Key Of Grey

Photo courtesy of vissago

June 25th, 2009 – This post is a continuation of one that I wrote almost a year ago.

It’s easy to obsess over having the highest possible quality; we all want the best. In the audio world, what does that mean? We see 24-bit here and 192 kHz there and I wouldn’t doubt that we’ll see 32-bit 768 kHz or something else ridiculous in the near future.

Let me quickly break down what bit depth and sample rate mean quickly. The higher the bits, the more dynamic range you have. The generally accepted rule is a 6dB increase for every 1-bit. This means that CD audio at 16-bits has a dB limit of 96.

The sample rate refers to the amount of times a sound is sampled in a fixed amount of time. The higher the sample rate, the better the resolution. Imagine a smooth sound wave. Now imagine building the shape out of rectangles (which is what digital sampling is). The more blocks you have, the smoother the curve is. The Nyquist theory posits that you require a sampling frequency at least twice the source material to accurately reconstruct the signal. This means that CD audio at 44.1 kHz can capture an accurate signal up to 22.05 kHz. The average human hears frequencies between 20 Hz to 20,000 Hz, so the CD easily captures the frequency range an average person can hear.

So if a CD is at the limit we can hear, why record with bit depth and sample rates higher? Well, even if we can’t hear beyond a certain point, we can feel it. I, however, think that we are being overly optimistic about the whole matter. As an audio engineer, we want to record the source material as well as we can. Digital sampling rates could hit close to infinity (the analog sample rate), and I still think that engineers would be unsatisfied.

Yes, it’s nice to know that we’re working at the full potential of our gear. I don’t have to worry that I’m not getting the best performance out of what I have. But at the same time, I have a little voice in the back of my head that tells me it doesn’t matter. I should be looking for converters that give me the best sound, not the highest numbers. High numbers do not equal great sound. That little voice is right.

Having the highest bit depth and sample rate available is not the first priority. As I’ve stated before, it doesn’t matter how great parts of your signal chain are, if one is lacking; you know the old saying “you’re only as strong as your weakest link”. Who cares that a piece of gear can record at 24-bit, 192 kHz if the converters are garbage? You’ll just have great dynamic range and resolution of crap. I would take an old school, proven 16-bit 44.1 kHz converter over a crappy sounding converter that posts up 24-bit 192 kHz numbers.

Besides, you have to keep in mind that, while you’re down-sampling an analog source, your listener is probably going to be down-sampling your work further. Sure you can record in 24-bit, but only listeners with DVD-audio set ups are going to be able to tell the difference and honestly, how many of your listeners are like that? Even worse, think of all the people that are compressing your music into 128 kps MP3′s. They won’t be able to tell that you used 24-bit 192 kHz converters, but they will be able to tell that you used crappy pre-amps.

When will this game end? It seems to me that the world of audio parallels the megapixel war going on in digital photography right now. Manufacturers are rushing to fit as many megapixels onto a sensor as possible, but with the side effect of real world degraded performance. I hope that audio manufacturers don’t rush for the highest performance numbers only to skimp on the important parts of the audio equation. I guess the thing to take from this is to be careful what you buy. Don’t get caught up on numbers, buy on sound.