Which headphones need an external amp?

YouTuber ClavinetJunkie did a vid (https://www.youtube.com/watch?v=K8zAko5G6eg) to demonstrate that an iPhone 6s amply drives an HD 600 – with no discernible sound quality loss compared to using that headphone with a dedicated dac/amp. His main thrust is that the iPhone has sufficient juice to power the 600s to a very reasonable loudness, including in the bass.

This is of course contrary to the conventional wisdom that a 300 Ω headphone needs dedicated amp power, almost regardless of its sensitivity.

The manufacturer specs the HD 600 as 300 Ω, 97 dB/1V sensitivity. The iPhone 6s apparently can do 0.969 Vrms out at 300 Ω. So it should hit approx 97 dB, which seems to support ClavinetJunkie’s thesis.

Intel released their HD Audio chipset and this was incorporated into many laptops, including MacBooks, starting in 2014. LG and some other makes of smartphones have undergone similar improvements. So perhaps the conventional wisdom as to where to draw the line re which headphones require an external amp pre-dates that shift?

Is it that a good external amp is needed for an appropriate level of sound quality even when not for loudness? I’d love to hear people’s experiences with this matter.

1 Like

The amp on the 6S is impressive if it can drive a HD600. Also Headphone impedance is not a good indicator of drivability. The K712 has an impedance of 62 ohm while the HD58X has an impedance of 150. And yet the 58X is much easier to drive. Lastly yes—amp power output is separate from other factors affecting sound quality.

1 Like

Thanks so much for replying!

When I read this I assumed the HD 58X must therefore have a higher sensitivity rating than the K712 and didn’t give it any more thought.

But just checked. The K712 has a sensitivity of 105 dB/V and the 58X weighs in at 104 dB/V – both per the official web sites. So according to that the K712s should be very nearly the same loudness as the 58X at a given “volume” dial setting. Very interesting…

I’m guessing the difference is the result of Sennheiser and AKG doing their measurements differently. I did another quick test and the 58X is definitely noticeably louder at set volume level. Your best option is to just look at reviewers say for this kind of thing.


I am still new to audiophile world, but I noticed a big difference when I went from 4x PCI audio card that came with my motherboard to an external dac + amp (aune X1S and X7S from massdrop) and I do not listen to my music at high volumes, but the sound felt more alive so to speak. Then I got balanced cable from Periapt for my hd58x…
Anyway, this pursuit of great audio is dangerously addicting once you get a first taste

1 Like

Adaril: thanks for replying! Your feedback is just the sort of thing I’m looking for. The degree to which there is a subjective improvement to be had beyond just achieving adequate loudness is exactly the sort of thing I’m trying to sort out. I’m not in a situation where I can easily try-before-buy or re-sell if disappointed, so experimentation is a real problem.

Most headphones have a sensitivity (dB at 1 mW, also written as dBm) that will allow you to drive them on nearly every source. That is to say they are efficient enough sound loud on most sources. This ability to drive an inefficient headphone is a different question from the quality of the sound and impedance.

The quality of the sound will change if the headphone is driven but underpowered. For example the Tin T2’s are fairly efficient IEMs and they sound ok over a phone, but they come alive in the details they can provide with a stronger and cleaner amp. If you feed them more power than a phone’s amp they sound better (I tested them on a FiiO mmcx balanced cable to a desktop amp). If you keep applying more power, then you will find out that the bass goes from a nice raised mid-bass to getting grainy as you begin to overwhelm them with power. This has to do with the designed dBm sensitivity of the Tin T2 and also the impedance… how it handles more power.

Where the 16 ohm Tin T2’s bass begins to fall apart, the 56 ohm Westone UM Pro 30’s (more sensitive at 124 dBm) sound better. They become fully detailed. The separation, imaging, and sound stage get better with more power than they would sound on an iPhone amp. The music comes alive.

This applies to most headphones that are generally described as ‘hard to drive.’ I’m sure there are exceptions where things driven on a weak amp still sound good despite the numbers, but I have not experienced this.

Edit: Even so, after thinking about it for a few minutes, the differences are probably like chasing the 5% or so of what you want to get out of the sound. So if an iPhone can drive a headphone well enough, then the differences found in a discrete amp are probably minimal and due to reducing noise.

1 Like

I think there are some misconceptions floating around when it comes to powering headphones. For example, just because you switched the same headphones to a more powerful amp doesn’t mean you’re actually “giving them more power” unless you’re also turning them louder than you were on the other amp. The only exception to this is if you were getting them loud enough on average (enough RMS power) on the other amp but not also during the loudness peaks (not enough peak power) on some highly dynamic music. That’s where a beefier amp can definitely make them sound better by actually feeding them more power (at the moments when it’s needed). But this wouldn’t be a night and day difference, so you could easily perform a test of this kind of case where there seems to be no audible difference (clavinetjunkie has done this kind of sloppy test before where he claimed Bluetooth audio via aptX was not audibly better than non-aptX, when really he was testing vs. AAC on an iPhone so of course the quality ended up very similar; he should’ve tested vs. SBC to get more interesting results and be able to make a stronger “aptX vs. the rest” claim).

But if we’re talking about the case where the same RMS and peak power was already being achieved by the weaker amp, the quality difference has to be down to the noise, distortion, impedance matching and what it does to the electrical damping, technicalities like that. Furthermore, THD+N is usually not constant across an amp’s volume adjustment range - there’s a sweet spot where the THD+N is lowest and where and amp is at its cleanest. Depending on how the amp’s power output matches a given headphone’s needs, you may be able to operate it within or outside that best-performance area, and this will determine the sound quality you’re getting out of it. (You can try to determine this before buying if you’re lucky enough to find a THD+N vs. Power graph for your prospective amp - like the ones Amir publishes on Audio Science Review - and relate it to the impedance and sensitivity of your headphones to see where you will end up on the curve.)


I agree with abm0, if RMS and peak are the same, then it is down to distortion characteristics at volume ranges, noise, and load loss.

These are usually minimal if the products are well engineered.

Also, one point of clarification from my post above, I was making things louder by adding more power to hear differences. I also think the files being played had more of an effect on my observations too. My iPhone just has stuff from iTunes and probably shouldn’t be compared against lossless files via Tidal.

Wow! Many thanks to JVerbit and ABM0. You’re just nailing addressing the information I was looking for, building on the previous responses.

I have a pair of DT 1990s. The 102 dBm (neat short notation, thanks!) together with the 250 Ohm impedance seem to put it right in the no-man’s land between easy and hard to drive. I listen to AACs and CD rips of EQ’d acoustic music from a MacBook Air through aged and somewhat hearing-damaged ears. I’m sure that sounds lo-fi, yet it sounds vivid, dynamic, detailed and realistic to me. And I hear no quality loss anywhere in the bass from an impedance spike (if there is one, can’t find that measurement online).

If there were anywhere near as dramatic a step up from that as is the step from my previous low-budget headphones to the 1990s, I’d be willing to sacrifice the simplicity and convenience of what I’m doing now by adding some external electronics. And hopefully this information will prove useful to those on a tight budget, who are trying to juggle how to divide limited funds.

So about that: because of the difference between RMS power and peak power, one of the big questions when deciding how much power you need is what kind of music you listen to and what your preferred loudness for it is. Modern (post-loudness-war) music tends to have a reduced distance from average to peak loudness (reduced dynamic range), something like 6-12 dB, while the craziest audiophile-oriented classical music can have as much as 30 dB from average to peak. So as NwAvGuy notes in his in-depth article about power requirements, if you’re someone who likes things loud (but still sane and non-ear-damaging) and you decide to listen at an average of 85 dB_SPL (the limit noise level for a daily 8-hour work schedule above which the employer is required to provide you with hearing protection), proper reproduction of the loudest peaks in the craziest classical music will require that your amp be able to power your headphones temporarily to 115 dB.

In the HD600 case, this means the iPhone 6 is clearly inadequate, as 115 dB is 18 dB above the 97 dB/V spec, with 18=3*6 meaning 3 doublings of the voltage are required to hit that, so a 2^3=8V output is necessary. But if you only listen to modern pop/dance/electronic/top100 music with about 12 dB of dynamic range, you will only need to hit 85+12=97 dB peaks, in which case you’re exactly covered by the 1V output. :slight_smile: Or if you do like to listen to classical but you tend to do that when it’s quiet around and you don’t actually crank the average level to 85 but maybe keep it at a reasonable 70 or even lower, you will only hit 100 dB peaks or lower, and again you will be covered even when listening to crazy-classical stuff. A lot here depends on what you listen to and how loud. (Another fact that Olaf sadly fails to mention in his rather superficial experiment.)


This reply is extremely relevant for me, and hopefully will help someone else. I did indeed read the NwAvGuy page (and many other of his pages multiple times). You’re addressing precisely what I couldn’t figure out for myself. He slants it that anyone listening to classical should at least assume 110 dB and better go with 115. Yet I know that’s ear-bleed territory for me. I rarely see a peak of 80 or 85 on my phone’s dB app. But I’ve always been insecure I wasn’t factoring something else in. (Headroom? Gain? Who knows what?)

NwAvGuy says my MacBook can hit 1.2 V and 45 mW at 32 Ω. Does it scale linearily to 1.2 V and 4.8 mW at 300 Ω? Dunno. When I make that assumption and do the math for my DT 1990s (meaning plug in numbers into some competent person’s spreadsheet), my MacBook can hit 109 dB SPL at the 1990’s 102 mW & 250 Ω. And even assuming a 500 Ω impedance spike in the bass, that’s still 106.5 SPL. Sounds like I should be fine with room to spare. All I really know is it sounds fabulous and I can’t crank to anywhere near max volume without stabbing my ear drums.

Also, of course, there’s the parallel question of the MacBook’s DAC. But since the music I listen to is rarely available in anything “better” than 16/44, I’m going to assume I’m getting at least a similarly competent rendition from the DAC. I can already hear and find a bit annoying the faint clack of a wind instrument’s metal keywork (never heard that live) and not just in the quietest passages. More detail than that I do not need.

Happy camper here – thanks all!

In my mind, every headphone “needs” a headphone amp. Bassicly every headphone improves with a external solution. If thats not in your budget then thats a different story,but you should still look into an amp in the future.
The power ratings dont matter that much, because basically every amp can run most headphones, there is only a hand full of headphones that need a lot of power.

Welcome aboard and thanks for weighing in!

Thanks! That’s just the excuse I need to break the piggy bank and buy a new toy. ;) Problem is there are alyways those pesky few nay-sayers who refuse to share in the sunny forecast. For example, YouTube reviewer Metal571 commenting on one of Oluv’s/ClavinetJunkie’s iPhone vs DAC/amp exposés:

“Exactly why I don’t review dacs and amps…this difference is so incredibly small compared to using different headphones that it just barely matters and definitely wouldn’t skew my own review of a headphone, at least in my case.” - https://www.youtube.com/watch?v=K8zAko5G6eg&lc=UgxSo80UG6EVfLQdy2F4AaABAg

Your welcome, I especially love class a amps, but even a dragonfly red sounds so much cleaner and has much more power then a phone. And you can get really great deals with used gear :wink:

is an external dac + amp better than a dedicated soundcard like Xonar or X-Fi ?

In general, yes. There may be exceptions among the really cheap external ones (especially the dongle-sized inline ones), but in general dedicated is better because it’s easier to space out the electrical traces and components properly and lower the noise and distortion in ways you can’t manage as easily on the busy PCB of a complex multi-function device like a computer or smartphone.

Classic Metal.