The System in Question
It was a typical New England day in early November as I drove through Rhode
Island. The once beautiful fall leaves had since turned brown and fallen from
the trees, the rain pounded the highway relentlessly, and on the set of Dan
in Real Life there was trouble. A new recording prototype developed for John
Pritchett had gone haywire. An audible clicking sound covered all digital tracks
streaming from John’s Cameo mixer. This problem hadn’t shown up in any of our
tests. The sky seemed all the gloomier as my silver PT cruiser raced down the
highway. What could have gone wrong?
First, a little background information: Up to this point, digital recording
systems using more than 10 tracks have been bulky, power hungry, and difficult
to use in the field. Most of these systems require a large desktop computer,
necessitating AC power and, for use in the field, a DC backup, for a combined
weight of over fifty pounds which needs to be put in a large case on wheels just
to be conveniently carted around. While this is a workable solution for many of
today’s large musical films like
A Prairie Home Companion, your average feature with a three person crew
and an intense production schedule calls for a truly portable solution.
This was the challenge that sound mixer John took on when starting Peter
Hedges’ new film Dan in Real Life. He already owned a Zaxcom Deva V
recorder and a Zaxcom Cameo digital mixer and wanted to have a complete and
exact multi-track back up solution for his setup that could be expanded if he
needed more tracks on a future job. This backup recorder also had to be able to
be carried anywhere the shooting day might call for. Armed with these
requirements, John approached Gotham Sound to create this new compact, portable,
expandable backup to his all digital Cameo / Deva V recording setup.
John’s existing setup uses the Zaxcom Cameo mixer feeding up to 8 digital
tracks into a Zaxcom Deva V. We decided to split the digital signal out of the
Cameo to make sure that the signal being sent to the new backup recorder would
have exactly the same level and track assignments as those sent to the Deva V.
Not only did this save time calibrating levels and potentially repatching if
track assignment changes were needed, but it also allowed us to send all 8
channels of audio on one digital 4 pair cable with one connector on each end,
rather than needing a bulkier 8 channel analog cable with individual connectors
for each channel. To split the digital signal we used an Aphex AES Splitter,
modified to run on 12 Volt DC.
Now that we had the signal split, the real challenge started: how to record 8
channels of AES into a portable and expandable solution. Expandable was made a
requirement because this new recording solution needed to not only be used as a
backup but also as a primary recorder on a job that required recording more than
8 tracks, a more than likely scenario for a professional mixer to encounter,
given the long takes, big camera moves, and large number of speaking characters
in a scene favored by many features. Since 10 is the maximum track count for any
portable location recorder available today, there was only one logical choice:
computer based recording.
At the heart of this recording system, John chose to use MetaCorder.
MetaCorder contains a number of great features that make it preferable for
location recording over a ProTools based system; it has repeatedly proven itself
reliable on other projects, not to mention the stellar technical support
provided by the manufacturer, Gallery Software.
The next choice we had to make was what type of computer. Using a G5 tower
was out of the question, as it was simply too big and unable to be DC powered.
That left either the Mac Mini or a Mac PowerBook; both are small and can run on
DC power, but in the end we went with the Power Book because of its built in
screen and ability to fold into a very slim package, where the Mac Mini would
require an external monitor and additional size and power considerations.
Now that we had the basic recording system designed, it was time to select
the interface. In the past, we’ve used the Lynx AES 16 PCI cards, but those were
not a viable solution as laptops are only expandable by USB, FireWire, or PCMCIA
(USB and FireWire being the most common solution). The big problem here is that
neither PCMCIA cards nor USB or FireWire interfaces are available with 8
channels of AES inputs. Most will have one input, if they have any at all.
However, there are many interfaces that have 8 channels of ADAT digital input,
so now the question became how to convert 8 channels of AES to 8 channels of
ADAT. To do this, we chose the RME ADI-4 DD digital converter sending ADAT into
the RME Fireface 400 which then sent the signal via FireWire into our laptop.
The great thing about these two interfaces is that they each occupy one half
rack unit and were able to be rack mounted together in one space, as well as
both being natively DC powered.
The digital signal flow of John
Pritchett's rig, Mark I
Initial tests using a Fostex DV824 in our shop showed that the system was
solid and everything was functioning properly. Despite all of our testing, we
could not anticipate what complications would arise once this new prototype went
into production. Now standing in a gusty rain storm, under a tent whose flaps
refused to stay closed, in the middle of a Rhode Island parking lot, John and I
went to work on tracking down the problem.
We tried everything from verifying pin assignments to swapping ADAT cables,
rebooting several times, and pushing every button and switch available. The
laptop was functioning well enough and everything at the Cameo end seemed to be
working as expected. After a few hours the tent was nearly lost to the gale-force winds and we had almost certainly caught pneumonia;
John and I called it a day. We knew that only a few things had changed since our
test situation: the power source, the digital source, and anything that may have
happened in shipping. It was there that we continued our search back at Gotham.
First we set out to verify that our previous testing setup still worked. We
plugged the DV824 into the mix and everything seemed to work just fine; this
eliminated the possibility that anything had been broken during shipping. Next
we decided to replace the DV824 with the closest thing we had to a Cameo, which
was our Zaxcom Deva V. Instantly the same problem reared its head, the RME just
could not lock on to the Zaxcom’s AES signal. We went a step further and plugged
the output of the Deva directly into the DV824 which the 824 handled flawlessly.
It was obvious we could no longer use the RME solution we’d developed, but it
was still unclear as to why.
After some research and a lot of asking around, there were some old stories
of this very problem happening when interfacing other machines with the Cameo.
It turns out that Zaxcom has a slightly different AES standard than the RME ADI
expects. Most professional equipment is prepared to handle this slight
difference in AES and can process or ignore it without a single glitch, but the
RME could not and had no setting to allow us to ignore what it would consider
errors, so now it was back to the interface drawing board.
As I mentioned previously, there are no PCMCIA cards or USB or FireWire
interfaces that have 8 channels of AES inputs available. The only computer
interfaces that have that many channels of AES-in are full size PCI cards made
for desktops. What we really needed was a way to interface one of those cards
with a laptop computer. The solution came from the graveyard of Gotham’s
basement, rising from three years of dust and disappointment. This ghost is
known as the Magma chassis, a PCI expansion system.
Before the age of USB and FireWire audio interfaces, we had originally
specified the Magma chassis in our very first computer based recording design.
While we were very hopeful at the time, the technology was simply too new.
Computers froze, tracks clicked, and eventually the project changed
directions and the Magma was buried. It was time to see if the technology was
The Magma connects via a PCMCIA (also known as CardBus) slot on the laptop
and expands it into a box that has up to 4 full size PCI slots. We chose to use
the 2 slot version in an attempt to keep the size down. In addition to allowing
the use of full size PCI cards, the Magma chassis is DC powered and has a space
to power a full size ATA hard drive and FireWire adapter. We placed a spare 80GB
7200 RPM hard drive in the chassis, the equivalent to the largest hard drive
found in the audio recorders available today. It should be noted that the latest
hard drives available are upwards of 400GB while still spinning at 7200 RPM,
allowing for a huge amount of external storage.
Inside the chassis, we used a Lynx Two-C master card and LS-AES slave card.
While we could have gone with one straight AES I/O card, we chose this system
because of its flexibility and expandability. The Lynx Two-C allows the use of
two slave cards providing up to 16 more inputs, in addition to the 8 inputs it
already has on it. It also provides 6 analog inputs and two analog outputs,
allowing us to send a two channel return to the mixer without having to use a
digital converter. Using the LS-AES as a slave card provided us with 8 solid
inputs of AES which then was fed by an internal cable in to the Lynx Two card.
The Lynx Two-C was then plugged into one of the available PCI slots on the Magma
The digital signal flow of John
Pritchett's rig, Mark II
This new system tested perfectly with the Cameo. The Lynx control software
actually has an option that allows it to compensate for an AES standard other
than what it expects. Once we enabled that option, the difficult task of getting
digital audio into a laptop had been completed. Now we could start making
modifications to this system that would make it more functional and user
The first thing that we did was to take an iGo computer power supply
generally used by travelers on planes or in cars and modify it to be used with
the Mac PowerBook and 4 Pin XLR. A custom 4 pin splitter box was made and
mounted on a pull-out shelf along with a G-Tech 7200 RPM hard drive to give us
the ability to record to two hard disks simultaneously without putting any
stress on the computers internal hard drive. Everything was rack mounted into
one small SKB case 4 rack units high.
The front and back of the rig,
Since this system was going to be placed on a different cart from the Cameo
and the Deva, we decided to mount a small DC powered Xenarc touch screen as a
secondary monitor in order to check track levels and enter scene, take, and
metadata information. The last step was to take all the cables used to
interconnect the primary cart and the new backup recorder case together and wrap
them in snake skin, keeping them neat and easy to manage. Inside this umbilical
there were several feeds: 1. 8 channels of digital audio, timecode, USB control
for touch screen monitor, VGA for monitor, and a 12V DC feed from the new Remote
Audio Power Master. This umbilical cable can quickly be disconnected from the
main cart and wrapped neatly in the rear of the backup recorders case, making it
easy to carry away.
In the end, we created a DC powered, compact, 16 channel recording system
that had the ability to take up to 10 digital inputs and can be easily upgraded
well into the future. It seems that technology has finally caught up to the
requirements of location recording for feature films and that the sun was
finally shining on the latest portable digital audio recorder.
about products mentioned in this article: