I love listening to music on speakers. There’s something visceral about feeling the sound fill the space I’m in. It also presents music in a more forgiving way. The detail is still presented, but by not feeling so close, I find problems are less attention pulling. When I was a child/teen, I would listen to music on my handed-down HiFi setup regularly. Even then, when it came to sheer listening hours, headphones reigned supreme. Portability and convenience are strong draws. As an adult, even stronger than these, the need to be bearable to live with has led to a further increase in percentage of time spent with headphones on.
For many of us, headphones have been the primary listening method for most of our lives. They have kept us sane in noisy shared houses; brightened dark, miserable morning commutes; and, as much of a proponent for speakers as I am, provided the best quality listening experience that can be found when in small and changing locations.
I think it is my combination of love of listening to music and interest in music production that have led me to spend, probably far too much, time thinking about how my listening setups should sound. The internet is not devoid of blame for this slight obsession. There are countless message board forums and YouTube comment sections where self-professed audiophiles will fight to the hilt over the differences made by tiny nuances of audio reproduction.
What are we looking for from listening on a pair of headphones? Do we want perfect recreation of the audio file being played, or do we want the speaker listening experience replicated?
One goal I have seen from several online aiming for is a feeling of realism from acoustic recordings, aiming to feel the artists are really there in front of you performing. Where you can close your eyes and imagine the musicians spread out on a stage, feeling as though you could open your eyes and find them to have been mystically transported to be there in the room. I find this kind of thinking aimless, and not just because the majority of the music I actually want to listen to does not fit in the box of super-clean recordings of acoustic performances. Consider this: which recording do we try and match our systems to? What if we get one song’s violins sounding perfectly like a violin does in real life, then another song’s acoustic guitar isn’t quite right? What about electronic music, will that sound ‘right’ on this setup? I’ll note here that I unfortunately have neither good enough sound memory, nor enough experience with violinists to trust I would recognise when it sounds realistic anyway.
In the production space, flat frequency responses are often focused on. This can at least be empirically measured. It does, however, only deal with one aspect of sound. You’ve still got to worry about distortion, imaging, soundstage, and the magical je ne c’est quoi that only a small mortgage and a good salesman can provide.
My taste in headphone sound has largely been defined by my set of Shure SRH840s. The ‘s’ isn’t part of the product code, but it feels weird to refer to a set of headphones as singular. I’ve had these headphones for over a decade, and though the ‘leather’ on the headband and the first set of pads is slowly flaking away, I still use them weekly for recording the podcast. These act as my personal ‘flat frequency response’. Music on these headphones just sounds ‘normal’ to me. It turns out, according to graphs I’ve seen online, they are relatively flat for a set of closed back headphones.
For the past year I’ve mostly switched to a set of Beyerdynamic DT 900 Pro Xs (again, to ignore the ‘s’ is your prerogative). This was partly due to the aforementioned slow disintegration of my beloved Shure pair. But I was also interested to try a set of open back headphones. I had heard people online raving about wider soundstage and clearer high frequencies. Plus, as much as I love the 840s, they do have a few shortcomings. They less cup my head and more linger somewhere close to the edges of it with a lack of clamping force and thin, awkwardly shaped pads.
The 900 Pro Xs needed a bit of taming. There’s a noticeable boost in the high end that makes voices and cymbals pop out more than I want and give them a (to me) unnatural sound. It’s not horrendous, but I might have returned them if I didn’t have the option of calibration.
I’ve been using Sonarworks SoundID with them which provides frequency response correction, adjusting the computer’s output to offset problem areas in the headphone’s reproduction. It’s done the job. The Beyerdynamics sound close to my 840s frequency-wise once calibrated, but the soundstage is much wider, and the comfort makes wearing them for long periods much preferred.
This set up has been consistent until last week. Beyerdynamic released a plugin called HEADPHONE LAB. Designed for producers, it provides two functions: frequency response correction, and speaker emulation. The former can replace SoundID for me. I haven’t done back-to-back comparisons, but I have no complaints listening to albums with Beyerdynamic’s plugin on, and don’t feel much requirement to go any deeper than that. One benefit is that for certain models, mine included, Beyerdynamic have used the measurements they take when the headphones are made in the factory to adjust for the profile of this specific set. The difference from the generic adjustment is subtle, but it is an improvement to my ears.
The other side of the plugin is the one I’m still evaluating. Loudspeaker emulation is designed to make listening on headphones simulate listening to speakers in a studio environment. It is probably worth mentioning now that this feature doesn’t require a pair of Beyerdynamic headphones to use. SoundID has a feature like this, but it’s marketed as a way to check how your mixes will translate in different locations, whereas Beyerdynamic present it as a full alternative to listening to the standard output.
The concept of a whole new way of listening to music on headphones raises a few interesting possibilities. The most obvious being, if it works, it could improve how enjoyable listening to music on headphones is. For producing music in small spaces, it could provide a better way of hearing what you’re creating. And, in a best-case scenario, it could for many instances, render speakers unnecessary.
So, what does it actually do? Beyerdynamic don’t provide much detail. What I can hear is: there’s some short reverb being added to the signal and there’s some amount of crosstalk (an amount of one side’s signal being sent to the other). Both parameters are adjustable in the plugin. There’s a dry/wet dial for the room noise and three widths of speaker placement that seems to change the crosstalk. I’ve got the dry/wet set around 10 o’clock and the width set to 60 degrees (the middle setting).
I’m incredibly sceptical of this kind of feature. I generally assume the fewer steps between the audio and the output device, the better. Plus, the audio industry has so many snake oil products, it’s fiscally prudent to assume everyone’s trying to rip you off. They usually boil down to washy reverb and EQ that makes it sound like its coming out of an FM radio’s built in speaker. You lose all the detail, and, moreover, it sounds awful. Beyerdynamic’s emulation feels different to me. It actually does a decent job of approximating the feeling of listening on speakers. You don’t get the feeling through your whole body, but the thump of a kick drum coming down the centre of the image gets much closer. It gives the impression of the centre image sounding more like it’s coming from in front of you than from the sides. Imaging, particularly on hard panned parts, feel more natural. I think this has made listening less fatiguing over long sessions. The sound does get a bit brighter, but not as an unpleasant peak like the default Beyerdynamic sound, it feels like a slight lift that still leaves the other frequencies clear underneath it.
I need more time to settle in with this new setup, then to compare to without it, but I do have a concern that I am losing some detail with the loudspeaker emulation on. I don’t think small details are made as immediately apparent. To be fair, that would also be the case when listening on a set of speakers. That might also not be the worst thing. While tiny hidden details can be useful to have something to talk about on the podcast, generally they make no difference to the music, or present some small issue in how it was recorded or produced.
I’ve tried, very briefly, using this feature for its actual purpose: production. My very early impression is I love it for composition. I often struggle in the early stages of writing finding that the sounds don’t sound cohesive, and I get pulled off track making minor adjustments to sounds before the key elements are together. This made drum machine samples sound closer to ‘right’ to my ear. I think for final touches, I’ll be switching back to traditional headphone output so I can hear more details when comparing with a reference track.
I’m going to stick with the loudspeaker emulation for a while. Finding it to sound good has helped me let go of the hunt for an ‘accurate’ representation of the music I’m listening to. On an intellectual level I’ve known for some time this feeling was unreasonable, but it’s taken something that seems technically imperfect that I’ve enjoyed listening to in order to shift the impulse.
Footnotes:
If you want to try it out, you can download HEADPHONE LAB for free from the Beyerdynamic website. It’s a VST, so loading it in a DAW for production is easy enough. For music listening, I have it opening in Foobar2000 using an extension. The lack of standalone mode does make it trickier when streaming from Qobuz. It’s definitely possible to do, just a bit more of a pain than SoundID which works well as a standalone app.
You might have noticed a slight inconsistency of my use of the term headphones across this piece. To open, I was comparing headphones to speakers and at least heavily implying the inclusion of in-ears. Well, how does the speaker emulation hold up with Sony’s WF-1000XM5s. It’s not good. Drums and bass sound strangely boosted and compressed. The reduction in soundstage becomes problematic when you had less to begin with.