• WANTED: Happy members who like to discuss audio and other topics related to our interest. Desire to learn and share knowledge of science required. There are many reviews of audio hardware and expert members to help answer your questions. Click here to have your audio equipment measured for free!

Why would I benefit from an amp when the headphones play loud enough without one? (or: why do headphones "scale well" with amps?)

Nantes

Member
Joined
Apr 8, 2019
Messages
5
Likes
3
I did a lot of research before getting my Massdrop Sennheiser HD 6XX (a much cheaper version of the HD 650 with a more convenient cable and jack). I was fully prepared to have to buy an amp for it, as that's what everyone online says it needs. However, it plays at a very good loudness with just 50% volume from my FiiO K1 DAC. Even my iPod is able to drive it loud enough (gasp!). I don't want to listen to it at deafening levels because, well, I don't want to go deaf.

Still, so many people claim it "scales very well" with amps, that it needs an amp to get the full benefit from it, it doesn't sound as good without an amp, etc. I'd like to understand why exactly that is, with a good scientific explanation. I thought the function of an amplifier was just that: to make the signal louder, and not to improve the potential of the headphones?

I'm also very keenly aware of bias effects when listening and how even a half-decibel volume difference during blind tests on otherwise identical gear will make people subconsciously believe the slightly louder song sounds better. Since I doubt those people did proper blind testing with their amp with perfect volume matching between amp and ampless (this is extremely hard to achieve anyway without a professional decibel meter), my current hypothesis is that those people are fooling themselves and what's happening is that they're playing the music with the amp slightly louder.

I'd love to be proven wrong though, as it means I can "unlock" my HD 6XX's, which currently don't sound much better than my DT 770 Pros. Does anyone know the science behind headphone "scaling"?

Some convenient info that I remember from my head:
FiiO K1 delivers 35 mW maximum
FiiO K1 max voltage is 1.1 V
HD 6XX impedance: 300 ohm
HD 6XX sensitivity: 103 dB (1 Vrms) at 1 mW or 97 dB (SPL) at 1 mW (decided to verify this one to be sure)

As per this calculator: https://www.headphonesty.com/headphone-power-calculator/ the FiiO K1 drives my phones at a reasonable loudness level (90 dB) rather than deafening levels (110 dB) which is for some reason recommended when calculating.
 
Last edited:

bravomail

Addicted to Fun and Learning
Joined
Oct 19, 2018
Messages
817
Likes
461
I would agree with your thinking in general. You need a good amp. But then Fiio K1 could be already strong enough amp. Since I don't own it, I cannot comment whether it is good or bad or weak. If you have spare money and your config is desktop like (not mobile) - then get some DAC/Amp stack. Schiit Modi3/magni3 or Khadas board with JDS Atom. If you are not willing to waste money - just enjoy and don't worry!
 

edechamps

Addicted to Fun and Learning
Forum Donor
Joined
Nov 21, 2018
Messages
910
Likes
3,620
Location
London, United Kingdom
I did a lot of research before getting my Massdrop Sennheiser HD 6XX (a much cheaper version of the HD 650 with a more convenient cable and jack). I was fully prepared to have to buy an amp for it, as that's what everyone online says it needs. However, it plays at a very good loudness with just 50% volume from my FiiO K1 DAC.

Well, the FiiO K1 is technically an headphone amplifier (just not a particularly beefy one), so there's no contradiction there.

I thought the function of an amplifier was just that: to make the signal louder, and not to improve the potential of the headphones?

To me, the most important reason to use an headphone amplifier is to have a low output impedance, which can be really important with lots of headphones. Most typical audio outputs in the wild are not good at driving headphones because they have a high output impedance. When combined with headphones with a low and/or varying impedance, the result is frequency response variations that can get pretty significant and can greatly compromise audio quality. So yes, in that sense, a good headphone amp can definitely "improve the potential of the headphones", in the sense that it will ensure you get the intended frequency response.

That said, there are also lots of badly designed headphone amps out there that have high (> 2 Ω) output impedance. IMHO these are pointless and I avoid them like the plague.

my current hypothesis is that those people are fooling themselves

Not necessarily. The difference between high output impedance and negligible output impedance can produce very audible differences in many scenarios. For an extreme example, we can look at the Ultimate Ears Triple-Fi 10 IEMs, which exhibit extreme low-impedance swings. If we do the math, we conclude that connecting these earphones to an audio output with a really bad 100 Ω output impedance would result in frequency response swings in the order of 13 dB, which is an extremely audible degradation.

That being said, when the impedance of the headphones is high, this is less of a concern. The Massdrop Sennheiser HD 6XX with its 300 Ω output impedance will do just fine with any output that has an output impedance of around 40 Ω or less.
 
Last edited:
OP
N

Nantes

Member
Joined
Apr 8, 2019
Messages
5
Likes
3
I would agree with your thinking in general. You need a good amp.

Wait, what? If you agree with my thinking then you agree that the differences in sound between amp/ampless seem to be mostly due to placebo effect, so I wouldn't need an amp...

Well, the FiiO K1 is technically an headphone amplifier (just not a particularly beefy one), so there's no contradiction there.

That's what is says in the Amazon page, but it seems to be just marketing, because there is not much difference in loudness between it and my laptop jack. Doesn't classify as an amp in my books, just a DAC.

To me, the most important reason to use an headphone amplifier is to have a low output impedance, which can be really important with lots of headphones.

Yup, I'd already read pretty much the entirety of that excellent blog (except for product reviews of course). The K1 has <1 ohm impedance so I'm totally fine there.
 
Last edited:

edechamps

Addicted to Fun and Learning
Forum Donor
Joined
Nov 21, 2018
Messages
910
Likes
3,620
Location
London, United Kingdom
That's what is says in the Amazon page, but it seems to be just marketing, because there is not much difference in loudness between it and my laptop jack. Doesn't classify as an amp in my books, just a DAC.

Well, I guess it's a matter of definitions. The reason why I consider the FiiO K1 to be an headphone amp is because it has a low output impedance and is capable of delivering a sufficient amount of voltage and current to drive most headphones that are not too demanding. This is not true for a DAC - a pure DAC would typically have a relatively high output impedance (because it's meant to drive a line input, not speakers/headphones directly). A pure DAC might also run into current (power) clipping when attempting to drive low-impedance headphones, something the FiiO K1 would have no problem with. So, while I can understand that the FiiO K1 might not "feel" like an headphone amp to you, it is definitely still better than a typical DAC when it comes to driving headphones.
 

restorer-john

Grand Contributor
Joined
Mar 1, 2018
Messages
12,588
Likes
38,288
Location
Gold Coast, Queensland, Australia
All things being equal, a low output impedance is better for headphones.

That said, there is far too much exaggeration of the dire consequences of higher output impedances. Headphones are benign and extremely easy to drive. Their power and control requirements are trivial.

Here's some actual impedance sweeps:

https://www.audiosciencereview.com/...dphone-amplifier-drive.2327/page-2#post-76501

Nothing to see here, you can drive all of those headphones with any source impedance and not see wild variations in FR. Compare to the speakers below and tell me which are actually 'demanding'...
 

edechamps

Addicted to Fun and Learning
Forum Donor
Joined
Nov 21, 2018
Messages
910
Likes
3,620
Location
London, United Kingdom
https://www.audiosciencereview.com/...dphone-amplifier-drive.2327/page-2#post-76501

Nothing to see here, you can drive all of those headphones with any source impedance and not see wild variations in FR. Compare to the speakers below and tell me which are actually 'demanding'...

That's because you've been lucky with your choice of headphones. Try for example the Etymotic ER4PT, Apple In-Ear, lots of Shure IEMs, other IEMs, but also full-size headphones such as the Sony MDR-1R, AKG K271 MkII, Shure SRH840. I could go on. Now look at these impedance vs. frequency response curves and tell me again that "you can drive all of those headphones with any source impedance".

Now, I'll admit that outside of IEMs, headphones that exhibit low and varying impedance are few and far between. However, when I buy an headphone amp, I don't want to be wondering whether I'm getting the correct response or not. I want an amp that will deliver the correct response with any headphone, period. (To me this is more important than getting enough level, because while the issue of not getting enough voltage is obvious, the issue of getting the wrong response is more subtle and harder to diagnose.)

(And yes, I do agree that loudspeakers are even worse in that regard. Yay for active speakers that take amplifier matching out of the equation.)
 

JohnYang1997

Master Contributor
Technical Expert
Audio Company
Joined
Dec 28, 2018
Messages
7,175
Likes
18,292
Location
China
Frist of all, simplest answer: loud =/= good.

Then the a little bit more than that: loud = voltage and most of the time is set by gain which causes illusion of high power but in reality don't. Rail voltage determines power output for high impedance headphones. Output current determines power output for low impedance headphones. Not enough of either will cause distortion for corresponding type of headphones.

Then the next level shit, a little bit of bro science: the control of amplifier is important, the counter current caused by coil inside the headphones will lead to distortion and not enough impact. Great amplifier will have very very low distortion across the frequency range with both inductive and capacitive load and real headphone load. Lowest output impedance helps too. No matter how people say they can't hear distortion below 0.01% Better is better. There is no excuse.
 

solderdude

Grand Contributor
Joined
Jul 21, 2018
Messages
15,891
Likes
35,912
Location
The Neitherlands
In your case of the HD6XX this might interest you.

Let's just say that the low impedance thing is not an issue for the HD650
hd650-new-pads-0-ohm-vs-120-ohm.png


HD650 via 0.1 Ohm via 120 Ohm so everything below 10 Ohm is moot.
 

RayDunzl

Grand Contributor
Central Scrutinizer
Joined
Mar 9, 2016
Messages
13,204
Likes
16,986
Location
Riverview FL
Even my iPod is able to drive it loud enough (gasp!).


A contrived experiment, to simulate an amplifier that plays "loudly enough" and one that has the power to deliver peak values.

Note that I'm not Golden Eared.

Using Benchmark DAC2 and HD650

In Audacity, take a track, amplify it by 6dB to clip it, copy that to a new track, attenuate it by 6dB. Switch between tracks while listening.

Peaks in the unmolested track require 4x the power (2x the voltage) of the clipped track.

Maybe 98% of the clipped track is not clipped - so, they both sound equally loud, and I have a hard time hearing a difference. Maybe you can. Give it a try. Even the spectrum is all but unchanged between the two.

So, for this example, a (pick a number) 1W amp plays just as perceptually loudly as a more powerful amp, and (maybe) without obvious problems, but a 4W amp is required to actually reproduce the waveform peaks at that loudness level.

1554787543117.png
 
OP
N

Nantes

Member
Joined
Apr 8, 2019
Messages
5
Likes
3
A contrived experiment...

First of all thanks for the experiment idea, it looks very interesting. Excuse my ignorance, but I haven't managed to get it working :(

Here's what I did:
1- Opened the FLAC of FM-84's "Never Stop", which seems to be a heavily compressed track. It even reaches 0 dB on Audacity's meter sometimes, which shows a red bar to mean that it's clipping already;
2- Created a new track, copied the original over, and amplified that by 6 dB (marking the checkbox to allow clipping) and then amplifying again by -6 dB to return to original loudness
3- Comparing the waveforms even at a very high magnification like in your picture, I cannot find any evidence of clipping whatsoever. Both tracks look identical.

What am I doing wrong?

Also: in case I can hear a difference when doing the correct test (I'll save the butchered track as a new file and ABX test it), what would that mean in practice? How do I know for sure which tracks will clip by virtue of my not having a sufficiently powerful amp? Doesn't that depend on what volume I'm playing the track at?
 

RayDunzl

Grand Contributor
Central Scrutinizer
Joined
Mar 9, 2016
Messages
13,204
Likes
16,986
Location
Riverview FL
What am I doing wrong?

Make sure you use 16 or 24 bits for the track, not 32 bit float.

Audacity lets 32bit clip visually on the +6dB step, but keeps the real values behind the scenes, and uses that saved data for the attenuation step, restoring the full unclipped waveform.

Example using 32 bit float (fail for this experiment):

1554851222394.png

Menu: Edit -> Preferences -> Quality -> Default Sample Format -> Set to 24, and repeat the experiment:

1554849915478.png





Also: in case I can hear a difference when doing the correct test (I'll save the butchered track as a new file and ABX test it), what would that mean in practice?

The experiment is to simulate an underpowered amplifier (the clipped track) vs an adequately to excessively powerful amp by showing (by listening) that the ability to play loudly (perceptually as loud as the more powerful amp at the same listening level) is not necessarily a true indication of the adequacy of an underpowered amp, especially when you cannot directly compare the two.

How do I know for sure which tracks will clip by virtue of my not having a sufficiently powerful amp?

If one does, then many will. Digital files have the same limit for maximum level, and that limit is regularly achieved via the modern recording practices.

Doesn't that depend on what volume I'm playing the track at?

Yes.

Lower volume = lower peak voltage demand = ability of a lower power amplifier to follow the signal peaks.
 
Last edited:
OP
N

Nantes

Member
Joined
Apr 8, 2019
Messages
5
Likes
3
Make sure you use 16 or 24 bits for the track, not 32 bit float.

Thanks, that worked (for future reference though, only after I closed and re-opened Audacity).

Voilà the test results:

clipped results 4.png


First two tracks were heavily compressed and were easy to tell apart by a distortion in the bass thump, but the third one (Milton Nascimento cover of a Beatles tune) played at around -12 to -9 dB and, after the amplification, only clipped in a louder part where it originally peaked at around -3 dB. After the butchering, that part peaked at -6 dB instead (hard cut), so I kept looping that part in the ABX test. Interestingly it does not sound any quieter to my ears than the normal version, so I didn't base myself on the perceived loudness; after several repetitions I noticed the clipped version makes a soft scratching noise at some specific points, and from there it was quite easy to tell apart.

So, as far as my limited brain understood from your excellent explanation, this clipping would only have a chance to be present if I listen to these tracks at or close to my K1's maximum volume output? If I listen to everything mostly at 50% volume it shouldn't ever clip and it would be no use to buy an amp?
 

RayDunzl

Grand Contributor
Central Scrutinizer
Joined
Mar 9, 2016
Messages
13,204
Likes
16,986
Location
Riverview FL
So, as far as my limited brain understood from your excellent explanation, this clipping would only have a chance to be present if I listen to these tracks at or close to my K1's maximum volume output?

No way to tell from here...

If what you listen to is loud enough and doesn't give the kind of sound problem you heard above, then maybe you have enough amp already.

Or run out and and buy one, they're cheap...
 

solderdude

Grand Contributor
Joined
Jul 21, 2018
Messages
15,891
Likes
35,912
Location
The Neitherlands
So, as far as my limited brain understood from your excellent explanation, this clipping would only have a chance to be present if I listen to these tracks at or close to my K1's maximum volume output? If I listen to everything mostly at 50% volume it shouldn't ever clip and it would be no use to buy an amp?

correct.
 

edechamps

Addicted to Fun and Learning
Forum Donor
Joined
Nov 21, 2018
Messages
910
Likes
3,620
Location
London, United Kingdom
So, as far as my limited brain understood from your excellent explanation, this clipping would only have a chance to be present if I listen to these tracks at or close to my K1's maximum volume output? If I listen to everything mostly at 50% volume it shouldn't ever clip and it would be no use to buy an amp?

The FiiO K1 will not clip into 300 Ω (nor any other realistic load impedance, for that matter). It is voltage-limited, meaning you won't be able to raise the volume high enough for it to start clipping.
 

solderdude

Grand Contributor
Joined
Jul 21, 2018
Messages
15,891
Likes
35,912
Location
The Neitherlands
max output voltage of the K1 is indeed limited to 1.15V and can't clip so in HD6XX = 4mW

Max SPL will be 104dB SPL so average levels of around 90dB are possible.
When it plays loud enough for you then you don't need an amp.
 
Top Bottom