Paying $1000 for a headphone amp to listen to $1000 headphones...why (from a technical perspective)?

I recently learned about this Sony product called a PHA-3 Headphone amp and I’m intrigued (from a strictly technical point of view, this is not IMHO) why some people would pay $1000 in order to listen to their iPhone (using presumably equally expensive headphones like the Shure SE 846 mentioned in the reviews).

What is it that this magic device does to the music that is so vastly superior to plugging headphones directly into an iPhone?

I understand that the digital signal is getting converted into sound by the chip in this machine rather than the chip in the iPhone, but what might be superior about this? What can this chip do to the signal that the iPhone chip can’t do and how might the result be objectively measured?

thanks!

The reason is that some people out there have a discretionary $2000 that Sony think would be better in their coffers.
The sound quality is unlikely to be objectively (or discernibly) better than equipment at a quarter the price but…hey, that’s never stopped the industry before.

OK, trying to avoid to much descent into golden eared lunacy.

First up, you really owe it to yourself to seek out and try a high end headphone setup. You will be gobsmacked. Not necessarily this setup, but any reasonably high end headphones, amp and source. What is referenced isn’t even all that expensive, but is probably well into the land of diminishing returns. Compared to a Sennheiser Orpheus, there is a way to go. And Sennheiser are a serious professional level engineering company.

If you use a USB digital input from your iPhone (which is now possible without much effort - in the past it was not), and you use lossless compression of the audio, you get bit perfect output of the audio, so as good as the original source material. The iPhone ceases to be anything that matters, it could just as well be any other digital source.

If you use the DAC in the iPhone you have a pretty low end DAC, it is built to a price, to a low operating voltage, and to a low power draw. All of these things result in compromises in quality. So to the same issues are found in the little amplifier that is sued to drive the headphones. And then you get to the environment that these chips operate in. It is electrically noisy, really noisy. There are RF signal and computer clocks and signals singing away, and this makes it hard for the audio components to work as well as they otherwise might.

So inside the DAC/Amp box you have a lot of freedom to get things much more right.

First up, you need to clean up the clock of the digital signal. This is one of the most critical things in digital audio, and for many years took on a level of deep mystique. Noise in the clock is known as jitter or phase noise. If you allow the clock that controls the output timing of the digitised signal into the analog domain vary in time, it changes the signal, and not in a happy manner. If the variations were just random noise (AIWN - additive independent white noise) or other purely stochastic noise it wouldn’t matter so much, but the machinations of the digital systems conspire to add all sorts of correlated noise into the clock, and these can and will cause weird distortion artefacts. Not the usual harmonic distortion one is used to in analog systems, but things like aliases of the output frequencies at sum and differences relative to the original. These are not harmonically related to the original, and thus the ear’s masking functions do not remove them from perception nearly as well, and they are peculiarly objectionable.

So the internals of the exteral box can contain a much quieter electrical environment, and will use specialised additional circuitry to attenuate the phase noise in the input clocks. Next there will be a much higher specification DAC chip. It can be fed with much higher voltages and be allosed to draw much higher power levels, and fed very clear stable power, power that is usually very carefully managed with extremely careful routing and decoupling so as to avoid coupling of switching noise into parts of the system where it isn’t wanted. The analog parts of the DAC system can be much higher specification, and use a range of techniques. Critically there is usually a current to voltage converter - DACs usually output current, and this need to be converted to a voltage, at the same time applying filtering to remove alias frequencies of the audio. This isn’t easy to get really right, and usually benefits from high specification op-amps (or when you want to buy an argument, custom discrete components.) Again, careful power supply design and circuit board layout is critical, as you are still working in the RF domain. Indeed IMHO and experience, circuit layout is more critical than the use of supposedly high specification components. Even a seemingly trivial mistake can couple significant amounts of logic switching energy into the wrong path.

Then you get to the amplifier. IMHO there is a lot of snake oil here, as building an essentially perfect amplifier for purpose isn’t difficult anymore. But there are lots of amplifier designs that are tweaked to a particular sound, and some can sound very good with particular headphones. More a musical tuning than HiFi, but can sound lovely. But good headphones do need a reasonable amount of power to drive properly, and your iPhone isn’t going to manage it. Not on the very low power rials and power budget available. A good amplifier design does warrant a carefully designed power supply, with very low noise, and given the circumstances of its use, getting it down to distortion levels that are essentially negligible at any frequency and power level not unreasonable. That requires power levels and voltages not available in any phone.

$1000 for a DAC/amplifier isn’t all that big a number. One of the well known such devices is the Benchmark DAC2. It costs double. There is no snake oil. Just very careful engineering. However, don’t worry, if you want snakeoil, there is plenty out there to empty your wallet.
Personally, I have a pair of Sennheiser HD650, an Apogee Duet I use as a DAC, and a PPA amplifier I built.

this thing is also a D/A converter which can take digital input via optical (PCM/DSD) and connect to a PC/Mac as a USB sound card.

headphone amps are useful if you have low-sensitivity headphones and need more voltage output to reach desired listening volume, or if you have low impedance headphones that your device is having trouble driving (headphone impedance is too close to the device’s output impedance.)

Still wouldn’t pay $1,000 for one.

Francis Vaughan: Thank you! That was a wonderfully detailed yet approachable answer. Its fascinating that there are so many “flaws” (shortcomings) in the iPhone audio circuitry that are almost never discussed outside of audiophile circles. I suppose the exact same could be said of the camera/light-capturing chips in the phone. They’re fine, but absolute shit compared to the state of the art or dozens of stand-alone devices.

This audiophilia is a fascinating subject, I think I’d like to learn more.

I’m going to take a second to caution you to be careful here, lest you be led down a rabbit hole of unwise purchasing decisions. While all of the phenomena Francis Vaughan speaks of do exist, that doesn’t mean they actually matter in practice. Audiophilia is rife with hucksters hoping to sell you junk to “improve” your sound system, and will do anything to convince you that you can hear the difference between a rat fart and a mouse fart from a mile away, if you just buy their $295 wood disc or cable elevators.

as far as the iPhone goes, it has one of the best analog audio outputs for a consumer device. the early (and I mean early) iPods weren’t that great, but Apple did their homework on later devices.

http://www.kenrockwell.com/apple/iphone-5/audio-quality.htm

practically ruler-flat across the audio spectrum and able to drive low-impedance phones with minimal sag.

I’m most certainly not advocating mpingo hockey pucks.

Flat frequency response into a simple resistive load is trivial, and any device that can’t do it isn’t just poor, but by modern standards, broken. It says absolutely nothing about distortion, or an ability to drive a reactive load.

A read of some of the technical work from Benchmark or Lavry might put some in a better understanding of the reality of proper engineering. Both of these companies mostly target their products at professional audio, not home audio where snake oil is a problem. Pro audio cares about sound, and they also care about money. Most idiot audiophiles get really upset when they are presented with pro audio products, as they don’t employ snakeoil design, but are simply properly engineered. But don’t kid yourselves. Proper engineering takes talent and usually costs.

Some iPods used a Wolfson DAC, and that was pretty good for its constraints. But no matter how much you do, running a DAC and amplifier off 3.3 volts is never going to get you to the levels a proper system can achieve. The Benchmark DAC2 uses the Sabre 9018S DAC, which is about as good as it gets. It uses 4. They cost $50 each chip.

had you read my link, you would have seen that he gets into the iPhone’s frequency response and distortion driving actual headphones. I work in audio, I don’t need lecturing.

jz78817: Thanks for sharing some of your considerable expertise with us. Reading through your link the expert listener writes that “Wasting time with “Audiophile” DACs and other fluff usually degrades the sound more than just using the self-powered iPhone 5 directly as a source.”

Yet there are some seemingly objective (the reviewers have histories on amazon to indicate they may be real people not paid by anyone) reviews on Amazon where passionate people report substantial noticeable improvements. Do you think what they’re hearing is entirely psychosomatic?

“I just received the PHA-3 and compared to my previous PHA-2 the difference are night and day. Using Shure SE846 in balanced mode. Instruments are better spaced, more defined,. Tighter base, wider soundstage. I’m not sure if the difference can be ascribed to the balanced vs. single ended or just the different dac chips, but I’m suitably impressed.”

“Wow, the amazing soundstage, sampling range enhancement (with DSEE HX on), and the richness of details PHA3 adds to my music, which are mostly flac and wave with some hi-res, are simply engrossing. PHA3 should be able to drive any power hungry headphones (with GAIN on), I never need to turn the GAIN on even for my power hungry cans. When connecting to the extremely clean Fiio X5 line out, the PHA3 produces amazing sound on all my phones. I gave my audiophile friend a taste of PHA3 (who is using Cowon Pleune 1 and UE custom pro 18) and he liked it so much that he bought one himself within days.”

Are these people delusional?

I can’t tell anyone what they’re hearing, certainly. I tend to be skeptical of such claims for several reasons:

  • Humans are extremely susceptible to the “placebo effect.” expecting an improvement often leads one to perceive an improvement.
  • of all of the countless audio tweaks, gizmos and doodads out there, not one seems to degrade the sound at all. Every single one of them is an improvement, which lends support to placebo.
  • most of the perceived improvement is described with meaningless fluff words like “detail,” “definition,” “dynamics,” or the stuff you saw on amazon here:

this is all fluff.

it would be unfair to claim so. I just won’t take stuff like that at face value.

I helped a client set up his Mac Mini with a Burson high-resolution USB preamp.

Even through crappy headphones, and listening to mp3s, I could tell the difference between the stock Mac output and the Burson.

Was it worth the money?
Not to me - but clearly it was to him.

Unless it was properly blinded you can’t say that with any degree of certainty.

And even if you could tell a difference it is another thing entirely to say that one was *better *than the other. You might even be shocked and find that you actually prefer the lower-end set-up.

If by “early iPods” you mean the first several generations that used the little miniature hard drives, I agree. Boy, do I agree! I had one of the newer generation of those and once did a direct A/B comparison using a decent pair of Grado headphones between this expensive iPod and a dirt-cheap no-name solid-state MP3 player. The MP3 player was hands down superior in clarity and flat response at both the low and the high end. The iPod sounded more like an AM radio in comparison. That, and the unreliability of the bizarre super-mini hard drives (the unit was replaced three times and then failed again!) totally put me off Apple after all the hype I’d been hearing about the iPods. If they learned their lesson, AFAIC it’s too late. Things like that stay with you. Guess who’s not getting an iPhone 6? Or an Apple Watch? :smiley:

Not true.
The difference was very noticeable. As part of the configuration, I had to test the amp and the Mac Mini output. I could easily tell the difference.
You don’t need a double-blind test to tell red from green.

It makes me wonder what the set of ears listening to these headphones is capable of discerning? The older you get, the less you can hear.

I struggle to hear much higher the 12KHz. Teenagers seem to be able to hear up to 20KHz.

I guess no fancy headphones are going to cure that.

All the more reason not to suffer distortions due to lossy compression like MP3 and stick to pure lossless formats?

I appreciate some of the other features of headphones that are an advantage to listening. Background noise cancellation makes a big difference, if it is done well.

When you knew there was difference you could hear a difference. That is a world away from there actually being a difference that you could pick out as “better”.

yes, you really do. Just witness the recent kerfuffle over the blue/white/black/gold dress.

That’s not at all the same thing.
The Blue/black dress is a perceptual/encoding (camera) issue. It’s not a case of someone saying “I like the blue /white dress better than the black/gold dress”
Anyway, I know what I heard, and I’ve listened to a lot of audio systems. There are some differences that are obvious.
Unless you think that you need a double-blind test to tell if it’s raining or sunny.

Audiophiles are like any other person infatuated with a hobby. You can enjoy it or you can really enjoy it. It’s an outlet, like any other, to obsess and tweak out. If not race others to have the most lavish system.

And, like any other hobby, most is snake oil. A good parallel is with cars and people stating that a particular exhaust system or intake will have significant gains, when, rarely do you see a dyno. And, when you do, it’s a rather dismal, incredibly inefficient bang-for-the-buck improvement. Audio is no different.

I believe a lot of this is rooted in superstition and some truth to old school equipment and technology. Like, with cars, back in the day, when they were turning out cars with huge displacement engines, they’d have these pitiful exhaust pipes and so-so carb/intake setups. So, yeah, installing a real exhaust system and intake manifold and subsequent carb had noticeable improvements. But, not anymore with the desire to have maximum fuel efficiency usually means getting every last drop of HP out of the engine.

With audio, that original stuff was just so basic in design. It really wasn’t on purpose as it took awhile before CAD programs for electrical/electronic engineering came into place. When they did first appear they were a little better than basic and involved a decent computer that few had. Now, pSPICE modeling is available to the masses, CAD programs for making PCB’s is also available to the masses. With an enormous knowledge/user base that is available through the internet, etc. So, the gap between mid and high level has closed and as you spend more and more its an exponential diminishing return.

Another thing that a lot of people neglect is the source material. Even the old LP’s, not everyone was at the master quality and the technology of creating the LPs and the material of the LP’s was inferior today. With lossless encoding of original source material, and digital transport until you get to the DAC, you can get some good stuff.

two different people looking at exactly the same picture reported the colours as wildly different. It is absolutely relevant.

Obvious enough to be measurable? Obvious enough to pick out of a line up of multiple set-ups and evaluated as your preferred one? If so, great but…

Experienced people fool themselves all the time. People talk themselves into perceiving the difference they are expecting. People still think that high-end cables make a difference that a human can perceive.
Double-blind testing is what we do in order to take that bias out of the process. When you can consistently pick out the “better” system in a properly blinded test then you can consider yourself vindicated. Until then yours is an anecdote and no more.

Why wouldn’t an expert test their abilities in this way? If they passed such a test then their reputation would be massively marketable. The audio world is notorious for steering clear of such testing methodologies and one can’t help but wonder why.

I’m completely in favor of double-blind testing.
The same way I’m in favor of taste-testing.

Sometimes, it’s not necessary to do a taste test to tell if your toast is burned.