Universal Audio UAD2 Massive Passive

Universal Audio's latest plug‑in, based on Manley Labs' classic Massive Passive EQ, is being hailed as the best emulation yet of a hardware unit. So how would it fare when we pitted it against the real thing?

Universal Audio's programming team have recreated both the standard (above) and the mastering versions of the Massive Passive hardware. The latter has less gain range and slightly altered filter frequencies.

Six months of tireless programming in Universal Audio's labs has resulted in a plug‑in that UA are calling their "most ambitious and detailed EQ model to date”: a faithful software reproduction of Manley Labs' monstrous tube‑based EQ, the Massive Passive. The interaction between the passive filter circuitry, audio transformers and tube gain stages have all been emulated down to component level, with help from Manley, who provided all the components that are used in the hardware.

Like the real thing, the plug‑in comes in a standard and a mastering version, with slightly different gain ranges and filter frequencies. Both versions have four bands of fully parametric EQ with stepped filter frequencies and individually selectable filter shapes — bell‑shaped or shelving filter — and when set to shelving, the Bandwidth parameter controls the shape of the slope. The frequency ranges of the four bands are 22Hz to 1kHz, 82Hz to 3.9kHz, 220Hz to 10kHz, and 550Hz to 27kHz, resulting in a healthy overlap. There are also stepped high‑ and low‑pass filters and, unlike the hardware, the two channels are linkable for stereo operation.

The Standard version has 20dB of cut and boost available on each parametric band, whereas the Mastering version has a range of ±11dB, which makes it easier to fine‑tune and provides more than enough boost and attenuation for mastering duties. The filter section of the Mastering version has also been tuned for mastering, with low‑cut frequencies from 12Hz to 39Hz and high‑cut from 52kHz to 15kHz (both in six discrete steps); the filters of the Standard version run from 22Hz to 220Hz and 18kHz to 6kHz, respectively. True to the hardware unit, the Mastering version has stepped controls for output gain, EQ gain and bandwidth (these make settings repeatable on the analogue original, not that that is an issue with a plug‑in!).

To further add to the authenticity, the plug‑in affects the sound even when the EQ sections are disengaged, simulating the coloration made by just running audio through the transformers and valve circuitry. The plug‑in can, of course, be completely bypassed, using the Power switch.

In contrast to the non‑interactive filter design of most EQs, the filters of the Massive Passive operate in parallel, which makes them interact with one other. In practice, this means that the gain settings of the filters don't always reflect the actual gain made, especially at narrower bandwidth settings, and they don't 'add up' like they would have done in a conventional EQ. The filter interdependency is a big part of the unique sound of the Massive Passive, which makes it even more important to trust your ears rather than looking at the settings, when using it.

Massive Attack

With the help of a local mastering studio, I was able to go head‑to‑head with the hardware, comparing it with the plug‑in in a classic 'analogue versus digital' shootout. The analogue combatant was the standard version of the Massive Passive, so it was matched with the plug‑in counterpart, and both stereo signals were fed through high‑quality digital converters.

My previous encounters with the hardware have been brief but joyful, and I began by reacquainting myself with its magic. When I spent some time turning the knobs, sweeping the filters, and altering the Bandwidth and Gain settings, my spontaneous reaction was that it sounded remarkably similar to the plug‑in. Even the rather odd sound produced by the high shelving filter at steep slope settings was reproducible with the plug‑in.

After some fooling around, I started matching the settings of the two combatants, sending a mix through both of them and recording the output back into Pro Tools. The first filter setting was just a 22Hz low‑cut and a bell filter at 12kHz adding a healthy dose of treble. Both units produced an open and very audible treble without sounding harsh. The next test had the emphasis on the opposite side of the frequency spectrum, using a shelving filter to boost quite a lot from 100Hz and downwards. With the same settings, the plug‑in produced audibly more low end, which slightly shadowed the brightest transients of the mix. That said, the bass frequencies produced were impressively firm and thick in a way that few plug‑ins are able to manage. The hardware had a somewhat more coherent sound, but this was mostly because the filter effect wasn't as audible.

For the third test, the shelving filter was kept at 100Hz but now with a moderate boost, and a 16kHz shelving filter was added with the broadest Bandwidth setting, boosting a couple of decibels. The difference was now much more subtle, and the results showed the quality of the plug‑in in a very impressive way. I later played the two sound clips for a couple of colleagues, and they all picked the plug‑in as the better‑sounding because it was slightly more forward‑sounding.

Next up in the shootout was a stereo‑dubbed acoustic guitar, processed with a high shelving filter at 12kHz with the broadest Bandwidth setting and a bell filter at 390Hz, both adding quite a bit of gain. Listening to the result on PMC speakers in my own control room, I must confess that I sometimes could not tell them apart, because both clips sounded excellent. On the surgical level, there was a slight difference in the highest transient response, and the stereo width of the plug‑in version was a little narrower, but at the same time, the plug‑in showed more firmness in the lower-mid range. The same scenario happened when I listened to male vocals processed with quite some boost at 390Hz, and a bell filter at 16kHz adding a healthy dose of 'air' and clarity. Its firmness in the lower-mid range and slightly reduced harshness made the plug‑in the better sounding — at least most of the time, because sometimes I chose the 'wrong' clip when performing a blind test.

My final test was an already heavily processed drum loop, which was filtered with a low cut at 22Hz and a large boost with a bell filter at 3.3kHz. By now I had figured out that the settings of the hardware and the software weren't entirely matched sonically when set to the same settings — especially not in the mid‑range. Ageing tubes and small variations in production quality of the components might have been the reason. So I tried to match the hardware unit's mid‑range filter by widening the filter of the plug‑in and lowering the gain setting. Interestingly, this worked great for the first part of the drum loop, making the two sound clips almost identical, but when the chorus part of the loop started, the hardware seemed to have a snappier push in the mid‑range, while the plug‑in sounded a bit restrained in comparison. That said, the plug‑in still sounded great considering the amount of mid‑range gain that was being applied.

Massive Impressive

When comparing analogue hardware to the digital equivalent in the past, I've always been able to pinpoint the flaws of the latter. Most often it's the lack of depth, width and general authenticity that gives it away. This time it has been really hard to tell the hardware from the software, and I've already confessed to picking the 'wrong' sound clip from time to time.

Universal Audio have gone to great lengths to mimic the sound of the Massive Passive, and a big part of it is emulating the characteristics of the inductors, which not only can saturate, but also display hysteresis. Given its complex behaviour and non‑linearity, hysteresis cannot be simulated with dynamic convolution. In effect, hysteresis adds a mild broadband distortion, and combined with saturation of the transformers and distortion of the tube stages, means that the Massive Passive adds a certain 'warmth' which is, in my opinion, equally audible in the plug‑in.

The authentic sound comes at a price, though: one stereo plug‑in instance occupies 60 percent of the DSP power of a UAD2 Solo, and even a Quad card can only handle four instances. Raising the sample rate increases the DSP load by a few percent, but at 96kHz you can still load four plug‑ins on a Quad.

Considering the complexity of the hardware, Universal Audio have really nailed it with their Massive Passive plug‑in, which raises the bar of what can be achieved with digital processing.