Archive for Techniques
Nuts and bolts of PA tuning part 2
Yesterday we covered the thought process in tuning the main speakers in our PA. Today we’re going to discuss delay fills.
After starting with the main speakers, I work outward from there until we get to the speakers in the back of the room. Often the first step after mains is downfills. I like to work my way out from the mains because as I add more speakers to what I’m listening to, they all start to work together and I have to follow that energy rather than fight it. Every change you make to one set effects everything else so you have to think holistically and experiment.
The key to making fill speakers work is two things – delay time and EQ. If the main speakers are our starting point and benchmark that everything else must align to, fill speakers such as downfill or delays will normally arrive at either slightly different or perhaps greatly different times to your ear depending on where you are standing in the room. The trick is to use delay so that speakers that are separated by 5 feet or 75 feet sound like they are all arriving at your ear at the same time and thus working together. I use my software program to help calculate the necessary delay times, often times with a bit of trial and error in real world to nudge things forward or backward just a bit so it feels right.
After the delay time is right, I like to first turn off the speaker I’m getting ready to work with and just measure how everything else we’ve already optimized sounds in the location of the fill speaker. This tells us what is missing sonically in this location from what we’ve already done and that is what we will focus on with EQ’ing the fill speaker. I will high-pass the speaker so the low frequencies that are already hitting the area from the main speakers are not competing with additional stuff trying to come from the fill. Now balance the sonic EQ of the fill so the response curve sounds as close to standing in front of the main speaker as possible. Continue this process with each set of fills and eventually all parts of the system are operating at the same time.
As I said, this is a very high level discussion because the reality is that the process of tuning a PA is as much art as it is science. For me, depending on the complexity of the system and the acoustics of the room it is installed in, there can also be a fair amount of experimentation & trial/error. One of our largest rooms at newlifechurch.tv has a distributed audio system design – 4 main arrays in the front of the room with downfills below them and then a ring of delay speakers 2/3′s of the way back. It took me three attempts at tuning this room before I arrived at a product that I’m reasonably happy with as a long-term starting point. Each time I did it, I learned more about how all of the speakers interacted with each other, the acoustics of the room, and the result was less and less needed EQ each time.
In my experience with trying to get a great board mix for broadcast from the same console mixing FOH, a well-tuned PA is absolutely required to have any chance of success.
Nuts and bolts of PA tuning
When I first visited newlifechurch.tv last winter, I knew immediately that one of my first tasks when we got here would be to start from scratch with PA tunings and system optimization across the board. The tell-tale sign this would be necessary came from looking at channel EQ on the consoles. Channel after channel showed LOTS of EQ and I heard complaints that the board mix that fed the internet campus never sounded as good as it did in the room.
I have a basic process that I follow when tuning a PA. Everyone probably approaches something like this differently so your mileage may vary, but here’s a sampling of my thoughts.
First and foundational for me is to not take anything for granted pre-existing in the system from past engineers, the installation company, or “helpful” volunteers. Crossover points in bi or tri-amped speakers, the manufacturer recommended EQ points on a box, amplifier processing either bypassed or enabled, every speaker functioning, balanced levels between boxes, and on and on. I like to begin by ensuring that every box is actually functioning, the array is balanced so that the level remains consistent as you walk the room, and that all DSP in the system is either flattened, bypassed, or disabled.
There is nothing worse then spending hours working on a set of speakers to then find that a circuit was engaged somewhere, thus coloring what you’re doing, and you didn’t know about it. In one of the newlifechurch.tv rooms, just this step alone brought huge improvements to the system because I found that one of the three boxes in each of 4 arrays was operating at 50% of the volume of the other two boxes due to amplifier trim. The result was unbalanced coverage front to back that had likely existed for a long long time. Don’t take anything for granted!
Next, I’ll always begin a tuning process with the main speakers. In some rigs, like one of our main rooms, this is all I need to deal with because there are no delays or fills to add into the mix. The main speakers will always carry the biggest load of the work in a system and put the most energy into the room. Because of this, I like to start here and then fit the other boxes around the mains. I’ll talk about the software I use in another post. For now we’re keeping it to a 10,000 ft level and just talking process.
The ear doesn’t hear things “flat”, especially as the volume level increases to concert levels. As such, I’m not looking to create a flat PA. Some guys named Fletcher & Munson did lots of research years ago on this hearing phenomenon, resulting in the Fletcher – Munson curves. (Google it if this is completely new to you – fascinating stuff.) In working with the main speakers, I’m looking to smooth them out sonically and ideally emulate a Fletcher-Munson curve (smooth cut in the PA that starts around 1k and has its deepest point at 4 or 5k before returning to normal by 10k).
I go back and forth between measuring a 2 second sine sweep in order to graph frequency response and listening to a playlist of room tuning songs from my iPod that I’ve been using for this purpose for years and KNOW how they should sound. Often times you can over-tune a set of speakers by going crazy with every little dip and peak on a frequency response curve, but the real test is how it actually sounds with music. The magic is in a healthy balance between the two – science and art. In the end, the ears always win.
Tomorrow we’ll cover delay speakers…
Audio for Video revisited
I’ve written over a year ago about the audio for video processing chain I developed at Kensington with a Venue console but thought it might be useful to revisit the process in a new place with a new console (this time, Yamaha M7CL’s). The cool thing is that I’m using the exact same techniques as last time and the results we’re getting are stunning. If you haven’t adopted a process like this for your outside world feeds, why not? Seriously. Ok, here we go.
First, it is extremely important that you be able to monitor your mix through headphones on the console or record the L/R feed, play it back, and it sound good. This is a biggie. If you’re working with a big room (more than 750 seats for this example) and sonically the mix you’re listening to just ain’t happening, you most likely need to revisit how the PA is tuned, something in the speakers themselves, etc. I have a series of posts coming in the next week or two about my philosophy when it comes to system equalization & PA/room tuning so we’ll dive into all of that later. For now, I’m assuming the mix you’re hearing in headphones sounds good, but its just dead – sounds like it was recorded in a studio.
We’re going to add two pairs of mics to the room. The first pair is a set of shotguns that will be placed somewhere along the front corners of the stage, out of the way, aimed out perpendicular to the front of the stage and in such a way they can throw out into the room without picking up too much of the first couple rows. I have ours on tiny floor bases that make them just poke up over the front lip of our stage, pointed up at probably a 15 degree angle so they aim over the heads of those first couple rows. These mics will be the primary pickup point for the audience themselves. Pan them to 9 o’clock and 3 o’clock. The exact make and model of these mics is not that important to me. At Kensington and now at newlifechurch.tv I have used Audio Technica 8035B’s from Sweetwater, the least expensive name brand shotgun I can get. I might feel different about the level of quality necessary if I ever tried better mics, but I’ve always had to do this project on a budget so bang for the buck rules here. I’m most interested in the pickup pattern rather than necessarily the sonic character.
The second pair of mics needs to be condensers hanging about half way back in the room. These can be high and also out of the way. I have mine about a foot below the lighting grid so they are very high. At Kensington I used some Audix small diaphragm units that were sitting on a shelf. At newlifechurch.tv, I’m using some existing Shure MX choir microphones that were already in the air. The purpose of these mics is to add space to the recording, so it will feel like its happening in a big room (which it is). These get panned in the console to hard left and hard right.
Now in the console we do two things. First, we split up all of our inputs to one of two main busses – left/right and mono. You must have a console that has left/right/mono discrete busses for this to work. M7CL’s do. Everything music related (band, vocals, playback, effects) routes to the left/right and we call this one music. Everything speech related (pastor mic, MC’s, and spare pastor mic) go to the mono and we call it speech. Now we need a few matrixes – a pair for the PA and a pair for what I call WORLD. In the PA matrix, route music and speech at the same level. But to WORLD, add about 6-8 dB to the speech side. This will effectively balance out the perceived difference between music and speech on a recording.
Step 2 in the console is really easy with an M7CL – add the two pairs of audience mics in to the WORLD matrix we just built. The ratio between the shotguns and hanging mics I’ve found sounds best in our rooms is almost 2:1 shotgun to hanging. The hanging mics will wash things out really quickly so the trick is to get just enough to add the depth and dimension without putting in so much that it totally collapses the mix. EQ both pairs of mics by adding a high-pass centered at least at 250 hz, pull out some 400hz & 2k, and I like to add a bit of sizzle to the hanging mics. The last thing to do is turn on the buss compressor over the WORLD matrix, 10:1, medium attack & release, and set the threshold so the band pulls 2-3 dB off the top of the mix when things hit hard.
The finished result sounds like this…
Audio clip: Adobe Flash Player (version 9 or above) is required to play this audio clip. Download the latest version here. You also need to have JavaScript enabled in your browser.
DISCLAIMER: This is the first audio sample I’ve posted from NLC. It’s getting there but still a long way to go! Figured I’d give you samples in process rather then just waiting until it is all perfect. Ok, resume…
On our M7CL’s, I’ve invoked user levels in order to lock out access to the room mic channels and the WORLD matrix so that everything happening behind the scenes to build this mix will be protected and can be relied upon.
I’d love to hear your modifications on this system and, if you try it, your experiences.
M7CL tricks part 2
Today’s trick for the M7CL is not quite as involved as last week’s, but I think it is very powerful for efficient workflow.
Everyone has their own ideas for how best to lay out the 12 user defined keys on the console. Up until a few months ago, my norm was for keys 1-8 to be direct sends on faders buttons for mix busses 1-8, with keys 9-12 reserved for a couple mute groups, talkback, and a scene advance button. However, I discovered an option for user keys that I used to utilize on the PM1D and never realized existed on the M7CL: page bookmarks.
The touchscreen on the console is a blessing and a curse to me. There are some parts of operating the software that lend themselves quite well to a touchscreen but sometimes I miss having more direct access buttons to things such as the EQ & dynamics sections, as well as FX processors that reside on the Rack page. Assigning a user defined key as a page bookmark is REALLY handy because it allows me to bring back some of that fast access I want to menus or screens I’m getting to all the time so I can remove a touchscreen key press and instead get somewhere quicker with muscle memory and button feel.
I’ve started setting up my user defined keys so the first row of 4 are set up as page bookmarks that get me in 1 button press to the EQ detail, Dynamics detail, and my VOX Verb & VOX Delay. It’s a simple thing but it feels like it speeds up my workflow around the desk because now I can select a channel, press my EQ shortcut button, and then start dialing away at that channel’s sound. Same for dynamics. I’m constantly going to my verbs to dial them in exactly how I want them to sound. Without this shortcut, getting to that edit screen is at least a 2 step process and might be 3 depending on what screen I’m coming from. Now 1 press of the shortcut brings it up and another brings me back where I was previously.
Check it out. You might find page bookmark shortcuts does the same for your workflow.
Making the M7CL sing
I’ve been blessed to spend the past 4 years or more mixing on some really great, really big desks. When I first arrived at Kensington, we owned a Yamaha PM1D that, while I’d had previous experience on years before, I enjoyed learning inside and out. Then 2 years later we changed directions and embraced the Digidesign (now Avid) platform. Along the way I also was able to get my hands wet a bit with Yamaha M7CLs.
There are some tricks I’ve learned along the way from absorbing content from accomplished engineers in the field that made life mixing on the Venue and PM1D a bit easier and gave me better mixes. A new challenge since coming to NLC.tv has been trying to find creative ways to get the same bang for the buck out of the M7CL. I had already started exploring some of this process in my last few months at Kensington but I’m working now to flesh them out a bit more. I’m going to share some of my favorite M7CL tips and tricks over the next few entries (there’s too much here for a single post).
Today we’re going to tackle parallel compression. I’ve written on this before, as have others, but in its simplest form, parallel compression means double bussing a set of inputs to two different signal paths on the console. In the first path, everything remains clean and unprocessed. In the second path, a nice compressor is placed over the signals and they are compressed as a group, usually pretty hard with variable attack & release times depending on the song. Then the clean and squashed signals are recombined before going to the stereo bus on the mixer, for me usually at a 2:1 ratio of clean to squashed. This is especially magic for vocals and snare/toms for me. Mixing the styles of music that I do, vocal intelligibility is normally one of the most important goals I’m fighting for and getting the vocal to sit nicely in its place with the rest of the band can be challenging. Since I started implementing this parallel compression trick last fall, it has done wonders to the ease with which I can accomplish vocal consistency I really like. It becomes an even more powerful tool the larger the vocal group becomes. At Kensington it was normal to only have a single lead vocal and perhaps a BGV or two. At NLC.tv, 5 to 6 vocals is the norm with sometimes as many as 7 or 8 on a given weekend.
Doing the parallel compression thing on an M7CL is really easy. First, I like to set up 2 busses as fixed busses instead of variable so I don’t have to worry about making sure the sends to them are all at unity. This can be accomplished under Bus Setup in the console setup menu. Next, as long as I have enough mix busses available, I like to set up one buss for the clean group and just call this one VOX. I unassign the VOX channels themselves from going straight to the L/R buss and instead route them to this VOX “subgroup”. While I plan to keep the processing here as clean as I can, especially the larger the vocal group you have, it can be really handy to have a single place you can grab an EQ and deal with a problem area that effects all of the vocals during the heat of mixing.
Now I also route the VOX channels to the 2nd group that I call VOX Smash. This group is setup just like the first one except on this one I engage the compressor on the buss, set to a 6:1 ratio with a medium attack and release. The M7CL has an excellent feature called automatic delay compensation so even though the same channel is going through two signal paths with different processing times, they stay perfectly in sync so that when they are combined into the master L/R buss, they are still in phase with each other.
If you’ve never tried this concept before, I can’t suggest strongly enough that you do. I always had the preconceived notion that a trick like this was only available to execute on larger desks but have been very pleased with the results I can achieve on our M7CLs and have started sharing the love with all of our engineers on this easy and effective mix technique.