When mixing digitally, which is often the case in a home recording studio, there are a few things that you can do to get your mixes to sound like the great mixes of the good ol’ days (when everything was analog). Before we can explore the options of warming up your mix it’s essential to understand how digital and analog mixing differ.

The first step to understanding the difference is to define actually what is “analog warmth”? Sound has always been one of the hardest things to describe in words, and warmth can be interpreted in numerous ways, but I believe that the following description is what most people are talking about when describing “a warm sound”.

At the most basic level, warmth is distortion. We call it distortion because warmth actually changes the waveform of the sound, thus distorting it from the actual natural waveform. Warmth or color is a particular type of distortion that happens naturally inside analog gear. When sound waves are converted to electrical waves using a transducer (microphone, generally in the audio engineering realm) it is never a perfect replication. This is what spurns frequency response curves, and why a condenser microphones sound significantly different to dynamic microphones, and why Neumanns sound different than SM58s.
So microphone circuitry is the first stage of analog warmth. Historically, certain microphone types have become very famous because of the desirable and enjoyable warmth or color they add to a sound signal.

The next stage is the mic pre-amp. This is often a critical part of the chain as it is the first time the engineer has control over the signal (aside from mic choice and placement). It is always part of the recording chain (meaning that once you record it, it cannot be undone). This is why people will spend big dollars on a mic-pre that suits their taste and style. Since many people tend to like tube or valve mic-pres now would be a good time to talk about tubes.

Tubes (specifically triodes) are often used in pre-amps, in mic-pres, guitar amps, bass amps, even some mixing consoles use tubes in their circuitry. This is because tubes have a distortion characteristic that is very pleasing to the ear. Simply, when a tube distorts the distortion manifests itself primarily in the second and third harmonics of the fundamental frequency. That is, if the note of the sound entering the circuitry is an E the distortion created around that sound will primarily sound like an E an octave above, and a B above that (so the distortion is primarily an octave and an octave plus a fifth, both of which are consonances with the fundamental). OK, so maybe that wasn’t so simple, how about this: the distortion from tubes sounds good with the original signal.

BJTs and MOSFETs, generally called “solid state” are not quite as nice when they distort. The distortion created by these devices typically resides in third and higher odd fundamentals (5, 7, 9…). Musically this is not as desirable. The distortion in the third harmonic (an octave plus a fifth) is often considered a consonance, but the higher order odd harmonics are not considered consonant, and therefore add tension to the signal. This tension is often described as “harsh” or “gritty.” The advantage of solid-state circuitry is that it is considered “cleaner” when it is not pushed into distortion. Cleaner, meaning less intrusive upon the original signal; a more accurate representation of the original sound (which also means less color).

Herein lies our issue as audio engineers. We now have the most accurate signal path to record through that the world has ever seen and the best part about it is that it is substantially cheaper to produce than the old electronics. This is all well and good except it is generally agreed upon that the old stuff “sounded better”. Now “better” may just mean we are more used to it; familiar with it, like comfort food. With the background of the problem acknowledged, let’s address the issue of how to add the analog warmth back into our digital mixes for the best of both worlds. (Please See Mixing Digitally: Adding Warmth to Clarity (Part 2: Practical Applications))

Equalization (EQ) is one of the fundamental steps in mixing audio. Wether it’s a live performance, a music record, or sound for film, EQ is used to along side volume or gain to create balance in your mix. It helps me to visualize the frequency spectrum of human hearing in my head when I begin to plan a mix. EQ is the main tool that I have to set certain instruments in particular frequency ranges. If we tried to push all the elements or instruments in a mix at all the same frequencies there would be not definition, and all the instruments would sound mushed up together. The first sign of a good mix is always the balance. It doesn’t matter how many cool effects are in a mix, or how much genius automation, if the balance is off it will never reach it’s full potential. So how does one achieve a good balance? Read on my friend.

Like I mentioned earlier I try and visualize the frequency spectrum in my head and then make a plan as far as which instruments I will place where. The first step when applying EQ is filtering. The recording process is not perfect, and I can’t see how it will ever be, plus I’m a subscriber to the theory that when one is tracking one should capture the whole sound (i.e. I don’t EQ signals when actually recording them to tape or disk) Filtering is the process of eliminating all the parts of a signal that you know you don’t want, and or won’t use.

EQ Video

I will typically roll off the bottom end of almost all the tracks (possibly not the Kick drum or Bass guitar… possibly). For instance the electric guitar (especially through and amp) rarely outputs anything useful below 80 Hz. Rumble and other non-audible frequencies like to hang out down there as well so it’s good to deep six those before they start taking up your precious headroom. As a general rule you can pass though each of your tracks with an EQ plug-in or use a desk/console if you have one, and roll off the low frequencies until you can hear it effecting the sound of the instrument and then back it off a little. This will do wonders for your headroom and tighten up your mix. On some tracks I will apply the flip-side of this tactic and roll off the top end, but one must be more careful with this. A lot of times it’s hard to tell when to stop rolling of the top and it’s easy to end up with a very flat sounding mix.

Next I will start to place certain instruments in certain frequency ranges. The video in this post goes over this as well. The electric guitar often encroaches on the frequencies that I plan on using for the bass guitar so I will cut some of those frequencies out of the electric guitar signal (typically somewhere around 200-300 Hz). The definition of an electric guitar, and most instruments for that matter, usually happen in the 2.5 kHz to 7 kHz range. This is also known as the presence range and these are the frequencies that humans hear “best.” That makes this band of frequencies very important, because not every thing can be there at the same time. Spend your time experimenting with you balance here in these frequencies. I tend to boost electric guitars on the low end of the presence range and save the bulk of it for the vocals. I also boost the electric guitars up around 7 kHz to let them sparkle and sizzle a little.

That’s the gist of it! The important thing is to be conscious of where in the frequency spectrum each element of the mix is going to live, and then of course getting good levels. Listening to a mix at very low volumes can often tell you what is hot in the mix and what is being drowned out and masked, I do this often when getting a balance.

Dynamics units are to either plug-ins or outboard gear (sometimes even inline in a desk) that are used to control the levels of sounds, or more specifically the range of levels a track can have. There are four main categories of dynamics units, but they are all closely related, these include: Compressors, Limiters, Expanders and Gates. All of these affect what is referred to as the dynamic range. The Dynamic range can be defined as the range of volume any sound has from it’s quietest part to it’s loudest part. Being able to control the dynamic range of a signal offers the engineer a powerful tool when balancing a mix.

A good starting point to explain dynamics units is compression. As the name implies compression compresses or shrinks the size a signal is allowed to operate in. The main advantage of being able to to this is that a signal can be made more “consistent.” By more consistent we mean that because the signal now has a restricted dynamic range the overall signal level will be more predictable. For example if a vocal seems get louder and softer during a song (and that is not the artistic intent) compression could be used to make the softer parts louder and the louder parts softer all at the same time. Once applied the overall level of the signal (the track fader or gain control on compressor) can often be raised to suit the mix.

If you haven’t realized it yet you will soon, audio engineering is all about compromise because “there is no such thing as a free lunch.” The compromise with compression is that by compressing the signal you are also raising the noise floor. Notice in the image below that by making the quietest parts louder we have also made the noise that was quiet, now louder. That is if the noise floor was originally between 0 and 2 dB, it is now between 10 and 12 dB. This means that if you compress very heavily you will start to hear more of the noise in the recording.

Now that we have discovered the balancing act that we must play with compression, let’s take a look at what we can actually control to keep our footing. Compressors will almost always have the following controls. Threshold, Ratio, Attack time, Release Time, Input level, and Output level (Gain). We will follow the signal flow through the compressor to explain what each control accomplishes.

Input level controls the level of the signal as it enters the compressions stages of the compressor. If the control is left at 0dB the signal hitting the compressor should be the same as what is coming off of the recorded media or previous effect. If you want to raise or lower the signal level prior to compression this is where you would do it.

Threshold is one of the controls you will use to control how much compression is to be applied. This control allows you to decide when (that what level) the compression will kick in. Setting a very low threshold will result in heavy compression, while a higher threshold will result in lighter compression. The threshold basically sets the point that the compressor says oops the sound is getting too loud, time to apply some compression to keep things under control.

Ratio, is the other control that determines how much compression is to be applied. If the ratio is 2 to 1 (2:1) that means that signal above the threshold is halved. That is if the threshold is set at 60 dB and the signal that is input peaks at 80 dB the amount of signal above the threshold is 20 dB. With a ratio of 2:1 the peak will be reduced by 10 dB (20 divided by 2) and the resulting signal will peak at 70dB rather than 80dB. A ratio of 5:1 would yield a 16 dB decrease or a peak at 64dB rather than 80dB (20 divided by 5 = 4dB, 20-16=4dB). If the ratio is greater than 50:1 the compressor is considered a limiter because at this point the compressor is basically “clipping” off the peaks and making them level with the threshold.

Attack time refers to the amount of time it takes the compressor to react to signals that breach the threshold. I could also be thought of as how sensitive the compressor is. The image below illustrates that a longer attack time allows more of the full transient (a sudden change in dynamic or level) through before clamping down on the signal. This is typically good for drums and percussive elements as the attack or transient of these sounds are essential to their character. A shorter attack time allows more control over the overall dynamic range this can be good for things like rhythm guitars and fingered bass.

Release time refers to the time that it takes the compressor to release control of the dynamic range back to the original signal. Say for instance the guitarist plays some very heavy loud stabs followed very quickly by some soft palm muted strums. The waveform may look something like the image below. Notice the effects of the different release times. A slow release time still compresses the quieter parts and does not allow them to be as loud as they would naturally be. With the medium attack time you can see that the quieter parts gradually increase in level to their natural level. Finally a fast attack time allows the quiet parts to return to it’s natural level very quickly. The slow and medium attack times in this case are examples of what is commonly referred to as a pumping or breathing compressor (sometimes this is a desired effect, but commonly it is a sign of over compression).

Output level or (make up) gain allows you to control the output signal from the compressor back into your channel. Often after compression the overall level of the signal is decreased. Most times this is the intent, if you used compression to tame an extra loud part of signal you can now raise the overall signal in the mix. Sometimes there is an auto gain setting that attempts to match the output signal to the loudest part of the input signal. This never seems provide a level I’m happy with so I prefer to set the output level manually. Remember the issues of the noise floor when setting the output gain, if there is too much of the noise floor in the signal after the make-up gain is applied, you may have to go back and re-adjust the ratio and threshold to make a compromise.

One of the most mystifying topic of the studio is impedance. Impedance is basically what it sounds like, anything that impedes the flow electrical current. Impedance manifests itself in three ways: Resistance, Capacitive Reactance, and Inductive Reactance, the most commonly referred to of these is Resistance. The letter assigned to impedance is “Z” which is why you’ve probably heard the terms Hi-Z and Lo-Z fly around from time to time. Hopefully the following will clear out some of the fog.

For the time being let’s call the microphone the source, and the mic pre-amp the load. A typical dynamic microphone like a Shure SM58 (R), will have a source impedance of approximately 200 Ohms. As a general rule you would like to have a load impedance at least five times the value of the source impedance, otherwise signal strength is lost. The mic pre of a Digidesign(R) Mbox 2(R) is 3.5k Ohms which is well beyond the desired 1k Ohms. Similarly if you would like to connect a synthesizer such as a Yamaha Motif to a line input on the Mbox 2 (R) is good to check the impedances. The Mbox 2 (R) line impedance is 10k Ohms and the Motif’s output impedance is XXXXXX. Some equipment (especially some vintage and specialty synths) need to be passed through a DI (Direct Injection) box to achieve an appropriate mic level.

This brings us the the issue of DI boxes and guitars. The nature of guitar pickups makes it difficult to adhere to the load five times greater than the source rule. Guitar pickups are inherently much higher impedance (10k to 50k Ohms). This means even though the typical instrument cable you connect to your guitar will also fit in the line input, you will likely get little or no signal. This is because you likely have a higher source impedance the load impedance, thus grossly breaking our general rule (load impedance > 5 X source impedance). The DI box was designed as a solution to this problem.

The DI box is an impedance converter allowing you to pass the same signal through, but creating a new source impedance. The Art Zdirect passive DI box offers an input impedance of 50k Ohms and an output impedance of 600 Ohms. Due to the fact that the signal is typically hotter coming from a guitar (higher in voltage, especially from humbuckers) the 50k Ohms input is acceptable as the mic pre will not be expecting the hotter signal off of the guitar pickups. Therefore it is possible to record a guitar straight into the Mbox 2 (R) through the DI box (600 X 5 = 3k Ohms < 3.5k Ohms, OK!)

Some audio interfaces offer instrument inputs as well. These inputs are designed to accept a guitar or other hi-Z source and often have a load impedance in the ball park of 2 Mega-ohms or greater.

Having a good grip of how impedance works in the studio is the first step to attaining a good signal to noise ratio.

There are two main types of connections used in a studio to carry audio signals. These connections are balanced and unbalanced. When possible it is always better to use a balanced connection.

Unbalanced connections:
Unbalanced connections consist of two contact points, one is considered hot, the other ground (0V). Unbalanced connections transmit the signal (the desired information) across the hot wire, and the information is interpreted with respect to ground (0V). The issue with unbalanced connection is that they are more susceptible to electro-magnetic noise than balanced connections. The best way to explain why, is to explain how a balance connection works.

Balanced connections:
Balanced connections typically (there are a few exceptions) consist of three contact points. One is considered hot, another cold, and the final one a common ground. Unlike the unbalanced connection the hot and cold connections both carry a signal and it is the difference between these two signals that carries the desired information. In this way both the hot and cold connections are exposed to generally the same electro-mechanical noise. Therefore when this difference between the hot and cold signals is extracted the noise on both lines will phase cancel yielding only the desired signal information.

*Note: Using a balanced cable alone does not create a balanced connection, both pieces of gear and the cable must support a balanced connection to create a balanced connection. (Example: Using a balanced cable to plug an electric guitar into and amp will work, but not create a balanced connection because the guitar and amp will most likely not support a balanced connection.)