Hifi, why bother?

[Mark Hennessy]

Well after reading through things here and noticing more and more complex approaches to audio reproduction I am still of the opinion on keeping things simple. I note Jeffrey's post "As I keep saying, we don't just listen with our ears " and agree.

Umm, he was quoting me!

I'm not normally precious about this sort of thing, but I've been saying that for years, so claim it as mine, for what that's worth

As an engineer examining the system I do notice that one thing requires addressing, although has been hinted at throughout: Assessment of the physical and mental capabilities of the user / listener. A simple point would be that all peoples' hearing is different. How can any electronic system be designed for correctness without addressing and assessing these differences individually; Noting here that simple tone controls or even the usual equalizers will not be sufficient. Like with hearing aids individual assessment would really be required. Perhaps still out of reach for commercial companies.

This is a good point, but consider this: a hi-fi system should sound as realistic as possible.

So while we definitely do hear differently, our hearing system will be working much the same, relatively speaking, whether it's in front of a real instrument or a replayed version.

If you used a comprehensive equalisation system to return your hearing system to some sort of nominally flat condition, then your hi-fi would sound radically different to a live performance. Is that really a good idea? Probably not.

The place for detailed equalisation would be in a hearing aid. Modern ones are very comprehensive. Having done that, there's a much better chance that your hi-fi might sound like a live performance. But of course, that assumes one is needed.

All that said, a bit of gentle equalisation can be useful as our ears age - that's why I think it's a bad idea that bass and treble controls were removed from most hi-fi amplifiers back in the '80s.

Though it's not just our ears that are the problem. Speech on the TV is often terrible. Now, I know what those lapel microphones sound like because I use them at work, and I equalise them appropriately (I only have to walk next door into the studio to hear what the presenter sounds like). Whenever I watch the TV news, what I hear is a microphone with inadequate equalisation, so lacking in HF and having a bit of a midrange "honk". The reason for this, of course, is extreme pressure on budgets that mean a lack of time to worry about such subtleties - it's "good enough". But then, I have a pair of LS3/5As for watching TV - if I was using the speakers in the TV, that would be the least of my worries

Of course, one of the big problems with hi-fi enthusiasts is that they rarely listen to live music, so don't know what real instruments sound like. This lack of a point of reference certainly explains some of the more extreme speaker designs I've heard

Covering Amie's point about the expensive ECC82. Yes it is probably expensive at £50 but when compared to the latest super mega processing chip, also costing £50, that will be obsolete by September and, although mathematically technically superior, for most listeners is neither here nor there? Sales and peer pressure will win out. Companies will sell all manner of 'latest gadgets' with brief lifetimes knowing full well they can push even harder the next model. 'Audio phools', as they are often termed, are a point to extreme but there are a lot of casual / everyday users who are, in many ways, led on in exactly the same way by the latest technology and it's supposed, and often underutilised in practice, benefits.

I am not suggesting it at all but exemplifying the greater scope of the problem. To each his or her own. For me it's 'simple'. For others it's complex processing etc.

Tracy

I've been looking at the sold prices for Mullard ECC82, and there's a really wide range. I honestly wouldn't have expected to pay more than a fiver for one - hence my original response to that question - but some do go for much more. Like I said subsequently, if £50 was a realistic asking price, then I don't understand why the question was raised. So I'm not really sure what conclusions I should be drawing. No problem...

Yes, I agree that simple is best. But today, complicated is cheap.

We've been given one of those Google smart speakers. No, we're not using it

It uses a 2" drive unit, and it sounds surprisingly reasonable for that size. However, a more careful listen shows that there's a lot of DSP behind it - you can hear the multiband compression they're using to give the illusion of bass and power. But to the casual listener, it sounds reasonable enough.

The "simple" alternative would have been a bigger drive unit (and perhaps a tweeter, now that the diaphragm is larger), a bigger amplifier (hence bigger PSU), and a much larger enclosure. You could probably do away with all the audio processing, apart from some gentle EQ that would amount to a few Cs and Rs in the analogue domain. It's what I would do. But then, I'm not afraid of analogue electronics - unlike most graduates that I see today.

But of course, your product is now substantially more expensive, and less marketable because of the bigger size.

So in this example, "simple" is not "best". "Complicated" means "more sales". It's compelling, isn't it?

I'm not afraid of digital signal processing - I use digital sound desks all the time at work, and own one personally for doing live gigs. But I worry that DSP is over-used as a "sticking plaster" to fix problems that could be solved by spending the time and money where it counts. For example, a bad loudspeaker drive unit can be improved a lot by complicated equalisation, and at first glance it will sound quite reasonable. But you have to remember that peaks and troughs in the response of a drive unit are caused by resonances, and resonance implies energy storage. So even if you bung in a notch filter to tame a spike, it will still draw attention to itself, even though it appears to be completely flat on a frequency measurement plot. That's something you often need to do to PA speakers with horn tweeters, BTW. You can still see it when you do a waterfall plot, and you can usually hear it too.

Ultimately, signal processing is a tool. It can be used wisely, or not...