Go here to read part 1 of this discussion.
Michael Lawrence: Jim, you and I have had a lot of conversations – and written a few articles together – about the concept of the spectral response of a sound system, or the “tilt” of a PA as we often call it. In the past we’ve focused mainly on the subjective elements of the discussion – why certain people might have a preference to mix into a flatter or more tilted PA – but this time out let’s dive more into some of the technical justifications for why someone might opt for one approach over another.
We’ll start with the topic of headroom. If we think about the “out-of-the-box” response of a large-format PA system before any EQ, we might reasonably expect to see a significant tilt that has the bass range 12 dB or more above the midrange. So we might grab some low-frequency EQ to cut that back and balance the rig back to “flat,” but now what happens if we happen to be mixing an act or a genre of music that calls for a very tilted, bass-heavy sound, like reggae or hip hop?
We then have to create the tonal signature of that mix in the console itself. And I’m a big proponent of that because it makes your streams and board recordings sound “correct” right off the desk. But if the mix needs to be very bass-heavy, the problem is that we’ve used EQ to effectively reduce the voltage gain of the PA in the LF, so we have to drive the desk really hard to get the PA up to level at LF. That 12 dB LF cut means that 12 dB more LF energy is needed in the board mix to get back to the same place, and that can really eat up a console’s headroom.
I don’t typically mix in those bass-heavy genres but I have a friend who does that type of music almost exclusively, and he’s told me that voltage gain and headroom can become a limiting factor when trying to get that kind of response out of a flat PA, particularly when the PA doesn’t have a ton of power to spare. That’s an interesting factor in this discussion that a lot of people might not have considered, and although I definitely prefer a flatter PA target curve, I think this is a really valid reason why it might not be the best solution for all acts, from a technical perspective. What do you think?
Jim Yakabuski: I can definitely see the logic in that and have experienced it to some extent while mixing at home during the break using nearfield monitors. To achieve a similar low-end sensation to mixing in a full spectrum PA system with subwoofers, I usually have to either beef us the low-end EQ on inputs like kick, bass and toms or add a global hump around 60 to 80 Hz on the main left-right. I can confirm that it adds considerable gain to left-right as evidenced at the meter.
As with many approaches in live sound mixing, I think the answer lies in compromise and content-based choices. If the target frequency response is hugely tilted in the lows and subs, as you might expect with EDM/hip-hop and some rock ‘n’ roll, then taking all of those frequencies out of the PA system’s tilted default preset – only to add them back in on the channel EQ or the main left and right – doesn’t make a whole lot of sense and without a doubt decreases the console’s headroom.
But your point about “linear” sounding mixes into an almost flat transfer function tuning of the PA has a lot of merit, and it’s something I’ve been shooting for a lot more lately. I want my console, all on its own into a pair of nearfield monitors or headphones, to sound totally full frequency – full, big and punchy – so that every mix I send out of it, whether video record feeds, lobby and backstage feeds, and PA matrix feeds all sound big on their own. Then I tailor the PA to translate that. This works great if the target show SPL is 90 to 95 dBA, not running hotter than 100, but if tons of level is needed, power and sub, you may as well just leave all of that in the exaggerated humpy PA curve and tailor the mix into it. Smoother levels at the main left-right and all the power needed in the sound system.
ML: Good point. I definitely think it’s fair to say that headroom is an important consideration here, and if you don’t have enough “rig for the gig,” you might be shooting yourself in the foot a bit by cutting back so much energy out of the PA. I have to say, I’ve really been enjoying the whole “accurate board mix” benefit of a flatter target curve, knowing I can just print the file and off we go without having to worry about what EQ I need to get it to translate.
Another aspect of this I wanted to hear your thoughts on is the idea of “convenience EQ,” which is something we’ve touched on in the past. When I was mixing into more heavily tilted systems, I found myself doing a lot of LF EQ work on a lot of my inputs – lots of high pass and low shelf, carving out a lot of that energy below 250 Hz or so. And once I started mixing into a flatter target curve, I wasn’t having to do that. I’d just roll up the high-pass filter and be pretty much where I wanted to be. So, in that case, the “convenience” comes in doing the EQ once (on the output side) so I don’t have to replicate it 24 times at my inputs.
But besides the convenience, is there a more technical benefit as well? I mean sure, who wants to use more filters than they need, right? But I’m thinking in terms of signal chain – a transfer function from the mic to loudspeaker, including everything in between. If we consider the time-domain effects of all those filters, if we’re using fewer filters, we’re doing less smearing to the impulse response, keeping the signal more intact, and that seems like a benefit.
I think about the board mixes done by my friend Chris Mitchell (FOH, Umphrey’s McGee) and how clean he’s able to keep the transients because he doesn’t use EQ on the inputs – he just gets a carefully-chosen mic in a carefully-chosen spot (on an instrument played by a skilled musician of course!), and his board mix sounds really tight and clean. Then there’s a single stage of EQ at the output side that takes care of getting the PA to reproduce his board mix properly, and you’re off to the races.
So in that sense, I’ll gladly go with whatever PA target curve leads me to need the least amount of corrective EQ on my inputs. For the acts I’m typically mixing, that’s more or less a flat curve with about a 6 dB bump up in the sub range, and a very slight rolloff above 1 kHz or so – this seems to be what I always come back to. But I have to imagine, if I were working for more bass-heavy acts, it might make more sense to tilt the system up a bit, which would probably still leave me in the situation where I didn’t have to do a ton of EQ on my inputs to get the mix together.
I know you’ve been experimenting over the last few tours you’ve done with getting the PA to a less tilted curve – did you notice that you were able to ease up and pull some filters out of your input EQs once you made that change?
JY: For sure! Without a doubt. The less exaggerated the tilt in the low-end, the less hacking away at the high-pass filters on inputs that don’t need to live in the subs (vocals, acoustic guitars, etc.). And the same applies to low-frequency and low-mid EQ cuts on inputs – I found I was cutting much less in this area with less boost in the PA from 315 Hz and down.
So, here’s another aspect of the tilt discussion I was thinking about the other day: What sounds better? Does my mix sound punchier, smoother, fuller and more coherent if I just leave the PA preset as it’s delivered to me by the manufacturer and make the necessary adjustments to the input high-pass filters to “mix into the system,” or do I use the best tools available to me (manufacturer loudspeaker processors and software EQ/shaping filters) and flatten out the low end first so that the transfer function of the entire system (including subs) when measured looks almost “flat”?
Well, the answer is: I don’t know yet. With all this time away from PA systems I just won’t know until I get a chance to do a listen test. But I intend to create two console and sound system snapshots, one with the PA tilted and the necessary input EQ cuts and higher high-pass filter frequency settings, and another snapshot with the PA thoughtfully “flattened” using the best tools available. Will a kick drum and bass guitar sound better with a bump in the 50 to 80 Hz range on the channel EQ and the PA flatter, or will it feel more pleasing to the ears and coherent (and transient) with the tilt left alone, starting with the manufacturer’s default preset? I’ll have to get back to you on this.
ML: That’s a great question. My gut feeling would be that whatever combination of choices results in the least applied EQ overall would keep the transients the most intact and sound “better,” however we’re choosing to define that. But a listening test seems like the way to go, for sure. That sounds like a lot of fun – I’m sensing a “Part III” to this discussion in the future!