Desktop audio is weirdly undervalued, and really easy to get right

TL;DR: Replace your desktop speakers with open-back studio monitor headphones, and thank me later.

Thesis

When you think about it, it's a little odd how little people are willing to spend on PC audio reproduction. They'll spend $1000 on an ultrawide curved monitor, $200 on a thocky mechanical keyboard with bizarrely-named keyswitches, $180 on a self-charging wireless RGB mouse, $3000 on a name-brand office chair, etc; and then they'll drop $20–$70 on speakers (considerably less than even one year of Spotify Premium) and call it a day. I think part of this is that inadequacy in HID is felt very sharply, while inadequacy in audio is well-compensated-for by the brain, so you can get by with a mediocre signal — that's why analog phones worked. There's also that passive listening reveals fewer flaws than active listening, most people have not experienced truly quality desktop sound reproduction and so don't know what they're missing, and it's difficult to actually find the right products if you don't know what you're looking for and are going off of consumer-focused advertisements. But even for passive listening, think about it: if you spend most of your day at a computer, you are probably also spending most of your day listening to music or podcasts or whatever else. Did you know that the human voice's fundamental frequencies, even those of women, are heavily attenuated or even completely cut-out by cheap satellite speakers? That you guarantee significant intermodulation distortion in "S" sounds when your speaker has only one driver? That speaker grills can meaningfully mess with high frequencies? That most speakers do not clearly reproduce high frequencies if they reproduce them at all? That you physically have to have a certain cone size to properly reproduce low-pitched sounds? You probably didn't. But these are all true, and your budget speakers are victim to all of these modes. Even relatively decent consumer speakers, like Logitech's $200–$300 2.1 systems, run into these limits, though far-more-gracefully. And so, if you don't invest in a proper sound system, you simply will not hear all of what you are playing, and what you are listening to will always be muddy, uneven, and incomplete: functionally equivalent to a display that is blurry, stuck with a permanent blue hue, and missing a quarter of the pixels. Somehow, that scenario is unacceptable in a display but accepted in speakers, and I think this is weird. Don't you also think this is weird?

That said, I don't actually think desktop speakers make sense for 99% of people. Yes, people get them; but you probably shouldn't, and it's because of yet more things that you probably don't know: The room you are in and even how much clutter is on your desk can heavily modify what sound actually reaches your ears and when it does so, your tabletop acts as a megaphone for some frequencies and a mirror for others, what you actually hear can vary significantly when you move your head, and how you position your speakers matters quite a lot. And so, ultimately, it ends up being that, in order to get a good speaker setup, you're looking at $200+ for a pair of studio monitors and then another $200+ for a good sub, pretty much as a minimum; you also have to invest in speaker stands, isolation pads, proper cabling, a DAC, and maybe a ground noise isolator, coming to another $200+ minimum; you then have to invest in room treatment (also not cheap) and measurement/compensation (more money, and a lot of time+skill); and that's to say nothing of the time it will take for you to learn enough acoustic theory to properly integrate these things. That's way too much to expect any normal person to spend or do. Thankfully, you don't have to do this to get really good sound: you just need a pair of good studio monitor headphones. Audio-Technica makes a good pair of open-backs for $150, which is a great price for gear you will use continuously for 8+hrs/day for the next several years. That's right: for $150, you can compete with a $600+ setup for minimal time-commitment and without learning anything about speakers or audio engineering. (That said, I will list out the items needed to build my full audio setup at the end of this article, for those who want to go the desktop speaker route anyway. But if I had to do it over again, knowing what I know now, I'd absolutely have gone with top-end headphones instead.)

Whys

"Whoa", you may say. "You don't have to go $600 deep; you can start with just a pair of good nearfield monitors and go from there." On this, I disagree. For example: I got a pair of Kali LP-UNFs — very good ultra-nearfields (around $350) with a truly full range; like, they're still putting out quality audio at 39Hz (just -10dB down) (the -3dB point is 54Hz, fwiw). Yet, my $200 Logitech Z625 (a 2.1 system) sounded better, as did my $150 Audio-Technica ATH-R50X headphones. Why? Because the Kalis were trying to do too much: producing 39Hz and 1950Hz in the same woofer makes for substantial IMD sidebands (distortion) at the high end, which just happens to be right at one of the most-sensitive parts of human hearing, so of course I heard it. When I high-passed both systems at 200Hz & 12dB/oct (the shallowest curve that caps excursion), the Kalis suddenly sounded significantly better than the Logitechs. High-passing by definition significantly limited my bass output, so I had to get a subwoofer: the Mackie CR8SBT, at around $200. It handles the bass so my Kalis can focus on mids and highs; and now the system sounds great. But that put me out around $600 (after tax) just in speakers, to say nothing of the cost of accessories.
The moral? If you're going to go the nearfield route, you really have no choice but to go all the way if you want it to be worth doing. Half-measures leave you worse-off in quality and money than full-measures. So if you're not willing to pony up $700 or so for that kind of a setup, don't — spend even half of that on a Sennheiser HD 650 ($380 at the time of writing) and you'll be living very large.

[Warning: technobabble; skip if not an audio nerd] "Why did Kali make full-range satellites if they still need a sub?" I'm not entirely sure, but I'm quite glad they did. Full-range sats synergize with a sub that XOs high and shallow: the shallowness lets the sats continue to contribute even deep into the bass range (in my configuration, they're down -12dB by 60Hz). That continued contribution permits a higher XO (120Hz+) than normal because the sats are able to keep stereo imaging robust. Going higher on sub-sat XO is good because it weakens IMD sidebands at the woofer-tweeter XO (1950Hz). REW says without EQ I'm flat ±5dB from 29Hz to 21.3kHz, which is fantastic for me not having done any room treatment as yet. The only remaining weak point in my setup is phase misalignment at the sub-sat XO; my sub only has a ±180° switch, which is… not ideal.
…Okay, enough of that nonsense; let's get back to language normal people will understand.

Even with an infinite budget, at-ear speakers vs in-room speakers are just entirely different categories of product, and the category that matches your needs as a casual PC user is overwhelmingly at-ear-speakers. Let's look at a quick breakdown of what traits each has that the other physically lacks:

• At-ear: perfect stereo imaging, privacy, unaffected by head position, unaffected by room, minimal parts, works out-of-the-box without integrations
• In-room: deeper floor for the low-end, social listening, whole-head bone conduction, nothing worn on head

You may be wondering: why specifically over-ear headphones instead of some other head-mounted option? A number of reasons. I'll start with what over-ears have as advantages over every other class, and then I'll get into the specifics of those other classes.

• Over-ear headphones, because they are so large, can produce real bass; and again because of their size, the sound they produce is transmitted to the cochleae (your hearing organs) via air-conduction, cartilage-conduction, and a little bone-conduction at the same time, just like native sound out in the world. No other class of wearable speaker can achieve these things as well as headphones. They're also surprisingly compatible with earplugs because they have non-air-conductive pathways and a high loudness ceiling; though of course this does impact their effective frequency response.
• Bone-conduction headphones are very frequency-limited (say "goodbye" to bass and air unless augmented by some other channel) and inherently smear stereo (the vibrations do not stay local). However, BC phones are unbeatable if you have conductive hearing loss; more, they're compatible with earplugs and suffer NO reduction in quality therefrom. They require clamp force to work well, though, necessitating some sort of neckband, which can be annoying in some contexts; and loudness, quality, and frequency response can vary fairly considerably by clamping force and exactly where on your head you have positioned them.
• Cartilage conduction headphones have inherently better quality than bone-conduction ones; but like BC, CC suffers in clarity because there's no room to fit multiple drivers without interference. More, CC is never pure CC — it's always partially AC (air-conduction), and that means you can't put earplugs in and expect to get unchanged audio quality. CC is very rare on the market and does not come cheap (same price as decent studio monitors); and like BC options, requires clamping force and some sort of neckband, and vary in characteristics by placement and clamping force.
• Out-of-ear earbuds work decently well, but they're limited in bass reproduction, muddy in highs, and lacking in air. They're effectively incompatible with earplugs because they are too weak to meaningfully reach the BC and CC pathways. Nevertheless, they can provide better audio quality than BC or CC at a fraction of the cost, for way less power, and in a much-more-compact form-factor.
• In-ear earbuds vary widely. As a class, they roundly beat out-of-ear earbuds on quality; but they still struggle with the fact that they are tiny and yet somehow must reproduce both bass and high frequencies in that space. Designs capable of doing this decently well are priced so high as to be totally unreasonable for consumer use — well-above the price of decent studio monitor headphones — and even at such prices, they still provide worse quality than price-matched headphones. In-ear earbuds also have the distinct disadvantage of preventing airflow to the ear, thus increasing humidity, thus increasing infection risk and wax buildup; they also have a nonzero risk of a dome falling off into the ear. These factors make them the least-healthy choice of head-mounted speaker. (If you are nevertheless still interested in them, note that the best ones are marketed as "IEM"s — "In-Ear Monitors".)

So, why open-back headphones? Because closed-back ones (unless they're superbly constructed) chaotically ricochet the sound trapped inside them, which causes comb-filtering, distortion, booming, wildly inconsistent volume, etc — the effects of an untreated room condensed into an ear-sized package. Open-back headphones, in contrast, avoid these issues by allowing sound to escape, thereby significantly improving quality at the ear. The caveat of this solution, of course, is that people nearby can hear your music, and you can in turn hear them. The latter trait is often desired — there's a reason a lot of people buy head-mounted speakers that don't block their ears — but if you do need to block outside sound or if you really don't want other people or your mic to have any chance of hearing your audio, then you should buy closed-back; but I wouldn't recommend buying them in any other situation, if at all possible. Also, it's worth noting that, if your budget allows, these are not mutually-exclusive options — I actually have a pair of noise-cancelling closed-back consumer headphones and a pair of open-back studio monitor headphones even though I have quality desktop studio monitor speakers, and I switch between these things depending on the situation: the closed-backs for phone calls and for protection when I'm playing instruments (recorders are dangerously loud!); the open-backs for when I need tight stereo imaging (mostly listening to ASMR) or I want to listen to music without bothering people nearby; and the desktop speakers for all other situations.
The other reason you should get open-backs is that, for all-day listening, they are simply more-comfortable to wear, and they're healthier for your ears too. This is because they allow airflow, thus preventing heat and moisture buildup.

Referata

It's important to note that not all headphones, be they open-back or close-back, are equivalent, and a good closed-back pair can still obliterate a bad open-back pair; open-back only wins all else equal. The biggest differentiators are typically price and whether the headphones are aimed at consumers or at pros. That said, even nice consumer-focused headphones, including some rather pricey Bose ones, pretty much by definition have an intentionally very uneven frequency response, typically heavily boosting bass and moderately boosting highs. This is basically a very crude caricature of Fletcher-Munson equal-loudness compensation corrupted into meme-tier bass-boosting (probably because it sells well to customers who are used to hearing literally 0 bass from their cheap speakers); this heavily colors the audio: think "obnoxious Instagram filter permanently baked-into your speakers". But you don't want your speakers to color what you are listening to, just like you don't want your display to tint everything purple — you want your speakers to disappear, and for the sound to feel like it's actually happening right in-front of you. And for that, you need speakers with a more-or-less flat (even) response. The term for products that aim at this goal, be they headphones or desktop speakers, is "studio monitors", or "monitors" for short. (Yes, displays are also called "monitors"; yes, this is confusing.) Not all studio monitors are equivalent; but in general, any studio monitor is going to be in an entirely different league than non-monitors; and the more you spend in either category, the more you typically get.
While we're on the topic, here's another downside of speakers with an uneven response: you have to turn the volume up higher to get the mids (the most-important range) to an acceptable level, and doing this can often make the lows too loud; this can then make listening to music tiring, or worse: actively dangerous for your ears. An uneven response is also more-fatiguing, and you don't want the thing you listen to all day long to be fatiguing. A quick note, though, to obviate nitpicks: Not every speaker that aims for "flat" is equally flat or has an equally broad range, and even the flattest speaker still varies by ±3dB in a soundproof room. So whenever you buy anything that reproduces audio, always check its frequency response, and make sure it is acceptable to you.

Since it's come up, I want to give a quick side-note regarding equal-loudness compensation: your ears hear some pitches better than others, and this effect is strongest at low volumes and weakest at high volumes. That means turning up the volume of your music doesn't just make it louder — it makes it sound different (fuller). However, going too loud can be dangerous for your hearing; so one way to have your cake and eat it too, is to (using your DSP (Digital Sound Processing) pipeline of choice (EasyEffects on Linux, APO on Windows)) enable ISO 226 equal-loudness compensation. This will boost insensitive ranges (like bass and air) to simulate what audio sounds like at a higher aggregate volume, while keeping things overall quieter (and therefore safer). This only really protects your hearing if you're using studio monitors, though, because unflat consumer headphones already have mediocre compensation built-in, so you'd be doubling up in a weird way with an end-result that's likely worse than doing nothing unless you first corrected the frequency response of the consumer headphones via EQ. Also, to do loudness compensation "correctly", you have to measure SPL; but you can calibrate it "good-enough" by just adjusting the strength of the effect to your personal preference. The downside of compensation is increased latency; possibly phase rotations (depending on implementation), though not at frequencies where you're likely to notice; and throwing away data right in the most-sensitive parts of your hearing (this is only theoretical if your DSP chain is 32-bit float).

Also: You will often hear people say high frequencies don't matter. These people can be conditionally correct: high frequencies aren't essential for speech comprehension or music, and most people have started to lose them by their 30s. But a lot of people haven't lost them by their 30s, and also there are a lot of people who are under 30. Mine are now slipping due to an unfortunate acoustic trauma last year, and yet I can still identify and follow songs I know well even when they are low-passed at 16kHz — though only if I boost my highs to make up for my recent loss; which, just for reference: I can just barely hear to about 19.3kHz when boosting, if I stay within safe levels — so the sound up there is definitely not dead or useless. So if you can still hear high frequencies, you owe it to yourself to enjoy them while you still can. Most speakers struggle to reproduce above about 15–16kHz (rolling off and/or distorting it); but most studio monitors are more-or-less flat to 20kHz (or at least they'll claim to be). My personal experience with sudden high frequency loss has taught me that the brain primarily uses high frequencies to get spatial information about the room you are in, rather than as a way of transmitting/receiving information; but they can nevertheless indeed do the latter. For immersion in videogames or enjoyment of ASMR, high frequencies are imho quite underrated. By investing in a pair of open-back studio monitor headphones, you will unlock access to these frequencies. To actually avail of that access, though, you need to make sure the audio you're playing hasn't had its high frequencies ruined — many lossy audio codecs, including MP3, deliberately mess up highs to save on data.

Additional benefits

• Headphones do not annoy your neighbors, headphones (even open-back) provide you with far more privacy, and headphones are far-less-likely to be picked-up by your mic. Society would be meaningfully improved if people didn't blast subwoofers in apartments and if your voice didn't echo back to you on other people's mics; and headphones are the solution to these problems.
• Headphones not only save you in CAPEX (upfront costs); they save in you in OPEX (ongoing costs) by requiring WAY less electricity than a full studio monitor setup. Heck, headphones with moderate-to-low resistance require less electricity than even USB-powered speakers.
  • By requiring less electricity, they are also better for the environment; and indeed, this is also true in terms of how many materials are required to build headphones vs to build desktop speakers.
• Headphones are a lot easier to move than bulky subs and sats, and they take up extremely little desk space.
  • In fact, headphones aren't just easier to move — they're outright portable, and that's true even of studio monitors that aren't designed for it. You can't take your $700 nearfield studio monitoring setup with you to the grocery store; you can take your $150 studio monitor headphones, and because they're open-back you don't even have to take them off when you're talking to the checkout people. Yeah, they're bigger than AirPods, but they're also way better. Normalize wearing headphones in public and spread around the joys of quality audio!
    • Heck, you might not even have to take them off in your car on the way to the grocery store, if the law permits — open-backs won't prevent you from hearing sirens, and they're probably significantly better than the speakers in your car. (I know my own car, a 2018 Chevy Cruze, has cheap paper cones for its speakers and they are… okay at best. An aftermarket speaker upgrade would cost several hundred dollars, which is a lot more than $150.) One pair of headphones, and you can solve your desk and your car audio problems at the same time.

Using headphones outside the house

(This section is tangential to the main point, but I wanted to mention it anyway, since I brought up headphone portability.)

Headphone use outside the house was much easier back when we still used walkmen or mp3 players, and when phones still came with aux ports. Unfortunately, for modern phones, you have to get a USB-C-to-AUX cable in order to use headphones. If at all possible, get a cable designed to handle Audio Adapter Accessory Mode (AAAM). If your phone doesn't support that, then you have to get a cable that contains a little chip (a DAC) that translates phone audio (digital) into headphone audio (analog). Cheap DACs can significantly degrade audio quality, and the most-common way in which they do this is by adding a constant "hiss" sound. If you want to use your fancy headphones from your aux-less phone, you owe it to yourself to invest in a cable with a decent DAC. Ideally you'll find one with a posted SNR of 96dB or better. If the products you see are not posting their SNR, that's already sketch; but you can try to use the following heuristic: look for ones that support 24-bit audio. Now, to be clear: you have zero use for 24-bit audio when you're wearing headphones — 16-bit caps out at 96dB, and the loudest you're ever going to be willing to put up with in is in the 80s dB SPL. But DACs that can only do 16-bit audio tend to be lower-quality, so looking for 24-bit can be a decent heuristic if you have nothing better to go by.

You may ask: "Why bother with cables when Bluetooth exists?" Well, you can use Bluetooth if your studio monitor headphones support it, but they almost always won't; again, this is because they're generally aimed at pros, and pros want a wired connection because it is consistent and nearly free of latency. Bluetooth, in contrast, is actually very slow, and is capable of dropping out. You normally won't notice the slowness of Bluetooth when you're just passively listening to audio, but it's not outright irrelevant: if you carefully watch a video, you can often see sound lag behind lip movements just from the Bluetooth alone; you're more-likely to interrupt people during calls when everyone's behind Bluetooth delay; and if each ear is a separate Bluetooth device, they can actually subtly mess with stereo imaging and clarity if the signal arrives to each ear at even slightly different times (even 2ms is enough), although a lot of buds nowadays are able to compensate for this last one. Nevertheless, a wired connection simply doesn't have these complications, and for short runs will "just work"; hence why it is generally preferable to Bluetooth if you can manage it.

Limitations of studio headphones

• Being wired, while optimal for audio quality and latency, has the undeniable drawback of tethering you to something. If a cable is a problem for a particular task, just use a different audio-listening device than your studio monitor headphones.
• Studio monitors and open-back headphones almost never come with builtin mics, and this is on purpose: studio monitors are for listening, not recording; and builtin mics on open-backs are quite likely to pick up on whatever you're listening to. Honestly, though, you're a lot better-off with a dedicated microphone than a headset mic, anyway; but this is nevertheless something to be aware of.

Notes on audiophilia

Getting good sound is not an endless journey. Once you get a known (measured) good system — full-ranged, flat, and low in distortion — reproduction is no longer your problem, and you should stop trying to min/max it. The remainder of your chain is your room (if your speakers aren't on your head), your media, your ears, and your brain, so your attention needs to move there. Play uncompressed audio, protect your ears, and don't let obsessiveness distract you; do these things, and your listening experience cannot be made better except by placebo.

(Yes, you can always invest in some new upgrade that can net you some theoretical benefit; but that's what it is: theoretical. A change in SNR from 100 to 130 simply does not get you anything for desktop audio listening; both situations are identical because you cannot hear the background hiss in either scenario.)

Regarding your ears: Protecting them is actually the single biggest, most-important thing you can do for audio quality: they can't repair themselves, doctors can't repair them either, you only get the one pair, and all audio quality hinges 100% on their function. If you spend any money, musician's earplugs and some means by which to carry them with you day-to-day is the single biggest ROI that exists for audio quality, and no amount of spending on audiophile gear nor hearing aids can ever claw back the quality you lose when your ears get damaged.

Conclusion

People should value their speakers as much as they value what those speakers play. Quality desktop speakers aren’t automatically a bad option, but they’re an acoustics project disguised as a simple purchase, and for most people, they demand far more money, space, time, and learning than they’re worth. Open-back studio monitor headphones deliver the vast majority of the possible quality at only a fraction of the cost and with minimal complication; so if you want good sound without turning audio into a hobby, they’re the boring, correct answer.

My setup

As promised, here is a list of my desktop audio equipment, in case you want a great setup without the trouble of designing it yourself.
(Prices are pre-tax, and what I paid for them when I bought them. They may be higher or lower now.)

Brains

Total: $321

• $46 USB Hub: USBGear USBG-BREC307PWS (Has a name-brand controller and comes from a reputable manufacturer.) (The DAC gets all the USB2 OUT frames, the mic gets all the USB2 IN frames, and the webcam uses USB3; so there's virtually no contention between devices. Gives maximally reliable audio.)
• $40 USB Ground Noise Isolator: DSD TECH SH-G01L (Importantly: Supports High-Speed USB2. If it didn't, audio would lose out on microframes.) (Completely eliminated some very audible hum that was coming from my aging laptop.)
• $130 DAC: Creative SoundBlaster X4 (It has a button that switches between speakers and headphones.)
• $125 Measurement mic: UMIK-1

Desktop Speakers

Total: $696

• $7 6' AUX-to-RCA Dual-Sleeved Cable
• $200 Subwoofer + Desktop Analog Volume Dial: Mackie CR8SBT (The analog volume dial this comes with is a major win.)
• $1 Custom 3D-Printed Legs (Gives the bottom port more room so it isn't choking.)
• $12 4×1" 20A Platinum-Silicone Domes (The math says these are the optimal size for my sub's weight.)
• 2×$20 15' Balanced TRS Cable
• $55 Speaker Stands (These ones are designed to sit right against a wall.)
• ½×$22 16×¾" 20A Platinum-Silicone Domes (The math says these are the optimal size for my sats' weights.)
• $350 Satellites: Kali LP-UNF (Full-range and designed to sit against a wall. I got them in black.)
• $20 8' Phoenix Cable (I Wago'd it to the one that came with the sats to get a 12' cable so that I could put my sats 180° apart.)

Headphones

Total: $374

• 2×$5 AUX Male-to-Female 90deg-to-Straight Cable (I plugged these to the front of my DAC and then routed them behind it; the end-result is quite aesthetic.)
• $6 AUX-to-TRRS Dual-Female-to-Male Straight-to-Straight Cable (This merges the DAC's front mic and headphone jacks.)
• ½×$7 TRRS Male-to-Male Straight-to-90deg Cable (This lets the headphone switch take up less space.)
• $30 Headphone Switch + Volume Dial: Cubilux AS-J3
• $149 Open-Back Headphones: Audio-Technica ATH-R50X
• $150 Closed-Back Headphones: Bose QuietComfort 45
• $35 Headphone Stand with Electrical Plugs
• $1 Custom 3D-Printed Dual-Headphone Top for Headphone Stand

Tips

Note that, when buying studio monitor headphones, you need to be wary of the connector and of the Ohms. Since studio monitors are primarily designed for professional audio work, it is not outright unusual for them to use exotic connectors (balanced TRS/XLR) or to have a high resistance (Ohms) requirement, which requires reciprocally high voltage. For your use as a consumer, you want monitors that can be plugged-in via 3.5mm aux cable, and you want their Ohms to be low-enough that your laptop's onboard DAC can drive them. Otherwise, you will need to fork out a bunch of money on a dedicated DAC. Now, I'm not saying that isn't a nice thing to have, but it does increase cost and complexity.

For anything you wear prolongedly: make sure it's comfortable and well-fitted. If it's not, get something that is. (Sometimes, replacing the stock cushions is an option if everything else is otherwise to your taste.)