Speech is one of the most difficult sound sources to record well. Why is this so? What should be done to achieve a perfect recording?

How to record speech to radio and audio book standards

Speech is one of the most difficult sound sources to record well. Why is this so? What should be done to achieve a perfect recording?

Familiarity

The main reason why recording speech is difficult to do well is that we are so familiar with how it should sound. How many people do we talk to during the course of an average day? With every encounter we become increasingly familiar with the sound of the human voice.

Playback through loudspeakers

 Recorded speech cannot sound exactly like a real person. One reason is that moving coil loudspeakers (which are the kind in most common use) always sound a little bit 'loudspeakery'. Even a highly accurate electrostatic loudspeaker has different directional characteristics to the human voice.

Microphone distance

Another reason is that the distance at which microphones are commonly used for speech is much closer than we would normally listen to another person from. For speech, a microphone is almost always positioned much closer than one meter away from the mouth. One meter is about the minimum listening distance in a real-life social interaction, other than between intimate partners.

Background noise and ambience

The combination of the human ear and brain has a remarkable ability to filter out and ignore background noise and ambience. Microphones however do not have such abilities. Any background noise or ambience is permanently printed onto the recording, and heard very clearly on playback. Worse still, speech has silent gaps and pauses during which background noise and ambience can be heard. Music on the other hand tends to be more continuous and can obscure a certain amount of noise. Music also commonly benefits from ambience, where speech is almost always better recorded quite dry.

Popping and blasting

'P' and 'B' sounds cause the mouth to emit a short pulse of air. When a microphone is quite close to the mouth, this pulse can strike the diaphragm, causing an audible and unpleasant 'pop' in the signal. Similarly, other speech sounds can create blasts of air that distress the diaphragm, causing unpleasant sounds that would never be heard in a normal person-to-person conversation.

The simplest way to guard against popping and blasting is to move the microphone from the direct line of fire of the breath, but still point it at the mouth. So, for instance, if the person speaking holds their head level, and the microphone is raised to the height of the forehead and pointed down at the mouth, popping and blasting will hardly ever occur.

This sometimes looks a little odd, but this is because the conventional microphone position directly in front of the mouth, so commonly seen, is in fact wrong.

Another method, although it is not 100% effective is to use a pop shield consisting of an open mesh material stretched over a wire or plastic frame. A third method is to use a windshield, which is an open-cell foam material shaped such that it can be placed over the grille of the microphone (including any openings behind the diaphragm). Once again this is also somewhat less than 100% effective.

Another solution, although more costly, is to use a more highly directional microphone, such as the Sennheiser 416, which works on the interference tube principle. Because it has a very narrow pickup pattern, it can be used at a greater distance from the mouth where any air blasts will have dissipated before they can strike the diaphragm. The drawback here is that the sound quality of interference tube microphones is generally not as good as other designs.

Sound insulation and acoustic treatment

To get a really good, professional sound on speech, it is necessary to have a recording area with good sound insulation, and acoustic treatment. Sound insulation will limit the level of background noise coming from outside. Acoustic treatment will reduce the amount of ambience within the room. Clearly it is vital not to have any sources of noise within the room.

It is quite common to use small booths for speech. However, it is difficult to acoustically treat such a small space without it sounding audibly 'boxy'. A larger, room-sized space, with acoustic treatment, can often sound better although clearly it will cost more to set up properly.

Choice of microphone

The most accurate type of microphone is the small-diaphragm capacitor microphone, particularly models designed to have an omnidirectional pickup pattern. This could be recommended to anyone who is keen to achieve the most natural sound possible.

However it is more normal to try and achieve the 'expected' sound, rather than the most natural sound. Large-diaphragm capacitor microphones are often used. They are not as natural-sounding, but they have the sound people expect to hear, so they are popular for speech.

Vacuum-tube capacitor microphones (sometimes called 'condenser microphones' - condenser is the old word for capacitor) can work really well for singing as they create a sound that is perceived to be 'warm'. However, they can sound a bit buzzy on speech, and the sound can be tiring to listen to over a long period. The same applies to vacuum-tube microphone preamplifiers. Having said that, the vacuum tube sound might be useful for radio, where a sense of exitement is often required.

Summary

Although it would be nice to be able to provide a summary in one sentence, in reality all of the points explained above require attention to achieve the best possible sound on speech. There is however one final point, which is that being easily satisfied is not the route to success. Only when you have achieved the ability to make recordings of speech that are fully equal to the best audio book standard can you say that you have truly mastered this difficult skill.

Publication date: Monday December 24, 2012
Author: David Mellor, Course Director of Audio Masterclass