Headphone Output Impedance and Mastering

[RedStone]

I've been researching buying a new interface since my old computer is aging and I've upgraded to a Macbook Pro M2.

I have an Apollo interface, which was a huge step up from what I had been using (Scarlett). For tracking and mixing, I'm Apollo all the way. I love those little tanks.

But for mastering, iI'm craving a more extreme level of linearity and exactness.

This has lead me down a rabbit hole of research, the results of which have surprised me (a lot!), Especially when it comes to headphone jack output impedance, dynamic range, and Total harmonic Distortion. This is where 95% of Prosumer audio interfaces seem to either lack or fail.

This YouTube channel by Julian Krauss has been revealing. Many of the interfaces I've wanted to buy because I thought they'd be great are no-gos now. Though, I'm looking for something pretty specific. This video in particular was really helpful (screenshot is below)

This is a readout of the headphone output impedance of a few common interfaces. What I've learned is that output impedance can be multiplied by 8 to roughly determine what headphone impedance will work well with the interface (it's not perfect, just a ballpark, but the closer to an exact match the better).

The closer the HP out is to zero ohm, as long as it has enough power, the more hifi headphones it is compatible with and can drive with a linear response instead of scooping the sound due to impedance mismatches.

A friend of mine loaned me a Steinberg UR22 Mk2 to use for recording on my Mac while I figured out what to do about getting a new audio interface. I'm thankful to have something, but man the UR22 mk2 is not great haha. The preamps are noisy, the channel balance is wonky on the outputs, and the HP output has an impedance of 91 ohm. That means the lowest impedance headphones that can it can supply with Flat frequency response and the tightest bass response is: 91ohm x 8 = 728 ohm.

Insane.

Now this doesn't mean that a 250 or 600-ohm pair of audiophile headphones won't sound good, it just means that the impedance mismatch will colour the audio signal. For mixing, it's not the worst if you're using 250 ohm headphones. But for mastering, I need the cleanest possible audio signal with no colouration, and inaudible levels of distortion across the frequency spectrum. As I've learned, this is why the likes of entry-level Lynx or Antelope interfaces start at over $2000 USD.

But I may have found a gem. The MOTU M2. It does have some harmonic distortion issues at certain volumes but overall doesn't seem to be all that noticeable. The frequency response is linear on all outputs, and the headphone out has an impedance of 0.2 ohms, near-perfect channel balance, excellent cross-talk, and transparently clean and quiet preamps. I've never used Motu Interfaces before, so I'm curious to know if anyone has experience with these, or what other interfaces people are using for mastering

 

[rob aylestone]

I suspect it’s because few people really master on headphones, so the prime use of headphones is monitoring while tracking. If you are really concerned about tiny details then your room and your monitors will make far more difference to what other people hear? I’m also not really convinced that the scoops you mention are significant. As all those headphones we use have very different sounds anyway, so if the interface changes with impedance how is this balanced against headphones that are far from flat? Because they’re more affordable than speakers, we often have multiple pairs of headphones, and accept the different sounds they have. With speakers we have a more common standard as we don’t have lots we work with often. Headphones and speakers all sound different, so doesn’t that make the impedance thing just more eq changes we have to live with?

 

[ecc83]

I have followed the vagaries of headphone amplifiers with regard to their output resistance* for a couple of years.

Historically HP amps had a physical 'build out' resistor in the OP circuit which performed three vital tasks.
1) It protected the amplifier from a short on its output (the amp of course had to be safe into just the resistor!).
2) It isolated the amplifier from load capacitance which can cause instability.
And 3) not vital but very useful, it helped to even the loudness into phones of a variety of nominal impedance.

Then came the "fashion" I shall call it, for HP amps to have an OP R of near zero Ohms or very much lower than the 33-100 Ohms that had been usual. Why? Well "better damping factor" was one cry...Hah!

"Damping Factor" is basically nonsense. What matters is the TOTAL resistance in the circuit and virtually all MC headphones have a DC resistance very close to their stated impedance. Check a few on your trusty DMM. So, it matters not very much if your HP amp OPR is 33R or zero, the current flow to effect damping is much the same with say 250R cans. Bit better for 32 R jobs but not much.

Then there is the fact that headphone manufacturers KNOW that most amplifiers have a significant output resistance and surely make their products accordingly? The number of near zero amplifiers is still quite a small percentage.

Lastly, to make an amplifier with an output of about a watt but with a very low internal resistance means you have to build in some other form of short protection. That is quite difficult to do without introducing glitching as the VI limiters kick in and all adds to the cost.

"Resistance not "impedance" because you don't want the output to be frequency dependent.

Dave.

 

[musicturtle]

I did this a while back because I got a set of Beyerdynamic 990s that are 250 ohm and settled on the UA Volt 276...here are there specs.

Maximum Output Power into 32 Ohms Maximum Output Power into 300 Ohms Maximum Output Power into 600 Ohms

84 mW per channel 22 mW per channel 12 mW per channel

 

[musicturtle]

After the fact, I bought a set of the Slate VSX headphones and I love those things for mixing and mastering.

 

[ecc83]

I did this a while back because I got a set of Beyerdynamic 990s that are 250 ohm and settled on the UA Volt 276...here are there specs.

Maximum Output Power into 32 Ohms Maximum Output Power into 300 Ohms Maximum Output Power into 600 Ohms

84 mW per channel 22 mW per channel 12 mW per channel

Right! Back of envelope calculation gives that interface an output resistance for the headphone source of 18 Ohms. Lower than most but still very much above the near zero that the 'followers of fashion' say is needed. Since the 250 Ohm 990 pros have a sensitivity of 96dB/mW you should have no trouble getting enough, clean level.

Dave.

 

[RedStone]

I suspect it’s because few people really master on headphones, so the prime use of headphones is monitoring while tracking. If you are really concerned about tiny details then your room and your monitors will make far more difference to what other people hear? I’m also not really convinced that the scoops you mention are significant. As all those headphones we use have very different sounds anyway, so if the interface changes with impedance how is this balanced against headphones that are far from flat? Because they’re more affordable than speakers, we often have multiple pairs of headphones, and accept the different sounds they have. With speakers we have a more common standard as we don’t have lots we work with often. Headphones and speakers all sound different, so doesn’t that make the impedance thing just more eq changes we have to live with?

yeah at this point for me it's more of a technical thing to remove possible issues. Output impedance can change the tone of a pair of headphones. It may be slight, but that slight difference makes a big difference later on as tonal decisions start piling up. And if those headphones are well-tuned out of the box, it's a real problem to have that tone get changed by the equipment. So, what I want in headphone amps is the ability to drive a wide variety of headphones transparently and with enough power to drive high-impedance headphones like HD650s or low-impedance ones like IEMs without perceptual change in frequency response.

 

[ecc83]

yeah at this point for me it's more of a technical thing to remove possible issues. Output impedance can change the tone of a pair of headphones. It may be slight, but that slight difference makes a big difference later on as tonal decisions start piling up. And if those headphones are well-tuned out of the box, it's a real problem to have that tone get changed by the equipment. So, what I want in headphone amps is the ability to drive a wide variety of headphones transparently and with enough power to drive high-impedance headphones like HD650s or low-impedance ones like IEMs without perceptual change in frequency response.

You are I fear on a hiding to nothing here. The frequency response of headphones will change slightly any way with 'ear fit'.

Before you rush out and buy one of the new fangled "zero impedance" cans amps it might be prudent to ask the headphone makers WHAT drive resistance* they recommend? They also might give you some data that shows that the response is largely unaffected by source resistance given that BY FAR the greatest DC resistance in a headphone circuit is almost always that of the headphones themselves. Those here with 'old knowledge' of tape head drive circuits will know what I mean by "constant current drive"!

But if you do go ahead and buy such a HP amplifier first get a cut in stone, Gold filled guarantee from the makers that the amp is TOTALLY short circuit proof.

*We are after all dealing with almost all resistance and no complex impedance.

Dave.













































/
?

 

[Papanate]

I've been researching buying a new interface since my old computer is aging and I've upgraded to a Macbook Pro M2.

I have an Apollo interface, which was a huge step up from what I had been using (Scarlett). For tracking and mixing, I'm Apollo all the way. I love those little tanks.

But for mastering, iI'm craving a more extreme level of linearity and exactness.

But I may have found a gem. The MOTU M2. It does have some harmonic distortion issues at certain volumes but overall doesn't seem to be all that noticeable. The frequency response is linear on all outputs, and the headphone out has an impedance of 0.2 ohms, near-perfect channel balance, excellent cross-talk, and transparently clean and quiet preamps. I've never used Motu Interfaces before, so I'm curious to know if anyone has experience with these, or what other interfaces people are using for mastering

If I understand you - you are going down a rabbit hole for very little gain - mastering is hard on Headphones - trying to find the right impedance is a needle in the haystack type of affair - Grado RSIx will work for you - they are very expensive too - or the Beyerdynamic DT 1990 - which are a little less expensive - anything less and you will just have to learn your headphones whatever you get.

 

[ecc83]

Headphones, no matter the quality, are not reproducing anything like the sound 'pattern' of live instruments. Those sounds exist in space and have place and direction. Even the very best monitor speakers do not quite reproduce the true 'experience' of a string quartet or come to that a rock band.

I suppose if all your music is generated electronically and never has physical form you can argue that this does not matter.

Headphones then to me at best are a 'tool' to be used when speakers cannot be. A live mic in the same room say and they are very useful for 'forensic' work exposing faults and noises you might miss on speakers. The latter have however become very much more analytical at the top end.

It is also hard to use headphones at a consistent loudness and 'mission creep' can happen and you end the session with partial deafness!

Dave.

 

[RedStone]

Headphones, no matter the quality, are not reproducing anything like the sound 'pattern' of live instruments. Those sounds exist in space and have place and direction. Even the very best monitor speakers do not quite reproduce the true 'experience' of a string quartet or come to that a rock band.

I suppose if all your music is generated electronically and never has physical form you can argue that this does not matter.

Headphones then to me at best are a 'tool' to be used when speakers cannot be. A live mic in the same room say and they are very useful for 'forensic' work exposing faults and noises you might miss on speakers. The latter have however become very much more analytical at the top end.

It is also hard to use headphones at a consistent loudness and 'mission creep' can happen and you end the session with partial deafness!

Dave.

Agreed, listening to headphones at loud levels is an ear-death wish.

I would differ on your conclusions about the fidelity of headphones as a general concept, though many headphones have been poorly designed and sound terrible, but have been well marketed and are therefore popular. The number of studio headphones I've tried that sound anything but hifi grows weekly haha! Beyerdynamic is one of the worst for how poorly they are tuned, even though their build quality is generally pretty great.

When deisgned and tuned properly, headphones can get us closely aligned to the sound of a neutral set of hifi speakers in a well-tuned room.

This is not just my opinion. There is some well-reasoned science to back this up ...

7Hz Salnotes Zero IEM Review

This is a review, listening tests and detailed measurements of the 7Hz Salnotes Zero IEM. It was kindly donated by a member and costs US $22 ($20 on Amazon now). The molded plastic looks cheep and the left and right markings barely visible. The included tip was extra soft/thin which may be a...

www.audiosciencereview.com

--
The link is to an unbiased, review and scientific analysis of a $20 IEM (which I now own) that sounds better than any "studio headphones" under $200 I've personally tried. They also blow the Sure 215 IEMs I used a few times clear out of the water - the Salnotes have great isolation and excellent tonality. Stage musicians and live sound engineers, take note

There are better sub-$100 IEMs than this (Zero:2 is one of them). But the zero, despite its too-good-to-be-true minimal cost, comes stupidly close to being tonally perfect out of the box. of course, it does have flaws ... the sound stage is narrow, and there are some issues in the upper treble. But the tonality of these headphones has given me a clear glimpse into audiophile sound - much clearer than sonar works did (I find their curve squashes the soundstage of my headphones). So, I skipped buying coffee for a couple of days to justify getting a pair of the Zeros. I was VERY skeptical that they would live up to what the review was saying.

But they did live up to the review, and from a pure tone perspective, I have used them to successfully create live mixes "in the box" that audiences are absolutely loving. So far, I'm 2 for 2 on doing a preliminary mix inside these headphones, then switching to the PA to deal with room resonance and calling the mix done.

 

[RedStone]

And an update on the interface. I went with an Apollo Twin X quad. Aside from being a bit annoyed that I am now locked into the UAD ecosystem forevermore, the fidelity of the mic preamps, headphone amp and line outs are all outstanding.

 

[rob aylestone]

I'm getting a bit lost here? I have several pairs of Shure 215s - original, some in custom moulds and the idea of mixing on them is crazy? I was using them on stage for years, and some were old enough for the sweat to get into the transparent jacket and show you the corroding braid! That said the reason I used them was because the bass was controlled. What I mean was that 'better' ones had more bass, extra transducers etc, but I struggled badly with tuning - especially on the bottom B string. I played one song in Eb, and that song was all bottom - nothing above a G. I played by muscle memory and did not notice I was perfectly out of tune. The 215's don't do the bottom at all, so I was really hearing the harmonics, not fundamentals. I couldn't mix on these, the same way I couldn't mix on my DT100 Beyers - the comments above are spot on - they're amazingly tough but nowhere near the sound of even modest real monitors. I used to have a pair of Sennheisers pre-2000, possibly 280's? and they were, from memory, much closer to the speakers I had back then - and easy to wear for hours, which DT100's are not.

 

[ecc83]

My point Redstone is that no matter how high quality* the electro-acoustic performance of headphones is, the sound they produce is still not "out there" as it is with loudspeakers and 'real' music.

There is a paradox here. If you have ever done any dummy head recording you will know that reproduced ON HEADPHONES the results are scarily real and positioning bang on through 360dgrs. Play the same track on speakers and the illusion falls apart. AFAIK the realism of dummy head recording cannot be translated to speaker reproduction (not with just two speakers at least).

The reverse holds true, recordings of real sounds in real rooms cannot be reproduced on cans with the same spacial effects. Yes, there ARE various clever cross-feed and delay systems around but all I think are bit of a kludge?

*Plus no two ears have the same 'pattern' and that will alter the response.

Dave.

 

[Steenamaroo]

Interesting info here.
I'll never knock anyone for doing some research and thinking something through,
and I have, on occasion, mixed on headphones where the source material is repetitive or very familiar,
but people often forget one fundamental issue and that's that your left ear hears the right speaker and vice versa.

This isn't the case with headphones.

I'm no mastering engineer and I'm always happy to be corrected, but I'd assume stereo image is a very important part of their job.

 

[RedStone]

it's very true that there's an acoustics component to headphones, much as there are with speakers. In the case of headphones, your inner ear canal is the "room" the speaker is playing into. This Means that you at least need to find ear gels that fit properly so that you don't wind up with anemic bass.

The drivers in any headphone can be tuned, and when drivers are tuned to a Harman curve (specifically, v2018), the result is a fidelity of sound that will surprise you immensely, as I was when I discovered it.

Headphones are a unique beast that way. They tend to have a narrow sound stage and up until not long ago, mostly bizarre frequency responses that are anything but neutral. Widening on-ears and in ears with a stereo width plugin can make them pleasingly spacious without sacrificing fidelity. This can help deal with the over-direct sound you get from most headphones.

People spend a lot of money to minimize the effects of room reflections, which is what makes your left ear hear the right speaker and so on.

To put it all together

• Impedance relates to power delivery (amplifiers) and power draw (headphones, speakers)
• High output impedance on an amplifier delivers less power
• Low output impedance on an amplifier delivers more power
• High-impedance headphones/speakers require less power - so can be paired with high-impedance amplifiers without much worry.
• Low-impedance headphones and low-impedance speakers require more power. Many headphones also have uneven power needs across the frequency spectrum, with higher power needs in the top end and lower power needed in the bottom end. The result of pairing such headphones with a high impedance amplifier that can't deliver enough power to drive the top end of the headphones at the same consistency of as the low end leads to an exaggerated low end.

Let's use the Beyerdyanmic DT990s. Their stated impedance is 250ohms, but sub 200hz, they go up to 350ohms. So their power draw is uneven and comes to a trickle in the low end due to the high impedance. If the headphone amp isn't able to serve enough power to drive these headphones at their lowest impedance point, you'll get an even more exxagerated low end starting in the low-mids and you'll need to correct for that even after using correction software like sound ID.

Here is more on that ...

 

[ecc83]

There are some assumptions and wrong thinking here.

First of all the output impedance of an amplifier can have nothing to do with whether it cab deliver "high current" or "high power"
The OP resistance of an NE5532 op amp follower AT LOW SIGNAL VOLTAGE will ne next to buggerall but it will struggle to deliver more than 20mA say at 10V rms.
A modest 20W valve amp might have an OPZ of 10 Ohms but could deliver over 2 amps into 4 Ohms.

So, the ability of an amplifier to drive current into a load is a function of the AMPLIFIER DESIGN. You can have weedy HP amps that struggle to drive 32 Ohm cans and high end class A jobs that can put out 5 watts!

Klaus mentions, albeit a bit under his breath, that the frequency response he demonstrates with a HZ amp and low Z cans is NOT the acoustical response of the headphones. If you were to measure the electrical response in a wire to a passive loudspeaker you would get curve all over the shop! Does not matter...WHAT matters is the response the speaker delivers into the room.

The downside of a fairly powerful HP amp with a sub 1R say OPZ is that high Z phones will be OK but low Z cans will blow your head off. Having a build out resistor of between 33 and 100R tends to even out the loudness. If you look at most headphone specifications they have a sensitivity broadly around 100dB/V and therefore lowZ cans will be much louder on a near zero OPZ amp than higher ones. Plus, if you SHORT a 'zero Z ' amp you could bugger it.

Lastly, does he not think Beyer, Sennheiser et al do not KNOW all this? They know the limitations amp drive imposes on their products and design them accordingly!

Dave.