When my wife and I go out to Italian restaurants, wherever we may be, she has a propensity to try the lasagna. Making a very excellent lasagna herself, she is always curious how various restaurants interpret the dish. Over the years she and I have come to the same conclusion – there are no two identical lasagnas, and each and every restaurant has its own take on the dish. Now, having identified the very best ones we know of, she is less likely to order it in an unknown place. If she wants a really good one, we simply wait until we are home and can drop into the local best of the best. It turns out that digital to analog converters (DACs) have something in common with lasagna, and it is something peculiar to this component alone.
Like all high-end audio components, DACs have their own sound, making the shopping process important, arduous, and time consuming. It is not everyone’s forte to get into the weeds of what goes on in their equipment, but when it comes to the DAC it can be helpful to know some of the basics, and it can allow the shopper to make some simplifications in the process that help get him or her to the finish line. It is even possible to have an enlightening conversation with the designers when a very elementary understanding is achieved. The goal of all of this is to make the best selection of the DAC without spending inordinate time and trouble in the process.
Let us digress a bit for introduction. Sound is nothing more than specific variations in air pressure versus time. Recording sound is the process of measuring air pressure over time and recording the result by one of various means. Every method begins with one or more microphones, which convert air pressure variations into small electrical signals. Unlike analog media such as magnetic tape and vinyl records that preserve a continuous record of these voltages, digital recording does not create a continuous record of microphone signal voltages. Instead, a digital recorder samples the signal at discrete points or intervals of time and stores the information as a binary number that closely represents the signal amplitude at each time interval. This results in a computer file that contains a large number of binary numbers. That file can be streamed off the internet, stored on a computer or music server, or it can be stored on a DVD as was common a few decades ago. At the playback end, in the case of a digital file, a music server streams the file from the internet, or from a local drive, and serves it to the DAC as a precisely timed sequence of numbers contained in the file.
In comes the DAC. Its job is to read the digital stream from the server and convert it back into an analog (time-continuous) signal that represents the air pressure variations of the original musical performance. Stated briefly, the process in the DAC is to clock in the binary numbers of the file at the precise rate at which they were originally sampled, convert them to voltages (what the preamp, amplifier(s), and speakers operate on), and reconstruct the resultant series of voltages into a time-continuous voltage like the voltage produced by a microphone. We won’t get bogged down in the nuances of the first two processes, but we will take a closer look at the last process because it has something to do with our present interest, why different DACs sound as different as lasagnas taste.
The graph below shows a simple sign wave for illustration purposes, as a blue dotted line trace. This may be thought of as a simplistic representation of the microphone voltage in the recording studio. The red circle markers represent the points in time at which the signal was sampled by the digital recorder. The green trace represents what would emerge from a conversion of the sample values back to voltage. Notice that it is roughly similar to the original sign wave, but with obvious differences. This is where the reconstruction filter comes in. Its purpose is to block out the content that gives it a sharp, stair stepped contour, so that it will more accurately resemble the original signal before it was sampled. A detailed discussion of this filter is highly mathematical and beyond our present purposes. What is important to realize is that the reconstruction filter is performing an algorithmic process that is not unique. In other words, there is an infinite number of ways to reconstruct the signal (reconnect the dots), with tradeoffs in filter complexity and accuracy, yet none is absolutely perfect. The reason for this is easily appreciated by considering that the filter is being asked to generate signal levels at points in time for which the original signal was not sampled. How can it possibly do that? The situation is not nearly as bleak as it sounds thanks to something called the Nyquist Theory, which states that complete recovery of the signal is possible with a perfect reconstruction filter. Yet, the required perfect filter is not attainable in the practical world and there is a significant element of art involved in the required compromises. Unavoidably, there is artistic interpretation in the design of the practical filter that results in sonic interpretation. Put simply, no two reconstruction filters sound exactly alike.

What does all of this mean to the audiophile? The main thing it means is that DACs from the same manufacturer tend to share similar, if not exactly the same, reconstruction filters. So, if you decide you like a certain company’s offerings in this space, you may want to stay with that company and explore their entire line of DACs. Liking the sound of a DAC means, to a large extent, liking that company’s reconstruction filter. Certainly, other factors enter into the sound of a DAC, such as power supply design, the type of converter used, the analog portion that prepares the signals for output (whether vacuum tube or solid-state), etc. Furthermore, a large number of DACs employ standard chip sets that contain most of the above-described functions, and as a result they may sound similar to each other to a degree.
So, what should a person do? Find a DAC he or she really likes, by finding a system that sounds great and taking note of the DAC therein. This might be done at an audio show, one of the few remaining audio salons, or at a friend’s house. On the recommendation of an audio reviewer, one might perhaps purchase a product with an option to audition. The point to this is to identify the DAC chipset and reconstruction filter (ask the manufacturer for this information) once you hear a DAC you like. If the DAC uses a standard chip set such as those offered by ESS Technology, for instance, you may want to note which specific chip you are listening to and then screen for DACs that use the same one. Standard chip sets usually contain the majority of the reconstruction filter on board, so the DAC manufacturer has done “one stop shopping” in these cases.
More sophisticated DACs at the upper end of the price range tend to depart from the standard chip offerings in deference to more specific and eclectic chips, even those designed in-house in some cases. There can be very extensive and proprietary reconstruction filters within these elite products. If you dare to take a serious listen to such products and find yourself favoring a particular one, bear in mind that you are favoring a very custom and proprietary reconstruction filter. At this point, provided you can afford the DAC, you may have a component you can enjoy for many years to come. But, if the itch to upgrade returns, consider going back to the same manufacturer, because a really great lasagna is easier to return to than to discover again.








