ATL M2820 Mixing Console


The worldfs first 192-kHz, 32-bit digital mixer—an incredibly intuitive and robust console that delivers exquisite, high-resolution audio with near-zero latency


Powered by a 40-bit FPGA engine, the new M2820 Mixing Console from Acoustic Technical Laboratory Inc. offers unmatched audio quality and standout reliability. Boasting a remarkable 192-kHz sampling rate and 32-bit I/O, itfs sound is simply second to none. Marking the next phase in mixer evolution, the M2820 features 28 input channels and 20 output channels, all connected via an advanced routing system that takes full advantage of ATLfs time-proven analog circuitry and high-performance A/D and D/A converters for near-zero, 0.5-ms latency. And while wefre on the subject, note that you also get eight channels of USB output.

The intuitive user interface and high-resolution LCD screen make it faster and more efficient than ever before to create any mix youfre after. Further, with proprietary FPGA technology at the heart of the mixerfs audio-signal processing, all channels are handled in parallel, audio and control signals are completely isolated from one another, and whatfs more, custom builds tailored to specific applications are also supported. And as if this werenft enough, parallel processing also significantly reduces the clock speed, all but eliminating the need for cooling and keeping power consumption at a low 200 W. Meanwhile, the mixer can also be powered by 24 V DC. All in all, the M2820 Mixing Console packs a formidable array of state-of-the-art technologies, providing everything required for a cutting-edge mixing solution.


Design Concept 1: Easy to use. For anyone

In designing the M2820, we wanted to liberate the sound engineer as much as possible from having to remember or look up complicated operations and procedures. The conscious decision-making required for such tasks is a left-brain activity, and it disrupts the creative processes of the right brain. Like any good tool, the mixer must feel like a natural extension of the body—something that the engineer can use instinctively, without thinking, to realize the sound in his or her mindfs ear.

The music mixer is the central element of a highly complex yet subjective process where many input signals are expertly mixed together, multiple mixed signals are output, and highly personal adjustments are made based on the resulting sound. As such, it relies on the fine-tuned sense of hearing of the engineer. And in order that he or she may focus on creating the perfect mix as efficiently as possible, our task was to minimize the number of actual operations required. To this end, faders all correspond directly to input channels, information is presented in a highly intuitive fashion for instinctive decision making, functions are logically assigned to switches and buttons, and screens are arranged for effortless navigation to the desired functions.


To be truly intuitive—to free the engineer from the realm of conscious decisions—we knew the M2820 would have to provide unfailing peace of mind, it would have to feel like a natural extension of the body, and it must allow for extended use without tiring. A key step in achieving this goal was minimizing the number of operations and procedures required to be remembered. And while human error can never be eliminated from the mixing environment, we wanted to make sure that any mistakes could be easily rectified. When the creative juices are in full flow, trying to recall each step in a complex procedure jerks onefs train of thought from the right brain into the left. This simply gets in the way of creating the perfect mix.

Meanwhile, the streamlining of operations and screen layouts also allowed us to achieve another usability goal—namely, ensuring that the M2820 could be quickly and easily put to use by anyone, regardless of their level of experience. Beginners and experts alike will find functions efficiently organized to suit their needs.

With the advent of high-resolution smartphone cameras, anyone could take high-quality photos wherever they may be, prompting a boom in the popularity of photography. Here, high performance and convenience significantly lowered the barriers to entry, which in turn has allowed new talent to blossom. Meanwhile, advances in digital technologies have made the PC ubiquitous, and with so many regular users, computer applications must allow complex tasks to be completed almost without thinking about them. Ease of use is key, even to the extent that the total number of mouse clicks must be minimized for optimum simplicity.

The same applies to the sound mixer. More than just a simple summer of audio signals, it can rightly be called the main control panel of the music studio or live-music venue. As a professional creative tool—not unlike SLR or TV cameras—it relies heavily on the senses, making the operator much like a performing artist. The starting line for any such device must be incredibly simple operation that allows the sound designer to recreate in the real world what can already be heard in his or her head.


Design Concept 2: Uninterruptible sound

All digital processing of audio signals is hardwired into the M2820, and as long as the power stays on, the sound will not cut out. Built using proprietary technology, a dedicated FPGA processes signals on all audio channels in parallel, manipulating this data using parameters as variables. Thus, when no changes made to the parameters, audio signals will continue to flow through the system in accordance with current settings. In this, we have realized a level of stability equivalent to that of an analog desk.

The current states of faders, encoders, and switches are detected in a sequential fashion, and the corresponding control data is written to specially designed registers in the FPGA dedicated to audio-signal processing. The use of registers in this way to transfer control data has made it possible to totally isolate control signal processing and audio signal processing from one another. Meanwhile, a separate CPU writes control data and monitors the operating status of the FPGA, and using an array of microcontrollers, values are rapidly acquired from faders, encoders, and switches. This also allows complete scenes to be easily stored and recalled whenever needed.

Graphics processing on the M2820 is achieved using a DIMM-PC equipped with a GPU and running Linux. High-speed rendering has been achieved through the development of special GPU code.


The field-programmable gate array (FPGA) is a type of integrated circuit that, using software, can be freely configured by a designer after production, thereby making it easy to build custom chips. Depending on the size of the chip needed, the number of logic elements in an FPGA can vary widely, and similarly, prices can range from several dollars to tens of thousands.

While software is needed to create programs in the FPGA fabric, the resulting chip is completely hardware—in other words, it does not need any software to operate. Calculations occur as a matter of course when current flows through the chip, making it just as reliable as an analog circuit.

Yet the FPGA approach offers many more advantages: For a start, high-density integration provides for smaller chips. In addition, a single FPGA chip will consume much less power than a comparable circuit board, and it can be powered using batteries alone. And if we consider that packaging accounts for roughly half the cost of circuit assembly, it is clear that VLSI designs made possible by FPGA technology allow for significant cost savings in terms of wiring, boards, wrapping, sockets, connectors, and casings. Finally, VLSA chips are fast—much faster than circuits assembled at the circuit-board level—and one of the reasons for this is the dramatic reduction in distances between individual components.

As evidenced by our development of a new method for parallel processing of audio signals in the M2820, the FPGA approach offers great potential going forward. This is particularly true when we trust Moore' Law and its assertions that (1) the number of logic elements in dense integrated circuits will continue to grow, boosting stability and performance; and (2) greater LSI production volumes will continue to pull down prices.


Design Concept 3: Stunning sound quality

The sounds we hear are not pure tones—they also contain a rich mixture of overtones and noise. And we do not hear them with our ears alone: our ear drums certainly vibrate, but so too do other parts of the body. What all this means is that the sensation of hearing is highly complex, and to faithfully capture these signal in the digital realm, we need the highest possible sampling frequency.

At the heart of audio-signal processing in the M2820 is a dedicated 192-kHz chip built using proprietary FPGA technology. This phenomenal sampling frequency, combined with the remarkable dynamic range afforded by the chipfs 40-bit architecture, delivers performance that far surpasses analog devices and ensures that digital processing is possible right down to the noise floor.

This means that the mixerfs sound quality is ultimately defined by its analog circuits and analog-to-digital converters (ADC). With this in mind, we brought ATLfs tried and trusted amplification circuits fully up to date for use in the M2820, pairing them with the Asahi Kasei AK5397—an ADC boasting a signal-noise ratio of 128 dB—in a 192-kHz, 32-bit configuration for stunning sound quality.


Design Concept 4: Support for 196-kHz, 32-bit multitrack recording

When connected via USB to a PC, the M2820 can feed up to eight tracks of superior quality audio to Pro Tools or any other digital audio workstation running on it. This makes it incredibly easy to create high-resolution recordings and masters.

Sound mixers typically combine samples from 16 or more input channels simultaneously, and internal processors need a high bit depth in order to handle all of this data. Summing 32 channels of 32-bit audio would require processing with a bit depth of 37. However, the summing of samples can produce nonlinear results, and when hundreds of these errors occur in a digital-to-analog converter, significant distortion may be audible in the output. It is for this reason that we developed the M2820fs audio-signal processor as a 40-bit system.


Key Features

  Highly intuitive user interface

  Audio input channels: 28

  Audio output channels: 20

  USB output channels: 8

  Sampling rate: 192 kHz

  Resolution: 32 bit

  Latency: 0.5 ms

  Complete separation of control and audio signals

  Time-proven analog circuitry combined with high-performance ADCs

  FPGA chip used for all digital processing of audio signals

  No DSPs used

  High-resolution LCD

  Support for custom builds


Block Diagram