Focusrite Liquid Saffire 56 Review
Focusrite Liquid Saffire 56 Review
Mike Rivers, July 2011
The History Lesson
Long ago (about 1999) and far away (Portugal, actually), a couple of Brits named
Mike (Kemp and Eden) demonstrated their new product, the Sintefex FX8000
Replicator, at the 107 th
Audio Engineering Convention. It was a real, working product, too expensive to take very seriously at the time, based on their patented
Dynamic Convolution algorithm. Depending on the software model loaded into its
DSP hardware, it could be a very accurate replication of an equalizer, compressor, mic preamp, effect unit, or really any analog device. As I recall, there as a ½” Ampex tape deck among the models they were demonstrating. The
Dynamic Convolution process involves measuring the impulse response of the device being modeled, then putting the signal to be processed into a convolution algorithm along with the impulse response. The resultant output is the input signal with the frequency and phase response of the device being modeled.
Convolution isn’t new. You might even have a convolution reverb plug-in for your
DAW. But what the Sintefex folks did was take impulse response samples at various different levels over the full dynamic range of the device being modeled, and for the full range of adjustment of all the controls (even nonsensical combinations) That’s a whole lot of impulse response samples and that’s where
“Dynamic” in the name of the process comes from. Since most audio devices worth modeling behave non-linearly, by applying, on a sample-by-sample basis, the impulse response sample that corresponds to the amplitude of the input signal to the device, and choosing the sample based on the desired setting of the modeled device’s controls, the output is an accurate emulation of the input signal passed through the real device.
A few years later, Focusrite partnered with Sintefex to produce the Liquid
Channel. The Liquid Channel was a single channel mic preamp, compressor, and equalizer that was capable of modeling a wide variety of well known preamps and signal processors. It was elegant and expensive. Since the sound of a mic preamp is largely dependent on the electrical interface between the mic and the preamp front end, Focusrite took the Dynamic Convolution model a step further and used actual inductors and capacitors to simulate the reactive load to the microphone that the preamp being emulated presented. As you might imagine, this was a bit over the top, but Focusrite has kept on with the process and applied it to a few other scaled-back devices over the years, including the subject of this review.
What You Get
The Focusrite Liquid Saffire 56 (which from now on will be called the Liquid 56 for short) is a two rack space multi-channel computer audio interface with Firewire
I/O and a built-in DSP mixer for near zero latency input monitoring. It sports two mic preamp channels with Liquid modeling of ten well known (and some not so well known) preamps in addition to six standard Saffire mic preamps. In order to count the 28 inputs and 28 outputs (that’s why it’s a “56”), I had to include the two
“Loop Back” inputs which don’t connect to the outside world, but are for routing audio between two programs. In addition to the eight analog inputs, there are two sets of ADAT optical inputs and outputs for 8 channels each and a set of stereo coax (RCA) S/PDIF input and output jacks. One set of ADAT ports can be switched to stereo optical S/PDIF. At 44.1 kHz sample rate, all 16 ADAT optical channels are available, and the number divides by two for each doubling of the sample rate. Counting Word Clock and MIDI I/O would have made it a Liquid 60.
All connections are on the back with the exception of two high impedance (DI) instrument pickup jacks and two headphone jacks on the front. Mic inputs are on
XLR connectors and everything else is on TRS ¼” jacks. This is a pretty good looking unit. The front panel looks like sculptured black anodized aluminum, but it’s really plastic. The universal power supply accepts any voltage between 100 and 240 vac, 50 or 60 Hz.
The front panel mounted power switch (thank you!) is a British “upside-down” rocker. A group of eight 5-step LED meters indicate input levels from the analog inputs, either of the two ADAT Optical inputs, or the S/PDIF inputs as selected from the software control panel. LEDs indicate status of such things as phantom power, line/mic/DI input source, pad, low cut filter, Firewire communication, clocking, MIDI activity, and what the level meters are displaying when they’re not displaying the analog input levels.
The Liquid 56 is obviously intended for rack mounting, but should you want to put it on a table or shelf, there’s a sticker with four tiny soft feet packed along with the software installation CD. There are dress covers to fill the rack mount holes (it’s why you don’t see any holes in the panel in this photo) when you don’t need them, but they were missing from my unit, a marketing sample that has been around the block a few times.
Under the hood, construction looks neat, though there’s not much you can do in the way of maintenance. The internal power supply is in the clear, but the front panel controls are well buried so they’ll be difficult to dose with DeoxIT when they get scratchy.
A CD contains Mac and Windows versions of the drivers, the Saffire MixControl software control panel, the manual as a PDF file, and . . . WAIT! THERE’S
MORE! . . . a bonus set of VST/AU Focusrite plug-ins – a compressor, gate, equalizer, and reverb, all as good as any that come along with the popular DAW programs, some maybe better. I didn’t spend much time with them other than to confirm that they loaded and worked. As expected, the driver/control panel version on the CD was out of date and I grabbed the current version from the
Focusrite web site. You should do the same.
Gozintas and Gozoutas
Most of the action is on the rear panel. Take note that the mic and line inputs for each of the eight analog channels are on separate connectors, not the combo
(XLR with a ¼” jack in the middle) connectors frequently found on similar units.
Even better, you can have both a mic and a line level source connected to a channel simultaneously with no interaction between them. This is really important if the Liquid 56 lives in a rack where it’s mighty inconvenient to unplug a mic cable before you can use the channel’s line input and vice versa. The XLRs are locking, too, so there’s no worries about inadvertently pulling a mic cable loose while fishing for another cable behind the rack in the semi-dark.
Selection between the Mic, Line and DI (channels 3 and 4 only) inputs is made via switches on the software MixControl. LEDs on the front panel indicate which input is active, but it’s still necessary to go to the computer to change the input source for a channel, It could be the result of a mild case of revieweritis
(attempting to try out everything at once), but at times I was distracted by having to crank up the computer when changing a channel from a mic to a line source.
Not all the analog input channels have identical controls. They all have a gain knob, phantom power switch, and high pass filter switch. Only channels 3 and 4 have input pads. The Liquid channels (1 and 2) and the odd-numbered standard
Saffire channels have polarity (“Phase”) switches. While I’d prefer to have a polarity switch on each channel, I can understand putting then on one channel of adjacent numbered pairs for sake of economy. When two mics as a pair, chances are you’ll connect them to adjacent inputs to keep your sanity. With the odd-even arrangement, you’ll be able to reverse the polarity of one mic relative to the other if that makes it sound better. The two Liquid channels, because of their flexibility, may be less likely used as a pair, hence the individual polarity switches.
All of the mic inputs have essentially the same sensitivity. At maximum gain, channels 3-8 require an input level of –42 dBu to reach maximum output
(0 dBFS). This is pretty typical for today’s general purpose mic preamps. When speaking in a normal conversational voice about 2 inches of a Shure SM57, the gain needed to be just below maximum in order to avoid clipping. At the minimum gain setting, it takes +7 dBu into the mic input to reach peak level, which should cover fairly loud sources without needing a pad.
The Liquid channels differ in sensitivity by a couple of dB depending on the preamp model and the setting of the Harmonics control. Focusrite specifies the gain as 60 dB. The closest I could come to measuring it was by using the routing matrix (hang in there, it’s coming) to route an analog input as directly to an analog output. That bypasses all the controls but the analog gain, but the routing is done digitally which means there’s an A/D and a D/A conversion between the input and where I can measure the output level. Anyway, I got 59 dB for the standard Saffire inputs, which I’ll call “confirmed.”
The input gain controls are touchy when you get near the top of their range. If the Gain knob was calibrated in dB of attenuation rather than from 0 to 10, it would look kind of like the knob to the right, I’ve observed (and complained about) finicky gain control action on a number of other products priced for the project studio.
Maybe the various manufacturers all buy the same pots.
There’s plenty of gain adjustment range when close-miking a strong singer, but when recording a quiet acoustic guitar player, you’ll probably end up running the gain wide open and you’ll still have a little headroom. I’ve found that this causes some people grief, on the assumption that if they need full gain for anything short of a flea crawling across the table there must be something wrong. But trust me.
It’s OK to crank it up when a quiet source calls for maximum gain. These preamps are nice and quiet. With the input terminated with a 150Ω resistor (a fairly common test setup), I measured quiescent noise at maximum gain as -72 dBFS peak. This is about 3 dB quieter than my Mackie Onyx 1200F.
Input impedance of the Saffire mic inputs (channels 3-8) is 3.1 kΩ, a little higher than what’s typical for mic inputs these days. You may find that naturally bright mics through these inputs sound a little brighter than you’re accustomed to
hearing them. The input impedance on the Liquid channels is close to that of the preamp being modeled. The manual, which has a pretty good one-sentence description of each of the modeled preamps, lists the input impedance of each model, however my measurements didn’t always confirm them. Using an NTI
MR-PRO generator I found Z in
ranging from a rather high 12.0 kΩ for the “Flat”
(no model) setting, a low 220 Ω for the “Old Tube” modeled after a Pultec MB-1.
The “Tranny 1H” (an API model) fell in the middle at 1.8 kΩ. For both the Saffire and Liquid channels, Input impedance was essentially constant over the full frequency range and full range of the input gain control, and was not affected by the Harmonics control, the input pad, or high pass filter. From this, I gather that the loading is predominantly if not entirely resistive.
Where there are mic and line inputs, there’s always the question: “Can I bypass the mic preamp electronics when I’m using my high end outboard mic preamps?”
Here, as is the case with most mic/line audio interfaces and mixers costing less than $50,000, the answer is that you can’t. The Liquid 56 line inputs are built the same as nearly everyone else’s – with a resistive attenuator to knock the signal level down and raise the impedance, a couple of capacitors to block phantom power, and then straight into the mic preamp.
That’s really a pretty sensible design since the mic preamp circuitry is very quiet, transparent, and provides a balanced interface. If Focusrite were to design a separate input stage for the line inputs, it would probably be very similar to the mic preamp circuit anyway, just with lower gain. Since much of the character of a a mic preamp comes from the interface between the microphone and the preamp, that’s unchanged when using your outboard preamp. Swallow your pride, plug in your high dollar preamps, and enjoy them. Spend the money you saved by not having to pay for eight extra line input buffers, and buy a better mic.
At minimum gain, the analog line inputs will take up to +30 dBu before clipping, while at maximum gain, they clip at –19 dBu. They’re differential (balanced) with an input impedance of 43 kΩ.
The DI inputs were a bit of a surprise. Typically, the DI input on a mixer or preamp has about the same gain as the line input, maybe upwards of 10 dB more. Pickups are really pretty hot, and they have a wide dynamic range. The instrument inputs on the Liquid 56 are a lot hotter than most. At minimum gain, the DI input clips at +7.5 dBu. At maximum gain, it clips at –42 dBu. The pad gives you an extra 8 dB of headroom, which I found, when engaged, made it easier to set the gain. The DI Input impedance is greater than 1 MΩ which is fine for any type of pickup, and these inputs are very clean and quiet.
Another unusual feature is that the DI inputs, like the mic and line inputs, are differential, something that I’ve only seen in the occasional custom built instrument amplifier and the long gone companion amplifier to the Gibson Les
Paul Recording model guitar which featured low impedance pickups and
balanced output. It may not be difficult to modify your guitar for a balanced output. When connecting it to the Liquid 56 DI input with a balanced (TRS) cable, you’ll have a real balanced interconnection with a likely improvement in hum rejection. It’s food for thought.
The high pass filter would probably be described as “100 Hz” (the spec sheet doesn’t characterize it), being 3 dB down at 115 Hz, 6 dB down at 85 Hz, 12 dB down at 56 Hz, and an honest 12 dB per octave on down from there. Nominal frequency response of the analog inputs is 20 Hz to 20 kHz ±0.1 dB at 44.1 kHz and at 96 kHz, it’s 1 dB down at 46.5 kHz. The low frequency response extends well below the 20 Hz spec, being about 3 dB down at 5 Hz. At maximum gain, the low frequency response rolls off a bit more, being about 0.5 dB down at 30
Hz which is nearly identical at the low end to a Mackie VLZ Pro mixer preamp.
One thing that’s conspicuous by its absence is analog inserts between the mic preamp output and A/D converter so a compressor or equalizer can be patched in for tracking with processing. Some (manufacturers, mostly) argue that this isn’t
“the DAW way” – that you should record the purest signal you can and then process during mixing using the wealth of DAW plug-ins available. But sometimes you just want to use a hardware processor up front. If you aren’t using all the analog inputs and outputs, you can kludge an insert by sending an analog input to a line output, connecting your outboard processor between that line output and another channel’s line input, then recording the channel that’s receiving the processor output (or record the bare channel, too, just in case).
Still, an insert jack on a couple of channels, or better still, a pair of balanced jacks, one each for send and return, would be nice.
Analog line outputs are on active balanced ¼” TRS jacks. Though you can use any output for whatever you want, outputs 1 and 2 (which are prominent on the rear panel with a white background) are designated as the monitor outputs, typically for the control room monitors during tracking or mixdown. The crosscoupled configuration of the output stage provides the same signal voltage when connected to either a balanced or unbalanced input. Output impedance is about
The line level outputs aren’t blistering hot. With a maximum output level of just under +16 dBu, the Liquid 56 won’t blow most monitors off the table. Should you find, however, that you can’t turn the Monitor Volume control up very far before it’s too loud, enabling the Active Monitor Pad from the MixControl Settings menu lowers Outputs 1-2 by 12 dB, giving the control more working range. I found this to be unnecessary with my power amplifier gain set for my normal monitoring setup, but it could come in handy for powered speakers or power amplifiers with no input level control.
There’s not much to say about the ADAT optical I/O other than that it works as expected. For click-free operation, I needed use fairly short optical I/O cables
with the Saffire Pro 26 io that I reviewed a while back, but the Liquid 56 worked fine with 8 foot garden variety Toslink cables.
When using any digital input, the word clocks of the Liquid 56 and the digital source must be synchronized. Either of the two optical inputs, the S/PDIF coax input, or the external word clock input (BNC) can be selected as the Liquid 56 word clock sync source, or of course the internal clock can be used if you choose to synchronize the connected device to the Liquid 56. The JetPLL
clock technology used in the TC Firewire chip does a good job of minimizing jitter regardless of whether the clock sync source is internal and external, so there’s no worry about using an external clock source if that’s the most convenient hookup. Because of the low jitter clock system in the Liquid 56, don’t expect an ear-opening improvement if you use a fancy external clock generator. That was then, this is now.
The word clock input on Focusrite Pro 26 io that I reviewed a while back presented a pretty hefty capacitive load that substantially rounded off the square wave word clock signal, but that’s not the case with the Liquid 56. The word clock waveform remained nicely square when connected to the Liquid 56, so paralleling another device’s word clock input if necessary should have a pretty good chance of working. The word clock input on the Liquid 56 has a built-in fixed 75Ω termination, so another device that shares the same word clock source should either be bridging (no termination) or be capable of locking to a word clock at a lower than standard level.
As an experiment, I “double loaded” the word clock output from my Mackie
HDR24/96 with the Liquid 56 and Mackie 800R preamp and both stayed locked with the word clock input on the 800R either bridged or terminated. The
HDR24/96, however, is pretty fussy and wants to receive a full level word clock signal. When using the Liquid 56 word clock output to sync the 800R and
HDR24/96, it was necessary to un-terminate either the 800R or HDR24/96 in order for both of them to be happy. While it’s usually possible to come up with a hookup that works when paralleling word clock inputs, a termination on/off switch on the Liquid 56 would be a welcome addition, just in case.
I didn’t give the S/PDIF (coax or optical) and MIDI inputs and outputs much of a workout other than to connect something and verify that they work as expected.
MixControl software version 1.8 and later includes a switch to change the operating level of the coax S/PDIF I/O to the higher AES/EBU level. Although it’s unbalanced, most AES/EBU devices should work just fine when connected to the
S/PDIF RCA jacks using a simple hard-wired adapter or cable.
Front Panel Controls MixControl Controls
I’ll bet you thought I’d never get here. The two Liquid mic inputs provide emulations of ten preamps plus “Flat” which, with the exception of the higher input impedance, is essentially the same both in specs and sound as the other six preamps, very clean and quiet. It’s nice to have that 12 kΩ input impedance, though, for use with classic large ribbon mics which work best into a load greater than 10 kΩ.
Like the original Liquid Channel, the Liquid 56 uses a combination of electronics and an integrated DSP chip to model the performance characteristics of classic preamps. Unlike the original Liquid design that switched in expensive transformers when modeling called for a transformer input, the Liquid 56 simulates the inductive characteristics of a transformer using an analog circuit known as a gyrator. Resistance and capacitance appear to be real components.
For the Liquid 56, Focusrite has made a reasonable choice of ten models (as opposed to the 40 offered by previous Liquid preamp products) both to make the technology affordable by nearly any project studio, and also to offer a reasonable range of sounds without spending all day deciding which one is just right. You’ll find the compete list of the modeled preamps at the end of this article. Apparently there are some trademark issues, since the models, oddly even Focusrite’s own
Red 1, have phony names, but the manual clearly identifies the original target for the model right down to the serial number of the sampled unit.
These models react very much like the real thing, so for the “character” preamps, distortion increases as you drive them harder. The Harmonics control (the only user control aside from Gain) introduces a variable amount of second, third, and fifth order harmonics. Different models have different ratios of these harmonics.
To give you an idea of what the Harmonics control does, following are some spectral plots of a 500 Hz tone going into the DEUTCH 72 model at three different settings.
Now, I’m not a vintage preamp maven, so I didn’t try to judge how close these models come to the original. I will say, though, that there’s a subtle, but definitely
noticeable difference to be heard when switching among the models, with some differences more obvious than others.
Over a couple of weeks experimenting with it, I found that regardless of the mic or the source, I gravitated to TRANNY
1H (API 312), DEUTCH 72 (Telefunken
V72), and NEW AGE 1 (Millennia HV-
3D) when looking for something different. While these may or may not be accurate emulations, I think it’s fair to say that the older models capture the
“vintage vibe” while the emulations of contemporary preamps such as the
Millennia and Avalon have the clear and open sound that’s expected of them. I found the modern “clean” preamp emulations were less different from
“Flat” than the vintage ones, and I would say that’s a good thing.
Wouldn’t it be nice to be able to route the DI inputs through the Liquid channel? Unfortunately, that’s not on the menu, but it’s possible with some creative patching. Route Channel 3 or 4
(where the DIs are) directly to a line output (hang in there!), patch that output jack to the Liquid preamp input using a TRS-XLR cable or adapter, then choose your preamp model. The low impedance line outputs can easily drive even the low input impedance preamp models, and with a little juggling between the gains on the DI and Liquid channel there’s no need for an in-line attenuator or matching transformer. Is it worth the trouble? Honestly, I didn’t think so, but it’s another tool that might give you just that little twist that you’re looking for.
There’s a mysterious item labeled “Liquid Transformer Loading” in the Settings menu of the latest MixControl that’s not covered in the manual, only in the on-line
FAQ (at least as of this writing). This Loading, apparently, has been found to cause level fluctuation problems when some microphones are used with the
Liquid preamps, so it can be disabled on a per channel basis if needed. I saw the menu setting before I read the FAQ and guessed that it might be a termination switch for a modeled output transformer if that was part of the modeled preamp. I listened for a difference with it on and off and didn’t hear any so whatever it does must be pretty subtle. However, since its potentially adverse effect seems to be microphone-dependent, it must have something to do with the input rather than the output. I didn’t encounter a problem with any of the mics I tried, so I can’t tell you what the potential problem sounds like.
The Mixer Section
MixControl is the software control panel where all of the mixing, routing, and hardware setup takes place. One of the features that sets the Liquid 56 apart from a straightforward Firewire I/O box is the built-in DSP mixer. You can think of it as 16 mixers with mono outputs, 8 mixers with stereo outputs, or an in-between number of a mix of stereo and mono mixers. At 176/192 kHz sample rate, all but four mixes disappear. Each mixer can have up to 18 inputs.
If you’re accustomed to working with hardware mixing consoles, you might think of it as a mixer with a 16 auxiliary sends which can be paired for stereo outputs.
The Liquid 56 mixer only affects monitoring - changing a level or pan in a mix doesn’t change the signal being sent to the DAW.
It’s a pretty basic mixer, just level and pan sliders, with an input level meter for every channel. There’s no built-in EQ, dynamics, effects, not even insert points for outboard processors. If you have enough mixers (highly likely) and enough channels (probably) you can use an outboard reverb unit to add the oft-requested reverb in the headphone mix, but if a guitarist asks you to make his instrument brighter or less boomy in the headphones, without EQ you’re out of luck.
There are a couple of undocumented mixer shortcuts. Control-click or Alt-click on a level slider brings it to its 0 dB (unity gain) position. Control- or Alt-clicking on a pan slider centers it. Overload indicators at the top of the mixer’s meters stay on until they’re cleared, (they’re even remembered after a restart) but clicking on the overload indicator area (even if it’s not red) on the Master meter clears all of the channel overload indicators as well as the master. The maximum input level to a channel appears as an itty bitty number below the slider. This gets reset along with the Clip indicators. I’ve really become partial to overload indicators that clear automatically after a preset time period, and I’d welcome that feature here in a future update.
Channel inputs are selected from a drop-down menu. Choices include all of the analog inputs, ADAT inputs, S/PDIF inputs, or returns from the DAW. Simply stated, if there’s an input connector, it can be used as the input to a mixer channel. Similarly, the mix outputs can be assigned to any output connector. The only restriction is that when you select an input source for a channel, it becomes that channel’s input source for all of the mixes, though this is really no different from a hardware mixer. If you plug a mic into Channel 1of a mixer, it will be in Channel 1 in the main mix and in all of the auxiliary sends.
For live tracking, you’ll typically have all your mics and direct instruments as sources for the mixer channel inputs. Typically you’ll use one mix for the control room, and as many other mixes as needed for the players’ headphones and effect sends. When overdubbing, you’ll need to hear playback from the DAW as part of the mix, but you probably won’t need all of the live mics. It’s really very flexible.
The mixer in the screen shot is one that I set up for tracking a string band. The tabs along the top select the mix for which the controls are active. These can be named, hence Control Room and the four headphone mixes for the band members in the studio. Channels 1 through 8 are from the eight mic inputs. No digital inputs were used on this session, however if I needed more than the eight mics, I could connect my Mackie 800R through one of the Liquid 56 ADAT inputs and used those the ADAT inputs as sources for additional mic channels.
(Focusrite would probably prefer that you buy one of their OctaPre series preamps, of course).
The slider at the far right is the output or “master” fader for the mix. In this example, the Control Room mix is routed to Outputs 1 and 2, the officially designated “monitor” outputs. A mix can be assigned to as many physical outputs as you want. In this case it’s going only to Monitor 1 and 2, but if wanted to record the control room mix through the digital input on my DAT recorder (hey,
I’m an old timer), I could simply add S/PDIF to the list of outputs for the Control
Room mix from the pulldown output assignment list. When the mix is assigned to more than one physical output, the legend in the output section reads “Many” rather than the assigned output.
While you can assign a mix to as many outputs as you want, you can feed a physical output from only one mix – that is, you can’t feed three mixes to one output. I’m not sure why I’d want to do that other than perhaps to make a mix from stems, but chances are I’d leave that as a DAW job. If I have a complex tracking session, it might be nice to have submixes for drums, keyboards, or background vocals, but with a lot of inputs to a mix, I’d probably wanr the flexibility of a real console rather than the Liquid 56 mixer.
In this example, the four Studio mixes feed the inputs of a four stereo channel headphone amplifier located out in the studio, giving each of the players an independent monitor mix. For a live tracking session, these mixes will usually be pretty similar, a fairly well balanced mix, with the occasional “more me” or “less cowbell” request. A very useful feature which you can see at the top of the Output section is the “Copy to Mix” button Rather than setting up the headphone mixes one at a time, after getting a workable control room mix, I simply copied it to each of the four studio mixes and then tweaked each one to the player’s taste. It’s a real time saver.
You can save and load mixer setups, which makes it handy both for returning to a previous session and for starting from scratch with a basic setup. A mix can be saved as a file (you’ll probably develop a library for various situations), but it can also be saved to memory in the Liquid 56 box. If you power up without a computer connected, it will comes up in the saved configuration. This is handy if you want to use the Liquid 56 without a computer as a simple stand-alone mixer or a set of mic preamps, as long as you remember to leave it with the proper setup stored. You can only store one as there are no controls or display on the hardware for selecting a setup from a list.
Front panel headphone jacks 1 and 2 are hard-wired to line outputs 7-8 and 9-10 respectively. This is very convenient if you’re recording yourself or a duo in the control room – you have headphone jacks handy which provide plenty of clean volume. But since they’re hard-wired to outputs that might be feeding studio headphones, if you’re the engineer in the control room, you might have to do some juggling if you want to put on the cans for a closer listen. Soloing any of the channels or mix outputs routes your selection to the control room monitors
(Outputs 1-2) but if you want to hear a studio headphone mix like the player is hearing it, in order to get it to the front panel headphone jacks, it’s necessary to assign that mix to Outputs 7-8 or 9-10.
In my example setup, Studio 3 and Studio 4 mixes are already assigned to those outputs so you can check those mixes simply by plugging your phones into one or the other of the front panel headphone jacks. But if you want to hear the
Studio 1 (Outputs 3-4) mix on your control room phones, you’ll have to also assign it to one of the pairs of headphone outputs, which will take those outputs away from the Studio mix. It’s not a big deal, just something you’ll need to remember to put back together when returning to tracking after getting the Studio
1 mix sorted out. A nice feature for the future would be to offer the option of switching one of the headphone jack inputs to duplicate the Monitor outputs (1-2) instead of outputs 7-10.
Note the “DAW Mix” channel immediately to the left of the output section. This is the stereo playback from the computer, typically consisting of the recorded tracks
mixed using the DAW’s mixer. You’ll use this to listen to playbacks and for a reference mix when overdubbing. During tracking, note that In the mixer screen shot above, the DAW mix channel is muted. You want to hear only the mix of inputs through the DSP mixer. To hear a quick playback after a recording pass, it’s only necessary to un-mute or solo the DAW channel.
When overdubbing, you can work with the DAW mixer to give the player what he wants to hear, but another approach is to set up a Liquid 56 mix which includes individual DAW tracks or stem mixes (drums, background vocals, and such) created in the DAW, plus a mic or two for recording the new track. You may find it more convenient to tweak the playback mix with the Liquid 56 mixer than switching to the DAW mixer when the player wants to hear more or less of a track, though using the DAW mix allows you to apply some EQ to help un-clutter a busy mix and get a better performance.
The mixer has two Solo modes. Both are post-pan, so you always hear the soloed channel in its panned position. PFL is pre-fader and Solo is post-fader.
PFL sends the selected channel to the Monitor outputs (Line 1-2) only and doesn’t affect the other outputs. Solo affects only the output of the currently selected mix. I mostly solo channels to identify (or confirm) the source and to listen for problems like buzzes, distortion, or odd sounding reflections, so for my purposes, PFL is the best choice.
The Mute function only affects the currently selected mix. Mute a mic in one mix and it’ll still be on in the other mixes.
The Monitor Volume control on the front panel works in conjunction with the
Monitor section of the MixControl. With the H/W Control button clicked on, the front panel Volume knob is active and the software knob does nothing but indicate the relative position of the hardware knob. When H/W Control is clicked off, the front panel knob is inactive and the Volume knob in the MixControl takes over.
When switching between software to hardware control, the monitor volume doesn’t change until you’ve turned the active
Volume knob through the present volume setting. If it’s too quiet, you may have to turn the knob down to “catch” the control before you can turn it up. The point of this dual control action is to avoid a sudden blast should the knob you switched be set higher than the current listening volume. The buttons below the volume knob mute or dim (-20 dB) the
Monitor mix, mute the left or right channels individually, or sum left and right for a mono (centered) mix.
The ten blue, red, and gray buttons above the software monitor volume knob correspond to the ten analog outputs to select what the monitor volume knob
controls and how. A red button indicates an output that is turned off. A blue button indicates that the output is controlled by the front panel or MixControl volume knob, whichever one is active. A gray button indicates that the output is turned on, but at full volume and not controlled by the knob. You can, of course, control it with the output level slider for the mix assigned to that output. In this example, I didn’t want the players’ headphone volume to change whenever I adjust the volume in the control room, hence the blue buttons for outputs 1-2 and the gray buttons for the outputs that go to the studio. I could tell you that the 4 th player was out for lunch, but, no, I just switched those buttons to the Off setting for the screen shot so you could see all the colors. Clicking on the buttons toggles them between red and blue, shift-click turns them gray (I had to read the manual to figure that one out).
Incidentally, the monitor level is actually controlled digitally whether you’re using the hardware or software knob. Give either Monitor Volume knob a quick spin and you can hear a bit of lag before the volume changes (my car radio is like that, too). An unadvertised bonus, I suspect as a result of the digitally controlled
Monitor volume, is that the outputs have a delayed turn-on, avoiding a thump in the speakers when the Liquid 56 is powered up.
The line outputs aren’t just for headphones or control room speakers. They can be used for monitoring a surround mix, in which case you’d want all of the outputs all to follow the volume control together. Focusrite has provided a few presets (what’s on and what’s off) for the volume control section to get you started with various surround setups, alternate speakers, and headphones.
These appear to be fixed and I could find no way to store others.
Lurking behind the controls for the mic and line inputs is the functional equivalent of a patch bay. Actually, it’s more like an old time telephone switchboard than the patch bay that you may have in your studio. I’ve mentioned routing inputs directly to outputs a few times already. Well, here’s where you can do it. Clicking on the leftmost button below the mixer section opens a pop-down menu to select either the analog input or router display. The button’s legend changes between “Pre-amps” and
“Router”, showing which control section is active (as if it weren’t obvious).
The Router control panel displays every possible output, 26 physical and two (the
Loop Back outputs) virtual. For each output, there’s a clickable button that pops up a menu from which you select a source for that output. A source can be an analog input, a digital input, a DAW stream, or a mix.
In this example, the Control Room mix (left and right) is routed to the Monitor
Outputs, and each of the for Studio mixes are routed to their respective Line
Outputs. Those assignments are made automatically when setting up the mixer outputs. The example above shows routing for two different studio monitoring scenarios. In the straightforward application, Line Outputs 1-10 carry the control room mix which goes to the monitor speakers, and each of the four studio mixes which feed a four-station headphone amplifier.
Now suppose we have a personal mix headphone monitor system with ADAT inputs such as a Hear Technologies Hear Back. Analog inputs 1-6 are routed to
ADAT1 Outputs 1-6, while the stereo DAW mix goes to ADAT1 outputs 7 and 8.
The ADAT1 output is connected to the input of the Hear Back hub. With this arrangement, each band member, using the Hear Back console, can make his own independent headphone mix. For tracking, since the DAW mix isn’t needed, two more mics could be assigned to ADAT1 Outputs 7 and 8.
The two Loop Back outputs, which appear as DAW outputs 27-28, can be used to feed two channels (or a stereo mix) to another program, or, by selecting Loop
Back 1 and 2 as the source for a stereo track in your DAW, you can record a stereo reference mix for the band to take home after the session. Another application for the Loop Back is to record the output of a media player playing an on-line streaming audio program from a web site.
Latency Issues and Answers
The Liquid 56 ad copy calls this a “Zero latency DSP mixer,” a claim about which
I’m always skeptical - “Yeah, for large values of zero.” There’s no such thing as zero latency when there’s a digital signal path involved, but input monitoring latency through the Liquid 56 comes closer to zero than any other digitally-based monitor mixer I’ve worked with yet. The delay between microphone input and headphone or monitor output is an impressively short 0.7 milliseconds at 44.1 kHz, and it’s only slightly longer, 1 ms, when going through the extra DSP of the
Liquid preamps. At 96 kHz sample rate, this drops to an outstanding 0.25 ms for the standard and 0.35 ms for the Liquid preamps.
It takes far more delay than this to affect musical timing accuracy, but a delay in the range of 1.5 to 3 milliseconds, which is typical of most “low latency” digital monitoring paths, can introduce a different problem. When you’re hearing your own voice through two paths, one through your throat and head, and the other through headphones, the difference in arrival times can result in comb filtering when they mix at your ear. With the earphone signal delayed slightly, a singer may notice that his voice sounds odd because of the cancellation of certain frequencies related to the time difference. Some singers are distracted by this and can’t give their best performance. With the very short delay through the
Liquid 56’s mixer, the “funny EQ” effect occurs at frequencies high enough so that comb filtering is practically unnoticeable. This is a low latency digital mixer that’s almost as good as a straight analog path.
Understand that these extremely low latency numbers are only about monitoring of analog inputs through the mixer. This signal doesn’t involve the computer at all, so the “Latency” adjustments (ASIO buffer size) doesn’t change it. When it comes to monitoring what’s coming back from the computer as is the case when playing a virtual instrument running as a DAW plug-in, all bets are off. The delay between playing a note and hearing it is almost totally dependent on computer latency, so it may not be any better than your old clunker. Stay with me here.
MixControl provides two different latency adjustments, one is the familiar ASIO buffer size (Windows only – Mac users have their own adjustments for this), the other is Firewire Driver Latency, something that I believe is unique to the TCAT
Firewire chip sets and the core driver that they supply to hardware manufacturers who use their components. (TC Advanced Technology is a branch of TC
Electronic, maker of fine signal processors, stomp boxes, and audio interfaces, among other things.) The Firewire Latency settings are Short, Medium, Long, and Extra Long. In other devices, I’ve seen it called names like Safe or Legacy modes. This is a buffer for the Firewire interface itself. A longer setting may sve you from fussing with computer optimization tweaks like disabling network hardware while recording – good for beginners who may not need extra low latency.
At the Short Firewire Latency setting, the minimum ASIO buffer size is 64 samples but for longer settings, minimum buffer size is 240 samples. The maximum buffer size is 3072 samples. I used the default settings of 512 samples and Medium Firewire latency for most of my work with the Liquid 56 and I had no problem recording 16 simultaneous tracks nor with overdubbing while playing those 16 tracks. This was on a modest 3 GHz Pentium 4 computer with 2 GB
RAM, running Windows XP Service Pack 3.
I know that some of you tend to dismiss drivers (or maybe the whole product) as
“junk” if a buffer larger than 64 samples is necessary in order to avoid clicks and stutters, I checked my 16-track 44.1 kHz test project using a 64 sample buffer and had no snaps, crackles, pops, or stallouts when recording, playing back, or overdubbing. Plug-ins? I put the Focusrite compressor that’s bundled with the
Liquid 56 on the two vocal tracks and nothing changed except the vocals sounded better. But there are limits, and I’m sure that some time you’ll find them.
Latency involves more than just a trip through the ASIO buffer, hence the actual throughput delay time is greater than simply the number of buffered samples divided by the sample rate. Here are some measurements I made on the Liquid
56 using Reaper running on my trailing edge of technology computer, over a range of buffer sizes with which you’ll probably feel comfortable:
FW Buffer Delay Delay
The Delay numbers are not exact
20 since my measurement method depends on my accuracy in placing a cursor on the edge of a waveform.
Med 512 1472 33
While I didn’t attempt to bog down the computer with plug-ins to see when it choked, I did put a virtual
Short 128 408 9.2 electric guitar plug-in on a track and overdubbed a part with it with neither playability nor crackles or pops when using the 64 sample buffer setting. I haven’t yet come up with a suitable setup for measuring this delay so you’ll have to settle for an “it’s OK.” Delay times and VST instrument performance on your computer will most likely be different, probably better than on mine.
I’ve already described many of the performance parameters, so this part will be mostly about installation, operation, and ease of use.
Installation and Setup
Like all products of this type, there’s software to install before the Liquid 56 will do anything useful, and, as usual, the driver and control panel application on the supplied CD wasn’t the latest version. There’s only a single installer file (one for
Windows, one for MacOS) so it’s a quick and easy download from the Focusrite web site. Windows Installation was painless with no surprises in Windows – run the executable installer file, then power up the Liquid 56, connect the Firewire cable, let Windows locate the drivers (ASIO, MIDI, I think WDM, and maybe something else – it goes through several “Found new hardware” cycles), and it’s good to go. I trust that Mac users will find installation just as straightforward.
Installation of the bonus Focusrite plug-in suite is a separate operation which requires a few steps. Once the files are loaded on to the computer, you must register the software, get a license key, then activate the license. Following the installation (I did this from the CD since there were no updates), on your first use of one of these plug-ins, a registration window pops up in your DAW with instructions on what to do. Registration and activation is most easily done on line if your studio computer has Internet access, though, if not, you’ll be prompted through an off-line registration process. In either case, after submitting the registration info, you’ll promptly receive a license file via e-Mail. Point the registration pop-up to where you’ve saved the license, then your new plug-in suite is ready to use.
I performed one MixControl update during my test period at the suggestion of a
Focusrite tech support person when we were trying to sort out an odd (and pretty inconsequential) issue. This was a beta version which included an updated driver and the process was essentially a full re-installation, which took just a couple of minutes.
Now You See It, and Now You Don’t
The front panel is mostly black with exception of the Liquid preamp controls which have a silver background. The panel legends are in a very delicate font and in what I consider to be normal light for a control room (hardly dim) I was unable to read anything on the front panel without the aid of a flashlight. I’ll admit to being pretty old and don’t have the greatest eyesight, but I found this panel to be particularly difficult to read. Admittedly, there isn’t all that much to see, and once I got the hang of where the mic/line indicators are there wasn’t much to do but reach for a gain control, but still, I like to see where I’m aiming.
I had a similar visual problem with the MixControl screen display. At my LCD monitor’s preferred resolution of 1600 x 1024 pixels, text was so small that I had to lean in very close in order to read it. The window is a fixed size so it won’t get bigger or smaller, it can only be minimized to get it out of the way. I found it a bit annoying to have to fetch the mixer again after making an adjustment on the
DAW as there wasn’t enough screen real estate to display them both. A dualmonitor setup would be a good idea, particularly when using the mixer actively during tracking.
The MixControl meters are clear, bright, and have a reasonably smooth dynamic response. Essentially they show a short term average but the red clip indicator at the top of the column responds instantly. There’s a meter for each mixer input channel as well as one for the output, so if you’re tracking with more than 16 inputs, some will go unmetered (and un-mixed, as well). There’s a partial solution
to the shortage of mixer meters – you can switch the front panel LED meters to display the analog or either of the ADAT input levels. If you’re clever with your input assignments, you can use those “overflow” inputs for things such as room mics, alternate mics on an amplifier or instrument, or miscellaneous percussion mics that you’ll want to check during setup but don’t need to hear in your tracking mix.
There’s a problem with the software meters, however
– they have no scale. The ∞, 0, and 6 that you see are for the sliders. The screen shots to the left show levels (as closely as I could set them using the front panel gain controls) of, from left to right, -20, -10, -5, and -0.2 dBFS. The front panel LED meters have scale values, but thery’re rather coarse. Compare the two sets of meters.
Looking at the software meters, my calibrated eyeball tells me that Input 1 is at about the right level for a safe recording, with 20 dB of headroom for peaks (I’m conservative, you may be more daring), however, this level lights up only the first LED on the hardware meter. Input 2 looks a bit hot for a normal level but still comfortably below clipping. I’d call it safe, but on the hardware meters, it’s only lighting up two LEDs, so if you were watching them, you might be inclined to push the gain up a little further. Input 4 is just 0.2 dB below clipping, but the hardware meter tells you that you have a safety margin of 3 dB, which you don’t.
I keep telling users that they shouldn’t be a slave to the meters, to set levels sensibly, and leave the gain controls alone until something changes, but on the other hand, I keep asking manufacturers for better meter resolution in the critical range between clipping and –10 dBFS. My recommendation with the Liquid 56 when tracking is to set the gain so that your eyeball average of peaks on the
MixControl meters is about 2/3 of full scale. This will put you in that comfort zone between clipping and “my tracks aren’t hot enough.” Or just check levels with the
The Dreaded Firewire and Software Compatibility
The ‘net is full of advice on which Firewire chip sets, computers, motherboards, and add-on internal or external bus host cards work and which ones don’t. Most of what you read is obsolete, though not necessarily wrong. There are still some valid recommendations. I could make this part of the review much more interesting if I had a tough time getting the Liquid 56 or other Firewire audio devices to work, but in the eight or so years that I’ve had Firewire audio gear around here from, let’s see, Mackie, PreSonus, M-Audio, TASCAM, oh, and
Focusrite, everything has worked without a hitch.
Conventional wisdom is that you need a TI Firewire chip set in your computer. I have several Firewire interfaces, but none with a TI chip. I have two laptop
PCMCIA Firewire host cards, one with an NEC chip and the other, a combination
Firewire and USB host (I never expected that one to work!) with a VIA chip. My desktop computers all have PCI bus Firewire cards, one with a VIA, one with an
Augere, and one with an unidentified Firewire chip. The Liquid 56 works just fine with all of them. I have no computers with a built-in Firewire port, and I expect that it won’t be long before it will be difficult to find one, so don’t fret about needing to use an add-on Firewire I/O card. When you buy a new computer, be sure that it has a place where you can add a Firewire host card should you need it. They’re available now in ExpressCard (newer laptops) and PCIe (newer desktop sized motherboards) bus formats and aren’t very expensive.
I’m still running Windows XP since I’m not doing anything that requires a more advanced operating system, and I don’t want to change what works. While Macs are slightly more uniform than PCs, no two computers are alike, making Firewire compatibility is a bit of a crapshoot, though in the case of the Liquid 56, I believe that TC has a pretty good handle on things. They supply Firewire audio chips to a lot of manufacturers so they get a lot of feedback, and they have a development team that’s constantly improving and updating the drivers for all the operating systems. Since Focusrite is in this loop, I’d say that your chances of plugging and playing with minimal trouble are pretty good on any Windows or Mac that’s newer than about the last five years.
I have some minimal tweaks to my computers to give them a fighting chance with audio applications. I disable wireless networking on the laptops, and wired networking on the one computer that has significant DPC spikes with networking enabled, I have the screen saver disabled, and use “Always On” power management so it doesn’t go to sleep after some period of inactivity. The one connectivity issue I experienced with the Liquid 56 was with my nearly eight year old laptop that’s set up for business office use and is allowed to go to sleep.
When that happens, it was necessary to power off the Liquid 56, shut down the
MixControl, and restart it. But as long as the computer stays awake, so does the
I only ran into one quirky software conflict, and that was while experimenting with the Loop Back output. I was attempting to play a streaming on-line radio program and record the audio using Sound Forge 9. I had done this without a hitch using
Winamp as the media player and Reaper or WaveLab as the recorder. Sound
Forge by itself works fine with the Liquid 56. But with a 44.1 kHz stream playing, as soon as I started recording in Sound Forge, the Liquid 56’s sample rate switched to 48 kHz, Winamp quit playing, and Sound Forge gave me an error message about that sample rate not being supported by the hardware – which isn’t surprising since Sound Forge was set for 44.1 kHz and the Liquid 56 was
(now) running at 48 kHz. Thinking there might be a problem with Winamp, I repeated the experiment using Windows Media Player and got the same result.
Same thing when I eliminated the on-line connection and used Winamp or Media
Player to play a locally stored file. I finally stumbled on the solution which was to first start Sound Forge, then open the media player and start it playing. I didn’t delve deeply into this but I suspect that Sound Forge may default to 48 kHz when first started, then switch to the requested sample rate. The Liquid 56 driver sees that something wants to record at 48 kHz, switches to accommodate it, and things came tumbling down from there.
Since I had only one Saffire interface here, I didn’t have the opportunity to try dual unit operation, however the Firewire driver can accommodate two Saffires.
Due to the bandwidth limitation of the Firewire port, it won’t handle all of the streams from two Liquid 56s, though a Liquid 56 can be paired up with a Saffire
Pro 24 (with or without DSP) or Pro 14 to provide additional computer I/O streams. The DSP mixer screen display doesn’t expand to fit the additional inputs of a second Saffire, however with two
Saffires cascaded via Firewire, clicking on the unit name allows you to select which unit’s control panel is displayed.
The two interfaces synchronize their word clocks automatically and appear to the DAW as a single interface with the total number of inputs for both units.
Since each unit’s inputs can only be routed it its own outputs (you can’t route an input on Unit 2 to an output on Unit 1) so you’re somewhat limited in how you can use those extra inputs in a near-zero-latency monitor mix. One approach to this is to create a single stereo mix of Unit 2’s inputs and send this to Unit 1 as a submix via the S/PDIF I/O. Another, which may offer more flexibility, is to connect the ADAT I/O between the two units and bring multiple submixes in from the slave to the master mixer.
There’s little to say about how the Liquid 56 sounds, so I won’t say much. It’s very good. There is no odd distortion from the converters, the mic preamps are clean (except when you want them dirty), and what goes in comes out. You can spend a lot more money and get a small improvement for sure, but there’s no reason to question the ability to make recordings of excellent technical quality using the Liquid 56. There’s nothing sound-wise to complain about, and that’s pretty high praise.
The fact that the monitor mixer doesn’t have a full channel strip or effects means that you’ll need to work a little harder getting a good sound out of the mics in order to get a good sounding monitor mix while tracking. But there’s nothing wrong with this – eliminating problems with tracks during the recording phase means fewer problems to solve when mixing.
The Liquid 56 doesn’t come with a printed manual. There’s a PDF manual on the installation disk that’s 30 pages long. I grumbled, but printed it anyway. I like to have paper in front of me when I’m learning a new device, and I like to sit on the couch and read through the manual before getting started. You might just want to load it on your eReader and be done with it.
You might get the impression, given the length of this review, that this is a pretty deep product. The manual provided is concise and a bit terse, but covers what you need to know to get going, and gives you a start on more advanced topics.
There are several application notes and FAQs on the Focusrite web site that cover such topics as stand-alone operation, two-unit hookups, interfacing outboard mic preamps, troubleshooting, new setup features added since the last revision of the manual (hey! It’s a downloadable file – why didn’t the just update it?) and there are set-up guide videos for several popular DAWs. This is a worthwhile resource with some useful content, but you have to know it’s there.
Now you do. http://www.focusrite.com/answerbase/en/category.php?id=93
“Pro Tools – Finally Focusrite Ready”
This was the catchphrase around Focusrite’s booth at the Winter 2011 NAMM show. The point, of course, was that Avid’s Pro Tools 9, introduced at the same show, no longer requires Avid or approved hardware I/O hardware, but will work with any hardware with an ASIO driver such as the Focusrite Saffire liine. But to further sweeten the pot, Focusrite designed an RTAS plug-in that makes the most frequently used functions of the MixControl software accessible directly from Pro Tools. Development and debugging took longer than anticipated, but the plug-in is now out for beta test and should be officially released soon. The plug-in doesn’t affect audio, and while it can be placed on any track, it makes more sense to me to create a separate track with nothing but the MixControl RTAS plug-in.
Since I don’t have Pro Tools 9, I was unable to try it and get a screen shot to show how it looks with the Liquid 56, but this will give you an idea of what to expect. Actual functionality, and therefore the GUI, will vary depending on the capabilities of the hardware. I’m just speculating here, but hopefully for the Liquid 56, there will be a button for selecting among the preamp models. I wouldn’t expect the full mixer to pop up in a plug-in, but there may be a simplified
version like the one shown here, with the Blend knob for adding the DAW mix to the input mix. The concept is not total replacement of MixControl with a plug-in, but rather to provide quick and uncluttered access to controls that you’re most likely to need during tracking. This may serve to at least partially alleviate my beef about the MixControl window not being sizeable. I haven’t heard talk of a
VST plug-in, but time will tell.
I’m very impressed with the Focusrite Liquid Saffire 56. It’s flexible, it sounds excellent, it set up quickly and without any fooling around, it’s stable, it’s not buggy. It looks good, it’s nicely built, and it has a sensible set of inputs and outputs that make it easy to integrate into any appropriately-sized studio. I have some niggles with the user interface but this is very subjective and I suspect wouldn’t slow me down once I got used to it. Focusrite has always been strong in the area of analog design and they have chosen good collaborators for this product. TC should keep the Firewire connectivity up to date, and Sintifex has provided a good choice of preamp sounds that are usable without being radical.
There’s a lot of bang for the buck here.
• “Character” preamps when you want them, clean preamps when you don’t
• Very low monitoring latency – exceptional for a digital mixer
• Mic and line inputs can be connected simultaneously (valuable for rack installations)
• Very flexible input and output routing
• Integral power supply – no wall wart or line lump
• Nice bonus plug-ins
• RTAS MixControl Plug-in (coming soon)
• Poor visibility of front panel legends and on-screen text
• No meter scale for on-screen meters make level setting kind of iffy
• Gain pots are touchy near the top end of the range
• No inserts between analog inputs and A/D converters
Street price: $1,000
Further Information at: http://www.focusrite.com/products/audio_interfaces/liquid_saffire_56/
Focusrite Novation Inc.
840 Apollo Street, Suite 312, El Segundo, CA 90245
TEL: (310) 322-5500
Silver 2 Avalon VT-737SP
FF Red 1 H Focusrite RED1
Saville Row Helios console
Dunk Manley SLAM!
New Age 1 Millennia HV3D
Class A 2A
Universal Audio M610
* Your assessment is very important for improving the work of artificial intelligence, which forms the content of this project