Choosing Kit: How to Make Sense of Audio
George Lucas once said: “I feel that sound is half the experience. Filmmakers should focus on making sure the soundtracks are really the best they can possibly be because… sound is where you get the most bang for your buck.”
If like George Lucas you recognise the importance of audio to filmmaking, you may find choosing cameras, microphones or portable audio recorders a frustrating experience. Many manufacturers’ specs ignore audio altogether; others give figures that are simply untrue. And even when decent audio specs are given, they aren’t standardised across different products, making comparison difficult.
But don’t be disheartened! In this article I will try to help you make sense of the audio specs you do come across, as well as offering a few alternative ways to find great-sounding kit. May the force be with you…
Top tip: When we talk about sound levels, we use the unit decibels, or dB. Decibels don’t work in a simple, linear way like centimetres; instead they tell us the logarithm of a ratio between the thing being measured and a reference quantity. More about the maths here, but the most important thing to note is that a difference of 9dB equates to a doubling of perceived loudness and -9dB a halving. For example if one manufacturers noise specification is 9dB lower than another’s the hiss on quiet recordings will sound half as loud.
Manufacturers’ specs: what to look out for
1. Frequency response
Many devices capture middle-frequency sounds well, but you may be disappointed to find that low-pitch sounds (like bass drums) and high-pitch sounds (like birdsong) are all but lost from your recordings. A frequency response spec tells you the accuracy with which a device captures tones within the range of human hearing. It is measured using an audio test set to feed a sweep of pure tones from low to high frequencies at a constant level into the device and back into the tester. A spec of ‘20Hz to 20kHz ±1dB’ is good; it means that within the range of audible frequencies, there’s no more than 1dB (10%) of variation. Often a graph of frequency response is given, especially
in microphone specs.
All devices have a level of ‘self-noise’, meaning that even if you record a perfectly silent scene there will be a certain amount of hiss on your recording. Obviously you want this to be as low as possible. Frustratingly, different specs treat noise differently; some give an unweighted measurement, while others weight their measurement in a variety of ways to take into account the sensitivity of the human ear to different frequencies (a very low or high hiss will be less bothersome than a mid-range hiss). This can make it hard to compare noise specs for different products, but this article should help .
Top tip: When comparing noise specs with different weightings, note that figures quoted ‘ITU-R 468-weighting’ are approximately 12dB higher than figures quoted ‘A-weighted’.
3. Maximum input level
All devices have a maximum input level; if you try to record sounds louder than this, your recording will sound distorted. The difference between the maximum input level and the noise level tells you the dynamic range of the device. The dynamic range of most microphones is far greater than the dynamic range of a recorders analogue to digital convertor. This is why good recorders provide an analogue preamp stage for gain adjustment, which allows you to match the range of levels used on the mic (for a given recording situation) to the range used on the A/D. This of course assumes that the noise on the preamp stage is lower than that of the A/D, which is not always the case. When buying a microphone and a recorder to use together, you want the dynamic range of your recorder (plus the usable dynamic range on the recorder’s preamp) to be at least as large as that of your mic. In reality, though, this is nearly impossible to calculate as specs fail to mention the gain setting at which the measurement was made, or the available gain range of the recorder, or the point at which the noise from the analogue gain circuitry exceeds the noise on the A/D.
Top tip: A well-designed 16-bit recorder will provide 96dB (A-weighted noise) of dynamic range at its A/D and another 30dB of preamplification. However, most recorders only provide 80dB at the A/D, regardless of whether they’re in 16-bit or 24-bit mode and 15dB of usable preamp gain.
You needn’t rely on specs alone
With many specs confusing or misleading, it’s lucky that there are these three alternatives.
1. Use your ears
Look out for videos with great sound, ask friends and colleagues, buy or borrow kit and give it a try. Many filmmakers are ill-informed about sound, though, and if you get into audio you may find that what your friend considers good sound is not actually up to scratch (unless of course your friend is George Lucas).
2. Buy an audio test set
If you get really into audio, this is your best option – if I were working on a feature film I wouldn’t commit to any piece of equipment without thoroughly measuring its performance first. Audio test sets start at a few hundred pounds and could be a good investment.
3. Look at the Lindos database
At Lindos Electronics we’re creating a database of audio test results. It’s still in its infancy, but anyone with one of our test sets can upload results and it should become a comprehensive resource for instant comparison of equipment.
Top tip: So far, the best recorders we’ve tested are those made by Tascam.
A passionate audiophile, Chris Skirrow works for Lindos Electronics, a UK-based company that has made audio test sets for 30 years. Lindos has recently entered the microphone market with the MiniSonic Mic Kit, which can be connected to your camera (or your computer or portable audio recorder) for professional-quality sound at an affordable price.