Recording Part 6 – Dynamics

The control of dynamics in a recording can be tricky. As guitarists we all know about dynamics – some passages need to be played more softly, others more loudly, etc…, but in a recording you don’t necessarily want those quiet passages to get lost by being ‘too’ quiet, and you also may want your track to be as loud as possible without losing any clarity or nuance (or you’re OK with finding a happy medium), and these issues are what we refer to in recording as ‘dynamics’.

I’m going to try to break it down here without getting too tech-y or using too much jargon, though there may still be a few unfamiliar words, which I will try to define as best I can. At the end of the article you’ll find some waveforms to look at and MP3s to listen to hear the effects of compression.


The Simplest way to make your recording as loud as possible without in any way affecting the dynamic range of your recording is through a process called normalizing. Normalizing looks at the peak level (loudest point) in your track and makes that as loud as possible without distorting it in any way, and brings the entire track up in gain the same amount. To keep things simple I’m going to talk in percentage terms (very un-technical). Let’s say that your peak level was 60% as loud as it could have been without distorting – you’d want to bring the track up another 40% percent to make it as loud as possible. So normalizing would bring up the entire track by 40%. Relative dynamics would remain exactly as they were, but the overall volume of your track would be 40% higher. Just of like turning up the volume knob on your stereo.

One thing to keep in mind, is that normalizing will bring up the sound of your computer’s fan, noise generated by your recording system, or anything else you don’t want to hear in you recording, by whatever level that peak level is increased (40% in our example), which is why you want to record nice healthy levels that don’t peak. You don’t want to just record a low level with the thought that you will normalize it later.  See my article on setting levels for more on this –

Normalizing is generally fine with a finished product – your solo track or a mix – but normalizing every track in a mix is not recommended, because you are just raising the level of noise in every track, and that noise can add up and suddenly become very audible.

Normalizing is a product of the digital age, and is always done with a plug-in. Whether you work in Garage Band, Pro Tools, Logic or any other DAW (audio recording program), you will find a plug-in for normalizing as part of your included suite of plug-ins.

N.B. – Call me paranoid, but I never normalize to 100%. It just feels too close to clipping for comfort, so when I normalize I generally do it to about 98% or less. This is possibly irrational when dealing with a final mix, but I can’t help it.


As I mentioned, normalizing does nothing to change the relationship between your loud moments and your quieter moments, so you may end up with a track with great sounding fortes and practically inaudible pianos. This is where compression comes in. Compression or limiting can reduce the volume level between your fortes and your pianos, making the overall level of a track louder, which may or may not be desirable. I should say that the common wisdom is that compression is not used on classical recordings precisely because classical players and listeners appreciate dynamic range (the difference between the loudest and softest moments), but the reality is that sometimes if you want your track to be audible on wimpy little laptop speakers you may wind up compressing your tack just to be heard. Flamenco tracks are compressed more often than classical tracks.

The simplest way to understand compression is to think of it as making your peaks (loud moments) quieter. This in turn makes your quietest moments louder in comparison. Most compressors can also then increase the overall gain (level) of the track, giving you a louder track with a more limited dynamic range. Compression comes at a cost, however – to some degree this compression will affect the quality of that audio that is compressed (the louder bits). In most pop and rock music compression is part of the sound we expect to hear. In critical recordings, or recordings of solo instruments, you probably want to find that line where the compression reduces your dynamic range while affecting the quality of your sound as little as possible. Understanding how a compressor works will help you find that point.

Compressors generally have a few basic controls that we’ll take a look at here:

Threshold – simply put, this determines how much of your audio will be affected. It’s called threshold because it is the threshold above which audio will be compressed. The lower the threshold, the more audio will be compressed. In many cases you can set a relatively high threshold and only affect your loudest peaks, and thus leave the bulk of your track untouched. The threshold is generally expressed in dB (decibels), so just know that the lower the threshold the less audio will be affected. You can use your ears to hear when it’s too high or low, or you can look at how the waveforms in your DAW change when you apply compression with various thresholds.

Setting a really high threshold can also be very useful if you just have a few stray peaks that are too loud but you otherwise want to leave your track untouched, and you may get a lot of extra level out of a clean track track by just taming a few really hot peaks.

Ratio – The ratio is basically how much the signal will be compressed. Ratios are expressed as 2:1 or 4:1 up to ∞:1. A 4:1 ratio means that for every four dB above the threshold you have set, only one dB will be allowed to pass over the threshold (so 8 dB over gives you 2 dB over threshold, etc…). You don’t need to understand decibels to get this, you just need to know that the higher the ratio, the more you will hear the effects of compression (which is generally undesirable in classical music). I suggest that if you want transparent (inaudible) compression, stay at or below a ratio of 3:1.

Attack – The attack is how long it takes to reduce the signal by the desired amount, and is expressed in milliseconds or seconds. I find these numbers useless, as I can’t wrap my head around milliseconds, and I have one compressor that simply goes from ‘fast’ to ‘slow’. You may want to experiment to see how audible the difference is to you. If your priority is to reduce the dynamic range as much as possible at all costs I’d play with a faster attack, and if your priority is to make everything sound as natural as possible I’d audition the slower attacks.

Release – Release is the flipside of attack, as it represents how long it takes for the compressor to ‘let go’ of the signal once the signal itself falls below the threshold you’ve set. Again, this is often expressed in milliseconds, so use your ears to judge. Generally a slower release sounds more natural, as the audio level doesn’t just pop back up.

**I should point out that not all compressors give you control over attack and release times. Some classic compressors have fixed or program-dependent attack and release times, so once again I recommend you just listen to see if you like what’s happening to your track.

Soft Knee / Hard Knee – This is a feature that some compressors have that let’s you decide if the compression will look more like a step or more like a curve. A hard knee jumps straight to your full compression ratio (4:1, for example), and a soft knee ramps up to it. Soft knees tend to sound more natural, and hard knees can be useful in when mixing drums or generally in a mix where you may not hear every instrument in complete detail.

Makeup Gain – Because compression reduces the peaks in your track, you inevitably wind up with a quieter track, and most compressors have what is called ‘makeup gain’ to simply increase your overall level after compression. Most compressors add a little to a lot of ‘color’ to your track, and if this particular ‘color’ sounds good to you, you can even use a compressor’s makeup gain without compressing the signal at all. If you are not after the ‘color’ the compressor offers, you can simply take your compressed track and increase the gain with a transparent gain plug-in or even normalize the compressed track.

Limiting – Limiting is a form of compression with a very high ratio (10:1 and up) that pretty much doesn’t allow the signal to pass above the threshold without much regard to how this affects the audio. Limiting has its uses, but not necessarily in the quest for natural sounding solo tracks.

**I should finally add that recording engineers commonly use compression when recording, but when it comes to critical recordings or acoustic instruments that are meant to sound as natural as possible I advise against this. Once you’ve recorded a track with compression you can’t un-compress it, so there’s very little upside but a possibly fatal downside to recording this way (and in any case most home digital interfaces won’t make it easy to do, if it’s possible at all). Many musicians use compression as part of their sound (remember that many compressors impart some ‘color’ to the sound), but classical guitarist are generally not among those who do. I’ve limited myself here to the use of compressors as a way of taming the dynamics of an existing track and making that track louder, but compressors can also be used as an effect, and that is a whole other story.

Following are some examples of a track that has been recorded with a healthy level, then normalized to make it as loud as possible without reducing dynamic range, then compressed a little, and then a lot. You’ll probably hear the difference right away, though depending on your speakers they may all sound distorted or compressed. If this is the case try listening on headphones.

Here’s the unprocessed audio – Clean

Here’s the same track normalized – Normalized

Here’s that track compressed with a high threshold and a 3.3:1 ratio (less compression) – 3.3 ratio

And here’s the track with a lower threshold and a higher ratio (more compression) – High Ratio

© Kai Narezo, All Rights Reserved

Both comments and pings are currently closed.

7 Responses to “Recording Part 6 – Dynamics”

  1. Humberto Santamaria says:

    Very good, thanks. PS. where may I go to see the previos 5 articles
    Regards. Hto

  2. Johnny Longrifle says:

    Thank you . . . Enjoyed and learned much from the series so far .


  3. Tom says:

    “Once you’ve recorded a track with compression you can’t un-compress it, so there’s very little upside but a possibly fatal downside to recording this way (and in any case most home digital interfaces won’t make it easy to do, if it’s possible at all).”

    I don’t understand this. Is it about hardware compression? Why would someone want to add compression to a recording ‘destructively’ (meaning irreversibly)?

    • Kai says:

      Hey Tom – yes, this is generally done with an outboard hardware compressor, and is actually quite common for taming peaks or achieving a certain sound on vocals, drums and a lot of other instruments. In many cases this is done when you know that the end mix will be compressed anyway and that is part of the sound you’re going for. For guitars I never compress while tracking precisely because I can’t undo it, so if I need compression I’d rather use it after the fact when I have time to hear if the results are what I want.

  4. Tom says:

    Hi Kai – thanks for your reply. Hardware compression during a recording is a contentious issue and with good reason outside of exceptional circumstances. I guess also that some might prefer the sound of hardware, over software, compression but that would be a very subjective choice and, for many, not worth the downside.

  5. Ricardo Giuffrida says:

    Hi Kai, thanks for the article. What kind of equipment are you using? Which preamp, mics, converter and compressor are you using? Best Regards. Ricardo Giuffrida.

  6. Ivan C. says:

    Thank you once again guys! this is another awesome series… really appreciate it.


Change language: