Dan Lavry has gained a reputation for creating some of the best A-D converters on the planet — and this time you get mic preamps as well!
Lavry Engineering are very highly regarded American manufacturers of digital audio converters. Dan Lavry's designs have a very strong engineering emphasis, building on his considerable experience of designing cutting-edge digital equipment for the instrumentation, medical and telecoms industries. In these markets, the subjective descriptions of perceived performance that are so common in the audio world are entirely eschewed; only hard measurable data and real-world objective verification is acceptable. Lavry also developed the state-of-the-art converters employed in products from companies like Otari, New England Digital, Wadia, Pacific Microsonics, Mark Levison, and several other high-end manufacturers. Before setting up his own business, he also played a pivotal role in moving Apogee from manufacturing only sophisticated analogue brick-wall filters into developing complete converter systems.
Lavry's current products are innovative, with several proprietary features, including novel dither and jitter-reduction technologies. There are currently three colour-coded performance levels: the flagship models are 'gold', while the 'blue' products offer configurable modularity, and the 'black' units provide the most cost-effective entry to the marque. All three ranges revel in no-nonsense engineering, and each model does exactly what it is supposed to do, with an impressive cost-performance ratio.
The AD11, part of the LavryBlack range, is a two-channel, very high-quality A-D converter incorporating ultra low-distortion mic preamps. It's housed in a half-width, 1U case (a rackmount kit for two LavryBlack units is available as an option), and is intended primarily for portable recording applications — for which its USB2 interface is particularly useful.
The preamp accepts either line or microphone inputs on XLR or TRS connections via 'combi' sockets. In 'line mode', the gain can be trimmed between 0 and +12dB in 1dB steps, and phantom power is automatically disabled. In 'microphone mode', the gain is adjustable between +20 and +65dB (in 1dB increments, again), with phantom power individually switchable for each channel. A soft-saturation feature is also included, to help control transient clipping, and it can be switched on separately for each channel.
In line mode, the TRS socket accepts balanced or unbalanced signals with an input impedance of around 250kΩ per leg. A nominal -10dBV input signal provides a digital output of -18dBFS, but because of the relatively coarse bar-graph scaling, the meter actually indicates -24dBFS. This socket is not specifically intended as a DI input, but its additional 12dB of available gain and reasonable input impedance (the same as a Radial J48 DI box) makes it quite practical to connect electric guitars and basses directly.
If the XLR connection is used, a nominal +4dBu input signal returns -20dBFS (again, the front panel meters showed -24dBFS during my tests). When operating in microphone mode, both TRS and XLR connections can accept an input signal, and phantom power appears on both!
The digital output is provided on three connectors simultaneously: an XLR and an RCA-phono socket provide AES3- and S/PDIF-formatted signals, while a USB2 B-type socket (the square one) provides a plug-and-play feed for a computer. The AES3 and S/PDIF outputs are always identical in audio data and channel status information, but differ in output voltage (7V versus 1V). However, a configuration option allows the channel status information to be switched between default professional and consumer formats, if required.
A pair of BNC connectors handle word-clock in and out and, unusually, the word-clock input (which is permanently terminated with 75Ω) will also accept AES3 or S/PDIF clock references (via a configuration option). Standard and double sample rates are supported (44.1-96kHz), but Lavry argues very convincingly (and correctly, in my view) that anything above 96kHz is detrimental to timing accuracy, so 192 and 384 kHz are absent.
Despite its compact size, the AD11 incorporates a built-in universal mains power supply, and is equipped with a standard IEC inlet able to accept mains voltages between 90 and 264V AC. There is no externally accessible fuse, and the unit consumes about 14W when operational.
The front panel is very clean, with only a black on-off push button and four silver toggle switches, plus a horizontal LED bar-graph meter and a two-character, seven-segment LED display — and a large part of this simplicity is because of the dual-function bar-graph. Legends above indicate signal levels, while those below translate the configuration status. The audio meter mode is scaled from -42 to 0dBFS, with orange and red LEDs in the right half covering the top 6dB in 1dB increments. The left half uses green LEDs with 3dB increments from -6 to -18dBFS, and then 6dB steps to -42dBFS.
When the AD11 is powered up, it goes through a self-test mode, during which each LED in the bar-graph is flashed on and off, followed by a display of the current configuration (recalled from the last session). Each option is indicated in turn, with the LED 'hovering' for about eight seconds over the right channel's phantom power status, then again over the left channel's, before the display reverts to being a level meter. These 'wait states' allow the user to change the phantom-power mode before the unit boots up completely. It takes about 20 seconds in all before the converter is ready for use.
All four toggle switches are bi-directional momentary types, and the two on the left are used for system configuration. Pushing the first automatically opens the configuration mode, and further up/down pushes step forwards or backwards through the available parameters. The second toggle changes the state of the selected parameter or exits the configuration mode.
When displaying the configuration status, the first four meter LEDs indicate the current sample rate and the next two indicate the external clock status (word clock or digital audio reference). If either external reference is selected, the LED directly above will illuminate when slave lock is achieved. However, it is necessary to select the same internal clock rate as the external reference clock; if the two differ, the digital outputs will be muted.
The next two LEDs indicate whether the AES3 and S/PDIF outputs are formatted with professional or consumer channel-status data. Phantom power is switchable individually for each channel, (the left channel shown by the upper LED and the right by the lower), and the soft saturate mode is configured for each channel in the same way. Strangely, phantom power can apparently be turned on even when the unit is in 'line mode' — although, thankfully, 48V doesn't actually appear on any of the input sockets unless the gain is switched to 20dB or more (so the display can be misleading in this respect). The final LED indicates the status of the meter's peak-hold function.
The operation of the configuration switches is clever, but configuring the unit can require a lot of switch clicks. I didn't find it very pleasant to use and I suspect this could well be a 'marmite' feature that potential purchasers will either love or hate. For example, it takes 12 separate toggle clicks to reach the peak-hold parameter from the first sample-rate parameter, and 12 clicks to get back again: the selection process doesn't cycle around the loop, sadly.
Of more practical concern, though, is the fact that the operational status and configuration are entirely hidden in normal use. Call me old-fashioned, but I like to be able to confirm the status of a product with a quick glance, rather than having to reach across and enter its configuration mode just to determine whether the phantom power is switched on, or the clocking arrangements are appropriate. The latter is a particularly good example: the display gives no warning at all if the internal sample rate selection is mismatched with an external reference, which isn't very helpful when it all goes quiet! Every other converter I can think of has a permanently visible 'clock lock' indicator, and every preamp a phantom-power warning light.
The momentary toggle-switches to the right of the bar-graph adjust the preamp's gain separately for each channel, the current value being revealed on the two-character, seven-segment LED display. Oddly, this displays an entirely pointless 'LE' most of the time, but reveals the gain if either channel's up-down toggle switch is pressed. The first press reveals the current gain without changing anything, and the switch has to be released and pressed again to alter the setting for that channel. Separate clicks provide 1dB increments, but if the toggle is held up or down, the gain will increment continuously and with increasing speed.
If the preamp is in line mode, the gain can be adjusted between 0 and +12dB, and when the +12dB level is reached, the display flashes. It continues to flash for a further five seconds after the toggle switch is released, and to enter the higher-gain microphone mode, the same toggle must be pressed again within three seconds after the display stops flashing! The reason for this protracted process is to make the mode change a conscious decision. There are entirely separate preamp gain stages for the line and microphone modes, switched with relays; and the phantom power ramps up automatically when entering microphone mode (if it has been enabled in the setup configuration). Phantom power is turned off automatically when entering line mode although, unhelpfully, this condition is not accurately reported in the configuration display! The same flashing display procedure is repeated when leaving microphone mode, when the gain is reduced to +20dB.
For those who like to the excitement of recording without a sensible headroom margin, the AD11 incorporates a soft saturation facility, which can be switched on for each channel individually. This feature is claimed to emulate analogue tape saturation, and basically introduces a progressive non-linearity in the transfer curve, to 'squash' the top 3dB of the converter's dynamic range, 'crushing' transients rather than clipping them. Signals below -3dBFS are completely unaffected, but their overall level is raised by 3dB, so the average level ends up being 3dB louder when the soft-saturation mode is engaged.
As a line-level converter, the AD11 performs extremely well, with a huge dynamic range (the practical equivalent of about 20.5 bits) and an extremely neutral and transparent sound stage. The soft saturation mode is also transparent for signals below -4dBFS, but traps occasional full-level transient peaks very smoothly, and while the process is, obviously, noticeable, it is also quite benign and could be described as 'tape-like'. Personally, I like to maintain a sensible headroom margin when recording and try to avoid transient clipping or crushing, but for those who feel strangely attracted to meters hitting the end stops, the saturation system provided here works well.
The ability to send digital outputs to three destinations at once — a computer and a couple of hardware recorders, for example — is very handy for backup purposes, and the converter performed extremely well when I had it clocked externally. I was unable to measure or hear any difference in sound quality or artifacts. The USB connection proved trouble-free on a PC laptop, although Lavry point out the importance of checking and/or setting the operating system's sound interface options and the recording DAW project to the same sample rate as the AD11. Although it works with the default WDM driver, Lavry recommend using the ASIO4ALL driver, which is what I normally use on the laptop anyway. When I plugged in the AD11, the ASIO4All driver recognised it, and I experienced no problems using it with several different DAWs, including SADiE, Reaper and Audition.
As a microphone preamp, the AD11's performance is also extremely good, with a very low noise floor, a healthy gain range and a fundamentally fast, clean and transparent sound character. The phantom power supply was spot on for voltage and has plenty of current capability. I am a little disappointed that phantom power appears on the quarter-inch jack connection in mic mode, but I doubt this will catch anyone out in practice, and at least the phantom is automatically disabled when the gain is below +12dB… although it would be better if the configuration display reflected this state accurately.
The Lavry Black AD11 ticks all the boxes for sound quality and technical specifications, and I can't fault it at all at that level. However, in situations where configuration is likely to need to be changed frequently, the user interface is likely to become frustratingly tedious — it did for me, anyway! Consider, for example, the simple requirement to change from line mode to mic mode and engage phantom power on both channels. Doing this on most stereo mic preamps would require dialling two rotary controls around to the required gain and pressing two buttons. To achieve the same with the AD11 could involve 13 toggle clicks to turn on the phantom power, and then at least four more to change the gain — plus two five-second waiting periods before you can change modes. Admittedly, if phantom power had been pre-armed it would turn on automatically when the gain range jumped to microphone mode, saving the first 13 switch presses, but if you wanted to check that condition first, it would take at least two switch clicks!
Of course, my view of the user interface may not be shared by others and the most important aspect of a converter or preamp is the sound quality. On that score, as I have said, the Lavry Black AD11 is a very impressive performer indeed, and a highly cost-effective product. As a compact, high-quality stereo interface for location recording with a laptop, it is hard to think of anything close, let alone better.
For line-level conversion duties, the LavryBlack AD10 is £200 less expensive and provides some interesting tube, transformer and clean signal-path emulations. Another line-level A-D converter with USB connectivity is the Benchmark ADC1 USB, which costs virtually the same as the LavryBlack AD11. As for combined preamp/A-D converters, there are several four- and eight-channel units such as the Audient ASP008 and Focusrite ISA428 at a similar price, but with arguably inferior converter performance. At higher cost are options like Neve's 1073DPD and the Grace M201.
Bench tests of the LavryBlack AD11 confirmed its impressive credentials, with the results matching or exceeding the published specifications. The dynamic range measured to the AES17 standard was 120dB, and THD+N was better than 0.001 percent for signals above -14dBFS in line mode (distortion rises rapidly above -4dBFS when the soft saturation mode is engaged, of course). The frequency response is ruler-flat to 5Hz, and is roughly 30dB down at half the sample rate, achieving maximum attenuation of over 110dB at about 0.56fs. Although theoretically non-ideal, this kind of performance is quite typical. Channel crosstalk at 10kHz in line mode is a very healthy 112dB. The maximum input level in line mode via the balanced XLR connection is fractionally shy of +24dBu.
You can find test plots taken with an Audio Precision Analyser during the course of this review on the SOS web site.