The Jam Corner

A Closer Look At The Recording Process

This is a continuation of the Digital Recording series. This first part examines the process of recording prior to mixing in further detail. Check out this article for some insight into the kind of equipment you might be using to record your songs.

When we hear music, what our brain is processing can be loosely referred to as sonic information. The information aspect of it is created by the musician and interpreted by the listener; however, the bare-bone sonic data travels in the form of acoustic energy, or sound waves.

Sonic information is encoded using this sonic data. One of the more unique properties of music as an art form is the relationship of tonal elements to time at which they occur respective to each other. This is precisely what creates movement and excitement through music. However, when we are recording a piece of music, the time must be represented on a recording medium, where we're able to circumvent the restrictions imposed by passage of time to piece together multiple sonic elements (pieces of sonic information) as one. This is underlying concept for fast forwarding, rewinding, and selecting tracks on a CD player.

What goes into a good recording?

The idea behind recording is to convert sound energy (the medium for sonic information) to other forms of energy that are easy to store on our recording medium. Traditionally, the common recording medium is magnetic tape; in the recent past, however, digital data storage has all but replaced traditional tape - especially in amateur home studios, such as those that come in every GNX3/4/3000 box.

There are several factors in play - excluding factors non-specific to recording, to be mentioned later - which we will discuss in detail here:

1. Acoustic energy is difficult to digitize; electric energy is not. Therefore, we will need an input transducer to convert acoustic energy to electric energy. Examples of transducers are guitar/bass pickups and microphones; speakers are also transducers - output transducers - they convert electric energy back to acoustic energy.
2. The electrical signal generated by an input transducer is generally very weak (usually ~-60dbu, called instrument level signal); it is not suitable for either analog or digital processing yet. As such, we require an amplifier to increase the signal strength to line level, which is usually either "true" line level of +4dbu, or a sub-line level of -10dbu.
3. To allow digital storage of sonic information, the electrical signal must be represented as digital data. Here, we will need an analog-to-digital converter (A/D converter). The opposite of an A/D converted is a D/A converter, which takes in digital sonic data and outputs it as an electrical line-level signal - you have a couple of these in your CD player and soundcard.

The basic scheme for recording, then, looks like this:

1. A transducer takes acoustical energy and generates electrical signal that represents the sound signal. Hopefully, the representation is accurate with respect to the sound source.
2. The resulting electrical signal is usually an instrument-level signal, so it is routed to an amplifier to be boosted to line level.
3. ****At this point, another transducer may be inserted to convert the electrical line-level signal back to acoustic energy. If so, steps 1 and 2 are repeated. This is common for recording mic'ed guitar amps and outboard insert effects.****
4. When a final line-level electrical signal is ready to be recorded, it is fed into an A/D converter. This device generates multiple snapshots, or samples, of the electrical signal. The GNX A/D creates 44,100 samples every second, at even intervals. Because the electrical signal varies in an indiscrete fashion (i.e., it doesn't "snap to grid"), the quality of the digital representation will depend not only on the number of samples, but also on the level of detail at which they're recorded. The level of detail is in dirrect correlation with the amount of memory allocated to each sample. GNX units have a high digital resolution - each of the 44,100 samples taken per second is allocated 24 bits (3 bytes) of memory.

Once the incoming electrical signal is digitized, it is suitable for Digital Signal Processing (DSP). We are also able to store it on a hard drive, which, unlike magnetic media, do not lose quality over time. Digital data is also easily transferred between storage locations with absolutely no loss in quality.

However, before we go into manipulating digital data, we need to make sure that it's the correct data. To this end, all three types of recording components must get the best-possible signal to work with. Furthermore, we must try to produce the best possible sonic information from the sound source, so that there is sound quality and integrity to preserve during the digitization process.

Two common setups for recording involve voice and guitar. Respectively, two types of transducers used are microphones and pickups (microphones are used the same way for drums). The following diagram illustrates the setup, incorporating the optional amp-mic'ing stage:

How do I get a good recording?

The original source of sound will largely dictate the quality of sound in the final mix. While proper mixing and mastering can minimize recording errors, things like "full-bodied tone" and "organic feel" are difficult qualities to fake in the end when they were never there to begin with. When recording an instrument such as voice, piano, guitar, or drums, the construction, setup, string type, and performance technique affect what the input transducer picks up.

Room characteristics is another important factor for the initial acoustic-to-electric energy conversion. Even if you're recording your guitar through a GNX digitally with headphones on, various room resonances can cause noise to be generated by the strings and transferred to the guitar's pickups. One time we were recording bass in a room where there was a machette hanging on the wall. We didn't realize at first that the sword was shaking from the output of the monitors, and even when we did, we didn't give it any significance. As a result, we ended up with an annoying resonance that muddied up the bass part terribly and had to be re-recorded.

Any object in the room can potentially cause unwanted (or unnecessary) noise, but pay particular attention to electronic devices. Things like hard disk/CPU fans cause physical noise that's easy to get used to when exposed to them for hours at a time on a daily basis, but quite unpleasant when heard on your lead vocal tracks. Other devices, such as computer monitors, can cause interference resulting in additional pickup noise (single-coil players: I'm talking to you!)

Once the sound is picked up by an input transducer, it goes through some electronics. It is important that these electronics are in good shape. If you have a mic that goes in and out when shaken, active pickup electronics with a dying battery, or poor-quality cables, your recording will suffer. It may not seem terrible when jamming, but realize that during mixing many softer parts are brought up in level, and this will include noise. When playing live, a minor interruption in signal due to a faulty volume knob might be unnoticeable; but once you commit it to disk, it'll be there for the life of your mix.

You may have to spend some time taking care of house wiring problems that cause ground loops or other interference.

The transducer itself will affect how well the resulting digital signal is represented. Many microphones and pickups color the tone and change the frequency response of the original sonic information. Depending on your situation and goals, these effects may be desirable, irrelevant, or detrimental.

Likewise, amplifiers affect the tone. Also, various types are available (tube-based, transformer-based, etc) and each extends particular influence on the output signal. Generally, any amplifier going directly to the A/D conversion should be transparent. Exceptions apply: tube amps (or emulations thereof) are preferred for vocal guitar aplications because they add "analog warmth" to the original signal, as an example. Modeling amps using analog outputs also color and/or distort the signal, but can be fed into A/D converters directly.

If multiple amp stages are employed (e.g., you're using the amp to a speaker, picked up by microphone and amplified again before the A/D), the intention is to capture the coloration coming from the earlier-stage amplifier (so, guitar amp). Speakers are output transducers that further color your tone - yet another variable to keep in mind. In most cases, you will not do this if your output amp is transparent - we want to avoid as many transduction (is that a word?) stages as we can.

In much the same way, A/D/A conversions should be minimized. Because a digital representation is never 100%-accurate with respect to the original pre-digital electrical signal (sort of like squaring a circle - close, but no circle), multiple A/D and D/A stages eventually degrade the sound quality. My rule of thumb for digital recording is: Once it's digital, keep it that way. Recording with a digi-capable unit such as a GNX allows you to digitize the signal after the first transduction stage and leave it that way for good. Take advantage of that if possible, and you'll end up with a much cleaner mixing job ahead of you.

The process of recording - Level 2 Zoom

Before we go on, I want to be sure of something. If you've gotten to the point where you are planning which tracks to lay down, you should already have a songmap and a clicktrack ready to go. If not, do it before you begin recording. No one likes to waste hours of recording, only to have to do it all again. It's a big inspiration killer.

Now that we understand the considerations that go into good recording, let's address some song planning issues. Generally, there are three types of recording that should be pondered separately: (1) the big picture, (2) overdubs, and (3) punches.

Our first step is to capture the "big picture". A common practice is to record some scrap tracks. Even if the scrap tracks are with a drum machine, off-pitch, out-of-tune, these will still give you an idea of how your mix is coming out. If you've never heard your song recorded in a studio setting, hearing a scrap mix lets you learn things you've never known (or noticed) before.

If possible, you should try to record all instruments on separate tracks (not always possible simultaneously in a home-studio setting); this way, you'll be able to to mixdown one stereo pair for each set of instrument tracks, excluding the instrument being recorded. One possibility (if you have 8 input tracks) is to record drums as stereo overheads + mono kick; guitar and bass in mono; keyboards and other padding instruments as stereo; and vocals as mono. Then, as each musician is doing the final takes, he or she can hear everything except what he or she is playing - or some other "rough mix" combination of scrap tracks. For instance, the bass player may want to hear very little vocals, but a lot of drums and a good dose of guitar; the guitarist may prefer no vocals at all, but strong bass and kick. Using the scrap track approach lets you get the right feel for your upcoming masterpiece.

Now, as we consider each instrument, our goal in this part of audio production is to capture the fullest, most authentic, strongest, cleanest digital representation of the original audio signal. Here are some simple guidelines for doing so:

• At each stage in the signal path, the signal should be as strong as possible. If you wish to record distorted signal, this must be handled by the analog (or simulated analog) stages - digital distortion does not sound good under any circumstances. If you do not want distortion, be sure that your output level at each stage isn't causing it.
• As you monitor the input signal, be sure to set the output fader level at unity gain. Then, vary the input gain (sometimes called trim or pre-level) until the record meter is reaching the -0.1dB mark. It's optimal to have as loud a signal as possible, but again with digital equipment, clipping is never desirable. Therefore, lower the input gain just a tiny touch, to be on the safe side.
• Know what tone you're going for, and be consistent with it - this makes better mixes.

Record the drums first, especially if they're acoustic drums. Next, take down bass and guitar. Vocals will be recorded afterwards - it's good to have rhythm instruments down before doing vocals, to give the vocalist a pitch/harmony framework. Alternate percussion and other incidental instruments can be recorded last. This comprises the final "big picture". At this juncture, don't worry about miscellaneous fills, random noises, or any kind of special effects or effects processing - all that can be taken care of later.

As the big picture is coming into focus, you may find it necessary to overdub some parts. Select a free track (or create one, if you're in software), set the same levels along the soundpath, and record. One suggestion that will save some headache is to begin recording just before the overdub begins, and through just a little after. You'll be surprised how much easier your life becomes with a measure of buffer on either side of an overdub, when you go to mix.

If you recorded a nice, long, continuous part with great feel, but a few measures of bull****, you can choose to punch that part in. Elsewhere on this site there is a discussion of punching in/out. The idea is to begin playing back what's recorded, play along with it, and record just the part that's no good. Software devices make this a breeze, and many stand-alone recorders handle this task perfectly.

If you have room, do the punches as overdubs - meaning, on separate tracks from the part being fixed. This allows you to (1) give your mixing engineer some buffer as discussed above, (2) keep the punches separate until you're ready to finalize them. This way, you can also overdub parts that came out OK, but may have been done better (you just don't know if you can pull off any better). Additionally, you can always take the punch and insert it in reverse at a later/earlier point in the song - you just never know!

Once you got all your tracks together, get a rough mix (use tracks selectively if all of them just don't fit together right off the bat, or bounce several mixes). You may also begin experimenting with panning and levels at this point. Take the rough mix(es) home with you, and listen to them on a different sound system, away from your instruments, studio equipment, and bandmates. You may think of other things you'd like to record, certain things you may want to redo, and solos for which you'd prefer to have an alternate take or two available. But remember: you'll have to end up satisfied in the end, so don't let minor imperfections set you back months in your production.

Final word

Let's go over some key points mentioned here:

• When recording, capture the original sound.
• Use the best-available sound source in a quiet environment
• Minimize the quality degradation during transit (see diagram)
• Record in the right place (have your clicktrack and paper ready)
• Buffer your overdubs
• Record every element you have any intentions of working with in the mix

One thing to remember is the bit depth of the digital data stored. Using 24-bit recording instead of 16-bit makes the result easier to work with - basically your data will be more flexible. In layman's technical terms, higher bit resolution results in fewer "rounding errors" when mixing, applying effects, and working with dynamics. But even if you can only record in 16-bit, your mix should be done in 24 - we'll talk more about this in the next article in this series.

Off to the studio.

1531 Views | 0 Comments | Rating: (0 rates)