Jun 18, 2011
Not numbers, not measurements, not gear – it’s just about great sound
I have a Google alert set up to find people talking about the loudness war. Every so often it turns up a blog post that’s worth tweeting, sometimes it turns up a forum discussion that’s worth contributing to.
But more often than not it turns up things that frustrate me. The top contenders are: people trying to justify the bad sound of their favourite act’s last CD by saying it sounds better in the car/iPod earbuds/wherever; people claiming that we actually prefer the sound of crushed, limited dynamic range music (research says we don’t) or the classic “you can only hear how great ‘Death Magnetic’ sounds on really good equipment” (!)
What frustrates me most though, is people using numbers and waveforms to try and “prove” that something sounds bad, when it doesn’t. You can’t tell if something sounds bad by looking at it, or measuring it, you have to listen to it.
And before I go any further – of course I’m aware of the irony of me saying that.
I use numbers to talk about music all the time. I’m forever quoting my rule-of-thumb about not mastering with a dynamic range of less than DR8, and the waveform images I used to illustrate the clipping distortion in Metallica’s Death Magnetic were re-used by everyone from Wired magazine to the Wall Street Journal.
But here’s the point – I use those technical details to illustrate visually what I hear. I didn’t get hold of a copy of Death Magnetic, look a the waveforms and write an outraged blog post – I listened to Death Magnetic, thought ‘holy crap’ and then took a look at the waveforms.
And the value of DR8 isn’t pulled out of a hat, it’s my personal judgement, based on 15 year’s experience as a mastering engineer. Whenever I consistently try to push something past DR8 the sound starts to suffer. And I’ve never heard an album with an overall dynamic range of DR6 that I didn’t think would have sounded better if it were DR8.
A classic example if this “judging by numbers” is on the Dynamic Range Database site, in the notes for Elbow’s ‘Build A Rocket Boys’, which was the album we chose to give the Dynamic Range Day Award to. The notes say:
This record won the “Dynamic Range Day Award 2011″ – rather ridiculous !
Was it ridiculous ? Well, if you only look at the numbers, you might think the person who made that comment had a point – after all, Track 1 measures only DR4 – it must be crushed and lifeless beyond belief, right ?
Wrong. It starts out soft and gentle, and gradually builds to a thundering climax – it has great dynamic range.
So why does it measure only DR4, then ?
- Because the end is very loud, and the TT meter pays more attention to the loudest moments of a track
- Because towards the end, there is a huge sustained organ sound high up in the mix. Organ or flute sounds – and in fact long sustained notes on almost any instrument – naturally have a limited dynamic range. They aren’t changing much, so they’re not very dynamic
- Most importantly of all – DR4 is just a number
It’s the attempt of a machine to estimate our bodymind’s reaction to the effect on our eardrums of the movement of a bunch of air molecules moved by a rigid cone in a box attached to a magnet powered by electrical current amplified by some electronics controlled by a computer chip based on the numerical input of a stream of digits read by a laser from a microscopically pitted piece of very thin aluminium foil encased in plastic…
That’s a lot to ask of any number, and not surprisingly, it’s not very accurate !
Even something more complicated like a waveform can’t tell us much more – there are so many factors that determine the way a waveform appears – horizontal and vertical resolution and the measurement scale used, for a start.
Does that mean the TT meter, DR measurements and waveforms are useless ?
Even that simple sequence of DR measurements for ‘Build A Rocket Boys’ can tell us a great deal – for example, there’s lots of variety. So even though some of the songs measure as having quite a limited dynamic range, others have much more – and, there’s variety from song to song. That means – dynamic range ! Even if I hadn’t heard it, my guess would be that it should sound better dynamically than an album where every song measured 6 – and I’d be right.
So does that mean that ‘Build A Rocket Boy’s is the most dynamic record of 2011, despite the numbers ? No, of course not. That wasn’t the point of the award – we just wanted to celebrate a mainstream release with great dynamics. As usual I think it could have had a couple of extra dBs more dynamic range and sounded even better.
But I still think it was a great choice. If you want to make a complex issue like the loudness wars straightforward to understand by a wide range of people in a limited number of words, you need a few simplifications and rules-of-thumb. I came up with the idea of DRD to promote the idea that dynamic music sounds better, but I’m not going to start pretending that things don’t sound great just because the numbers aren’t as big as they could be.
And of course dynamics are only one part of great sound ! There are so many other factors – the writing, arrangement, performance and mix – everything we talk about here all the time.
So what am I saying, at the end of all this ?
It’s all about great sound.
Not numbers, not waveforms, not dynamics. This site is full of rules-of-thumb and numbers – compression ratios, EQ settings and guidelines, but at the end of the day, you just need to listen, use your ears, make your own choices, and aim for great sound.
Image by Roger Smith