Sound is what matters most in classical music

What matters most in a classical performance, whether in concert or recorded, is not the age of the audience, the hype surrounding the musicians, the gender of the conductor, the number of Facebook 'likes', the backstory of the composer, or the many other next big things manufactured by the marketeers. What matters most is the quality of sound, which must pass a threshold of acceptability for the message of the composer and performers to be effectively communicated to the audience. The rishis who composed the Vedic Hymns nailed it when they declared nada brahma - sound is god (Shatpatha Brahmana X.20.43). The overriding importance of sound is recognised in the search for the perfect concert hall. The results of this obsessive search include the 'acoustically' perfect' Elbphilharmonie - or perhaps not so perfect. It also triggered Simon Rattle's demand that the perfectly serviceable Barbican Hall in London should be abandoned and an 'acoustically perfect' replacement built next door. (Presumably the new £288 million hall will come complete with selfie hot spots'.)

However sound is not god when it comes to recorded music. Which is both puzzling and a serious oversight; because the consumption of music from recorded sources - home audio systems, mobile devices etc - is far greater than from live sources - concert halls, opera houses etc. It is doubly puzzling because critics' reviews of recordings very often influence the status of musician's in the concert hall. Yet the very same critics who complain they cannot hear the cello section in the Barbican  are surprisingly coy about the quality of the audio systems they use to pass judgments on recordings by the same ensembles*.

Many years ago when I worked in the classical industry I discovered that critics of the day often listened on distinctly sub par equipment. And candid photos on Twitter suggest that some of today's reviewers are using distinctly sub-Barbican, yet alone sub-Elbphilharmonie quality audio systems. (Call me a cynic, but I suspect this disparity in expectations between audio system and concert hall sound quality is because critics don't pay for their trips to the Elbphilharmonie but do pay for their domestic audio equipment.) Incidentally, headphones are not a substitute high quality speakers. A definitive classical mix should - but does not always - reproduces the sound stage of a concert hall, and only a properly positioned pair of speakers can replicate this. Headphones do not produce stereo sound: they produce an 'in-head' binaural sound image which is the equivalent to sitting in the middle of the orchestra.

Classical music's pursuit of excellence feeds on marginal gains, and music critics will doubtless argue that it is marginal gains in acoustics that are important - Barbican versus Elbphilharmonie. Which is quite true; but, of course, marginal gains are equally important in audio systems. Let me use my own domestic system as a case study to illustrate this point. But before cutting to the chase let's talk about costs. The final bill for the Elbphilharmonie came in at €789 million, and acoustic excellence in audio equipment can also be very expensive. The equipment I listen on may seem to cost a lot. But, for instance, the Rotel RA-1592 amplifier I will write about cost me the equivalent of several visits - including transport, overnight accommodation, food, and tickets - to the country house opera performers which the classical fashionistas plug so heavily. And that amplifier almost certainly costs less than one visit to Bayreuth or Salzburg. My Bowers & Wilkins speakers are even more expensive, a decent new small car could be bought for their list price. But this pain can be mitigated by buying ex-demo models from a reputable dealer, as I did**. (Incidentally, I also save a lot of money by buying almost new low mileage cars on which someone else has taken the considerable depreciation hit.) So having disposed of that particular canard let's talk audio systems.

Classical music's great reference recordings from EMI and Decca in the 1960s and 70s were balanced using a correctly positioned stereo pair of studio monitor speakers. So this is the model I use in my main listening room. Until recently Bowers & Wilkins Nautilus 803 speakers were driven by an Arcam Alpha 10 amplifier, with inputs from an Arcam CD37 CD/SACD player, Thorens TD125/SME turntable, and Denon TU-260L II FM tuner.
With an output of 100 watts RMS per channel the Arcam Alpha 10 delivers plenty of punch. But for a while I had several concerns. Firstly, age and extended use of amplifiers causes the capacitors to dry out with resulting compromises in sound quality. Second, the impedance curve of the Nautilus 803s is notoriously challenging.

As this post is the antithesis of click bait I will now digress into the arcane topic of speaker impedance. Speakers, including the B&W 803s, are rated at a nominal 8 ohms resistance. But this resistance varies with frequency: for instance the impedance of my B&W 803s dips as low as 3 ohms. The lower the impedance the greater the current (watts) required to drive the speaker at that frequency. Halving the speaker's impedance doubles the amount of current demanded from the amplifier. So an amplifier rated at 100 watts into 8 ohms needs to deliver 200 watts into 4 ohms, and 300 watts into 3 ohms to produce the same sound pressure level (loudness). Although the volume on an amplifier is never turned fully up, the ability to deliver very high current spikes for short periods is essential for delivering transient peaks - eg cymbal clashes - that are an intrinsic part of the classical sound. This ability to deliver current peaks is an important factor in distinguishing good from truly excellent amplifiers.

At 26 by 15 feet (8 by 5 metres) my listening room is larger than average and it is quite heavily furnished. So new higher output amplifiers were considered. Historically my approach to amplifiers has been minimalist - a straight wire with gain. But digital technologies have thrown that philosophy into question. It must be accepted that at some point in the not too distant future new recordings will not be released on CD, while the proliferation of online music resources makes network connectivity highly desirable. Then there is the fact that at my age being able to cue music from my iPhone using Bluetooth is very useful when the cat is asleep on my lap - even if the file format is lossy AAC.

So six months ago I rejected the minimalist options and opted to upgrade to the Rotel RA-1592 seen in the photos. At around £2000 this amplifier costs significantly less than the minimalist audiophile alternatives; probably because it is made in Rotel's own factory in China. The Rotel RA-1592 is quite a beast: it weighs 17 kilos which is enough to give anyone that most dreaded of injuries, a slipped disc. At 200 watts RMS per channel it ticks the power output box. It also offers a comprehensive range of digital and analogue inputs, including an asynchronous USB input supports PCM files up to 24/192, DSD64/128, and DoP, and coaxial (RCA) and optical (TosLink) inputs for resolutions up to 24-bit/192kHz. (However, it does not have an analogue moving coil phono cartridge stage; which meant adding the beautifully engineered made in Austria Pro-Ject Phono Box external phono amplifier to play LPs). As can be seen from the header photo, the Rotel RA-1592's rated 200 watts into 8 Ohms is underpinned by a massive toroidal transformer flanked by robust output stages mounted on chunky heatsinks.

Regular readers will know CDs are my primary listening source. The Arcam CD37 incorporates the acclaimed Wolfson 8741 DAC (Digital to Analogue Converter). Because this is more highly regarded than the AKM 32-bit/768kHz DAC used in the Rotel, my initial auditioning was done using the Arcam's DAC with analogue connection (phono leads) to the amplifier. But experimentation quickly proved that taking a digital output from the CD player and using the Rotel's onboard DAC resulted in an audible sound quality improvement. Initially the TosLink connection was used, but this was quickly rejected in favour of a sonically superior coaxial digital audio cable (QED Performance) which keeps the digital data in the electronic domain.

An email exchange with Rotel UK's very helpful technical department (Rotel and Bowers & Wilkins are part of the same group and share technical support) confirmed that the circuit topology surrounding the AKM DAC in the RA-1592 is 'voiced' to optimise sonic definition. This presumably explains the counterintuitive preference for the Rotel's DAC: the DAC integrated circuit is important, but what it is connected to is equally, or more, important. (One disadvantage should be mentioned here: the AKM DAC cannot decode SACD discs - sorry about the excess of acronyms - so the analogue connection must be used for these.)

My search for a domestic Elbphilharmonie produced astonishing results. Upgrading from the undoubtedly excellent but venerable Arcam Alpha 10 to the Rotel RA-1592 produced tangible marginal gains. Predictably orchestral showstoppers such as Mercury's truly iconic 1959 Dorati/LSO Firebird leave visitors slack jawed in amazement, and, of course, purely electronic sounds from Jean-Michel Jarre and others are deeply (bass deeply) impressive. But unexpected was the improvement reproducing solo instruments. Pianos are arguably the most difficult test for an audio system, and well-recorded solo piano discs, for example Frederic Chiu's Gurdjieff/Hartmann recital and Randy Weston's Marrakech in the Cool of the Evening sound almost like the real thing, while listening to violinist Sayaki Shoji's Bach & Reger mixdisc takes you right there into the acoustic of la Chapelle de l'Enfant Jésus.

Conversely poor recordings are ruthlessly exposed; which is not surprising because that is what monitor speakers are for. But a bonus with the Rotel amplifier are the tone controls. These are considered heretical by audio purists as they interpose additional circuitry in the signal path. But integrated circuit switching now means tone controls can be switched in or out of the signal path with zero sonic impact. (In the photo above it can be seen on the display that the tone controls are bypassed.) Many archive recordings are of otherwise excellent quality but suffer from top-end roll-off. An example is the Portfolio of Ahmad Jamal album dating from 1959. With nuanced treble boost this archive classic sounds as good as, if not better, than many recent jazz releases. (Why is jazz so underrated?)

Not everyone has the inclination or funds to build a truly high-end audio system. But the quality of a system affects the quality of reproduced music, and must therefore influence critical judgments. For this reason I advocate that critics should disclose the equipment used in a review. However the reverse side of this argument also applies: just as there is - despite the hype - no such thing as an acoustically perfect concert hall, there is also - despite this article - no such thing as a perfect audio system. As has been explained before, the true sound of music is what would be heard if it is played in an anechoic - acoustically dead - chamber, and that is a very nasty sound. Every concert hall, whether perceived 'good' - Elbphilharmonie - or 'bad' - Barbican - alters the original sound beyond all recognition by reflections and resonances. An excellent recording captures all those reflections and resonances, and then even more acoustic compromises are imposed even by an audio system costing an eye-watering sum. Like so much in classical music, we are chasing mist when searching for the acoustically perfect concert hall or audio system. But at least an article about it makes a change from yet another one about the alleged sex lives of celebrity musicians.

* Let's dispose of this nonsense about 'stalking' on social media. If you put your views in the public domain on Twitter don't complain if those views are discussed publicly. If you don't want them discussed then change your privacy settings or kick your social media habit. And before complaining about selective comment moderation, please try commenting on certain other high profile music blogs. Please remember that On An Overgrown Path is a personal blog not a public forum. If you want a platform from which to express your own divergent views without selective moderation, there is a very simple solution: start your own blog. Then after fifteen years of dealing with comments you may start to realise how tiresome it is moderating classical music's rampant egos.

** When considering the relatively high price of top end speakers, their longevity must be considered. Speakers are like wine: they improve with age. The predecessors to my B&W 803s was a pair of the excellent KEF104aBs. These cost what was a small fortune when I bought them in 1974. But they are still performing strongly almost fifty years later in my AV room.

New Overgrown Path posts are available via RSS/email by entering your email address in the right-hand sidebar. Any copyrighted material is included for critical analysis, and will be removed at the request of copyright owner(s).


Chuck Stark said…
Thank you for this excellent piece. I couldn't agree more about the importance of sound in classical music. I go to a large number of concerts - probably one a week on average during the fall/winter/spring season - and one of the great pleasures of hearing them is the pure visceral sound of the various instruments. At home I listen to recordings on an excellent very "high end" sound system, as well as on a pair of excellent headphones.

Which brings me to two comments on your piece. The first is the omission of any comment on the adverse effect room acoustics have on listening at home. No matter how excellent the recording or the playback equipment, by the time the sound from the speakers reaches your ears it has been distorted by the acoustics of the listening room, and the reflections and reverberations of the concert hall have been supplemented by the reflections and reverberations of the listening room. Of course, it's possible to minimize the effects by acoustically treating a dedicated listening room. But if, as is the case for me and for most people, you listen in a room that serves other purposes and cannot be, for aesthetic or other reasons, turned into an acoustically perfect space, you're stuck with an acoustic deficit no matter how excellent the other elements in the chain.

My second comment is on headphones, which you dislike. I fully agree that soundstage is best reproduced through speakers. But I find that some headphones - and I have compared quite a few excellent ones - can be superb for listening to classical music. A principal advantage is that they are not subject to the listening room distortions to which I referred above. Indeed, I find that, for this reason, I often hear more of the acoustic of the space in which the music was recorded through headphones than through speakers. However, some headphones are better than others in creating the illusion that you are in the space in which the music is being performed, rather than in the middle of the performers. I have found that Sennheiser open-back headphones excel in this respect (my current headphone is the Sennheiser HD800S, but the same has been true of Sennheiser's open back headphones for generations of models).

And of course, as you mention, the ability to hear the acoustics of the space in which the recording was made depends on the way in which the recording was made. The simple two or three microphone technique of the best recordings - many of the early stereo recordings, and some recent ones - conveys a far more natural sense of space than those made with microphones stuck in front of every other instrument and emphasized or de-emphasized throughout.

As always, I enjoy your blog.
Pliable said…
Chuck, many thanks indeed for that thoughtful contribution. I would make one small point. My comment about headphones refers solely to their inability to reproduce a stereo sound stage. I do not dislike headphones. In fact I find myself listening to them more and more, as discussed in a post last year -
Pliable said…
It is also very pleasing that in this age of Twitter and shortened attention spans an article of 2179 words can reach such a wide audience.
mathias broucek said…
Thanks for sharing. Whilst not all 24bit/hi res files sound amazing, at their best they are really something: the mid-range detail is really something. I get mine mostly from Qobuz, eclassical and 7digital.

My main complaint about modern approaches to hifi is that they need to controlled with an app on a phone or tablet - the very kind of device that can distract one from listening properly....
Pliable said…
Since writing this piece I have experimented with a QED Reference 40 interconnect between CD player and amp; which means using the Arcam's Wolfson DAC instead of the AKM in the Rotel amp. After a burn-in period for the QED interconnect I have decided this is the optimum signal route - the QED cable seems to do the Wolfson DAC justice. So I now listen using the analogue signal from the CD player.

Recent popular posts

How to reach a big new post-COVID classical audience

I put a spell on you

Requiem for my vinyl

Untold story of the counterculture's Islamic connection

Vonnegut gets his Dresden facts wrong

The Berlin Philharmonic's darkest hour

Classical music must face the facts - click bait pays

Berlin Philharmonic's first woman conductor

Benjamin Brittten's relationship with children

Philippa Schuyler - genius or genetic experiment?