The Sony Oxford is an imposing sight when first viewed. With its multiple LCD monitors (either four or up to seven on the fully configured 24C24) and sleek design, it has a look reminiscent of something found on the space shuttle. Its “new from the ground up” architecture doesn’t mean Sony is reinventing the operational wheel, however. While being slightly different than your average console, the Oxford uses many familiar tried and true features that make this 21st Century desk feel comfortable to any user used to other high-end mixing desks.
Product PointsApplications: Studio; post production
Key Features: 120 input; 48 busconsole; 96 full channels; LCRS; 5.1 and 7.1 mixing
Price: Customized; price varies
Contact: Sony at 800-686-7669.
+ Complicated setups are easily recalled
+ Excellent sound
+ Version 2 software
– Awkward fader-centered button
The Score: A well-implemented and supported digital audio desk.
As with most large-scale digital mixing consoles, the actual desk is merely a giant remote control – the brains reside in the machine room along with the A/D-D/A converters (ADC/DAC) and AES and MADI ports. No actual audio runs through the console, except for the analog signal from the talkback mic.
Every fader, pot and switch is merely a tool for sending data to the main processor, which then performs DSP on the signal (EQ, gain change, routing commands, etc.). This does not really affect the way the desk performs for the operator, but it’s important to keep this in mind.
Digital consoles tend to have many shared controls, such as EQ and dynamics. While this saves space on the desk surface, it can be frustrating to operate compared to the more familiar analog console layout, where each module has dedicated control for EQ, dynamics, etc.
The OXF R-3 is a 120-input, 48-bus console with eight stereo sub-groups, 24 auxiliary sends and multiformat capability. There are 72 full I/O channels, 24 mono and 12 stereo return channels. The surface is fairly simple to navigate and one can perform sophisticated tasks with a minimum of prep work.
Two standard sizes are available – the 24C24 and the 24C0. The former consists of a left bank of 24 faders, a center section containing monitor, talkback, stereo sub group, aux send and automation group masters and then a right bank of 24 faders.
The 24C0, currently installed at Sony Music Studios, has one bank of 24 faders making up the left two-thirds and a center section comprising the remaining third of the desk. While capable of the same functions, it has less flexibility as to how many faders can be simultaneously accessed – the larger console permits two banks of 24 faders on the surface at one time (e.g. 1 through 48), whereas the smaller control surface allows only one bank.
The console’s small footprint (64″ wide, 48″ deep and 41″ high) and adjustable angled meter bridge contributes to a less obtrusive acoustic signature. There are four LCD monitors (seven on the 24C24) built flush into the surface of the desktop used for displaying an array of information including routing, I/O assignment, labels, EQ curves and dynamics transfer function graphs.
The main LCD monitor in the center section is reserved for displaying project, mix and snapshot directories, as well as the preferences for linking and copying channels. Basic housekeeping is also monitored here: user name, time code frame rate, synchronizer setup, sample rate, backups, etc.
A built-in trackball can be ushered across all the screens for data modification. A keyboard, that can be hidden while not in use, is built into the armrest. The labeling function allows for an electronic scribble bar for each channel, auxiliary send, insert point, line, mic, and multitrack input, group fader, externals, and just about anything else on the console. This is a great feature – no need to save the console strip anymore – it’s saved with the automation data on disc.
Along with each project/title/snapshot window there are places to store notes – such as outboard gear settings and specifics on patching for the project – all of which is stored with the mix for later recall. The backup medium is a 520 MB MO disc that easily holds multiple backups of large projects.
Back to the shared controls mentioned earlier – although there are only 24 physical channel faders, there are 120 input channels located on five pages, which can be recalled at any time. Besides controlling the channel level to the main stereo bus, these 24 faders can also be used to control input gain, send level to the 24 auxes, group trims and multitrack send to tape.
Above the faders are 24 rotary controls – also switchable – that can be used to control pan, input gain, sends, group trim or multitrack send. While at first the shared controls seem limiting, they are extremely flexible and efficient. Different functions can be used at different times – perhaps while tracking, the mic level could be assigned to the pan pot for ease of reach, while the channel fader could control either level to tape or the monitor mix or, for that matter, a headphone cue mix out aux 17/18.
Once the panning is set, the input gains can be selected in place of the pans while the panning remains unchanged. To check on the panning, it can be easily recalled back to the small pots for viewing or further adjustment. This is a helpful technique for quickly setting one aspect of all channels. I find it helpful to check how much of each channel I am sending to each reverb. I simply call up any aux and see graphically across the faders the levels of each channel out to the selected Aux bus.
While mixing, one must carefully view and adjust all parameters one signal at a time. To set one channel with a more complex amount of processing, the access switch is depressed on that channel and the main surface of the console becomes dedicated to the accessed channel. Shared controls or assignable channel processing exist for pan as well as all other functions located on an analog console module: input gain, routing, dynamics, EQ/filter, aux sends and multitrack send level.
The routing section controls accessed signal routing out to multitrack busses 1 – 48, the stereo subgroups and the main stereo bus. Group trim, input level and pan, as well as surround bussing (via the multitrack busses) and surround panning are also available in this section.
It is also possible to choose the point in the path at which signal is sent to the multitrack busses. For example, when running a live mix while simultaneously recording to multitrack (broadcast mode) the multitrack send level can be split off before the channel fader so that the two signals can run independently.
The multitrack send level can also follow the channel fader, which is an easy way to expand a stereo mix into surround by merely setting separate multitrack send panning and fine-tuning the multitrack send levels to the surround outputs. The position of the send in the signal path can be placed before or after any of the EQ/filter/dynamics/insert/delay chain, allowing simple routing with EQ and effects “to tape” or only on the monitor path. The EQ, filter, dynamics, insert, and delay can also be placed in any order in the chain as required.
Twenty four aux sends can be read concurrently for any accessed channel and each aux can be sourced from any of the same points as described above. Auxes 17 through 24 default to eight channels of foldback or cue and the levels can be copied from the two-mix in a one time move and sourced prefader for a separate, but updatable, cue mix. The sends can also be set at nominal and sourced post-fader for a true two-mix to cue (which can also be refined). Any odd/even pair can be made Stereo with the odd control as level and the even control as pan.
The EQ section features five bands with the LF and HF switchable to shelf and a high and low pass filter for use on all full channels. The return channels have three bands while the master fader has only two bands. In the five band channel version, the center frequencies are continuously variable with a tremendous amount of overlap: LF 20 – 400 Hz, LMF 30 – 600 Hz, MF 100 – 6 kHz, HMF 900 – 18 kHz, HF 2 k – 20 kHz.
All EQ controls can be changed during program without zipper noise. Q factor ranges from 0.5 to 16, and gain is +/-20 dB. Filter rolloff is adjustable in 6 dB increments up to 36 dB/octave. The EQ section also has an overall EQ gain control that controls the total amount of EQ applied to the signal rather than being a simple gain control.
An A and B side to the EQ allows two separate EQs to be set and compared, or even switched between using the EQ automation. Each band is switchable in or out separately, and a frequency response graph of the applied EQ can be viewed on any of the three LCD monitors on the console face above the channel faders.
This screen displays a combination of the five EQ bands plus filters as a default, but any band can be displayed individually in a separate color from the combined response curve with a simple track-ball move.
All full channels have dynamics available, including gate, expander, compressor and limiter functions that can be viewed on screen like the EQ. A two-band side chain EQ can be inserted in the side chain only or in the signal path for additional EQ, or in both the signal path and side chain. This side chain EQ permits onboard de-essing and other specialized signal treatments.
Side chains can be linked in a group, or one side chain can be linked to control another allowing multiple compressors to be linked even if the channels are not adjacent. The return channels have gate and a three-knob compressor only. The master fader has a full featured stereo compressor available.
In all cases, the compressor section offers three choices: Normal, a six-knob compressor with linear time constants; Classic, a six-knob compressor with exponential time constants; and Custom, which emulates a simpler three-knob vintage type compressor.
There is delay available on all full channels up to 1.2 seconds with mix control from direct to 100% delay as well as feedback control. There are also several open positions on the free assign area for future additional effects. Insert points are available throughout: All 120 channels, 48 group outputs, and one on the Master fader.
Since it is a software assignable console, every I/O port must be configured from scratch, involving numerous elementary steps, such as assigning a D/A to the main monitor output as well as configuring MADI, AES or analog ports for each aux, insert point and external device. Once a basic configuration has been set, it can be saved under the Studio Library snapshot directory.
These more common templates can then be recalled and updated for each new project. A basic template can be recalled in less time than it takes to put up the tape. The tape machines and remotes need to be assigned separately from the snapshot configuration, an additional task that I believe could have been included in the snapshot recall options, but it is nonetheless simple to execute.
Aside from the Studio Library snapshots – the basic templates from which each project begins – snapshots can be stored under the project/title directories and can include the entire console setup.
There is a snapshot mask where the type of snapshot can be configured, such as EQ only, or just the channel I/O. Although this can be set on a store as well as a recall, I find it makes sense to always store a full snapshot, and then only recall what is required. Information can also be recalled from one channel onto another, for example an interesting EQ can be moved from one title to another and recalled onto a different fader. It is not necessary, however, to recall a snapshot for every mix recall, since the SAVE MIX function includes all EQ, sends, I/O etc. As long as a snapshot has been recalled that includes the desired installation (Studio specific items such as monitor DACs, external devices, etc.) the LOAD MIX command will recall all other aspects of the desk.
A great advantage of a digital desk is the ability to link faders, EQ, and dynamics, with or without an offset. The Copy function allows a one-time copy of any or all parts of a channel to any another. This is handy for matching gains on similar inputs, or an elaborate compressor setting across a series of drum channels. Linking works in the same manner, and can be undone just as easily. It is also possible to link and copy in one move, which is useful for ensuring that two linked functions also have the same setting.
Timecode is only recognized by the automation through the use of a Motionworker synchronizer, which can be configured from the console. Various multitracks, two tracks and video machines can be assigned as Master or Slave or Independent and controlled from the console face transport panel.
The Oxford’s automation is similar to SSL in that the basic modes are absolute and trim, with options for fader ballistics: The fader can either snap back to an original position when released, or ramp at an assignable speed in milliseconds up to 10 seconds, or be written at the released level to the end of the mix. There is also an auto-takeover mode in which the fader drops out of record as the new move passes the old fader position.
These same ballistics apply console-wide. For example, the butt or snap mode can be used on the high-pass filter frequency control for filtering mic pops out of a vocal track simply by raising the cutoff frequency before the pop and then letting go immediately afterwards. The automation can then either snap back or ramp to the original position.
A button in the center of each fader, designed to begin and end the writing of moves, is unique to the Oxford. While there are many applications for this feature, it generally seems a slightly clumsy way of mixing. Fortunately, touch write option is available for those of us used to writing the fader from the moment it is touched until the moment it is released. Touch modes also include touch hold, which parallels the write switch on the fader knob so that when the fader is touched it stays in write until touch clear is depressed.
Each fader can have its value displayed in dB. In trim mode the amount of trim is displayed from the moment it is touched, in plus or minus dB. Pan pots can also be read by value (degrees left or right), as well as all other controls. The ability to display any control’s value is a practical function that should be standard on all digital consoles.
Cuts can be automated using the absolute or trim modes. In trim mode the in and out points of the cut can be advanced or cuts can be made to last longer. This is an incredibly intuitive way of working with cuts. In the off-line window, cuts can set and adjusted by timecode.
All aspects of EQ and dynamics, panning, delays and auxes can be automated. There are many different styles of mixing accommodated by the console. One obvious missing feature is the relative mode with the match switch of the Flying Faders and the glide switch of the Capricorn automation, which was a favorite mode of mine.
This type of pass can be achieved through the trim mode and the touch hold feature along with the touch clear or global dropout feature to match or glide at the end of the new data. This involves adapting one’s method of mixing, but it is well worth it. After spending just weeks on the Oxford I caught myself using the trim mode just as much on the Neve Capricorn.
Each mix pass is remembered separately in RAM as a new dynamic layer and each layer can be undone up to last time the mix was saved. Each time the mix is saved to the hard drive, the separate dynamic layers are lost, which means no messy mix tree. Of course, the newly saved mix need not overwrite the previous save, since every saved mix is named differently by default.
Everything from LCRS to 5.1 to 7.1 is possible on this console. There is a dedicated monitor section for surround, with six external returns of up to eight channels each and an eight-channel volume control. The stereo monitor section overrides this section, however, so a simple CD or DAT playback can be accommodated in the middle of a 7.1 mix with one switch.
It is also possible to run a parallel stereo mix. The surround mix can either be based on the stereo fader moves (big time saver in DVD mixing) with separate panning/joystick moves (all automatable), or it can be completely unique on a channel by channel basis.
Multichannel panning has four controls: left to right, front to back, surround left to surround right and divergence, or spread. Except for divergence, these aspects are all controllable in combination on the joystick, which is great fun.
The sub channel has no gain control from each bussed source, but it is possible to use an aux out (e.g. aux 24) to adjust the level of sub contribution of each source, which can then be brought back on a fader and routed solely to the subgroup.
George Massenburg has recently approved alternate EQ and compressor software written exclusively for the Oxford. It digitally models his famous five-band parametric EQ and compressor with the old crest factor and hysteresis controls.
These digital versions exactly parallel the originals, even in their distortion characteristics. This is welcome news, as a digital version of his compressor is long overdue. Perhaps you haven’t done the math – a simple software update adds 72 channels of GML EQ and compression to this already well-equipped desk (for the latest in console upgrades, see the sidebar on p. 18).
I have recently completed several projects with producer Steven Epstein for Sony Classical – A Fiddler’s Tale, written by Wynton Marsalis for classical chamber music ensemble, and Spain written by Chick Corea and featuring his sextet, combined with full orchestra.
These projects presented different mix problems that we solved thanks to the OXF R-3. Both recordings were made on the Sony 3348HR 24-bit multitrack, with its high resolution making it the perfect companion to the console. Epstein has been mixing a great many of Marsalis’ jazz recordings with engineer Todd Whitelock on the same desk with great success. The Oxford has a transparent, open sound much like the Capricorn, with a friendlier EQ, since it can be set at any frequency.
While the Oxford is 32-bit fixed point internal processing, they really sound quite similar. The Oxford is much simpler to operate, yet as flexible. Very complicated setups are quite easy to implement. I highly recommend this excellent piece of equipment, currently my favorite item at the studio.
(While Sony Studios carries the corporate moniker it is by no means required to use Sony gear. The choices are still made by engineers concerned with good audio and not with toeing the company line.-Ed.)