If you have read any audio newsgroups on the Internet, you may have noticed that certain topics spark endless and heated debate: Mac vs. PC, analog vs. digital, Pro Tools vs. everything else. Though the number and intensity of those holy wars ebb and flow, they still rage on because there is no single correct answer.
The debate over analog summing has evolved into another such contentious topic. A quick look at high-end pro audio gear reveals dozens of analog summing devices, some with prices exceeding $5,000. Much of the marketing for those boxes implicitly or explicitly suggests that you need to sum your digital audio workstation tracks through their analog circuits for your mixes to sound their best. Naysayers insist that the boxes have no tangible effect on sound quality and that their venders are selling snake oil. As in all such arguments, many knowledgeable people with opposing viewpoints are all convinced that they are absolutely right.
If you are just tuning in to the debate, the signal-to-noise ratio of the arguments can make it difficult to get answers to simple questions. What exactly is analog summing? Why does it spark such heated debate? Does it require expensive equipment? And the most important question of all: will analog summing really help your mixes sound better?
Sum Great Reward
The idea of summing tracks using an analog mixer is nothing new. In the early days of digital recording, tracks were always summed together using a traditional analog mixer because no other options were available. With the rise of DAW software, recordists gained the ability to mix multitrack productions entirely within their computers.
As personal and project studios moved to multitrack digital tape, standalone digital recorders, and computer-based DAWs, it became clear that digital audio technology had won the tracking war. Regardless of how good analog recording might sound, digital recording became ubiquitous. But the idea remained that analog hardware was useful for adding warmth and life to digital recordings, and soon a huge market developed for tube mics, analog preamps, and compressors with classic tube, transformer, and optical circuit designs.
Some recordists felt that something was missing from the sound stage of their DAW mixes. They began to send as many tracks as possible, either individually or in small submixes, from the computer's audio interface to an analog mixer and then back to the DAW. They felt that routing individual tracks through an analog mixer improved stereo separation and the width of the stereo field. Some advocates reported that better clarity, a more open or airy sound, was possible when they summed their DAW mixes through analog mixers.
FIG. 1: With the Dangerous 2-Bus, -Dangerous Music became the first company to serve engineers who wanted a simple analog summing box.
A new type of hardware device, the analog summing box, arose to serve the specific niche created by such assumptions and observations (see Fig. 1). Put simply, an analog summing box contains the guts of a mixer minus any unnecessary extras. Typically, such devices offer multiple inputs, stereo outputs, and a clean signal path in between. Analog summing boxes often provide only an output gain control and sometimes individual pan knobs or level trims, but always fewer options than a complete mixer. Their manufacturers assume that all of the effects processing, submixing, automation, and metering happens in software; consequently, the hardware has only the minimum circuitry necessary to sum the channels.
Analog summing advocates feel that hybrid digital-analog systems give them the best of both worlds. In addition to virtual tracks and the advantages of computer editing, processing, and automation, users gain what they perceive as better headroom, definition, and stereo imaging from analog summing. And most advocates insist that analog summing is the one true way to achieve those results.
Not everyone believes that analog summing is necessary or even desirable. Some recordists insist that mixing digitally has no inherent disadvantages. They hold that any lack of clarity or stereo separation is caused by the person doing the mixing and not by the nature of digital mixing.
Moreover, digital summing advocates point out that even recordists who espouse analog summing do much of their summing digitally. Most analog summing boxes or small-format analog mixers have no more than 24 discrete audio channels. Even the most basic DAW offers at least 32 audio channels, and many DAW projects easily use two or three times that many tracks. To sum DAW tracks to analog, you need to group them together and submix them down to the number of inputs that the analog summing device offers.
For those reasons (and others), recordists who are satisfied with their digital mixes are adamant that analog summing boxes have nothing to offer. The only difference between summing a mix digitally and summing a mix through analog hardware, they insist, is the number of steps you take to achieve the same result.
You can see why this topic gets people on both sides of the fence so excited. Advocates of digital summing reject the idea that anything is wrong with their mixes and feel that analog summing advocates are calling all their output inherently inferior. Advocates of analog summing feel that digital summing advocates are claiming their equipment is worthless and their ears are lying. And when people feel they have something to prove, the truth often gets buried.
Nothing beats firsthand experience. You don't need to have thousands of dollars' worth of extra gear to hear the results of analog summing for yourself. You might even have the right equipment sitting around your personal studio already. Inexpensive gear doesn't necessarily mean bad results.
FIG. 2: If you need extra analog outputs to connect to a summing device, you can use a -digital-to-analog converter box such as the Apogee DA-16x (pictured above), which offers inputs in almost every digital audio format and 16 analog output channels.
First, make sure that your software can send either individual or bus outputs to individual outputs on your audio interface. Your software will need to send one output to each of your mixer's inputs. Some entry-level DAWs limit how many channels they send to your audio interface; Apple Logic Express 7, for example, can send only 12 audio channels.
If your DAW is up to the task, you'll need an audio interface with enough outputs. Very few audio interfaces have more than eight analog outputs, but many also have ADAT or other digital audio outputs. You can also use digital outputs for analog summing if you connect additional digital-to-analog (D/A) converters (see Fig. 2); just make sure you properly clock the devices to each other.
FIG. 3: You can use -affordable analog -mixers, such as the Mackie Onyx 1620, very effectively for analog summing.
Finally, you'll need to have an analog mixer or a summing box. It doesn't need to be expensive; any analog mixer can sum its channels to a stereo master output. You can use any budget mixer that has enough channels (see Fig. 3); there's no point in using a 4-channel mixer, because you'd probably have to submix almost the entire song, but any 8- or 16-channel mixer will do. If you can afford a dedicated analog summing box, many are designed just for this purpose.
Wiring your system for analog summing isn't difficult. Connect all the analog outputs from your D/A converters to your analog mixing device. If you're using relatively expensive interfaces, converters, and summing boxes, they may offer one or more DB25 connectors, which can carry as many as eight balanced analog channels over a single cable. Other interfaces and converters may require individual XLR or ¼-inch cables for each output or input. In those cases, you may want to use a cable snake. Also, if your converters have unbalanced outputs, you'll want to keep the cable runs as short as possible.
We Got Sum
Because you can sum signals using analog equipment in any price range, I wanted to test how analog summing would affect a mix with a reasonably priced system and with a more expensive system. I assembled an expert panel of four audio professionals to assist with the summing tests and to serve as the listening panel.
Producer and engineer Don Gunn, Jr., contributed two songs he recorded as well as contributed to the critical listening panel (see the sidebar “Sum Test Songs”). Recording and mastering engineer Barry Wood, proprietor of The Other Room mastering studio (www.otheroom.com) in San Clemente, California, meticulously calibrated our analog equipment to ensure the fairness of our listening experience (see the sidebar “Sum Calibration Information”). Producer and composer David Das and producer and engineer Jeff Haddad, who manages Beach City Studios (www.beachcitystudios.com) in San Clemente, California, rounded out our listening panel.
Each song was recorded in Apple Logic Pro 7 and then summed through three systems. The first system was the original digital mix, mixed and bounced entirely within Logic Pro.
FIG. 4: For occasions when you need additional analog outputs, the Behringer Ultragain Pro‑8 Digital ADA8000 is a very effective and affordable A/D/A -converter.
The second system was the budget system. I connected two Behringer Ultragain Pro-8 Digital ADA8000s ($289 each; see Fig. 4) to my RME Fireface 800's ADAT ports and clocked them both to the Fireface. I then took 16 balanced XLR outputs from the Behringer D/A converters into the ¼-inch TRS inputs of a Mackie Onyx 1620 mixer ($999). The total retail price of the Behringer and Mackie gear was just under $1,580, not counting the cost of cables.
Our third system was more expensive. We connected an Apogee DA-16x ($3,495) to the Fireface's ADAT ports. Then we used two DB25 cables to connect its 16 outputs to the 16 inputs of a Dangerous Music Dangerous 2-Bus LT ($1,499). The total retail price of the Apogee and Dangerous gear was $4,994.
The panel and I tried to be as rigorous, meticulous, and impartial in our listening tests as possible. To get the most out of each summing process, we remixed the levels of the stems for each mix. For both analog summing tests, we sent the main outputs of the summing mixer to two inputs on the Fireface to eliminate any A/D conversion differences on the way back into Logic Pro. To minimize any possible prejudice against a particular version, the file names were changed (with Barry Wood keeping the key), allowing us to listen to and comment on each mix without knowing how it was summed.
Although we took pains to be fair and unbiased, we understood that (as with any hot-button topic) those looking to find fault in our process would undoubtedly be able to find errors. There were enough variables in the process that we could control only so many, and we had to make choices about what our ultimate goals were.
Some might argue that our comparisons were less valid because we balanced the mixes to get the best out of the summing hardware. Still others might wish that we had tested each summing device through multiple converters, or that we had run a “control” test in to and out of the converters to identify how much of what we heard resulted from the conversion process. In addition, users of different DAWs might feel that our results don't apply to their DAW of choice. We understand and accept all such criticisms as valid; even the panel didn't always hear the same things in each version. Our goal was to conduct the most balanced listening tests of what we had and not to make absolute pronouncements.
After Sum Listening
Our listening panel did hear differences in the mixes. We were struck, however, by just how subtle those differences were. Far from being resounding, the differences between mixes were actually only matters of degree.
We all noticed a somewhat wider stereo image from the Mackie-summed mix of “A Boy.” Three of us also detected a bit more presence in that one, and Don noted that the low end was moderately tighter. Barry felt that the high mids were slightly more pronounced and heard the boost most noticeably in the tambourine. Jeff thought that the Logic Pro mix had a somewhat better bottom end than the other mixes. David heard a little more sparkle and high-end detail in the Dangerous 2-Bus LT mix.
The three versions of “Cricket Song” were so similar that they were almost too close for any of us to call. Don felt the Mackie-summed mix offered just a hair more low end. I heard a bit more low end in the Mackie and the Dangerous 2-Bus LT mixes, and the kick drum sounded slightly more defined. Jeff thought the Dangerous 2-Bus LT mix had slightly better high-end definition, a little more ambience, and a pleasing stereo field. David liked the stereo image and the sense of space around the guitars in the Dangerous 2-Bus LT mix and in the Logic Pro mix. I also liked the sense of space around the guitars in the Logic Pro mix, and I felt it had more energy.
To Sum or Not to Sum?
These listening tests demonstrated that busing tracks out from a DAW, summing them in the analog domain, and then busing the mix back into a DAW has a noticeable effect on the sound. The members of our panel were all satisfied that analog summing is not just a gimmick. But the question remains, is it necessary?
Clearly, there was no best or worst version of each song — no winners or losers, just slightly different colors. Ultimately, it comes down to personal preference. As our results show, one of us might have preferred the low end of the Dangerous 2-Bus LT mix, another the Mackie mix, and still another the digital mix.
If you're interested in how analog summing might affect your mixes, try running your own tests. If you have an analog mixer and a few converter boxes or the money to experiment with a dedicated summing box such as the Dangerous 2-Bus, mix with them for a while and see how those mixes compare to your digital mixes.
If our results prove anything, it's that even with less-expensive equipment such as Behringer converters and Mackie mixers, you can get pleasing results. The Apogee converters and Dangerous 2-Bus LT took far less effort to configure and calibrate, but we did find aspects of the budget mix that we liked as well. In the end, how well the mixing engineer knows the gear being used is the most important factor. You can create a great mix digitally, using budget mixers and converters or using expensive converters and summing boxes. It's all a matter of how you use them.
Orren Merton is a regular contributor to EM and the author of several pro-audio software books from Thomson Course Technology. He would like to thank Chanda Cook, Derrick Davis, Max Gutnik, Bob Muller, and Cathy Wagner for helping this article come together, as well as David Das, Don Gunn, Jeff Haddad, and Barry Wood for their assistance.
Apogee Electronics Corporation
Sum Test Songs
The song “A Boy,” by Randy Lee Fader and the Magnettes, was recorded in 2005 through Sytek and PreSonus mic preamps and an RME Fireface 800 into Logic Pro 7. The recording features one guitar, one bass, a minimally miked drum kit, three vocalists, and a tambourine. The band's Motown-inspired sound has plenty of air around the parts, which I felt would allow obvious differences in the summing to shine through. As a contrast, the song “Cricket Chant,” by the band Tarnish, was recorded back in 2001 using two MOTU 896s and Digital Performer 3.1. The dynamics and instrumentation were varied enough and the mix was dense enough to force us to really listen for subtle differences in the summing. The recording had more than 10 guitar tracks, 2 bass guitar parts, 14 tracks for the drum kit, and several layers of samples.
Sum Calibration Information
Accurately evaluating audible differences between the three signal paths in our test required that we have a level playing field. I therefore created an identical baseline level for the paths used in the experiment.
It was important to eliminate any discrepancies between the channel levels to ensure that the relative balance and panning of the mixes would not be altered. I achieved consistent levels by running test tones through each channel and measuring the resulting signals using Metric Halo's Spectrafoo metering software. The levels of each mix were matched as closely as possible to each other (again using Spectrafoo) as they were returned to Logic.
Calibrating the Behringer and Mackie signal chain required some care. I set the input trims by eye to their unity points. Then I adjusted the channel faders until they all were within a few hundredths of a decibel of each other. Additionally, because the Mackie did not have enough mono channels, we used its stereo channels. I had to address any left/right imbalances on the stereo channels by making slight adjustments to the pan knobs. (Mackie points out that the Onyx 1640 has 16 completely mono channels if you need them.)
The Dangerous 2-Bus comes aligned from the factory and has no facilities for user calibration. The Apogee DA-16x allowed for calibrating individual channels and calibrating the entire unit by increments as low as 0.01 dB, using front-panel buttons and menus. Luckily, though, I didn't need to calibrate the DA-16x or the Dangerous 2-Bus LT, because I found virtually no level differences from channel to channel.