Reflections after Munich | Stereophile.com
A great hi-fi show is a great place to see and audition new hi-fi equipment. Of at least equal importance to me, it’s also a good opportunity to talk about hi-fi with personalities in the field.
Early Thursday afternoon, the first day of the show, I visited an off-site “service apartment” temporarily occupied by people from Purifi Audio, the company best known for its advanced low-distortion class D “Eigentakt”. Purifi is also developing loudspeakers, which they demonstrated in Munich.
Amplification modules and loudspeakers may seem like an odd combination, but both are basic elements of a hi-fi system and both are major sources of distortion. As such, both are obvious areas of interest for people looking to improve hi-fi.
But haven’t people been working to improve loudspeakers and amplifiers for over 100 years? Surely all the fruits at hand have already been harvested, right? I guess it depends on your size, metaphorically speaking.
Bruno Putzeys, who co-founded Purifi with Peter Lyngdorf and Lars Risbo (the latter also present, as did co-owner/director Claus Neesgaard), told me that when they started Purifi, the main idea was to correct hi-fi problems with DSP—digital signal processing. They quickly changed course and started using science and engineering to tackle some of these problems at the source. The result is the Eigentakt module used in some of today’s best Class-D amplifiers. Another is their speakers, seven of which are already on the market, with more in development. Purifi’s drivers are backed by extensive performance data, far more than that provided by most other driver manufacturers.
Many advancements in loudspeakers involve new, sophisticated and expensive materials, all of which make the cone stiffer and lighter. Purifi’s new speakers use old-school cone materials: aluminum and paper. Instead of actual trial and error, Purifi solves many problems silicone before building the first prototype. So begins a cycle of measuring, listening and refining, but fewer iterations are now required as their simulation work brings them closer to the optimal solution. Such an approach has led to changes in the shape of the cones used in Purifi’s loudspeakers which see subtle but have significant sonic consequences.
Less attention was given to the environment, those “soft rubber bits” as one of the members of Purifi – I think it was Putzeys – called them during our meeting. Because they sit on the outer edge of the speaker’s radiating surface and because the area is proportional to the square of the radius, the surround can make up almost 20% of a speaker’s moving surface. You can’t ignore this if you want low distortion.
There’s nothing special about Purifi’s environment material either; it’s just rubber, but because of the way they’re molded, Purifi’s contours are less flexible than most. I’m running out of space to detail the technology (footnote 1), and I’d probably be wrong if I tried – so I’ll just point out that while Purifi’s cones look ordinary, their computer-optimized environments look pretty funny. Not that you would care after hearing them.
Purifi’s driver work isn’t limited to the parts you see. Innovations in the motor system reduce hysteresis distortion and address several other important issues.
But it’s not their specific innovations that I want to draw attention to. These are rather some big ideas that the Purifi approach illustrates:
• Simple, smart thinking combined with simulation can still, in 2022, achieve major performance gains in even the most basic hi-fi technologies.
• Easily audible improvement can result from reduced distortion that is already at levels far below what you (or at least I) imagine audible, often being well below signal level and deep in the noise. Under certain circumstances, reducing distortion from –80 dB to –100 dB (relative to signal level) can produce audible gains. And it’s not a maker of intuitive, artistic and poetic tube amps; it’s Purifi, the home of serious engineering and class-D modules that solve the problems of this technology by flooding them with feedback.
• Speakers that sound (and measure) well with single tones, such as, say, a frequency sweep, can have serious problems with more complex signals. Like, say, music. Intermodulation, via various mechanisms, is where things get interesting (footnote 2). What happens in the bass can have profound effects on the mids. But these problems can be solved, and some of them already have been.
The folks at Purifi are serious engineers, but don’t make the mistake of thinking they’re mindlessly siding with the simple-minded objectivists who populate some online chat rooms. They understand that science and engineering must be deployed with care. The members of the Purifi team seemed to me as impatient with the measurement-satisfied reductionists as they were with the radical anti-measurement crowd. Of course, it’s depressing when a forum post asks, “Is it possible for a component that measures well to sound good?” But owning an Audio Precision analyzer doesn’t make you an expert, and the unguided application of analytical tools can lead you astray.
Knowledge What measuring is key, and getting to that point is hard work. As it is written in a Purifi blog post – it is unsigned, but the voice and some biographical details resemble those of Putzeys – “We don’t hear the levels of distortion so much as the mechanics of distortion. You have to understand the mechanism before we can design a test that will reasonably quantify it.”
Footnote 1: If interested, see patents.google.com/patent/WO2020208070A1.
Footnote 2: See, for example, this Purifi blog entry. If you’re a tech-conscious audiophile, I encourage you to keep an eye on the Purifi blog, at purifi-audio.com/blog/. New posts aren’t frequent, but they’re meaty.