Note also that input overloading is a strong function of the preamp’s gain control setting. Most preamp manufacturers measure the maximum input level with the gain control set at minimum. This means there is a real danger that even after carefully matching the output and input levels of a mic and preamp, the mic can still overload the preamp.
This happens when the system needs the preamp gain turned up (correspondingly reducing input headroom) and the microphone is used for a wide dynamic range source. Not only do you have to worry about matching your mic and preamp, but also about real-world sources and gain settings.
Individual Noise Floors
Microphones and preamps each have their own noise floors. When selecting a preamp, it’s important to know to what degree the preamp’s noise degrades the noise of the mic.
Different mic technologies use different terminology to describe noise. Dynamic microphone data sheets rarely list noise as a specification because there is no active circuitry to generate noise; there is only a magnet and a coil. This category of mic is properly called electromagnetic or electrodynamic.
Output noise is very low—so low it’s not listed. However, some noise is generated, and this can be calculated by knowing the mic’s impedance.
Obtain the dynamic microphone impedance rating from the data sheet and use Table 2 to convert that into units of dBu, A-weighted. This noise is the white noise generated by the resistance of the wire used to create the coil, plus a correction factor of 5 dB for A weighting. (This is somewhat arbitrary, as true A weighting may decrease the level anywhere from 3-6 dB depending upon the nature of the noise, but agrees with Holman’s article—noted later—and measured results.)
The noise of the measuring standard 150 ohms (200 ohms for Europe) source resistor makes a good noise reference point. In Table 2, it equates to -136 dBu (A-weighted; -131 dBu when not). This means that you cannot have an operating mic stage, with a 150 ohm source, quieter than -136 dBu (A-weighted, 20 degrees C/68 degrees F, 20 kHz BW). Looking at Table 2 confirms that dynamic microphones, indeed, are quiet.
Use Table 3 to compare microphone output noise with preamplifier equivalent input noise (EIN). As an example, if your dynamic microphone’s output noise equals -136 dBu, and you are considering a preamplifier with a rated EIN of -136 dBu, then the difference between them is 0 dB.
Table 3 illustrates that this preamp with this microphone will degrade the total noise by 3 dB. That is, the combination of mic and preamp adds 3 dB noise to the total. More on how this table works shortly.
Condenser, capacitor, or more properly, electrostatic microphone technology involves a polarizing voltage network and at least a buffer transistor built into the microphone housing, if not an entire preamp/biasing/transformer network—all of which contribute noise to the output. Electrostatic microphones are quite noisy compared to dynamic designs, but are very popular for other reasons.
Different manufacturers use different terminology on their electrostatic microphone specification sheets for noise: Self-Noise, Equivalent Noise SPL, Equivalent Noise Level, Noise Floor, and just plain Noise all describe the same specification. Microphone noise is referenced to the equivalent sound pressure level that would cause the same amount of output noise voltage and is normally A-weighted.
This means the noise is given in units of dB SPL. A noise spec might read 14 dB SPL equivalent, A-weighted, or shortened to just 14 dB-A (bad terminology, but common). This is interpreted to mean that the inherent noise floor is equivalent to a sound source with a sound pressure level of 14 dB.
Problems arise trying to compare the mic’s noise rating of 14 dB SPL with a preamp’s equivalent input noise (EIN) rating of, say, -128 dBu. Talk about apples and oranges!
Luckily (again), tables come to the rescue. Table 4 provides an easy look-up conversion between a microphone’s output noise, expressed in equivalent dB SPL, and its sensitivity rating, in mV/Pa, into output noise expressed in dBu, A-weighted.
Using Table 4, a direct noise comparison between any microphone and any preamp is possible. The example shown by the blue column and row is for a mic with a noise floor of 14 dB SPL and a sensitivity rating of 20 mV/Pa, which translates into an output noise of -112 dBu, A-weighted.
Now, time to return to Table 3. Unfiltered electronic noise, whether from a resistor, a coil, an IC, or a transistor is white noise consisting of all audible frequencies occurring randomly. Due to this randomness you don’t just add noise sources together, you must add them in an RMS (root mean square) fashion. Mathematically this means you must take the square root of the sum of the squares—which is why Table 3 is so handy—it does the RMS conversion for you.
Use Table 3 to convert a mic’s rated noise output into units of dBu. Find the difference in dB between the mic’s output noise and the preamp’s input noise. Find that difference in the left column of Table 3 and read what the preamp added noise will do to the mic’s noise in the right column.
For example, if the mic’s output noise translates into -120 dBu, and the preamp has an EIN of -127 dBu, then the difference between the mic and the preamp is -7 dB. That is, the preamp is 7 dB quieter than the microphone. Table 3, at the row marked -7 dB, tells you that this preamp will degrade the mic’s noise by only 0.8 dB. Looking at Table 4 tells us that after about a 10 dB difference, the noise added by the preamp becomes insignificant.
Similar to Table 1, you can use Table 4 to map out a preamp’s A-weighted noise to show the combinations that add insignificant noise. If you use a -10 dB difference figure as a guide, then the preamp’s noise amounts to less than 0.4 dB increase.
The red-shaded triangle area in Table 4 shows an example of this. The areas not shaded represent all possible combinations of mic sensitivity and noise specifications that can be used with Rane’s MS 1b Mic Stage, for instance, and add less than 0.4 dB of noise.
If you allow 1 dB net added noise, then even more combinations are possible. (The shaded area is figured by taking the EIN of the MS 1b at -128 dBu, reducing it to -133 dBu with the 5 dB factor for A weighting, and using the -10 dB difference found in Table 3 for 0.4 dB added noise, resulting in all combinations less than -123 dB being blocked out.)
The author would like to point out that this note was inspired by an article authored by Tomlinson Holman, published in September 2000 Surround Sound Professional magazine, titled “Capturing the Sound, Part 1: Dynamic Range.”
Dennis Bohn is a principal partner and vice president of research & development at Rane Corporation. He holds BSEE and MSEE degrees from the University of California at Berkeley. Prior to Rane, he worked as engineering manager for Phase Linear Corporation and as audio application engineer at National Semiconductor Corporation. Bohn is a Fellow of the AES, holds two U.S. patents, is listed in Who’s Who In America and authored the entry on “Equalizers” for the McGraw-Hill Encyclopedia of Science & Technology, 7th edition.