I predict that in 5-15 years, the best new amplifier designs will look an awful lot more like Class D than Class A.
Class A was pushing the boundaries with high-efficiency speakers in many applications, but with newer low-efficiency models, it is just a non-starter. Class D designs become the logical answer.
First: when I say “Class D”, I’m really referring to the whole family of “digital” or “switching” amplifiers. There are a few common design variants. This also includes things like “Class T” (Tripath) amplifiers, and hybrid designs that use Class D to produce most of the power, and another amplifier type mixed in to improve sound quality.
Class A amps really can’t get too much better than they are. You can keep working on power filtering and hunting for components with slightly tighter tolerances, but these improvements are really really tiny compared to the current state of the art.
Look at the difference between a Class-D amp 20 years ago–crappy car stereos and boom boxes–and the stuff that Hypex, NAD, Bluesound, and Devialet are putting out based around Class-D today. They’ve moved the state of the art by leaps and bounds, but Class A/AB have seen comparatively minor progress. Little tweaks, a % here and there. That is just a sign that those technologies are mature.
There is big funding available for developing Class-D amps–because we need them everywhere. Car stereos, Bluetooth/wireless speakers, Soundbars, mass-market integrated speaker products from Bluesound and Sonos. As the technology matures, it is “trickling up” to higher quality applications. And this is possible because of the huge R+D dollars available to work on this stuff.
Class A/AB research happens in small HiFi companies with limited budgets and team size. The AV Receiver companies dabble in it sometimes, but it is not their focus. Mature companies tweak designs that they’ve been shipping for decades to make them a little bit better every few years. There isn’t a whole lot of large-scale R+D funding going into the old methods.
There’s another force pushing towards Class D–Hoffman’s Iron Laws say that you can only have two of: “small enclosure”, “efficient speaker”, “good bass extension”.
Aside from the set of people willing to put big stacks of HiFi kit in their homes, everyone else wants lots of compact speakers with good bass extension. That tells us that these speakers must be lower-efficiency designs.
So what you used to do with 20W, Class A, and a high efficiency speaker, is now something you might do with 50-60W, Class D, and a low efficiency speaker. Class A was pushing the boundaries with high-efficiency speakers in many applications, but with newer low-efficiency transducers needed to meet consumer demand for small, bass-heavy speakers, it is just a non-starter.
Another trend with smaller powered speakers is to use a digital crossover–which means you need 2-4 amplifiers per speaker instead of one. This places even more pressure on the amplifiers to be small, relatively cool, and energy efficient–more reason for class D.
And that need for more power is a real problem for technologies like Class A and AB–which consume too much power and radiate too much heat to be embedded in a small low-efficiency speaker.
All of this is forcing Class D technology to improve at a rapid pace as the market for these products is huge.
Then there is the energy efficiency legislation. This is continually making it harder to produce Class A designs. Class A stuff likes to be kept “warm”–so it doesn’t get along with the EU low power standby rules. And it consumes more power all the time, which makes it likely to see collateral damage as further environmental legislation plays out.
So, we’ve got politics, R+D dollars, market preferences for small, bass-heavy speakers, and concerns about energy efficiency. And they are all motivating Class D technology to become really mature and achieve very high quality.
So, when you see a company using modern, intensively engineered Class D stuff–like what Bluesound, NAD, Hypex, Devialet are doing–it’s worth giving it a listen before discounting it because it’s class D–and if you don’t like it, give it another shot in a few years–because this technology is going to continue to improve quickly. These people are taking the state of the art in a fast-growing little corner of the world and applying that research with quality as a top priority. They are making sure that the global move towards Class D style amplifiers does not leave audiophiles behind by actually working on making it practical for us on an ongoing basis.
This is already a long post, but I wanted to point out three other areas where the same kinds of trends are happening.
One is sigma-delta architecture chipdacs. They started as a way to reduce cost in CD players, but have grown into the dominant DAC technology. Today, the best chipdacs have technical specs that meet or exceed the best discrete DACs, at a lower cost, in a smaller package that emits less heat. ESS DACs go into ultra-high-volume applications like mobile phones and are used in reference grade studio applications, too. They have so many more R+D dollars to work with, and plenty of incentive to work on cost, energy, and quality in parallel–and they are doing a fantastic job.
Resistor Ladders are a lot like class A–small improvements from year to year coming out of small, modestly funded boutique manufacturers. Lots of focus on high-tolerance components and careful assembly–since that’s where the remaining minor improvement potential lies.
Another example is switching-mode power supplies. Outside of specialty applications, linear power supplies are nearly extinct–whereas 30-50 years ago they were all over the place. EU efficiency regulations are making it harder to manufacture and sell them–because like Class-A amps they are very inefficient and radiate a lot of heat.
R+D dollars are pouring into making Switching Supplies smaller, more efficient, and (electrically) cleaner. There’s not much of a reason to focus on a clean power feed for a mobile phone charger, but switching power supplies are going to displace linear power supplies in all of the specialty applications–scientific, medical, test equipment, audio–over the coming years, too. And this will force higher quality options to be developed and for their prices to drop.
Yet another example is DSP. At first, it was bad equalizers, crappy “bass boost” features and stuff like that, and it was associated with quality bottlenecks. But over the past decades, we’ve seen digital crossovers completely change the shape of most inexpensive speakers, and many expensive ones too (Meridian, Dynaudio, KEF, Steinway-Lyngdorf, …). And of course technologies like RAAT are popping up to bring the audio stream to the hardware DSP stage digitally to avoid the analog->digital->analog conversion common in some older architectures.
We think of Dirac as room correction suitable for our highest quality rooms, but a big part of their business is automotive DSP–because making high quality sound in a car is very difficult, but there is a huge market for it. The research that they are putting into automotive applications is going to make it possible to do great things in weirder rooms with smaller speakers too. Dolby ATMOS is also working on the “making lots of small speakers placed carelessly sound great in a weird space” problem. So we will see more flexibility and convenience at a very high quality level as these things mature.
In Closing
This is basically all good news for any audiophile willing to be open-minded to the new approaches. We’re going to get better sound quality in more rooms with less heat, cost, and power consumption, because the powers that be are pushing everything that way.
For an audiophile’s main listening room, I expect there will still be a lot of people who prefer a big stack of discrete components that work like a small space heater. Nothing wrong with that–those systems sound spectacular. I think they will become less popular over time, though.
The main reason I took the time to write this up is because I want to encourage open-mindedness and respect towards some of these newer methods. It’s easy to hear about an awful Class-D amp and write off the whole category, or upgrade to a linear power supply and write off switching supplies forever–but given the pace of change here, those are not necessarily good conclusions to hold for a long time.
There are companies putting all of these technologies into very high quality applications, and really, those are the companies that are making sure that we will have great options in the future.