Wind Revival (part 2)

(Sample audio from the session – a quick balance with no percussion or synths added)

It’s been a couple of years since I chronicled a Fannin Musical Productions recording session.  John has a thriving business selling marching band arrangements to high-schools nationwide, and the live performances of show options he provides to potential clients is something of a distinction in the marching band biz. For me, FMP recording sessions have become almost routine. But with plenty of down days to set up for latest session (a luxury not always available) I decided to try a few new tricks.

We’ve always used Lovett Auditorium on Murray State’s campus for recording high-priority sessions. It’s a big, wonderful hall built in 1926 with a basketball court-sized stage, several rows of heavy theater curtains, and 2000 seats for the audience. One of the very first recordings I made at Murray State was on this stage: a violin and piano concert. It didn’t take me long to realize that this hall was probably the best location for recording a wide variety of projects. There are several other recital halls, theaters, and auditoriums, but Lovett is by far the largest. I have enjoyed the soundstage environment for everything from Americana sessions to classical performances. No changes in venue!

The thirteen wind players were also a consistent element. We had eight brass players: tuba, euphonium, trombone 1 & 2, trumpet 1, 2, & 3, and mellophone. These folks travelled from Memphis, Texas, and beyond to be with us. There were also five woodwinds: flute, clarinet 1 & 2, and sax 1 & 2. Excellent musicians all around. We planned to get two good takes of everything and stack them so our 13 players would sound like 26. No changes there either; so far so good.


The analog time keeper.

The first significant change was pretty simple: a brand new audio interface. This was my second session ever using the Apogee Ensemble Thunderbolt (which replaced our well-loved and utterly spent Ensemble Firewire after many years of service). The technology between the two Ensemble designs represents about a decade of innovation. I really like this updated interface! The Control software is eons ahead of the old Maestro software, and the hardware unit provides a lot more control right from the face plate, not to mention offering a total of 30 inputs compared to the old model’s 18. Plus it’s jet black, which is really nice after staring at brushed aluminum for so many years! Using the new Ensemble is very intuitive. I found myself thinking, “I want to change THIS…,” and simply clicking on the name of any channel, source or destination and making it happen without trying to think like a software engineer. And of course working in Logic means the integration with the Ensemble is tight. I could tweak parameters on the hardware interface (like soft limit or 48v) from inside my Logic project file! It’s always a little risky to try to work a new piece of relatively unfamiliar gear into a session, but it turned out to be easy to keep the Apogee playing nicely with the monitoring system and two extra banks of digital inputs. And yes, I do think this new Ensemble sounds superb.

The second deviation was the most obvious visually and seemed potentially risky (from a morale perspective). It involved splitting up the winds into two lines: brass up front and woodwinds in the back, with 30 feet and two thick theater curtains hanging between the groups. Previously everyone had been in one long line stretching across the stage, but I hoped this new setup would reduce brass bleed in the woodwind mics and help everyone sound better. It definitely made a difference by creating enough isolation to actually use a distant woodwind section mic, whereas before there was no point given the spill. Aside from potentially making the woodwinds feel like second class citizens tucked away in the back of the hall, the video feed over which they watched the conductor had just a tiny bit of latency as well. However, given that they had a click track keeping time in their headphones this seemed to not cause any serious issues. Who watches the conductor anyway? They were good sports, and I think we’ll continue this practice in the future.


Looking down the brass line.

The 8 brass players were aimed out into the cavernous auditorium, which helped to minimize the acoustic energy bouncing back into the mics (and back through the curtains to the woodwinds). My AEA R88 stereo ribbon mic was in front of the first row of seats with a makeshift reflection filter behind it in an attempt to reduce the ambience of the hall. The R88 is a huge part of the sound for this project. It may be only one mic, but it is probably 50% of the sound for the whole thing. I regularly use the R88 as a room mic for almost any project (the list of things that don’t greatly benefit from some R88 blended in is very, very short). But there is something really special about the way it works on brass in a big room. Once you hear it blended in with the close mics, it just sounds like you want it to. I love all the mics used in this session for different reasons, but the AEA is the one mic that has to be out in front NO MATTER WHAT.


LDC on Bone 1; dynamic on Bone 2.

With several of the 2 or 3 member small instrument groups I tried to use different types of microphones. For example, I used a condenser mic on trombone 1 and a dynamic mic on trombone 2. I have found that using the same mic all over the place can sometimes lead to a build up of that one sonic signature. And no matter how great that signature may be, having it on every source usually isn’t so great. Think of it as a way to help differentiate similar sources (not to mention avoiding that condenser sensitivity at every station).


Ribbon mic on trumpet 1, SM7b on #2, and Miktek PM9 on #3.

The last big change was a major focus on phase relationships. I acquired Voxengo’s PHA-979 phase plugin this Christmas, and per my MUS 338 textbook, Mixing Secrets for the Small Studio by Mike Senior (chapter 8), I set out to examine how small timing and phase rotation changes might preserve more of the tone with all these different mics bleeding into each other, not to mention their relationship with the stereo room mic. There just wasn’t time to listen and move the R88 around during the very abbreviated sound check. Back in the day that’s how it would have been done. But with 13 people waiting on me, I had to settle on a distance for the room mic that has worked before and plan on fixing individual phase issues in post production.

The PHA-979 definitely made a big difference in correcting thin, nasal, phase-cancelled tone. Even the 3:1 spacing rule can’t prevent phase cancellation. And while I made sure there was 6 feet of space between players and no more than 2 feet from mic to intended source, the PHA-979 helped mop up the problems. It’s not an enticing plugin like vintage gear simulations, but it sure solves some core issues. For example, among the three trumpets I found that trumpet 1 was really bleeding into trumpet 2’s mic (trumpet 1 is LOUD). So upon combining those two mics, trumpet 1’s tone began to suffer due to his own bleed spilling into trumpet 2’s mic and being out of phase. I adjusted the phase of trumpet 2 (and therefore the phase of the bleed from trumpet 1) and got trumpet 1 sounding great again! Powerful stuff; thanks Mike!


Avenson STO-2 SDC on the ever-important mellophone.

The mellophone was the only brass instrument – aside from the tuba and trombone 1 – to get a condenser microphone. Actually, if you look closely, you’ll see there is also a Shure SM57 up as well. I auditioned both and decided the Avenson STO-2 small diaphragm condenser had a deeper, rounder sound. Mellophone (or horn, depending on who’s available) is a super important element to the brass line. It provides the middle ground, or the glue, that holds everything together. It has to be prominent, and given it’s easy-to-pick-out range, it needs to sound nice. Next time I’ll try the Avenson and some other mic to see if I can find a mic that can beat out the STO-2, but so far the little Texan omni is in the lead.

I also used several MikTek PM9 vocal microphones (on euphonium, trumpet 3, and bone 2). These are bright-ish, somewhat less midrange-y dynamic mics that sound a little different than the standard SM57 or SM58. I have also used MikTek’s drum kit pack (the PM10 and PM11) in these sessions before. They are great for splitting the difference between a dynamic and condenser type of sound.


Miktek PM9.

I’ve had good results running low brass through the Millennia HV-3D preamp for a long time. If it’s good enough for Skywalker Ranch and the LA Phil, then I think it’ll probably take care of me. I like to keep things on the low end fairly clean, and clean is the HV-3D’s bread and butter. The 4 channels were spent on euphonium, trombone 1 & 2, and horn. The Buzz Audio SSA 1.1 two-channel preamp is another clean machine, but it has a great extended low end and lots of gain. Just a touch less detail than the Millennia, but that’s actually what I want from my room mic. Perfect for the AEA R88. I used one Audient ASP880 as an ADC for theses two preamps, bypassing it’s internal preamps altogether and just piping things into the Apogee digitally. The second ASP880 was in full swing, though, routing some sources out and back from my 500 series rack of EQs and compressors, and handling preamp duties for the trumpets and some distant section mics.  Lastly, the on-board Apogee Ensemble preamps handled the woodwinds.


Sputnik tube condenser – NOT looking directly down the bore, but rather at the bell’s edge.

I tried running the tuba through the Focusrite ISA 430’s opto compressor for a thicker, more syrupy low end. If I was really on my game I would have alternated between VCA and opto compression between every double take for maximum variation. Sometimes the tuba can poke out, so I used a very gentle setting just to take a few dB off the peaks. The ISA 430 is just a little bit juiced-sounding. It seems to have just a touch more bottom and top than a more neutral preamp. This is usually the sort of sound I want for tuba in a marching band. I also engaged Focusrite’s “air” button, because I wanted as much brightness as I could get without resorting to EQ on the front end (and the opto compressor took a little edge off, too).


The woodwinds watching John conduct on TV, thirty feet behind the brass line.


Fathead Ribbon on sax 1.


AKG C414 XLS large diaphragm condenser on flute.

The flute and trumpet 1 ran through a Radial Engineering EQ set flat, just for the nice little punch that 500 series unit gives. They don’t seem to get enough love, but Radial makes great-sounding hardware. Then said high-end sources went through a couple of dbx 560A compressors to roll a few dBs off the peaks. I was conservative with the compression, but next time I’m clamping down harder on both these two! I really like the AKG 414 mic on just about anything, and the flute is a great fit. As for trumpet 1… you’d think that a dark ribbon mic (off axis from the bell) going through a compressor might be in danger of sounding too dark. Yet there were times when he really punched it and I thought I should have used that Radial EQ more aggressively.


The woodwind line with diffusor/gobos covered in John’s old drapes.

I tried a little experiment with the clarinets. They were both on sE Electronics RN17 Rupert Neve designed small diaphragm condensers (with the omni caps). These are mics I usually use as a stereo pair in front of the choir or as outriggers for the Wind Ensemble. They sound wonderful. I knew there would be bleed from other sources with the omni caps, but I decided that on one of the two double-takes I would pan each clarinet hard left and right. The idea was to increase the stereo spread with two ambient-heavy tracks. I could’t do this with both takes, as there were places where the extreme panning didn’t work, but it seemed to work well when the two takes were combined to make 13 players sound like 26.

I sang the praises of the PHA-979 already, but there is another plugin that saw action on even more channels. The Brainworx bx_console N is an emulation of a Neve VSX analog recording console. All 72 channels of it. That is to say, each channel is slightly different-sounding (just like the real console). No, I didn’t wade through 72 options and find the “perfect” sound for each source. But I did click through 5 or 6 channels at random and chose my favorite from that small sampling. Unlike the PHA-979, the console N IS the enticing sort of plugin that makes people drool. The N was used on all the tracks for the second take (the stacked parts) for the whole ensemble. Think wide, warm, smooth… a nice analog juicing which, when compared to the uneffected signal, was clearly preferable. Because I was stacking two takes, it was easy to make the choice to only treat one of the two. I was trying to get the best of both worlds and not over do it, and this also helped further differentiate the takes from each other. Another cool aspect to this plug is that the Neve EQ and compression emulations are built-in, so if I decided later in the process that I needed a little of one or the other, it was already loaded up. No need to drag another plugin over. Besides that, having a virtual channel strip for any track you choose is pretty cool!

Mix bus plugins included the Virtual Tape Machine, FG-Grey bus compressor, Brit 4k E console bus emulator, and FG-X limiter – all from Slate Digital. EastWest’s QL Spaces also found a place for its nice soundstage ambience.

I’ll have to wait until I arrive at the finished product to feel certain, but so far I think the sound of this session is a significant improvement. The biggest contributors to this being the extra separation between the woodwinds and brass and Voxengo’s PHA-979 phase correction plug. The PHA-979 seems purpose-built to help correct the excessive bleed from a session such as this. Looking forward to using it on drum kit! Cheers!