7 Simple Tips for Improving your Audio Recording Quality - By Pascal Roobrouck of Alcatel Lucent

This list is the result of experience collected at Alcatel-Lucent University. It contains 7 tips to improve your audio recording quality and make professional sounding recordings, even with affordable microphones.

1. Keep the background noise down: Noise is part of the audio signal that you don’t want to hear. The more noise in your recording, the more difficult it becomes for the listener to ‘filter out’ the good part of the audio, and as such the more difficult it becomes to listen to your recording. First you need to take measures to introduce as little noise as possible into your recording. (Further below are tips on how to deal with noise already into your recording). So when you are into the location where you’d like to record, stand still and listen to the background noise that is present. Normally you are not aware of this noise, because our brain tends to filter it out. If you start actively listening, you will notice that there are things like noise from air-conditioning or ventilation, radio’s – TVs, machines, people talking, phones, etc. Try to get to a location with a minimum of this noise, and if possible try to shut down (during your recording) equipment that makes noise. As an example I often need to shut down the refrigerator in the room. Also the fan of a computer can generate noise, so you could use a long cables and put your computers as far away as possible from the microphone. One last kind of noise, is caused by mobile phones. They emit very strong signals and those are picked up by the microphone, so keep phones away from your recording spot. See also here and here.

2. Keep the levels correct : Each microphone has a maximum audio signal that it can record without distortion. Recording audio that is louder will result in so-called ‘clipping’ : the signal is limited to its maximum level, causing a terrible distortion. On the other hand, each microphone also picks up noise. The noise is fairly constant, so the lower the audio signal you record, the smaller it becomes versus the always present noise. Therefore you should try to put your microphone in such a position (or tune the sensitivity or gain if you can) that the signal you record, results in the maximum level it can record, but not more. As audio-levels of speech vary all the time, the easiest way to check your levels, is to do a test-recording, and then look at the wave-form in your audio mixing software. If there are peaks reaching the maximum level (often called the 0 db level), then you are recording too close . If your waveform only uses 20% of its vertical range, you are probably recording too far away from your source. One final tip here is that often, a speaker will use his ‘normal’ voice during the test recording, but raises his voice as soon as there is an audience. You should try to anticipate this. In the ideal situation, you can monitor the levels during the recording, and slightly adjust them when needed, eg. when changing speakers in the middle of a recording.

3. Microphone handling sounds : Inside a microphone, there is a so called diaphragm that is moved by the airwaves. Any movement from the body of the microphone itself, will result in some movement of the diaphragm as well and as such result again in a kind of noise. As this noise is not constant, it is also much harder to filter out later. So you should carefully position your microphone in such a way that you never need to move it or touch it. On a table is OK, but not perfect, because people tend to move their hands on the table, which moves the table , and hence moves the microphone on that table . A better way is to have the microphone hang on a beam above your table. Also there are solutions that suspend your microphone with rubbers, which to a certain extent keep it ‘floating’ and isolated from movements and vibrations. If you have several people to interview, it is best to provide each of them a separate microphone. In case you just have one microphone, be aware that moving it to the next person will introduce noise, so leave some time between the move and the next question to allow you to properly cut the noise during the move.

4. Include 10 seconds of silence at beginning and end : Even if you take precautions to minimize background noise, reality is that with affordable microphones and real-life environments, you will always have some noise. But today’s Audio Processing software (such as Adobe Audition) allows to filter it out to a large part. This works better if the software ‘knows’ what is the kind of noise you have in your recording. So simply record a 10 second long silence at the beginning (and/or end) of your recording. This part contains just noise. From this part the software will learn your ‘noise profile’ and be able to filter out the noise from the remainder of your recording in the best possible way.

5. ‘Normalize’ and ‘Process Dynamics’ : After recording, you will mix several parts together in a so-called mixdown. To be able to mix properly, all different recordings should have the same level. If you took care of tip 2 (see above) this will almost be the case. Still the first processing you can do on your recording, is to ‘Normalize’ it. This will amplify the complete file in such a way that the loudest part of your file is amplified to a chosen maximum level. The absolute maximum you can set is 0db, but in my experience it is more reliable to normalize to a maximum level of (eg.) –1db, which is just a little under the absolute maximum. This leaves a bit of margin for rounding errors. Now, having normalized all your recording files, you may experience that still they do not all sound equally ‘loud’. The reason for this is that there could be a peak in your audio file, which is very loud, so the file as a whole cannot be amplified, because then the peak would be more than –1db and result into distortion. A solution for this is to ‘selectively’ amplify the signal : amplify the low parts more than the loud parts. The technical term for this process is called ‘dynamics processing’. If you apply this onto a voice recording, the peaks of the recording will be ‘flattened out’ to some extent, so it is then possible to amplify the whole file without distortion at the peaks. Feel free to experiment with a sequence of ‘dynamics processing’ followed by ‘normalization’ followed by ‘dynamics processing’ etc. See also here and here.

6. POP Filtering screen : While pronouncing the letter P (as in the word Pop) your mouth exhausts a sudden blast of air. This results in an air wave which is sometimes too strong for the microphone to digest, and so again it distorts the recording. A simple way to prevent this, is to stay further away from the microphone, or to place a so-called pop filter between your mouth and the microphone. See also.

7. Clap a Marker : While recording, if something goes wrong, you can stop the recording, discard it and start again. You can also just continue recording and later edit out the mistake. If the recording was good until the mistake, I suggest to just keep recording. In order to easily correct the recording later, you can simply make a loud clap with your hands. This will result in a visible spike in the waveform, and make it easy for you to locate the re-take. Be sure to remove all the bad parts, as well as the claps themselves before doing the normalization and dynamics processing (see above).