Digital Audio 101

The purpose of the article is to explain some of the basic fundamental concepts to users new to digital audio and answer a few questions that are often asked by novice computer recording enthusiasts. In my experience most people entering this discipline have many of the same questions, run into the same road blocks and have to come to grips with the same concepts.

What is a line level signal?
What is MIDI?
What kind of computer do I need?
How many channels do I need?
What is a virtual instrument?
What is the difference between a synth and a sampler?
What kind of sound card do I need? Will my on-board one work?
What is latency?
What software do I need to get started?
What is phantom power?
How do I record midi to audio?
What is a plugin?
What is a balanced signal?
What is an Insert?
What is the difference between a Send Effect and Insert Effect?
What does print to audio mean?
Why do I need headphones to monitor with when I multi-track?


What is a line level signal?

Believe it our not, understanding the significance of the term "line level signal" will help you understand a lot about digital audio and be of immeasurable assistance in hooking up your gear and getting it to work properly.

Line level is a term used to define the strength of an audio signal used to transmit analog sound information between audio components such as CD and DVD players, TVs, audio amplifiers, keyboards, most computer sound cards and mixing consoles. It is the basic signal level from which all other audio signals at compared.

Microphones on the other hand do not output line level signals so if you have a line level input on your audio interface you would need a pre-amp between the microphone and the line level input.

Electric guitars, electric bass guitars, acoustic guitar pickups do not output a line level signal. They operate on what is known as a high Z output. Again if you want to match an electric guitar to a line level input you need the proper device to match the different levels. One such device is known as a DI Box. DI stands for "direct injector". Our DI Box would sit between the guitar and the line level input on a sound card. Some mixing boards and some audio interfaces have High Z inputs on them so to facilitate the easy matching of these input types.


What is MIDI?

MIDI (Musical Instrument Device Interface) is a control protocol. Midi data contains no audio. An excellent analogy is a player piano. In effect MIDI is a digital equivalent of a player piano. Every bit of data in MIDI is either switch info and/or register info with data.

An easy example is a MIDI keyboard. Each note on the keyboard has been assigned a number. When a keyboardist plays a note, MIDI data that is sent to the computer. The data sent to the computer contains the note number, the time the key was pressed, the time the key reaches the bottom of the key stroke, the difference between the two is simple calculation. This calculation gives us the note velocity. This is assigned a value. Notes with a lower velocity are notes that are not struck as hard, so the volume would be considered lower, a softly struck note. We could also use this information in a sampler to choose one sample or another. The notes of a grand piano sound a bit different depending on how hard the key is hit

MIDI can also record what the settings are for the different switches/buttons, knobs and sliders on your keyboard. These "controllers" select what voice or sample our keyboard will play. These would be the sound patch, filter settings, volume levels etc.

MIDI can also control the data being sent form various "continuous data" devices. The mod wheel and the pitch wheel on a keyboard are very good example of this type of "controller". Continuous data controllers can send a lot of data because they read the controller info continually. The pitch bend controller is a perfect example. As you move the pitch bend controller the pitch data is being continuous monitored and the pitch is being changed on a moment to moment basis.

Of course since we can record all of this data we can also play it back. This is the beauty of midi. Also since  it is the control data that we have recorded it is easy to modify and edit this data on playback. The best example is the ability to change the sound patch and edit the key. Let’s say we record a piano part. After the recording it is very easy to change that piano part to an organ part or a vibes part just by changing the patch. This is something that you can't do with audio. Similarly changing the key or transposing the recorded part is just as easy,


What kind of computer do I need?

The computer that we use as a Digital Audio Workstation (DAW) is the basically the same computer we use for other tasks. Of course, the more we ask the computer to do, the more powerful a computer we need. Recording is one thing but applying real time effects and using the computer as multiple synths and samplers is another. Funny enough, applying reverb to a track can be one of the most CPU intensive tasks in digital audio production.

Lots of memory is a good thing to have in an DAW computer because many of the tasks we ask a computer to do in this field run better with memory. Audio data takes up lots of hard drive space. It is considered by many to be a good idea to have a second hard drive dedicated exclusively for audio data. One of the most memory intensive applications in digital audio work is using the computer as sampler. More memory allows the sampler application to keep sample files in memory. Access to ram memory is much faster than hard drive memory.

One desirable feature in a DAW computer is that it runs quietly. This is especially important if you are working in the same room as the computer. Quiet operation is a matter of degree. There are many things we can do to make the computer run quieter Some of these are inexpensive options, some can cost money. To quiet fan noise we use larger fans that run slower but still push enough air. With fan noise, it is the speed of the fan that sets the noise of the fan. Slower fans are quieter fans. One trick that can be used very effectively with case fans is to run them at 5 volts instead of the standard 12 volts. At 5 volts the fans run much slower but they still push air. This is easy to do just use a "y" cable for the hard drive power cable, cut off one end and wire your fan into it. The color-coding is pretty standard. Black - Negative, Red - 5 Volts, Yellow - 12 Volts.

There are products available to sound proof the case. They consist of acoustical foam that you use to line the insides of the case. For rattles, using a nice thick tape to line the contact points will often completely eliminate rattle noise.

For the "money is no object" kind of people there are power supplies available that have large fans, some with speed control features. Some of the higher end power supplies offer super heatsinks so that the fans hardly ever run. There are also huge CPU fan and heat sink assemblies that use large copper and aluminum heatsinks and large fans. These can run very quietly.


How many channels on my sound card do I need?

The question you have to ask yourself here is "how many tracks do I want to record at the same time." What I mean here is if you are working mainly by your self in a home project studio and you are multi tracking or using loop based software you might only need to record 2 channels at a time.  On the other hand if you wish to record a full drum kit you might need 8 channels or more.

Even if you only have a 2 In / 2 Out audio interface you can still work in a multi-channel environment. The limitation is you can only record 2 channels at a time. Of course you can still use multi tracking techniques to layer multiple takes to produce you own 32 (or 132) track masterpiece


What is a virtual instrument?

A virtual instrument is an instrument that runs in software on your computer. They can be sample based, synthesizers, physical modeled instruments, drum machines etc. If there is a real instrument, there is probably a virtual version of that instrument.

Why would we use one? Cost savings are one. What is an outboard sampler or a synthesizer other than a computer? They are usually limited by design to the task they were designed for. On the other hand your home computer is completely open-ended. It can be configured through software to perform a wide range of functions. Since you already have the computer, the cost of virtual instruments in much less than purchasing traditional electronic ones.

Another reason is that since the instrument is internal to the computer, any sounds that are produced by the VST instrument are already in the digital domain and can be rendered and mixed within your sequencer software. Another bonus we get from virtual instruments is the fact that we don't have to record them in the traditional sense of the word.

For example let us look at a Grand Piano. A real grand would cost thousands, requires frequent tuning and is a bit heavy to pack around. For arguments sake lets say we had one that was in tune, and it was in our studio (luck us to have a home studio big enough to fit a grand). We would still have to record it. Recording a grand piano requires good microphones (that's plural), perhaps acoustic isolation and know-how. There are many, many virtual pianos and more coming out every month. Some are priced very cheaply and some of the more expensive ones can be bought for hundreds of dollars or about the cost of one good microphone. We haven't even considered playing the thing. With software, midi, step time and quantization many home studio musicians can laboriously assemble a good sounding piano part. Not nearly as many of these musicians could sit down at a grand and play that part in real time without error.


What is the difference between a synth and a sampler?

A synthesizer is a device that uses electronic or digital signals to create either a replication of a natural sound or to create an entirely new sound.

A sampler on the other hand uses a series of recorded samples of a real instrument mapped across the full range or part of the range of a keyboard. Since we used a grand piano as an example above, lets use it again. If we wanted to sample a grand piano we would record a series of recordings or samples of each note or at the very least one sample every few notes. If we wanted to get greater accuracy and truer sound we would make several samples for each note, sampling the sound at different volume levels. Once all of the samples are made the samples are then looped so that the note will be able to be sustained longer that the original note was recorded.


What kind of sound card do I need?  Will my on-board sound card work?

Modern on-board sound cards, the ones that come built-in to most motherboards on today’s computers have come along way from a few years ago and are quite adequate for listening to music and video on your computer. In fact they are adequate for all but the most demanding gaming applications and recording. Unfortunately they fall short when put in service for audio recording. The two biggest limitations are sample depth and the drivers for the card.

In an audio recording card we are looking for 24-bit recording. Most on-board cards only offer 16-bit recording. 24 bit recording offers greater signal to noise ratio and greater headroom. Digital audio differs from traditional tape in that is very unforgiving to signal overload. For those of you old enough to have used tape, this is when the VU meter needles spike up into the red at the top end of the meter scale. With tape, modest overloading of the signal gave the audio a desirable saturated sound, with digital audio signal overload (or clipping) produces some very undesirable artifacts also known as "digital thwack" which kind of describes these said artifacts.

Of equal or greater importance is the quality of the drivers. For digital audio work we are seeking well-written drivers that offer low latency (see the topic below for discussion about latency) and support for one of the popular driver models used by today’s sequencers. In the PC world the most popular driver model is ASIO. (Audio Stream Input Output). ASIO is a specification developed by Steinberg, the developers of the Cubase sequencer. ASIO allows the use of multi-channel audio cards and low latency. ASIO is supported by virtually all PC audio software.


What is latency?

Latency is basically the time it takes the audio signal to travel through your sound card. If you wish to multi-track over top of previous tracks it stands to reason that the new tracks need to line up with the previous tracks. To do this the audio card needs to be able to read the existing tracks and record the new tracks at virtually the same time. Most experts agree that 10ms is the magic target for latency. Lets talk about what this means. Sound travels at aprox. 1.1 ft per millisecond. 10ms of latency is the delay you would experience when listening to someone talking from 10ft away. For all intents and purposes the sound appears to arrive instantaneously from the individuals mouth to our ears.

Latency is also important when using the computer as a synthesizer. If you do not have a low latency sound card there could be a huge delay between when you hit the key on the midi keyboard and when the sound comes out of the speakers. Again 10ms is the magic target.


What software do I need to get started?

This is another area where you have lots of choice. What software you choose will depend a bit on what type of music you are interested in, how you work, how much money you want to spend and possibly what was the last recording magazine you have just read. Most of us are looking for a sequencer at the very least. Many neophyte-recording enthusiasts have told me that they "just want something simple that will do everything they want". We don't want much do we?

The paradigm for a sequencer is the recording studio itself. If you just walked into a recording studio for the first time you would probably be overwhelmed by all the gear and gadgets. A recording studio by definition is a pretty complicated place. Any sequencer worth it's salt will have it's own complexities. I don't think this can be avoided but don't despair, once you get your bearings it is not that complicated because they are all basically doing the same thing, and that thing is pretending to be a recording studio. Once you understand one sequencer you will understand a lot about all of the others and to a great extent a lot about recording studios.

Some of the big sequencers names for PC computers are Cubase, Sonar, Pro-Tools, Ableton Live. Reason is a very popular music creation tool but in my mind not really a full sequencer because it only deals with midi and does not record audio. There are many lesser-known programs that target the budget user; one of note is PowerTracks by PG Music. This was the software I started with and it was very good and gave me a great start. Another one of note that some very knowledgeable audio dudes have raved about is called N-Tracks. One strategy that I can't find fault with is to start with something, anything and get your feet wet. Play around for a while and get acquainted with the basic concepts and practice of audio recording, these are universal. Once you have some experience you will be in a much better position to pick and choose your dream setup.

Besides a sequencer, you might also want software synths, samplers and various plug-ins. When I first started to get into the digital audio scene I religiously read an UK magazine called Computer Music, an excellent mag. It came with an included CD that contained a full set of music software, sequencer, sampler, drum machine, and synthesizer, all for the sticker price of $20.00. A great deal and certainly enough to start learning about recording in a home project studio.


What is phantom power?

Phantom power is (usually) a 48 volt DC current that is supplied over a XLR microphone cable for the purpose of powering microphones containing active electronics, most often condenser microphones. Since large diaphragm condenser microphones are the most popular microphones used by the home studio crowd, having hardware that will supply a phantom power signal will be very important. Most microphone preamps and most mixing desks will offer phantom power. Some condenser microphones have options to be powered by batteries but in my experience are not common.


How do I record midi to audio?

The simple answer is you have to play the MIDI data through a synth or a sampler to produce the desired audio and then record the output, either by jacking the output of said synths into an audio input or if using soft synths, routing the internal audio signal through your sequencer and render it to an audio track.  


What is a plugin?

Plugin is a term applied to software programs designed to be added to a sequencer to add a new or better feature to that application. A plugin can be either an effect like reverb, or compression or it could be a new instrument. The term plugin signifies that this effect or instrument is not a stand-alone program and needs to be hosted by a compatible sequencer, also known as a "host".


What is a balanced signal?

Balanced connections use three-conductor connectors, usually a XLR or TRS jack plug. XLR connectors, for instance, are usually used with microphones because of their durable construction, while jack plugs are usually used for mixer inputs and outputs because of their smaller profile.



Many microphones operate at low voltage levels and some with high output impedance (hi-Z), which makes long microphone cables especially susceptible to electromagnetic interference. Microphone interconnections are therefore a perfect application for a balanced interconnection, which cancels out most of this induced outside noise. The longer the cable run the bigger the benefit given by balanced cables.

The idea behind the balanced signal is that two copies of the signal are sent 180 degrees out of phase along the balanced cable. At the end of the signal path the two signals are summed back together and brought back into phase. The original signal is now amplified because the two waveforms combine and enhance each other. Any noise that was picked up along that path is canceled due to the fact that when the signal is brought back into phase any noise picked up is now 180 degrees out of phase and the their waveforms cancel each other out.


What is an insert point?

An insert point is a point in the signal chain (often a TRS jack on each channel strip) that allows the signal to be routed from the original path, through a device (often a signal processor of some sort) and then returned to the original signal path at the break point to continue it's journey through the channel strip.

If the send and receive are not used the signal flows through the channel strip as normal. If the send and receive are used, the signal path is routed out the send through the reverb and back to the return to then continue through the signal path.

There is a technique to use an insert point on a mixer to send a copy of the signal out to your audio interface. This is useful if your mixer does not have a direct output on each channel strip. This technique requires a modified balanced cable. In effect what we do is short the tip and the ring on one side of the balanced cable. This allows the original signal to pass through the channel strip unaffected by this process. It also allows a copy of the original signal to travel through the modified balanced cable to your audio interface.


What is the difference between a Send Effect and Insert Effect?

Most sequencer programs allow the use of both Send and Insert effects. The difference is that with an Insert effect the entire signal is routed through the effect. A common use of an Insert effect is a compressor because if we do use a compressor of course we want to send the entire signal to pass through it. A good example of a Send effect is a reverb. With a Send effect we only send a portion of the signal through the effect and also send an unaffected signal through the channel strip. We can then mix the two together to achieve a balance or the wet and dry signal. The wet signal refers to the part of the signal that went through the effect.


What does print to audio mean?

When we print to audio we are rendering a signal with all of its effects, EQ, automation settings etc. to a wave file so that we have a "hardcopy" of the track(s). The reason we do this is that in a very complex project with many virtual instruments and multiple effects even the most powerful computer will eventually not be able to handle the entire computing load. By printing some or all of the tracks to audio we reduce the load on the computer. Once we print the track to audio we can mute the original track and disable all of the effects and thereby reduce the computing load on our DAW (digital audio workstation) machine. It requires much less computing power to play an audio track than to run a midi track through a virtual synth then process the signal through a reverb, compressor, delay and distortion plugin. Remember that we haven't erased anything, we have simply muted the tracks in question so that the processing overhead for that track is negated and allows our processing power to be used elsewhere. Should we at a future time decide we want to modify the muted track, all we have to do is un-mute it, engage the plugins required and reprint the track in question. Since most sequencers have virtually unlimited track counts this is an extremely effective way to deal with high processing demands. It takes a bit more time, and is a bit tedious but often beats dropping a few thousand on a new monster machine.


Why do I need headphones to monitor with when I multi-track?

In a home studio, project studio environment we often record in the same room we monitor and mix in. We are also usually the musician, engineer, producer and gopher all at the same time. Lets say we have recorded a keyboard track and have a drum machine track down and now want to record a vocal, it stands to reason that we need to hear these tracks at the same time as we record the vocal. If we play the tracks back through our big honking monitors and try to record our vocal we will obviously also get spill from those tracks in our vocal because we have the speakers playing in the same room as the microphone. This is why we need headphones, so we can isolate the already recorded tracks from the microphone.

Headphones come in two types, open and closed. We often use open headphones for casual listening because they do not acoustically isolate use from the outside world. We can hear through them. We can hear a car honk at us, or hear someone speaking to us (depends on how load you have the music turned up). These types of headphones are great and comfortable and convenient but if we can here sounds from outside, the outside can also hear some of what we are hearing too. This can cause spill from the headphones into the mic.

 With closed headphones we are acoustically isolated more or less. There might be some spill but by and large the sound stays in the cups around our ears and does not leak into our microphone.


Closing thoughts

The field of home and small project studio recording is constantly changing with new gear and software being released weekly. The amount of new stuff is staggering. We have also seen in the past few years an unprecidented increase in computer power and at the same time enormous drops in the cost of computer gear. We literally have the world at our fingertips. The power and capabilities of a well equipped home studio now rival or in some cases surpass that of big studios of a few of decades ago when I was a young musician.

It is worth remembering that there is one huge difference between our home studio and the big studios of the past and that is the recording room itself. A professional studio room is scientifically designed machine to mechanincally deal with sound. Is this a big impediment? It depends. For the most demanding acoustic recording we might be at a disadvantage. That being said, we can still do fantastic work and achieve outstanding results. It is a sad truth but no matter how pristine our recording and sonic fidelity 99% of the public that will listening to our masterpiece don’t have the ears to differentiate great from outstanding. Most will not be listening to our music on systems capable of reproducing the fine detail that separates the great from the outstanding, heck probably for very good to great if you think about it. Many will be doing something else while they listen to our work. Eating, talking, reading a book, you know, you do it yourself. With the advent of the MP3 phenomena there is also the reduction in fidelity that is a byproduct of low bit rate MP3 encoding.

All of the above sad realizations work in our favor. The army of the home recording enthusiast is winning. Big studios are not as popular as they once were. People are using their own equipment. In the world of house music, rap, drum and bass trance music, acoustic fidelity is not a big issue as most of the sounds are electronically produced and the high sound levels very effectively mask any acoustic imperfections in any of the audio tracks.

I always tell people that in this game where you start is not where you eventually end up. Digital audio is a journey, not a destination.