Step By Step 2: Tuning

In the last installment we looked at the design process for this system; now let’s tune it. I have three mics: a wired at FOH and two wireless. I will follow my usual color-coding convention: Green, Orange, Pink from top to bottom of the venue. That means the FOH wired mic is orange, and the two wireless are green and pink.

Step 0 - Mic Compare

Any time I use multiple mics, I start by putting them all in the same spot and comparing the traces to make sure everything is matched before we move them around the venue and compare them. Since my wired FOH mic is also used for sound level measurement during the show at minimum gain, I readjust the gain every day to match the level with the two RF units. (Figure 1)

Step 1 - Mains EQ

We have two operations to perform on the mains - EQ them to our FOH engineer’s desired target curve for this act, and also address the HF shading from front to back. In practice I will often do these two tasks in parallel when I have multiple measurement mics. So mic 1 goes up into the 200 level to measure the top of coverage, mic 2 stays at FOH and mic 3 goes down front towards the bottom of the array. Before we add any filters, we can see how we’re doing with our mechanical design by comparing these front-middle-back traces. (Figure 2)

*FIGURE 2 - Mains On-Ax @ Back - Middle - Front*

We see the expected high frequency loss at the rear of the venue, which we’ll address shortly, but let’s pause to address the seemingly concerning pile up at 150 Hz or so down under the array at position 3 (pink mic). Fear not - it’s just combing caused by a ground reflection. Usually ground reflections are not so pronounced as to warrant my attention but this one is made more significant by the combination of significant down-angle on the bottom box (-71°), which means a lot of the energy bouncing steeply back upwards off the floor and into the microphone thanks to Snell’s Law, and also the high trim height (bottom box 39.5 ft off the ground), which may be counter-intuitive so let’s explain.

The severity of a comb filter depends on the relative level of the two signals - in this case Direct vs Reflected. When you have a low trim, as you might with a line array hung in a small club, the D:R ratio will be larger and the reflection less severe. Let’s say by way of example the bottom box is 11 feet off the ground. If the measurement mic is at 4’, that’s a 7’ direct path to the mic, and a 15’ reflective path from the bottom box, to the floor, back to the mic (for the sake of simplicity we’ll assume it’s aimed straight down to make the numbers easier). D:R level difference is 20log(7/15) or about 6.6 dB.

In this design, the path from the bottom box to the mic is about 55 feet and the reflected path length is about 62, for a level difference of less than 1 dB. As we trim higher, extra length of the reflective path is a smaller and smaller portion of the overall path length and so causes less level drop, making the floor bounce more disruptive to the measurement, and the steep downward angle means more energy heading back into the mic instead of past it.
(Flip this logic around, and you’ll see why a rear wall reflection from an underbalcony delay speaker isn’t as troubling as a rear wall reflection from a main PA much farther away.)

Mocking up a similar mechanical geometry in MAPP XT allows us to prove this theory by removing the reflective floor, something we can’t do in real life: Figure 3 below shows the response at this mic position without (red) and with (blue) a reflective floor surface in play. Looks quite a bit like the real-world data, wouldn’t you say?

*FIGURE 3: Isolating the effects of floor bounce under the array by disabling the floor for the red trace.*

The obvious question here is why not use a ground plane / boundary mic position by placing the mic on the floor, as it will eliminate the floor bounce. The answer is that it also changes the overall tonality of the measurement in a way that is less representative of what a listener hears 5 feet off the ground. Besides, we’re concerning ourselves strictly with HF shading here, and the top few octaves are sufficiently undisturbed by the ripple that we can see clearly what we need to do. So I do the same thing I always do with floor bounce: ignore it.

So we need to push the HF on the top boxes, and EQ the entire array to our target. Easy enough. (Figure 4)

*FIGURE 4 - Left: Overall Main EQ. Right: HF shading on top 4 and bottom 4 elements.*

This is all rather gentle and non-offensive. Let’s see the result (same mic positions, now that we’re shaded and EQ’d to target (Figure 5):

*FIGURE 5 - Mains Post-EQ @ Rear, Mid, Front*

Step 2: Subs

We are off to a good start, and note that this is primarily accomplished mechanically with only relatively gentle filtering. Now we will add in our subwoofers, and the main + sub combination will establish the broadband tonal footprint for the entire system. The top left of Figure 6 shows the extremely high level of front to back consistency in the sub range - this is why flown subwoofers are such a valuable commodity. Subs on the ground cause an unavoidable buildup of excess level in the front rows that is often undesirable and inappropriate depending on the type of performance. I fly subs to completely eliminate this buildup and then use small “front fill” subwoofers in the pit to dial back in some artificial build up as is artistically appropriate.

I flatten the response with a gentle filter at 45 Hz and add my typical “de-crudding” filter around 110 to get rid of any nasally overtone that most subs exhibit in this range. Combining these with the mains yields the front-to-back uniformity shown at the bottom of Figure 6. (The main-sub timing doesn’t change from the default value in this case because their relative positions have not moved from the default design.)

*FIGURE 6 - Top left: subwoofers front to back. Top right: subwoofer EQ. Bottom: system response at positions 1,2 and 3*

Step 3: Side Hangs

Now the two RF mics (green and pink) go to two on-axis locations on the side hang: green up top in the 200 level, and pink down in the 100 level. The shorter line and wider splays means we don’t have to remove the lo-mid buildup we did on the mains, and they land pretty much on target without any electronic help(although our friendly floor bounce returns for the same reasons as above). We must remember to check the response with subwoofers unmuted as they are part of the target response over here on the sides as well. (Figure 7)

Step 4: Main-Side Timing

Now that we’ve set level and EQ for mains and sides, we will drop a mic on the seam between them and set their relative timing. In a venue as large as an arena, “seam” is a stretchy concept, as it’s over a hundred feet long. I typically approach this by going about halfway up the seating and dropping the mic on the seam where mains and sides are equal in level. Turn on the mains, set the measurement delay to the mains arrival so the peak is centered in the IR plot. Then add the sides. The new spike shows the relative arrival - in this case, 6 ms early. In case you’re not sure which arrival is which, mute and unmute the side hang while watching the Live IR plot and you’ll see it pop in and out. (Figure 8) This approach is much faster than the traditional approach of using the delay finder twice and noting the delta value.

*FIGURE 8 - Setting main-side timing using the IR.*

Here’s all those traces together - we’re in very good shape here. (Figure 9)

Step 5: Front Fills

Our front fill EQ and level shouldn’t change drastically from day to day unless the stage height or barricade distance changes (or we’re in a theater situation where there is no barricade and the front row is closer than typical). On this show I set them a few dB hot to make sure the vocal stays on top of the guitar amps coming off the stage. They are light blue in Figure 10 below for clarity. The one thing that does change from day to day is the delay, because they’re timed to meet the bottom box of the main hang, which changes in height. This seam is typically a few feet back from the barricade, and I set the timing in the same way as above, using the IR, only I set the front fills to be about 2.5 ms early so listeners near the seam will cleanly localize to the stage.

*Figure 10 - front fills trace and front fill subs trace*

Step 6: Listen and Walk

If we’ve been paying attention we should be very close to a final tuning by this point, but our ears are the final judge. I start my tuning playlist and walk carefully through the entire coverage area of the system, listening critically for transitions and changes in level and tonality between sources. Adjustments at this point are typically on the magnitude of 1 - 1.5 dB and 1 - 2 ms. If it’s more, I probably missed something in the measurement process and will go back and try to figure out where I went off the rails.

This entire process typically takes around 15 minutes, a bit less in a small venue and as much as 20 minutes in a larger venue, as the vast majority of the time is spent walking mics around, and sometimes getting up to 200 level takes a bit of a circuitous route, or finding that dog with the keys to let me into VIP level.