A guide to using microphones in some of their common applications in recording. Covers good practice in setting up and connecting microphones, and includes tips on getting the best from your equipment and subject.
When recording with microphones, attention to good positioning and signal management can yield massively improved results. This guide will give the new user some important tips and caveats to improve their recordings, and will give the more advanced user some insight into more complex techniques. We will also give you a few pointers on getting the best from your subjects and their performance - be that yourself, a colleague, or any other speaker.
Subjects will sometimes feel nervous and uneasy, or act unnaturally when they have a microphone placed in front of them. This is only natural, especially if they are relatively inexperienced in being recorded or having their voice amplified. The best way to help deal with this is to offer some simple and positive advice and reassurance, and to treat their (or your) concerns with sensitivity. No matter how great your technical expertise, an uncomfortable subject will rarely sound good. Conversely, confidence and good microphone technique can count for much more than expensive equipment and processing.
Processing recorded signals is dealt with in depth elsewhere, and if you want to know how to start enhancing and correcting your recordings once you have them on disc, then you should begin by looking at our Introduction to Signal Processing and guide to Digital Equalisation. There is a very short appendix at the end of this paper explaining a few basic groups of effects, but it is nothing more than an overview, designed to give you an idea where to look next. The emphasis here is in maintaining the integrity of your microphone input signal and getting the best raw materials into your recording environment.
We'll begin with the process of connecting your microphone(s) to the computer and Public Address (if used), and running through the chain which the signal follows, looking at each element it may encounter along the way. Once the signal path is established, and you're confident that you are getting a good quality signal into your recorder, we'll look at how microphones can be correctly and creatively used to enhance the performance of your subject, and make them sound their best.
The difference in level between the input and output signal of an amplifier circuit is called its gain, and is measured in decibels (dB). The single most important factor in getting the best recording from your microphone is maintaining a healthy signal level throughout its path to the recorder, without exceeding the limits of the system. Throughout the recording chain, and in both the analogue (pre-digitisation) and digital (inside the computer) domains, there may be several stages at which the signal can be amplified or attenuated, and the same rule of ‘enough, but not too much gain' applies each time.
Microphones need considerable amplification to raise the level of their output (usually quite low) to a level where it can be input to a digital recording environment, or amplified further over a public address system. Correct management of this process will enable you to get the best from your digital converters, maximise your recording's dynamic range (the difference between the quietest and loudest sounds you can record), and minimise background noise. Excessive gain leads to the signal level exceeding the capacity of the circuit, resulting in the extremities of the waveform being literally ‘clipped‘ off - hence the term ‘clipping', meaning distortion of this type. Conversely, insufficient gain will lead to a low level recording, which will use a smaller proportion of the digital bits available to record it, giving a more low-resolution sound. Also the additional boost needed by the signal to make it audible will amplify any background noise, leading to more hiss and hum in your recordings. Neither of these is a very desirable result, and both are easily avoided by performing a few quick soundchecks before pressing the record button.
Distortion cannot be removed once recorded, so if in doubt, always err on the side of caution (i.e. better to be too quiet than too loud)
Many of us are familiar with audio meters, where each audio channel will use either a moving needle or a row of coloured lights to indicate signal strength. Often these meters will incorporate a section coloured red in their upper regions, which indicates that signal clipping is imminent or already in progress.
There are two common types of meter:
Bar graph meters. Image © iainf - used with permission
Some bargraph meters incorporate a "peak hold" function, where the highest level reached will be displayed either for a few seconds, until resetting to the current level, or until manually reset by the click of a button.
Once you have all the stages of the input chain connected, you need to set the input and output levels of each device. Setting these levels correctly is very important, and will affect objective sound quality more than almost any other procedure in the setup process, so take the time to do it accurately.
A good starting point is to turn all input levels to a very low level (just enough to be audible) and all output levels near to full, before beginning.
Use closed-back headphones for setting levels, with all speakers turned down to zero, to prevent any feedback artefacts. As you are starting with very low input levels, which you intend to increase, you should start with a quiet monitor signal too. It will get louder. If you are setting levels for a vocal mic, you may need assistance at this point - ie somebody to speak forthrightly into the mic! If at any point your monitoring level gets too loud (and always adjust levels incrementally, and gently!), then turn it down from the headphone output control (be that a physical volume control, or in the software control panel on your computer), do not use any other level control, as that will defeat the point of the exercise !
Once you have a monitoring signal (ie you can hear the voice or reference tone faintly through your headphones) you can begin to wind up input gains. Start at the beginning of the chain. Increase microphone gain until the signal approaches but does not exceed 0dB (usually about 75-80%, if not marked) on the meters, when fed a high level signal - one that you judge to be as loud as is likely during recording.
Repeat this procedure with the next device in the chain, then the next, etc. Finally, set the input level for your recording software/device similarly, and make a short test recording. If this is distorting, or is too quiet (inspect the recorded waveform if possible), then go through the chain again, checking all levels.
An analogue meter will generally be marked at 0dB - the maximum level beyond which the waveform starts to be clipped, it is possible to exceed this level, but the more you do, the more audible distortion will be present.
Digital clipping is however a different matter. Clipping a digital input results in a sharp-edged truncated waveform, which sounds extremely unpleasant and should be avoided at all costs. A digital meter will have a maximum level of 0dB, indicating that no further headroom is available. Exceeding this level will be immediately audible ! [Some better digital units have a system for limiting the signal to avoid this digital distortion, but this is a failsafe mechanism, and should not be relied upon to forgive incorrect gain levels, as it will still have a detrimental effect on the recorded signal, though considerably less offensive than the digital clipping it prevents]
One property of directional microphones is that as a sound source moves towards the microphone, the bass end of the audio spectrum becomes more and more dominant. This is because of a complex relationship between the phase and amplitude of different frequency bands within the capsule of the microphone, so bear in mind that - especially at distances in the order of a few centimetres - moving the microphone towards the signal source may result in disproportionately more bass being present in the output signal. This is not necessarily always a bad thing, but something you should be aware of.
The sound of any microphone will be affected by the way in which its output signal is amplified. Because microphone outputs are (compared to most electronic audio signals) very quiet, the delicate process of amplifying them will also tend to magnify any flaws in the pre-amplifier design.
Many esoteric (and correspondingly expensive) models are available, incorporating vintage circuit design, valves, or digital modelling etc. For critical voice-over and music recording applications these will have different benefits, but for general use many budget and semi-professional audio interfaces and mixing desks offer perfectly competent mic amplification.
Voice recording techniques
In a professional environment, narrators are almost always recorded in mono, using a single directional microphone, at a fairly short distance from the speaker. Narrative recordings will usually benefit from a dry or ‘dead' acoustic, with minimal reverberation, so be aware of the acoustics of the room when choosing your recording space, and try to find a position in it which gives as little echo or ‘ringing' as possible.
Different locations within a room will have different reverberant characteristics, so try out different positions and directions for the narrator to face, and monitor the response of the microphone on closed-back monitoring headphones until you find a combination you are happy with. Soft furnishings and heavy curtains will help to absorb reverberation, and - in particularly difficult spaces - hanging up absorptive materials (heavy velvet curtains, duvets etc) will help to control unwanted reverb. Digital reverberation is always easily added at the post-production stage, but excessive natural reverb cannot be removed.
The best microphone type for voiceover is the large diaphragm condenser, as it gives the most detailed and ‘warmest' quality of sound, with low background noise. A distance of between four and eight inches between mouth and microphone should give good results, though this may vary with the speaker, microphone, or room. Many good quality vocal microphones are available - some with direct-to-computer USB connection - and vary in price from the budget range (£60 or so) to the astronomically expensive (£2000+). Non-USB condenser microphones will usually require battery or phantom power.
Lapel microphones (interviews + lectures)
Clip-on lapel microphones are a discreet solution for recording in a situation where a larger mic might be intrusive, or where the speaker needs the freedom to move about without having to hold a microphone. In a static situation they can be wired directly to the recorder or preamp, but often are used in combination with beltpack radio transmitters to allow complete freedom of movement.
Lapel mics are usually omnidirectional, and this, in combination with their mobility and the additional gain often needed to amplify their signal, makes them quite prone to feedback. If the intention is to ampliy the signal in the venue with a public address system then lapel mics may not be the ideal solution. A fixed mic for this purpose, and then a lapel mic for recording may be better. It is almost inevitable that a lapel mic will pick up more of the ‘room noise', be that reverberation, additional amplification, or background noises, and little can be done to avoid this. Always try to amplify them as little as you can get away with through the P.A., to avoid feedback and ‘ringing'.
Preparing the subject
As well as the technical business of setting up a good recording system, the performance of the person speaking will greatly effect the quality and professionalism of the result. A couple of simple techniques will help them feel at ease and perform well. If you are recording yourself speaking, then they are equally applicable, so set a little time aside to ‘prep' yourself before recording starts.
It is vital to soundcheck your equipment before recording, not only to confirm that it is all actually plugged in and working, but also to set gain levels in the recording chain, as detailed above. While you do this, it is an opportunity for the speaker (if they are present) to familiarise themselves with the sound of their voice through the P.A. (if applicable) and to warm up. As previously mentioned, ask them to say a couple of phrases in a more forthright tone (not shouting), to make sure that the recorder is not overloading, and that feedback is not being induced.
2. Microphone Awareness
Speakers need to be aware of the position of microphones; for example, if you are using a microphone on a podium, and they turn round to look at a screen or blackboard, then sound will be momentarily lost. Try to make performers aware of the position of mics and their relationship with them, without compromising their presentation style unduly.
Something also to be aware of is the transmission of sound mechanically through the floor and podium. Banging of a hand or foot on a surface physically connected to the mic can transfer a loud ‘thump' to the recording, so make sure your performer doesn't get too carried away and start abusing the furniture!
3. Just Act Natural
Some inexperienced performers will adopt a different voice if they are aware of being recorded or amplified, or may be tempted to speak louder or slower. The best thing is just to be relaxed, natural and confident.
Interviewees will generally be more relaxed and animated if they are interacting with the interviewer as they speak. However, a stream of acknowledging chatter ('yeah', 'uh huh', 'mmm', 'i see' etc) is not really what you want on your recording. Professional interviewers are trained to nod silently (and smile, even!) as this let the interviewee know they are listening, but doesn't mess up the recording.
Some pre-amps incorporate analogue equalisation and/or dynamics processing to process the signal before digitisation. Effects processing is covered in depth in our Guide to Digital Processing, but briefly, these a few effects which are commonly applied to microphones, and especially vocal recordings:
Equalisation allows different frequency bands to be cut or boosted, and will enable enhancement or reduction of presence (mid frequencies), clarity and ‘air' (high frequencies) or depth and richness (low frequencies).See Digital Equalisation for more details
Compression sets a threshold level for the sound, above which the volume of the signal will be reduced by a preset ratio, and at a variable speed, to enable control of variations in the loudness of the recorded signal. It can also be used to increase the perceived volume of a sound, by quietening down the loud bits, and allowing all the quieter sections to be turned up (the reason that adverts are always louder on TV and radio is due to high levels of compression). All this is accomplished in measurements of milliseconds.
Limiting allows a ‘brick wall' maximum level to be set, beyond which the limiter will not allow any signal to pass. It is equivalent to a compressor with a ratio of ∞:1
Gating again sets a threshold volume, below which the sound is muted, at a given speed. This can enable background hiss to be cut when the speaker is silent, for example. This is known as a ‘noise gate'.
Enhancement is usually a term for a particular implementation of equalisation, to enhance a particular type of signal - often concentrating on vocal frequencies.
Reverb is the addition of digital reverberation to simulate different acoustic spaces.
Be aware that these effects can be difficult to remove once applied, so be judicious in their use. Equivalent effects can be easily applied non-destructively in the digital domain, in post-production, so unless you are completely comfortable with their use it may be best to record a ‘dry' (uneffected) signal, and then experiment non-destructively with effects later.