Tech's Files: The Time and Space ContinuumUNDERSTANDING AUDIO PHASE AND POLARITY IN STUDIO ACOUSTICS 6/01/2009 8:00 AM Eastern
There's a joke about a technician who, after years of retirement, gets called in to solve a problem that no one else has been able to repair. After inspecting the machine, the tech marks a small “x” in chalk to show where the problem is and then submits an invoice for $50,000. When the company demands an itemized accounting of the charges, the engineer responds, “One chalk mark: $1; knowing where to put it: $49,999.”
Dedicated pros who invest a lifetime in their craft have a wealth of priceless knowledge. Nowhere is this more applicable than the elusive science (and art) of acoustics. A good acoustician can walk into a room and point out the most obvious errors without any diagnostic tools other than eyes, ears and experience. The on-site consultation fee from an acoustics expert can be money well spent.
Not all of us have the luxury of a well-designed workspace, and even “good” spaces should be re-evaluated from time to time. If that sounds like an invitation to go on a D.I.Y. journey for the terminally curious, welcome to this month's column, where I'll offer an overview of the most commonly used sonic diagnostic tools, plus a new dance: the oscilloscope twist.
In control room acoustics, one of the primary goals is to minimize reflections between the monitors and the listener. If all of the flat surfaces were mirrors — such as the desktop/console, sidewalls and ceiling — then reflections would be obvious, as would the optimum location for acoustic treatment. With mischief in the presence and “air” regions managed, “imaging” would be the most notable improvement — that is, the “realness” and depth of the phantom center image. Better acoustics translates into a more full-bodied center image. Analyzing and solving problems below 250 Hz can be more elusive, but let's start with the basics.
An acoustician's tool kit may seem impressive, but these tools are only as good as the user's ability to interpret them. Optimizing the control room environment may be unfamiliar territory as compared to the process of getting a good drum sound, but both applications share some common ground. The pre-DAW engineer's tool kit was “limited” to mic position, polarity reverse and EQ — all still valid. Sometimes, the either/or polarity option yields interesting and different results rather than a definitive “better,” and that's when workstations reveal themselves as more than just mixing and production tools.
Via DAW, the ability to freeze a moment in time allows for a greater understanding of how a single sound source arrives at multiple destinations — mics and ears — where the resulting combinations can make or break a recording, as well as your ability to monitor accurately. Now that we can zoom in on a multitrack drum recording, it becomes obvious that the 180-degree polarity-reverse option was, at best, a compromise.
We might not be able to time-travel, but using a DAW, time can be manipulated — from coarse sliding of individual tracks (in milliseconds or microseconds) to ticking off samples, “time aligning” by shifting the phase in degrees. It should also be noted that equalizers do their sonic sleight-of-hand by manipulating phase. An all-pass network like the Little Labs IBP Junior is a prime example of a tool that can improve the relationship between, say, a mic and a DI or an inside kick mic and an outside kick mic by allowing continuous 0 to 180-degree phase shift. This is great on a live gig where real time is essential.
The phase relationship between two loudspeakers is also critical. Tweaking a digital crossover network on a KRK control room monitoring system was a revelation for me, as it allowed me to manipulate the phase relationship between the midrange cone driver and the dome tweeter. When the timing is wrong, the sound is like every bad monitor you've heard, and when it's right, your ears just know it and no other piece of test equipment is necessary. Considering that the crossover frequency is typically in the middle of the ear's most sensitive region, optimizing crossover phase can also significantly reduce ear fatigue.
A Slice of the Sonic Pie
Just as an equalizer can manipulate bass, mids and treble bands, a spectrum analyzer divides the audio band into many smaller slices to reveal the energy in each band. This can be useful when the user is sorting out a mix, but when combined with a pink-noise generator and an omnidirectional mic, it can also help unravel the mysteries of a control room's frequency response. As pink noise is random, the display can be a bit erratic. A good spectrum analyzer allows the information to be averaged — like watching in slow motion — which is especially useful for observing LF response.
Initially, it's better to look at one monitor at a time, yet it's also important to find the room's acoustic center. Start with both monitors on, place the mic on a boom in the approximate center and then slowly pass through the “center” (between both monitors) to get the best, smoothest high-frequency response. If the response looks like a comb filter, then the mic is not in the center. Any untreated reflections will complicate the process (think mirrors). Once the mic is centered, look at one monitor at a time. The response should be the same, assuming the room design and monitor placement are symmetrical and the monitors are correctly wired and properly functioning.
Back when large studio monitors were the norm, a third-octave EQ was typically inserted into the monitor chain. While this was called “room tuning,” the more correct term might be “voicing.” IMHO, the cure was often worse than the disease, especially when radically different EQ settings were applied to each channel. The treatment should be identical on both channels; otherwise, left/right phase anomalies can seriously degrade the stereo image.
When a spectrum analyzer shows positive “bumps” in the curve, it doesn't indicate the bigger problem of resonance. Sure, a little subtractive EQ helps and in some cases may be enough. Similarly, no amount of EQ can fill in the holes. Understanding why requires more than the two-dimensional analysis of amplitude and frequency, but suffice to say, trying to force a solution with EQ is not the way to go.
You might expect relatively flat response from a close-miked monitor, yet the room will interact with the monitor in a way that the spectrum analyzer can't fully reveal. Spectrum analysis is great at assisting the ear in finding the trouble spots and for documenting the before/after curves. The third dimension is reverberation time for large spaces and decay/resonance for small spaces.
With eyes closed, the ears know when a space is absorptive, reflective or diffuse by how sound bounces around (or doesn't). Enter the Waterfall Plot, a 3-D representation of frequency, amplitude and decay time. Instead of random pink noise, an impulse is used to stimulate the space. This can be electronic or mechanical, such as a simple handclap, a balloon popping or two drumsticks clicked together.
Resonance implies that some frequencies take longer to decay than others, whether due to room dimensions and their ratios or to construction-related sympathetic vibration (windows, walls, cavities, etc.). Mid- to high frequencies behave similarly to light, but low frequencies are more squirrelly. The Master Handbook of Acoustics, by F. Alton Everest, states, “All room modes terminate in the corners of a room.” Surely, you've noticed how bass frequencies are more intense in the corners, and other boundaries often louder than at the monitors.
Here's an unusual way to reveal the direction of the offending waves and ultimately determine the treatment location in real time. Back when disk mastering engineers were obsessed with out-of-phase, low-frequency information, there was always an oscilloscope set to X-Y mode to show the phase relationship between the two channels. (For more info on this, see the “Audio Science” sidebar.)
Start by setting up a pair of cardioid mics in X-Y mode at the primary monitoring position (the sweet spot). Connect the mic preamp's outputs to the 'scope's inputs, also in X-Y mode. Most workstations have a multifunction audio generator capable of pink noise, square and sine waves. Start with pink noise to get levels and X-Y balance. Prove to yourself that everything is as it should be by reversing the polarity of one mic channel as shown in the graphic of the 'scope displays.
Switch the audio generator to sine wave and slowly sweep from 250 Hz down to 40 Hz. Along the way, your ears are likely to notice bass bumps and holes — it's useful to document these — and, hopefully, the scope's X-Y display will concur. Set the oscillator to a “hole” frequency, preferably above 100 Hz. Grab a portable bass trap, like Real Traps' MiniTrap or equivalent (2×4 feet by 4 inches thick), and move around with the trap until you find a position that has the most dramatic positive effect on the 'scope's X-Y display. The trap will vibrate when it intercepts the wave and may do so more when at a 45-degree angle. Note that the position of the room's door — open, closed or in-between — can also affect the path that the sound takes around the room.
The goal is to find a location and orientation for the bass trap that improves the phase at the listening position. Once you get that far, you can experiment with the density of the trap. Good luck!
Eddie's acoustics toolkit includes Smaart Version 5 (for its 24th-octave spectrum analysis) and Wavelab 5 for its ability to turn an impulse recording into a Waterfall Plot.
An oscilloscope is typically used to measure amplitude and frequency over time, much like a DAW's waveform screen. But in X-Y mode, each audio channel becomes a horizontal or vertical line that individually shows only amplitude, but together opens a window into the relationships between them, generating a Lissajous pattern. In-phase mono is a 45-degree diagonal line. Reverse polarity on one channel flips the diagonal direction. A phase difference of 90 degrees yields a circle.
— Eddie Ciletti