Focusrite Liquid Channel

End of the line for analog preamps?

By Lynn Fuston
Never do I recall a product that has simultaneously fostered such great anticipation and trepidation.

The product, Focusrite's Liquid Channel ("LC" henceforth), is a single-channel preamp/compressor/EQ with A-D and D-A converters. While the "channel strip" design is familiar, Focusrite's implementation certainly isn't. Their combination of a configurable hardware matrix for the front end along with "dynamic convolution" (impulse modeling) of other hardware devices allows the LC to offer "replicas" of vintage and current analog hardware, both preamps and compressors.

The eagerness for the LC is easy to understand: It will allow studio owners to audition up to 40 preamps and/or compressors that they otherwise might not be able to hear or use. And all that functionality is in one box, so no more repatching to audition different gear. And even though the LC isn't cheap, if you compare it to 40 channels - even 10 channels - of its analog counterparts, it seems a bargain.

Why then the trepidation? From a manufacturing standpoint, digital emulations have been making rapid inroads into the market for analog hardware. Some engineers are selling their infrequently used hardware to replace it with software. Digital may not have completely closed the gap sonically (depending on whom you speak with), but many feel that the time is close at hand.

When I pulled the LC out of the box, I was shocked by how heavy it was - it feels more like a power amp than a preamp. Looking inside, a large transformer for the power supply and another for the preamp section constitute most of the weight. It's an impressive unit, with 14 rotary encoders, 26 pushbuttons, and no less than 235 LEDs on the 2U front panel.

A front-panel LED indicates whether the transformer is in or out of the circuit. Another switch controls a unique feature called "Session Saver" that allows the LC to monitor the signal at the input and output. If the input is overloading, the gain is automatically reduced at the preamp. If the output is overloading, the makeup gain is lowered. An "activated" LED shows if gain was adjusted. This ingenious feature involves no limiting, just gain attenuation in single-dB increments. It's like having a second engineer watching your levels all the time and keeping you out of trouble.

The large LCD shows lots of information, and has corresponding knobs for many of the preamp and compressor parameters. A nice touch is the 12-segment gain reduction meter, with the first seven segments dedicated to the first 3dB of reduction (very useful). For purity, you can bypass unused processing, leaving only the preamp in the circuit.

On the back are XLRs for mic, line, and AES input, and line and AES output. Word clock in and out (BNC), and the Digital Link Bus (RCA) connections are next to the USB port, which connects the LC to a computer.

The first question most people ask (including me) is how a digital box can claim to replicate a variety of vintage preamps, with all the varying impedances of hardware units and different mic/pre interaction characteristics?

Through the use of a purpose-designed transformer and complex relay circuitry, and resistors and capacitors, the LC's analog front-end can actually change its impedance characteristics and transformer status for each replica.

In order to judge whether Focusrite has achieved their goal of replicating analog hardware, I took several armloads of preamps into Nashville's Classic Recording and set them up next to the LC. The signal source was a female vocal recording fed to a speaker in front of a Neumann U67, which then fed the preamps in the control room. I chose the reproduced vocal instead of a live vocalist to eliminate performance variables.

I calibrated everything to within 0.1dB. In doing so, I discovered that the gain settings from replica to replica aren't consistent. So 33dB of gain on one replica may be the same as 27dB of gain on another replica, making it challenging to do comparisons from one replica to another. Focusrite suggests matching gains and writing the resulting levels to memory, then recalling them for auditioning - a time consuming procedure every time a new source is placed in front of the LC.

I proceeded to compare the replicas with the hardware that was "modeled." (Focusrite will not state which replicas came from which hardware though the replica names are extremely suggestive.) I recorded the original and compared it to the LC version, and then did two more passes with the harmonics settings maxed out. Both even and odd harmonics share a single knob, which allows adding even harmonics at values 1-8 and odd harmonics at 9-15. These settings, even maxed, are subtle, with odd being the more pronounced effect; it adds more fuzz and grit to the signal. Adding even harmonics at anything less than max was very difficult to hear.

Now we get to the part you've been waiting for: What did the preamp replicas sound like? Honestly, they were far closer to the originals than I ever imagined they could be. I was shocked. Some were almost indistinguishable, while others were very good imitations.

The replicas of tube preamps seemed the farthest from the originals. One of my favorite tube preamps has an impressive, wrap-around quality; a low-end signature in the original that was lacking in the replica. It's not the kind of thing that you could measure on test equipment, but something you just feel, like a warm hug. (I hope you know that feeling.) But on several of the solid-state preamps, whether transformer or transformerless, the two were hard to differentiate. I suspected some digital grit would be the telltale giveaway for the LC, but I noticed none.

Adjusting the harmonics added a nice variety to the replicas and more than once made the replica sound even closer to the original. In the end, I preferred the originals in most cases - with the qualifier that I was listening to a solo voice in a quiet room. In a music track, I'm not sure if I would have been able to tell them apart.

I listened through the remaining 32 preamp replicas (that's a lot of preamps!) and found that, though distinguishably different, there seemed to be a similarity between them - the differences being less pronounced than you would hear between the hardware versions. But the LC's range is far greater when you add the option of the adjustable harmonics, plus many of the originals don't have 80dB of gain on tap.

I was fortunate to have two LCs and was able to audition all the compressor replicas in stereo. Strapping two LCs together is a simple matter of connecting the Digital Link Bus of the master to the slave, identifying the master and slave and turning on stereo link.

Once assigned, the slave mirrors everything from the master front panel.

I noticed a difference in update time between master and slave as the settings were changed: When switching the compressor in and out at 192k sampling, the difference was 400ms. At 48k, the difference changed to 301ms. I experimented further and found that when changing a parameter on the master, the slave unit may not "get the memo" that the value is changing until you stop turning the knob. This can result in a slow ramp on one channel and a rapid adjustment on the other. If you're accustomed to reaching over and grabbing a knob while tracking, this could be significant.

Like the preamps, there are 40 replicas of analog compressors. The compressors do sound quite a bit like the corresponding original hardware. Several compressors with which I am very familiar were well represented by the LC. To its credit, the LC replicas that I liked the most were of compressors that I like the most. Some of the comps I am unfamiliar with interested me so much that I have pursued hearing the original hardware versions to see what I think of them. (Not what Focusrite expected, I am sure, for the LC to serve as an expensive auditioning system!)

The equalizer offers extreme overlapping frequency bands. The high shelf band sweeps from 20kHz down to 200Hz, and the low shelf band stretches from 10Hz all the way up to 1kHz. The mid band is peaking with switchable Q setting and is sweepable from 100Hz up to 10kHz. The EQ section is switchable to a sidechain, which has a "listen" switch. Boost and cut on each band is &lusmn;18dB. The EQ can be routed before or after the compressor. You can choose to show all EQ parameters on the LCD screen.

The manual says it's "loosely based on the classic Focusrite sound of the original ISA 110" and is primarily "designed to allow small amounts of corrective shaping." That sums it up fairly well.

One nice thing: the boost/cut increments for the EQ are in very small steps near 0dB and increase at the extremes. For someone who is frustrated by coarse 1 and 2dB steps in an EQ, this is great.

I ran a calibration tone through both review units and immediately noticed something was askew. The output levels from the LCs weren't identical to the input levels. Not only that, but they weren't identical to each other. The master unit output was 0.56dB hotter than the input. The slave unit was 0.73dB hotter than the input. Focusrite designer Rob Jenkins confirmed that the difference was in the transformer windings and there is currently no calibration for it. According to Rob, the gain difference is consistent because the unit uses a relay-switched discrete resistor network. With a single unit, this won't be an issue, but for stereo work it may be noticeable as an image shift.

Then there's that inevitable side effect of digital processing: latency. The LC latency from analog in to analog out is admirably low. (I measured 41 samples at 192k and 179 samples at 48k.) For mono sources, no problem. But in a multimic setup, you may have to compensate to avoid phase cancellations. The engineers that tried the review LC for doing overdubs had no issues with latency at all.

For those who like control but don't like reaching for knobs, there's remote control software called LiquidControl that runs on Mac OS X or Windows XP. You can control up to eight different units, backup or restore data, add or remove replicas, and more. Speaking of adding replicas, Focusrite is committed to expanding the library of available replicas and within a month of releasing the LC had uploaded 16 new preamp and five new compressor replicas for free download.

So is the LC a winner? For those engineers who want lots of options but don't have lots of cash, the LC will fit the bill. For those unfamiliar with the sounds of many hardware units, the LC will help train their ear to hear differences. The sound of the preamp is very good, even without the replicas, and all the functions of the LC are very useable. The build quality and design are first rate. The digital encoder knobs are comfortable and work well. The "LEDs per square inch" ratio is sure to impress artists and clients alike. All in all, I think the LC is a useful and revolutionary box.

For innovation, I give the Liquid Channel five stars. Nothing like it has been done before and it certainly succeeds where other modeling software has not. Oddly, I don't think that sales of the LC will sacrifice many hardware sales. It will serve as a great starting point for many engineers who will eventually desire more and more of the actual hardware. Ultimately, the LC will be another valuable item in a studio's toolbox and will enable people to experience options they've never heard and variations that have never existed.