After collecting my favourite 100 albums on good old, 16-bit, 44.1kHz, uncompressed, compact discs, I want to treat myself to a nice set of speakers that’ll play the music I love as close as possible to what the artist who made that music wants me to hear it.
(still waiting for about a dozen more CDs to arrive to complete the top 100)
Listening to these CDs on my current setup (Polk S10 bookshelf speakers, Richter Krakatoa subwoofer & Loxjie A30 integrated amplifier & DAC) got me wondering if what I’m hearing really is Jack White’s beautiful guitar tone as he painstaking designed it? Am I getting the full glorious audio picture that Bjork worked so hard to create? Are those bleeps and boops in Kid A what Thom Yorke agonised over or is it a muddy mess that sounds nothing like what he intended?
That’s what I’m aiming for with what could be the most expensive thing I ever buy except for a car and house. Unfortunately for me, I have no idea what I’m doing. This post is me learning as I go about what makes a “good speaker”.
Humans can hear down to 20Hz and up to 20kHz. That doesn’t mean most music has all those frequencies in it, but a good speaker setup will be able to reproduce all the frequencies in that range. Have a listen to this:
Personally, I want a speaker as close to neutral as I can afford. There’s loads of audiophile terms like fatigue, crispness, warmth, tonality, clean and so on to describe sound, but essentially I don’t want to hear a speaker designer’s idea of what the music I listen to should sound like. I want to hear the what artist and producer of the music I’m listening to want me to hear.
This is called a “flat response”, or “transparent”. No frequency exaggerated or diminished either due to a design fault in the speaker or a decision by a speaker manufacturer.
To sum up - a good speaker for me is one that has a flat frequency response as possible all the way from 20Hz to 20kHz.
Audio engineers are also after a flat response as they want to hear the instruments and vocals just as it was recorded. Monitor speakers are designed with this in mind and try their best to achieve a flat frequency response.
Flat is not how most people like to listen to music, so the common knowledge is that unless you’re producing music, don’t get monitor speakers for general music listening, but that’s old and often incorrect advice. Monitor speakers can be great and bad, just like any other speaker. There’s no right or wrong when it comes to speakers. Listen to whatever you like! But me, I want that flatness and allow the music reveal itself. That’s the theory at least.
Another benefit of monitor speakers are that they’re designed to be listened to close (aka near-field) which is good for me as they’ll be used on a computer desk where I’m sitting ~60cm away from them and have built-in amplifiers - one less thing you need to decide on. Someone’s already done that hard work for you.
Golden ears, $5,000 cables, $100,000 speakers, tube amps and so on are the domain of audiophile wankers from the 90s. Some people enjoy it and good luck to them, but there is absolutely a point of diminishing returns with audio gear.
Sites like Audio Science Review, Erin’s Audio Corner and No Audiophile dissolve the snake oil that’s so rampant in this industry to give you hard data on these products. I have very little idea what that data means, but they have shitloads of it and more importantly, people with way more knowledge than me interpret it and provide recommendations.
Here’s a site that takes all the measurement data and lets you compare them to one another. Look at this comparison of the Genelec 8030C vs the JBL 306P MKII:
I don’t quite know how to read it properly, but I think you want the “listening window” line to be as flat as possible around the 0db mark. The solid line is the $1000 Genelec and the dotted yellow line is the $250 JBL. They’re not that far apart, are they?
Someone else has taken that data and converted it to an “Olive” score using Sean Olive’s “Multiple Regression Model for Predicting Loudspeaker Preference Using Objective Measurements” research paper (explained in this forum post), to give a number rating. They even put it on a chart aligned by price and score.
This rating number, based on Dr. Olive’s research is as the main dude on the ASR Forum summarises, “a set of objective measurements with no notion of what humans like or don’t like, is able to highly predict what we like in the sound of speakers!” - sounds impressive enough for me.
Going by that score, the top result ever is the Genelec 8030C when paired with a subwoofer. Combining the data from Audio Science Review with the research from Dr. Olive, the 8030C plus a subwoofer should be the most pleasant speaker to listen to music with.
It’s good that this stuff exists as it helps point someone like me in the right direction as to what to buy. God bless these pedantic audio nerds!
Long story short - yeah, probably. I wanted to try and get away without a sub to avoid the setup and integration with the room (ugh), but if I wanna hear sounds down to 20Hz (and I do), then a subwoofer is non-negotiable. Just look at this chart of common musical instruments and how low they can go:
I’ve got a 10" subwoofer in my current setup and a 15" subwoofer in my home cinema. The difference they make is extremely noticeable. As if I’d be able to hear The Beastie Boys, Chemical Brothers, Jay-Z and dozens of other artists I enjoy at their best without a subwoofer. I was a fool for thinking I didn’t need one.
Any subwoofer I get needs to be competent at 20Hz too, just getting down to 30Hz or 35Hz isn’t enough. If I’m going to fuck around with a subwoofer I may as well get a proper one. It’ll also ideally have balanced XLR inputs so I can connect a DAC directly to it and set an appropriate crossover on the subwoofer’s amp. Some fancy subwoofer and monitor setups has “bass management systems” that work it all out for you automatically. That would be cool.
All my music is digital. Be it from a computer or from a CD, it’s 1s and 0s that need to be turned into analog signals and that is the job of a DAC. Luckily this is a solved problem in 2022 - DACs with perfect conversion are now cheap and cheerful. Audio Science Review has dozens of DAC reviews and the Topping E70 seems to be the best bang for buck.
$525 locally, perfect digital to analog conversion with balanced output, and a plethora of inputs like USB/Optical/Coax and Bluetooth (AptX Adaptive & LDAC). It’s also got a built-in power supply and a remote control for changing output signal (i.e: volume) and changing inputs.
I could spend around $1,200 and get a Topping D90LE - the DAC with the highest SINAD measurement on record at Audio Science Review - but the E70 is incredibly close and less than half the price. It’s arguable nobody would even notice an audible difference.
Some fancy studio monitors have digital inputs, which means they have a DAC onboard and all I have to do is hook up my inputs directly to the speaker. Unfortunately most of the digital inputs are AES, which means I’d need to buy some sort of audio interface to go from USB on my computer or SPDIF on my CD player to AES. These things are not easy to find, nor are they cheap or come with remotes so I can switch inputs (Bluetooth, CD player, computer) easily. As nice as it would be to skip a separate DAC, it’s just not practical for a non-studio setup.
Having a perfectly flat speaker in an anechoic chamber (where these speakers are tested and the charts generated) is one thing, but the room you’re listening to the speaker in impacts the sound dramatically. Put the same speaker in two different rooms and they’ll sound different.
To combat this you can fuck around with “treating” the room (sticking foam around the walls, more curtains, etc.) and you can use what is called “digital signal processing” (DSP) to alter the frequencies the speaker puts out so when they hit the walls and objects in the room, they come back to your ears sounding more like what you expect.
You get a calibrated microphone, hook it up to some software on a computer or on a dedicated piece of hardware, play some tones through your speakers and a fancy algorithm works out what you need to change. If you’ve ever set up a Sonos speaker and used the Trueplay feature or used a home theatre amp with Audyssey EQ, that’s room correction.
There’s multiple options out there - Dirac and Room EQ Wizard being the most popular on the software side. Some speaker brands have their own systems built-in to their products, like Genelc’s GLM and Neumann’s Automatic Monitor Alignment.
The most interesting however is the MiniDSP range of hardware and in particular is the MiniDSP Flex. The Flex is not only a DSP, it’s also an excellent quality DAC with balanced output. It supports Bluetooth, coax/toslink and USB inputs. It’s got Dirac Live support, which is one of the best auto-magic room correction algorithms and can do bass management - just plug any subwoofer (even if it doesn’t have balanced input) in to one of the outputs.
The Flex a very tidy setup, but has to be shipped from Hong Kong and has a total price of US$879 - almost A$1300 and I’m not sure if it’ll get slugged with GST on import. It’s pretty fucken good though.
A common problem noted in many reviews for monitor speakers is “tweeter hiss”. ASR notes this in their reviews of some otherwise excellent bang for buck speakers:
Adam T5V: “Quick testing for hiss showed that I could not hear much past a few inches away from Adam tweeter. So really not a problem – at least not in my sample.”
Kali LP-6: “I tested for the audibility of hiss. Spectrally the noise is not as annoying as it is with some other active speakers. I measured with a ruler when the noise subsided and it was about 24 inches/60 centimeters. Even in my close in listening, it was not a problem. That said, I do wish that the noise was not there.”
JBL 306P MK II: “The usual hiss was there by the way and did not go away until about 2 feet when nothing was playing.”
Dynaudio LYD 5: “And oh, there is tweeter hiss that is a bit lower than JBL LSR305. Like that, I could not hear it from my seating location and certainly during music playback. Once sensitized though and in a more quiet space, you may hear and be bothered by it.”
Here’s a video from EEVblog trying to diagnose it on some KRK monitors:
It seems to be that if you’re listening at around 1m or further away, you aren’t going to hear the hiss. But if you’re in a near-field situation, like me, with the speakers roughly 50cm-70cm away, it is audible. I’m sensitive to this kinda thing as my study is very quiet during the day (suburban living!) and there’s no other background noise as my computer is silent and all my servers and network equipment go in the garage.
Perhaps I could buy these speakers and realise the hiss isn’t a problem, but seeing as the issue is so widespread - just look at Reddit and YouTube for the hundreds of complaints - I’m likely wasting my time with them.
Here’s what I need from my setup:
- Actively powered monitor speakers rated highly by people more knowledgable than me
- Subwoofer that’ll do at least 20Hz properly
- Topping E70 DAC is awesome if the speaker system has a built-in DSP/room correction
- MiniDSP Flex is a great DSP and DAC all in-one, albeit a bit expensive at $1300
I’ll pop in another post actually listing some gear with prices as this post has gone on long enough.