by a kick drum) and de-esser are 3 examples of side-chaining.
Explain what side-chaining is and give 3 examples. Side-chaining is the technique of c-___________ one's signal dynamics using another signal's interaction with the compressor's t________. Instead of the c__________ listening and reacting to the input signal, the processor is told to listen to another signal.
Gates with ratio controls are called expanders. Instead of completely cutting off a signal below the threshold, expanders simply reduce the volume of the signal below the threshold by the specified ratio. In this case they're truly a reverse compressor because they expand the dynamic range of the signal instead of reducing it. In an expander, the signals that passed over the threshold are unaffected and the signals that fall under the threshold are attenuated. This creates a greater dynamic range between the softest and loudest parts of a signal. Expanders are perfect for scenarios when you don't want to kill the bleed but just want to turn it down a bit or separate it more from the main signal.
Gates with r____ controls are called e________. Instead of completely c______ off a signal below the threshold, e________ simply reduce the volume of the signal below the threshold by the specified r____ In this case they're truly a r______ compressor because they expand the dynamic range of the signal instead of r_______ it. In an expander, the signals that passed over the threshold are unaffected and the signals that fall under the threshold are attenuated. This creates a greater dynamic range between the softest and loudest parts of a signal. Expanders are perfect for scenarios when you don't want to kill the b____ but just want to turn it down a bit or separate it more from the main signal.
Your ears are your best guide. The attack and release settings are based on the signal you are trying to process, but at a minimum, the attack should be set fast enough to grab hold of a signal before it completely decays, and the release should be quick enough for the compressor to recover before the next note or beat. Remember, fast and slow in terms of attack and release is relative to the signal you're processing. A 50-millisecond attack on a vocal or bass may work great but do absolutely nothing on a snare- or bass-drum hit that lasts only 100 milliseconds in total.
How do you work out what the optimal attack and release times should be on a compressor? (Part 1) : Your e___ are your best guide. The attack and release settings are based on the signal you are trying to process, but at a minimum, the a______ should be set f___ enough to grab hold of a signal before it completely decays, and the r_______ should be quick enough for the c________ to recover before the next note or beat.
Shorter attack times control transients and plosives, but settings under 1 millisecond can take the bite or brightness out of signals.Longer attack times allow more of the signals transient to get through. This helps add bite or punch to drums as the transient passes through uncompressed-- 1 to 10 milliseconds on drums or 10 to 50 milliseconds on non-percussive instruments.
How do you work out what the optimal attack and release times should be on a compressor? (Part 2) : Shorter attack times control t________ and plosives, but settings under 1 millisecond can take the bite or brightness o___ of signals.Longer attack times allow more of the signals transient to get through.
Multiband compressors are extremely powerful tools, giving the engineer total control over both the dynamic and frequency response of a signal within the same processor. Multiband tools are most useful when processing individual tracks or small groups When I do use multiband compression on my entire mix, I often disable most of my bands to focus my work on a specific range of frequencies, like the top end or sub bass. Be careful not to flatten your mix out.
How should you use a multiband compressor and what should you be cautious about? : One compressor might be working on the b___ notes of a signal, while another works on the mids and h_____________ content.
Digital systems remain linear in the response, all the way to the clipping point at 0 dBFS. While tape tube and other analog gear might push back and introduce natural compression to a signal as its headroom is breached into distortion, a digital system will simply clip the signal hard, creating a nasty square wave out of your waveform peaks.
In reference to the term Headroom how do Analogue and Digital recording systems react differently i.e as volume levels increase and signals near the point of distortion what happens in an analogue setup compared to a digital .
1) If compression isn't controlling a tracks dynamics, try automating its volume 2) Use compression creatively as well as for utility 3) Check your work against reference mixes
List 3 final tips to consider when using compression
1) Control transients, don't eliminate them 2) Create punchiness, don't completely squash your tracks just for more loudness 3) Be careful not to flatten out a tracks EQ response by over compressing 4) Make use of the sonic differences between compressors 5) Rely on your ears , not presets
List 5 tips to consider when using compression
compressors, limiters, expander gates, de-essers, and transient shapers.
List 5 types of dynamic controllers
Two of the most common vintage-modeled compression plug-ins are the Universal Audio 1176 and the Teletronix LA-2A. Many companies have modeled these classics into plug-ins. Waves CLA compressors as an example of the UA 1176 and LA-2A. The 1176 is a classic example of a FET, or Field Effect Transistor, compressor.
Name 2 famous vintage model compressors that have been emulated as software plug-ins. U________ A______ 11__ and the T_________ L___
Drums and percussion. The compressor reacts immediately and it gives greater control. Great for controlling snare and kick.
On what instruments would you use a hard knee?
Soft-knee compression is great on less percussive material like vocals, bass, and guitar, where you might not want to hear as much of an edge, but still have a nice firm signal sitting in the mix. Soft-knee compression can also work well on full mixes, where you're trying to subtly glue signals together without a lot of attitude from the compressor.
On what instruments would you use a soft knee?
A compressor's attack and release controls determine the reaction time of any gain reduction once the threshold is breached. When a signal is determined to be too loud and shoots over the threshold, the attack time is how long it takes for the compressor to grab hold of the signal and turn down the volume.
Refer to the diagram and explain how the attack parameters on a compressor work.
A compressors release time is when that signal falls back below the threshold and stops being too loud, the release time is how long it takes for the compressor to let go of the signal and return it to its uncompressed level.
Refer to the diagram and explain how the release parameter on a compressor works.
I like to use my de-essers before applying compression and EQ. So I can get the bad stuff out of the way before it hits my other processors. (Take a listen to this passage as I dial in too much full range de-essing, giving the singer a lisp... see Lynda.com Get in the Mix: De-essing a vocal track)
Should you insert the de-esser before or after compression and EQ?
In a digital audio system, values above 0 are simply cut off, or clipped, creating digital distortion; thus there is no higher value than 0.
Sometimes the dBFS can confuse people because it counts up from negative numbers and ends at 0. Why does it do this?
Turning up an audio signal's volume by 1dB is basically imperceptible. For most of us, it takes at least a 3dB increase for us to notice it. However, because of its logarithmic scaling, the loudness escalates quickly. A 10dB increase represents a perceived doubling of loudness, and an increase of 20 dB is about four times as loud.
The decibel scale is a logarithmic measurement scale. Explain ...
Inside our digital audio workstation, we will use dBFS, or Decibels Full Scale. This scale defines our system's maximum and minimum amplitude values, otherwise known as dynamic range. This helps us measure and control the relative amplitude of audio signals within that range.
There are many ways to measure amplitude. Inside our digital audio workstation, we will use dBFS. Explain the term dBFS scale
A soft knee is a more gradual form of compression where the signal is eased into the compressor's ratio setting over a larger threshold range. In a soft-knee compression curve, the compressor would begin to attenuate the signal earlier, at around -30, with a ratio of 2:1. Maybe you're amping up to 4:1 at -25, 8:1 at -20, and finally reaching 10:1, just pass the threshold, at around -15 dBFS.
This is a diagram of a soft knee with a threshold setting of -20dB and a ratio setting of 10:1. Explain : A s___ knee is a more gradual form of c___________ where the signal is eased into the compressor's r_____ setting over a larger t_________ range. In a soft-knee compression curve, the compressor would begin to attenuate the signal e_______, at around -30, with a ratio of 2:1.
Two of the most common vintage-modeled compression plug-ins are the Universal Audio 1176 and the Teletronix LA-2A. Many companies have modeled these classics into plug-ins. Waves CLA compressors as an example of the UA 1176 and LA-2A. The 1176 is a classic example of a FET, or Field Effect Transistor, compressor.
Two of the most common vintage compressors are the U________ A____ 11__ and the T________ L_-2A. name one company that has emulated these compressors and give the name of the software plug-in.
Insert a limiter before the compressor to create a layered-compression signal flow. The limiter will be set with a higher threshold to catch and tame the hot spots of a vocal for example, while the compressor will work more on the true average level of the signal.
You could insert a limiter, before a compressor on a vocal track. This is an example of layered compression. Explain what this meant by layered compression and why you might do this.
In simple terms, gain reduction tells us how much the compressor is turning down our signal, measured in decibels. The further to the left the needle goes (i.e. the more into the negative numbers) the more the signal is being compressed and the greater the gain redcution
What does the gain reduction meter tell you?
Limiter (sometimes reffered to as 'blockwall' limter). Ratio = infinity : 1 i.e. any value inputted will result in a fixed output value. Ratios of 10:1 or more normally refer to limiters
What dynamic controller is represented by this graph. Explain.
Bus compression is the practice of using a compressor on a group of tracks sub-mixed together, as opposed to only compressing each track individually. For example, I may use a compressor on my kick, snare, toms, and overhead individually, but I might also use a compressor on the entire drum sub mix to achieve an effect that's not possible through adjusting the individual compressors in isolation. Bus compressors are generally associated with the phrase "gluing the mix together" . ...if you're going use a master bus compressor over your entire mix, use a light touch. ratio 2:1 or 4:1
What is Bus Compression and why is it used?
Parallel compression is the process of combining an uncompressed version of a signal with a compressed version and blending to taste. The peaks of the uncompressed wave form are preserved, while the body of the signal is raised. For example to make it cut through the mix without flattening out the transients. This trick works on almost any kind of material where you want the benefits of compression without a lot of the artifacts. Try it on vocals, guitars, drums, or even entire mixes. Try automating the level of the compressed track, up and down throughout a mix.
What is Parallel Compression and why is it used? Parallel compression is the process of c________ an uncompressed version of a signal with a compressed version and blending to taste. The peaks of the uncompressed wave form are preserved, while the body of the signal is raised. For example to make it cut through the mix without f_________ out the t_________. This trick works on almost any kind of material where you want the benefits of compression without a lot of the artifacts. Try it on vocals, guitars, drums, or even entire mixes. Try automating the level of the compressed track, up and down throughout a mix.
Speech example : There's a girl named Sally who sells seashells ...if we want to engage the compression only when the hard S of a vocal is detected, but not anywhere else? That is the main purpose of the de-esser. De-essers are frequency- specific compressors because they're set up to react only to the amplitudes in a specific range of frequencies (above 5K) in a signal, rather than reacting to all the signal's amplitude uniformly.
What is a de-esser and give an example of how it might be used.
A multiband compressor follows the same idea as a de-esser, but works across the entire frequency range. It's simply multiple single-band compressors, each working on their own dedicated frequency range. One compressor might be working on the bass notes of a signal, while another works on the mids and high-frequency content. Each band will have its own threshold, ratio, attack, release, and makeup gain, and in most multiband processors, these controls can be linked together for ease of use.
What is a multiband compressor? A multiband compressor follows the same idea as a d-_______, but works across the e_____ f________range.
The secret to using compression presets is to understand that the preset has no idea what kind of signal level is coming in. Therefore, the default threshold and makeup gain settings are almost always going to be completely wrong, either giving you too much compression or too little compression. Therefore we start by adjusting our threshold and makeup gain to suit our signal's input level
When using presets what 2 parameters should you set to z___ and adjust m_______ : The secret to using compression p_____ is to understand that the preset has no idea what kind of signal level is coming in. Therefore, the d______ threshold and m_____ gain settings are almost always going to be completely wrong, either giving you too much compression or too little compression. T