New MILSTD B&K Impulse Noise System Exceeds Military Requirements

For measurement of SPL and noise intensity, the sampling rate is only a restriction on the frequency of sounds you can accurately measure (you need to be sampling at a frequency high enough above the frequency you'd like to measure to ensure there is no signal aliasing or other data artifacts). The REALLY challenging part of getting accurate measurement of sound intensity (and the main limitation of most consumer-level devices) is the dynamic range. This is a measure of the difference in magnitudes of values that can be accurately measured by the system. So for instance a ruler might have a dynamic range of 11.875 inches (smallest value it can measure accurately is 1/8", largest values is 12", difference is 11.875"... this is a gross oversimplification). In the case of audio measurements, its the difference between the quietest noise and the loudest noise it can accurately measure. Any SPL meter must be capable of measuring down to at least 0 dB-A (the quietest detectable noise by the human ear) which sets the floor of the measurement range. For some context, if you wanted to be able to measure down to 0 dB, and only up to 120 dB, you need a device that can accurately measure magnitudes ONE TRILLION (120 dB = 10^12) times larger than the lowest magnitude its capable of measuring. To measure up to 165 dB, you now need a system capable of measuring something 32,000 TRILLION TIMES (3.2x10^16) times louder than the quietest noise it can measure. This would be like having a set of calipers that could accurately measure the width of a piece of paper to 0.001", and was also capable of measuring the distance from the Earth to Mars with the same degree of accuracy. THATS whats challenging and impressive about a system like this. NOT it's sampling rate.

SAMPLING RATE ISSUES WITH AUDIO ANALYZERS

POWER = ENERGY/TIME & HUMAN HEARING TOLERANCE
There is a HUGE misunderstanding in the gun loudness measurement area relating energy and time. Sound ‘Power’ is loudness over a given time. Sound ‘Pressure’ is simply its level without a time value. For example, a simple static shock can be 10,000 volts from a door knob it has extremely low power due to low time. But, get hit with 120 volts continuous by a power line and you are toast! Much like the body accepting high voltage levels over very short times, do this over seconds and your hearing is gone. There is a ‘teeter-totter’ effect where you can handle a couple loud shots and still recover your hearing. On the flip side, even with much lower levels over time, this constant pounding takes its toll and you still lose hearing even after using ear protection.

SAMPLING RATE
As for sampling rates, using a common B&K 2209 or Larson Davis unit, you can make loudness (energy) measurements but these are all deceptive since most analyzers simply are not fast enough to ‘integrate’ the energy over a very short period of time to give a true reading how high the value goes. If the sampling rate is too slow it is like opening and then closing a camera shutter quickly. If you take a photo every second of a fast moving car, you might see it in shot but if you take 30 shots/second (typical video speed) you must much, much more.

Extending the video analogy, if you take 30 photos/second this sampling rate (frame rate) is 1/30 or 33 milliseconds. If that car is in view of the lens for at least 33ms you will catch something but it is NOT guaranteed you will see a full photo of the car! Using a common rule in sample rates, you really need to have a factor of 5-10 the frame rate to ensure a good understanding of that image (sound and video). If we have a gun shot with most of its energy inside a 1ms time window, you need a sampling rate of at about 0.1ms or 10,000 Hz. If most of the action happens within 0.1ms, you need a sampling rate around 100kHz. Some rounds are VERY fast thus the Military demanding 192kHz rate for accurate data thus an impulse of 0.05ms is going to get sampled accurately. The 256kHz rate for the B&K is a little more safety margin.

REAL WORLD DATA
An older/slower, analog analyzer like the 2209 works fine for RELATIVE data when comparing similar products to one another and can measure insanely high levels. The 2209 is a WONDERFUL steady state analyzer and does run respond quite fast but like a Dodge Hellcat Hemi engine, it can only accelerate to top speed so fast (torque limited). In contrast, like a Tesla in Ludicrous Mode, the new B&K unit can out-accelerate the Hemi reaching its top speed much faster even if not as fast at top speed.

With a 2209 you can gather excellent ‘family’ data comparisons such as silencers on 9mm. But, when jumping from one extreme range to another (e.g., .223 to 50 cal) you will see huge changes in loudness and length of time the sound carries on. The 2209 takes a certain amount of time for all its circuitry to “integrate” the energy (charge and decay its circuits) so this “Integration time” is its limit in acceleration which impacts its ability to reach full swing data. For some rounds the impulse is just too fast thus the 2209 is not able to “gather” all the sound info and integrate it into one collective value. Generally, the 2209 will simply measure ‘low’ compared to the new B&K system but this exact value will have a lot of variations, probably 3-6 dB depending on rounds.

SOLUTIONS & OTHER ISSUES?
The ONLY way to collect data that can be directly (fairly) compared across many guns, rounds, and environments is either via the faster B&K system or via a fast field recorder (e.g., Sound Devices MixPre) which can sample at 192 kHz and keep this data for later analysis. I haven’t even touched microphones yet – this is a HUGE issue which I want to crawl through the computer to strangle people using cell phones for measurements!
(FYI, I am a BSEE, have done high SPL audio measurement work for 40 years. It is VERY tedious work and often difficult to make true measurements which people can trust.)