Memoirs of a Sound Mixer: "Jersey Girl"
by Glen Trew I’ve never won a lottery. I’ve never even won “dollar day” on a film set. But if I’m not the luckiest guy in the world I must be close to it because when I’m not working in the Trew Audio “candy store”, I’m a sound mixer working in production. And, as if that’s not lucky enough, fate landed me on the best Kevin Smith film yet produced, “Jersey Girl”. Kevin Smith is a writer first, and his stories are told almost entirely through the voices of the actors. There’s no gunfire, no crashes, no explosions, no fires; just people talking and going through life. And not just any people. The words were spoken by the likes of Ben Affleck, Jennifer Lopez, George Carlin, Liv Tyler, and Matt Damon (to name just a few). The Director of Photography, Academy Award winner Vilmos Zsigmond (“Close Encounters of the Third Kind”, “Life as a House”, “Witches of Eastwick”, “Deer Hunter”), not only brought an incredible look to the film, but from my perspective he was fantastic for the sound department. If we needed help with a shadow or reflection, or whatever, he would always try to give us what we needed and was usually able to deliver. Too bad I found this out well into the movie! Glen Trew with "THE Cart" on Park Avenue in NYC. In elementary school when I was asked why I wanted to become a Sound Mixer for film and television, I answered, “Because I want to help people”.  Probably the way I helped this production most was by using a non-linear digital format. But long before I had the job, it was decided by postproduction supervisor and assistant editor, Phil Benson, that the film would take advantage of the Deva and Cameo non-linear recording system. For those not familiar with this system, I’m referring to the Zaxcom company’s portable hard disk recorder (Deva) and companion digital mixer (Cameo). There have always been many good reasons to use this system, and many films have already put it to good use. However, I believe “Jersey Girl” is the first to take full advantage of what the system has to offer: Magless dailies, text info recorded into each take, full 24 bit depth, multi-track, DVD-RAM media, no DAT backup, dual copies (from one recorder) at break-off, super-fast transfer (4 track-hours in about 8 minutes), etc. (More on the Deva and Cameo system later.) Prepping for this show was not so much different than other shows, the possible exception being that there was a fair amount of music recording and playback. So, my compliment of equipment included: Deva Recorder (no DAT or other back-up recorder used) DVD-RAM drive with Direct Drive 12VDC powered enclosure Cameo Mixer Z-Tap battery system Fostex PD-4 DAT recorder (for music playback) 6 Lectro wireless bodypack mic systems (using Sanken COS-11, Countryman B-6, and Katamount lavaliere microphones) 1 Zaxcom “plug-on” wireless mic system (for wireless boom) 2 Flat panel video monitors 2 Comtek transmitters and 10 receivers 2 Sennheiser MKH-60 short shotgun microphones 2 Sennheiser MKH-50 hypercardioid microphones 2 Sanken CUB-1 boundary microphones Playback system consisting of four Bose 802 speakers, one Bose subwoofer, two Yamaha power amplifiers. 6 Comtek “earwig” induction earpieces My fully enclosed custom cart Magliner Jr. “follow cart” for utility See the Jersey Girl Photo Album for more pictures! The New 216mHz Stuff from Comtek Wireless monitoring for the “Village” (director, script supervisor, and special guests) was better on this show than I’d ever had before. There were several reasons for this. First, I used the new Comtek 216mHz system, which operates on what seems to be a clearer band than the older 72mHz system (now often interfered with by business data communications). This system used frequency agile transmitters and receivers, and can operate wide-band and companded for much improved fidelity. I had the “BST 216” base-station transmitter with a remote dipole antenna. The system sounded so good that the director used full sized Sony MDR-7506 headphones and actually had a pretty accurate, high quality monitoring signal, that was very close to what I was listening to, so it could actually be used to judge sound quality. This is quite a departure from what has been expected from wireless monitoring in the past, which has basically been used only to verify the words said by the actors. When wireless boom was used, the headphone quality was adequate for boom operator monitoring. (For more detailed information, see the mini review elsewhere in this newsletter.) But for now, suffice it to say that during three months of shooting, we never had a single problem with range nor an instance of dropouts. For more information in this issue on the Comtek 216 System see our Top Ten List. Playback, and Then Some The playback scenes were no doubt the most complex of the show, because not only did we segue from sync dialog to music track, but we always recorded the actors singing live to tape as well. Even though the actors were pre-recorded on the playback track, the hope was that we would be able to use the actual live singing, thus avoiding the “studio sound” and lip-sync problems. Here’s how it went:   THE Cart (or life-support system for the sound department). Click for larger picture with detailed description. Using the Deva’s 4 tracks, we were usually able to track the principal actors separately, even though there were sometimes more than four actors. For instance, for one scene track #1 was the mono mix (as always), track #2 was actor “A” and actor “B” (they never sang at the same time), track #3 was actor “C”, and track #4 was two remaining actors singing backup together. For the wide shots (where lip sync was not critical so neither were the live tracks), the playback was done at full volume through full range speakers and a tightly filtered subwoofer (also used as a “thumper”). For the close-ups, Comtek earwigs were used (wireless earpieces that receive their signal from an induction loop). Instead of having the actors wear receivers and neck loops, I looped the entire stage and drove the loop with a power amp. If the actors were standing anywhere within the loop, they would hear the track inside their ears. (For mics, I used the Sanken COS-11s for everyone except for a little girl whose wardrobe was such that the mic had to be sewn into it. For this I used the tiny Countryman B-6, which became invisible even though it was in plain sight. All of the bodypack transmitters were the Lectrosonics 200 series. All the mics sounded fine and there was not even a single RF problem.) So, when it was time for the close-ups, we would start the track through the entire speaker system, and at the right moment duck the full range speakers and leave only the “thumper”. The thumper did not produce any frequencies that would interfere with the vocal tracks. The rest of the track was heard by the actors through their earwigs. Whether the live tracks or the pre-recorded tracks will be chosen is yet to be determined. However, most of the live tracks that I checked back were certainly clean enough and of a high enough quality to be used (and, of course, they will be in perfect lip sync). But most of what will determine which tracks are used will be the performance. During the pre-recording session, the actors have plenty of time to get their best performance and are not out of breath from dancing, etc. So there’s a good chance that, despite our valiant efforts, the pre-recorded tracks will be chosen. Oh well… Zaxcom Wireless Boom I normally insist on hard-wiring the primary boom mic. Quality and reliability are two reasons, but also because I rely on my boom cable system with talkback for communicating privately with my boom operator. However, when it becomes overwhelmingly more convenient to go with a wireless rig, I’m all for it.  On this show I finally got to use the Zaxcom digital system, and since I was using it for wireless boom, I chose the plug-on transmitter, which had built-in 48V phantom power. Truthfully, I was amazed by it. Even when I used it for second boom, able to compare it directly with the hard-wired boom using the same mic (Sennheiser MKH-50) there was never an instance when I could tell the difference between it and the hard-wired boom. No noise, no limiter action, no companding, no distortion, no dropouts, 24 bit digital quality. Sanken CUB I had heard so much about the Sanken CUB-1 boundary mics for car scenes that I was looking forward to using them. However, the only car scenes we had that involved significant dialog were in a fully operational enormous street sweeper. Since the noise level was way loud and there was no dashboard or headliner to plant them, I had to resort to wireless lavs. But I was able to find a terrific use for the CUB-1s as plant mics for a boardroom scene. Planted a foot or two away from the actors on the wide shots, they cut amazingly well with the overhead boom during the close-ups. I highly recommend one or two in the kit. Lectro Digital Hybrid Though still in its development stage at the time, I wanted to try out the newest generation from Lectro, the 411 series. Dubbed the Digital Hybrid because it is a combination of digital and analog technologies, the end result is in improved audio quality due to the removal of the traditional compander. When this unit was used as a bodypack, the increase in quality was obvious. The familiar “pumping” sound of traditional systems was gone. I did discover a low frequency fluttering noise when using it as a wireless boom during very quiet scenes. But after bringing this to Lectro’s attention, their DSP tech wrote a firmware upgrade. It was emailed to us from Lectrosonics, we installed it in the comfort of our own shop, and the problem was gone. This firmware is now standard on all new units. DC Power: The Only Way to Go One of the most profound improvements for me on this show was that I finally got around to converting my cart to full DC power operation; everything being powered from a single 4-pin XLR. This was no easy feat because I had always run everything from AC-DC supplies in the past, but it was certainly worth it in the end. The first thing I had to do was convert my 2-in, 16 out audio distribution amplifier. I ripped out its internal AC power supply and replaced it with two 24V DC-DC converters that could supply +/-20V from an external 12V input. Next, the cart power had to be distributed from the single input to nine devices. For distribution I used one Remote Audio “Hot Box” and two Remote Audio “BDS” boxes. Since the Hot Box has six individually protected output circuits, I used five of its outputs for the critical equipment (Deva recorder, DVD-RAM burner, wireless receivers, mixer, and DA). Sound Department (L-R) -- George Leon (boom), Richard Murphy (cable), Glen Trew (mixer) The sixth output of the Hot Box was used to feed the BDS boxes which fed power to two video monitors, Fostex powered speakers (with special 12V mod), Littlites, Fostex PD-4 (for music playback), three additional wireless receivers, Comtek transmitter, and even a built-in charger for my cell phone (!). The remaining outlet was used for a panel-mounted voltmeter so I could monitor the condition of the battery. All of this equipment required about 12 amps and was powered with a Z-Tap battery that could be turned off and on with a single master switch. This was absolute joy. We never had to wait for the electric department (which they liked as much as we did). On occasions when we had to relocate our rig across the set, it was done without having to power down anything. During night exteriors when the main lights went out as soon a wrap was called, I was still up and running, and even had enough light to wrap and finish my sound reports. OK, Now for the Deva and Cameo System First, a quick review:  The Deva, made by Zaxcom, is a small portable four track recorder that records to internal hard drive. It also “mirrors” the recording to an external device through its SCSI data port (like most productions using the Deva, I mirrored to DVD-RAM disks). These mirrored copies on DVD-RAM disks are what I actually turned in, just like I used to with the traditional Nagra tape or DAT cartridges way back in the old days. The Cameo is the Deva’s companion digital mixer, also made by Zaxcom (the Cameo can be used with analog recorders and the Deva can be used with analog mixers, but used together they bring even more to the table). In addition to maintaining 24 bit quality audio from the time it leaves the preamps straight through to the recording, and in addition to its phenomenal routing abilities, the Cameo also has the ability to enter text information onto the recording. To cover in detail all the ways that this system improved the production would simply take too much time and space for this article. So here are some expanded highlights: Four tracks The Deva can be either a 1, 2, or 4 track recorder with the touch of a button, but anytime I used more than one microphone, I chose to record four tracks, and here’s why:  Track #1 was always my mono mix. This was the default for dailies viewing and will often be the track used in the final mix. If I was using two booms, for example, “Boom A” would be on track #2 and “Boom B” would be on track #3, both of these isolation tracks would be sent prefader. If a plant mic were also used (for example), it would go on track #4, also prefader. This configuration will allow the scene to be completely re-mixed during postproduction. In other mic scenarios, the assignments for tracks 2, 3, and 4 would change accordingly. The ultra flexible signal routing options and capabilities of the Cameo made track assignments easy, and last minute changes could be done very quickly. Instant playback cueing Imagine this: A scene requiring three microphones where one actor jumps their cue (or maybe I made a mistake and forgot to bring their fader up). The mono mix would not have this actor’s mic recorded, and the Director would not hear it in his headphones. In the course of two seconds after the scene is cut, I could playback that actor’s isolated track (in my headphones and the director’s) to verify that it was clean. Then, if only halfway through playing back the track the director is satisfied and the AD calls to “ROLL SOUND!”, all I have to do is press the red record button and I’m recording a new take at the end of the previous one. There is no possibility of recording over the track I was checking back. Try doing that with a DAT or ¼” tape! Have Your Take and Send Them Two! Possibly the best thing about the Deva is the redundancy of the system. There is simply nothing as safe and nothing with as much backup as the Deva system offers. At the end of the day I had turned in two copies of the day’s work (on DVD-RAM) and still had the original recording on the internal hard drive. Making two copies is very unusual, but it was requested so that the editorial department would not have to wait for the telecine crew to finish before building the dailies reel. One copy went to New York for telecine transfer (the DVD-RAM disks were played back with the Fostex DV-40) and the other copy went to the editorial office in Philadelphia. This allowed both departments to work with the same scenes at the same time in different cities. At first I was afraid that making two copies would be a real time hog and cause delays, mistakes, and me finishing by myself in the dark. As it turned out, while it did keep my brain exercised (for a change) it was never a problem, and I would not hesitate to agree to it again. Being able to keep the original recording on the internal hard drive was a huge benefit. Several times I was asked to check select takes from previous days. In a matter of a few seconds, I could cue to and play any take done within ten days, and quickly start recording again. With one 40G hard drive I was able to keep 10 days of back up at all times. I could have swapped hard drives when the first one became full (allowing 20 days of backup) but I decided not to because by that time there were at least seven perfect duplicates of each take: DVD-RAM #1 DVD-RAM #2 The Deva hard drive The dailies ProTools hard drive The AVID hard drive My personal PC hard drive The telecine DigiBeta tapes. The chances of needing another were just inconceivable, so I used the single 40G drive for the entire show. "Jersey Girl" invades the New Jersey Turnpike. Text Entry, or “e-Sound Reports” Sound Mixer: “How did you like tracks 2, 3, & 4?” Editor: “You recorded 4 tracks?” Since the introduction of the Nagra D, portable 4-track recording has been considered more and more on the set. However, one problem with using four or more tracks in production is getting the information to postproduction in a practical, expedient manner. Thumbing through pages and pages of multi-part sound reports that are often scribbled during the heat of the battle in the rain - with questionable accuracy - has often been the reason good isolated tracks were never used. When the Deva recorder is used with the Cameo mixer, informative text can actually be entered along with each take. The editorial department of “Jersey Girl” made full use of this capability, and came to rely on it heavily in streamlining the process. With the Cameo mixer, a single mouse click reveals the text screen. In this screen are fields for “SCENE”, “TAKE”, and “NOTES”. The screen also has a virtual keyboard where you click on the letters, numbers, or symbols you want to enter. When I first considered adding this task to the traditional handwritten paper sound reports, I recoiled like a beaten dog: I already resented having to take notes, and I’m terrible at it. My hand written notes are always plagued with errors, and I feared that mistakes I might make entering text with the Cameo would just turn into more complaints from the editors. Plus, it’s another distraction that I don’t need. So I thought. Well, I was wrong. Entering text with the Cameo was easy, it was helpful not only to post production but even to me on the set, and it actually helped make my handwritten notes more accurate than they have ever been. I’ll explain: First, the current firmware revision of the Cameo includes fields for the user to create and store words and phrases that will be used repetitively. Once stored, these entries can be placed into either the SCENE or NOTES fields with two mouse clicks. This drastically reduces the need for “typing” every word or phrase each time an entry is needed. Second, the scene and notes fields stay the same until they are manually changed. For instance, when doing twenty takes of scene 102A, you only have to enter it the first time. Third, the take number advances automatically each time you roll, and resets to take “1” whenever you change the scene number (theoretically, you could never have to enter a take number). Each field is limited in the number of characters that can be entered, so I created a sort of “meta-shorthand” (examples below). Pre-stored words or phrases can have a variety of uses and I’m sure different users will adopt different styles. But here’s an idea of what my system evolved into, and some of the pre-stored entries: “WT1001” This was used to speed the slating process for wild tracks. The first wild track is “1001” (the idea being that scripted scenes will never be as high as 1000, so any scene labeled 1000 or higher is a wild track). Whenever we did a wild track, two mouse-clicks entered the correct wild track scene number. At the next opportunity I would update the stored field with the next wild track number (i.e. “WT1002”) so it was ready when needed. This helped me keep up with what the next wild track take number was (something I could never manage to do in the past). “AMBIENCE SC”  This was used in the NOTES section for describing wild tracks that were ambience (room-tone) and what scene they were for. I would click it into the field and add the scene number, the end result looking something like “AMBIENCE SC101”. “WALLA”  Same as above, but for “walla” wild tracks (background people milling about, murmuring, etc.) “ID TONE” and “-20DBF=0VU”. This was used in the scene and notes field during the head-slate for each “roll” (not really critical, but why not… it’s a nice touch). The following were used for tracking information to inform postproduction what was recorded to which tracks: “MX1” (mix on ch 1), “BA2” (Boom A on ch 2), “BB3” (Boom B on ch 3), “OL” (character’s name “Ollie”, followed by track number. For instance, “OL4” meant that Ollie was isolated on track 4). Other characters names were similarly abbreviated such as “MY” for Maya, “BR” for Bart, “GR” for Gertrude, etc. With this text, when an editor looks at the audio file in their workstation such as an AVID, they can arrange their system to see with a quick glance, the Roll #, Scene #, Take #, and track assignments (or other short notes). It could look something like this: DATE TIME ROLL SCENE TAKE NOTES 05/11/02 09:48:12:22 054 120A  2 MX1 OL2 GR3 BR4 05/11/02 10:05:06:10 054 120A 3 MX1 OL2 GR3 BR4 05/11/02 10:12:18:22 054 120B 1 MX1 ONLY 05/11/02 10:15:16:18 054 WT1033 1 WALLA SC120 05/11/02 11:07:20:13 054 33B 1 MX1 BA2 BB3 MY4 For me on the set, an unforeseen advantage to using this system was that the Cameo actually kept up with which take number was next. You know how when (not if) there is a discrepancy between the camera assistant, the sound mixer, and the script supervisor about what take number is next, sound is the last one they believe? Well, all the sudden I was the authority!  Boy, was that a switch! Mag-less Dailies One major reason for the postproduction authorities to insist on Deva, was because they wanted to project “Magless Dailies”. The traditional way to view dailies is to project the image from the film, and playback sound from magnetic perforated stock. Using the Deva system, the sound files (all four tracks) were transferred directly into the ProTools hard drive, and while viewing the dailies, the sound was played directly from the ProTools. The advantages were many: The time needed to transfer select takes to mag stock was eliminated. The cost of mag stock is eliminated. A synchronous mag player is eliminated. The sound quality is pure 24 bit. The four tracks could actually be mixed during dailies if desired. Two hours of sound can be transferred into the workstation in less than ten minutes(!!!). Mixing playback during dailies is a rather odd luxury, but I always had a small mixer at my seat in the screening room and we managed to benefit from it a time or two. For instance, there was a bar scene involving actors watching video playback of a football game on television. I recorded their dialog on clean tracks and isolated the sound of the TV football game on its own track. When we watched this scene during dailies, the first take was “dry” because only track #1 was being heard. But halfway through the second take I brought up the clean video track. The reaction from our small audience (including director and producer): “Cool!” and “Right On!” The Trew Audio entry in the Jersey Girl Halloween Party pumpkin carving contest. (We were robbed!) Reality Check Since that’s all the technical stuff I can stand for one article, I’m sure most everyone else has had their fill, too. So, let’s wrap this up with some perspective:  Aside from the fun we get from working with all of the above techno-stuff, it should really exist for one reason only: to help us contribute as much as possible to making the production more successful. Maybe it saves money. Maybe it saves time. Maybe it allows us to do useful things we couldn’t do before. (Maybe it even sounds better.) But whatever our contribution is, it should always be directed toward the creation of the final product, whether it be news, television, commercials, or, in this case, a feature film. To that end, I have never felt more motivated to contribute to a production than I was with “Jersey Girl”. It seemed that everyone involved with this show, from the actors on up, had a sincere interest in the film; an energy that originated from the director/writer/editor Kevin Smith. His writing was clever beyond mere surprise. Many times I witnessed a member of the cast and crew light up as they suddenly understood how one scene related to another in a way that may have not been obvious even by reading the script. As the production progressed, the cast and crew began to sense the power of this story and the way it was being told, then took on a more personal interest in it, and were moved to do their jobs that much better because of it. OK, so I guess I did win the lottery. See the Jersey Girl Photo Album for more pictures!