The Future of Microphones
Microphones and preamplifiers are drastically inefficient devices in the modern age. But before we explore that thesis, let’s review the development of microphones.
Early condenser microphone design
We can very crudely divide the world of microphones into dynamic and condenser microphones (let’s ignore the exceptions for a moment). Dynamic microphones are characterized by very low output (voltage), but good current drive capability (low output impedance). Condenser microphone capsules are the opposite: good output level, but extremely high output output impedance, such that condenser microphones require an onboard amplifier circuit to act as a current buffer.
The first solution was derived before the invention of the transistor: the tube condenser microphone. By its nature, an inefficient design, due to the lost power in the tube filament. But when the first FET condenser microphones were invented, the power demand was small enough that a dedicated power supply was no longer required. Phantom power came into common use as its main competitor, T-power, faded from use (unlike T-power, a phantom powered microphone input is also compatible with dynamic microphones; a major design advantage).
Early FET designs, such as the Neumann U87, used phantom power mainly as a source of the capsule bias voltage. The internal FET needed only a small amount of current, and an output transformer was required to yield the necessary low output impedance. The Naiant X-T and X-M/T microphones used the same basic approach.
So for either FET condensers or dynamic microphones, a significant amount of voltage gain was required for any purpose, whether live sound reinforcement or magnetic tape recording.
Modern condenser microphone design
Fast forward to today: many things have changed. Most condenser microphones are transformerless, electret condenser microphones of excellent quality do not require a high supply voltage for capsule polarization, and digital recording is the most popular choice. In spite of these factors, there has been a trend towards increasing demand for phantom power supply current.
At the same time, portable and battery-powered devices are growing rapidly in popularity. Full-specification phantom power is extremely expensive from a portable perspective: a 9V battery can supply about 5Wh, and full-spec P48 at 10mA per microphone costs 1W for a pair of microphones–far too expensive for a portable power budget.
What’s more troubling is that the vast majority of that power is wasted. When you draw 10mA from a standard P48 supply, that means 0.34W gets dissipated in the phantom supply resistors, and only 0.14W actually does any work inside the microphone.
Let’s think about how much power is truly necessary. Many digital converter ICs run off of a +5V supply rail, at most. Let’s assume the load of that IC is 10K (it is often higher), that means the IC only needs a tiny 300μW to run full scale. And yet, our microphone could be using 1,500 times that much power!
But wait, we haven’t accounted for the preamplifier yet. Most professional preamplifiers are designed to output a signal at +4dBu with over 20dB of headroom. The +4dBu is fine; that’s pretty close to the maximum level our converter wants to see. But wait, aren’t we supposed to leave digital headroom? Yes we are! So we don’t really need much more than +4dBu, do we? That extra unnecessary headroom creates wasted power in the preamplifier, and that too-hot signal is just padded down in the converter before it hits the ADC IC.
Condenser microphone design in a digital world
It’s time for a new way forward, but one that still respects the need for backwards compatibility. It already exists! The P12 standard is far more appropriate for the modern portable digital world than the P48 standard. Yes, +12V is too low for direct polarization of externally biased condenser microphones, but a step-up circuit is trivial to implement, and electret capsules of high quality that do not require a high polarization voltage now exist in a range of sizes and polar patterns. Without the large loss of power to the supply resistors and microphone circuits, efficiency is massively improved.
Next, preamplifiers should be regarded as a technology of the past. Condenser microphones already possess increasingly sophisticated amplifiers; they should be designed to output a high sensitivity that always ensures that their maximum dynamic range can be captured without requiring external amplification. Analog to digital converters should be designed with variable input attenuators to cope with loud sources, and should supply the required P12 phantom power. Dynamic microphones should also incorporate low-noise P12-powered amplifiers.
Naiant is ready for the future today–every Naiant X series microphone supports the P12 phantom power standard, and can be configured to generate a -10dBV line-level output. Of course, Naiant microphones are also fully compatible with the P48 standard, maintaining backwards compatibility with past equipment.
What about digital microphones? To date, these exist in two classes: AES42 and USB microphones. AES42 is an important concept, but for the time being is a high-cost and high-power solution. AES42 allows 250mA at 10V, which is 2.5W, or five times what an analog phantom powered microphone may consume. USB microphones are not regarded as professional tools, partially because many implementations to date have been less than robust, but also because the USB transmission protocol limits it to relatively short 5m cables and extenders. Power consumption can either be high-power (5V 500mA, the same 2.5W) or low-power (5V 100mA, 0.5W).
I believe that as USB microphones improve in quality and the cost of AES42 interfaces drop, these two systems will ultimately merge into a single microphone design that is capable of dual USB/AES42 operation, using the USB low-power minimum requirement of 100mA, whether powered at 10V or 5V, and perhaps can also support analog phantom-powered connections as well.
A newer standard is AES67, specifying standards for audio over IP operation. Given the low implementation of AES42, some commenters have felt that perhaps AES42 will become obsolete in favor of AES67. I am not entirely convinced of that; while audio over IP can greatly simplify system interconnection, I don’t feel that a further increase in complexity at the microphone is as helpful, since it is only at most a stereo single-direction audio signal.
There could be some potential for a standardized instruction set (since the AES42 instruction set is unique to that standard), but until the market decides on a single standard for AES67 devices, that benefit will not be realized. AES67 allows use of standard CAT-5 cables, but microphone cables must run in environments where durability is critical. A CAT-5 cable with the durability of a 110 ohm AES42 microphone cable would be the same cost, so the benefit is limited to cable standardization (and uniformity of connector gender).
There will be increasing demand for wireless digital microphone solutions that are interoperable rather than proprietary. There isn’t a clear leader in the market for this solution, but there are obvious benefits. Improvements in compression algorithms could make technology such as Bluetooth or its successors compelling, but allocation of bandwidth will also be a critical issue. Will the 2.4GHz band remain empty enough for multichannel audio, even with low-loss compression? Will the UHF TV bands currently used for traditional wireless microphones all eventually be reallocated to mobile services?
Although it is currently difficult to pick a winning standard, it seems a likely bet that the next ten years will bring about fundamental changes to the existing standards and practices for microphone interconnection that have endured for the last 40 years.