Moulton Laboratories
the art and science of sound
Taking Stock: How Important Is High Resolution Audio, Anyway?
Dave Moulton
September 2000

An epic discussion of high resolution digital audio.
Dissonance Resolved Records
Inspiring Music, Transforming Souls
www.dissonanceresolved.com
iZotope
Audio processing technology, tools, and plug-ins for Mac & PC
www.izotope.com
Total Recording
Comprehensive guide to audio production and engineering.
www.kiqproductions.com
< 1 2 3 4 5 6 >

Part II: Why All The Fuss?

October 2000, TV Technology

As We Were Saying . . .

Last month I had the temerity to suggest that 16-bit audio works very well, thank you, and that any reasonable consideration of the ACTUAL differences between 16-bit, 24-bit, and pure analog audio signals will reveal that those differences are going to be REALLY small. So small, in fact, that they are very difficult to distinguish and mostly, we won’t be able to.

Why have we rushed, willy-nilly, to embrace all these extra bits (and extra bandwidth), especially given the attendant costs and problems? It’s a puzzler, all right. I promised this month to discuss why there is such a fuss about hi-rez signals.

The simplest part of the answer is this: we’ve added all this extra resolution in the digital realm because, well, we can. And, happily, we can now afford to, as well, thanks to Moore’s Law. In doing so, of course, we’ve fooled ourselves. We’ve pretended that digital audio is a separate, independent system, that it stands alone as a parallel realm, and that it can be better than analog. And that idea, as I mentioned last month, simply isn’t true. Let’s take a closer look.

Digital Audio As Subsystem

Digital audio is a subset of audio that exists within the confines of analog audio production. To get to digital, we must go through analog, and to get out of digital we must go back to analog. We have no Acoustical-to-Digital or Digital-to-Acoustical converters. Therefore, the audio signal that exists in the analog realm, prior to conversion, is as good as it gets. The digital signal can’t be any better, only less worse. And as I mentioned last month, we’d like it to be enough “less worse” than analog that we can’t tell the difference.

But the point remains. Digital can’t be better than analog. Analog quality defines the upper limits of overall signal quality.

Now, the limitations in the analog realm are pretty fundamental. On the soft end of things, we have noise and hum. Thermal noise provides a bedrock bottom limit of about –126 dBV and the accumulated noise and leakage artifacts that are present in all active analog circuitry constrain us to a noise floor of approximately –110 dBV at the very, very best. On the loud end of things we are constrained by the limits of the power supply, which under the best of circumstances yield a limit of approximately +24 dBV. Taken together, these yield a best case dynamic range of 134 dB.

There are two things to keep in mind about this. First, this is a BEST CASE range, and the REAL dynamic range of a signal will be determined by the range between the SMALLEST power supply voltage and the NOISIEST signal stage the signal encounters during its flow from being an acoustic signal to an electronic signal and back to an acoustic signal. That REAL WORLD range is typically about 70 dB at best. Second, it is extraordinarily difficult AND expensive to increase that range. It takes some REALLY meticulous signal handling and levels management, some serious architectural work, some really anal-retentive behaviors plus a serious wad of dough (figure a couple of orders of magnitude more than we usually spend – instead of $10 grand for a small system, figure on a , for instance!). It ain’t easy, in other words.

Meanwhile, in the digital realm, we can accomplish all this comparatively cheaply and easily. All we gotta do is add more bits. Instead of budgeting sixteen ones ’n zeros for each chunk o’ audio data, we simply budget twenty-four ones ’n zeros. Costs a little, to be sure, but in comparison with the pain encountered in the analog realm, why, it’s hardly even a pin prick!

The Emperor’s New Stereo

So all we have to do is pretend that digital audio can be better than analog audio. Once we’ve inhaled that particular mindset, why, then we can have incredibly enhanced audio. 24-bit is totally awesome! We’ve expanded the frontiers. We can now have a level of resolution that has never before been heard by man! All we’ve gotta do is keep adding bits. 33 bits exceeds the theoretical limits of air itself! Holy molecules, Batman! 33 bits has gotta sound terrific! How about let’s try 40 bits (that’s about 240 dB of dynamic range)!!

See how easy it is? Throw more bits at the problem and it looks like you’ve solved the problem once and for all, so long as you ignore the physical realms of reality. For an engineering staff, it’s really tempting. For an advertising department, it’s, well, absolutely irresistible!

And who among us is going to admit that he or she can’t hear that “huge” digital difference? Not me! Nosiree Bob! So long as JUST ONE of us murmurs that some 24-bit box “seems to sound a little better than” some other 16-bit box, we’re all gonna sign on for the ride. We can’t afford not to! To admit that we cannot perceive a difference that others have pronounced “sounds better than” is out of the question. It’s professional suicide, plain and simple.

Bits’n Bandwidth For Fun and Profit

And that’s why there’s all the fuss.

We have created a 24-bit/96 kHz. digital audio window that is larger, by some amount, than the analog window that it supports. We like to believe that we can distinguish that larger window even though it is obscured by the smaller analog window, which in turn is obscured by the much smaller acoustical window we normally encounter.

We do this in the search for signal quality, which we equate with accuracy. We know that there is stuff wrong with our recordings, that the illusion is less than perfect. And so we tackle the easiest parts of it, which turn out to be in the digital resolution realm. And we keep expanding that resolution because it is so easy and cheap, and none of us are really gonna say, out loud, “No. This doesn’t matter. What we really need to work on is other parts of the problem, parts that aren’t gonna be so easy, or so cheap.”

And those of us who do know this for a fact, have learned to keep our mouths shut, play the game, and work on the other stuff, knowing that nobody’s gonna figure out the secret to our successes. They’re too busy worrying about bits!

Next month, we’ll come clean, tell you what we can REALLY hear. Uh-oh! Should be fun. Refill the Prozac prescription.

Thanks for listening.

< 1 2 3 4 5 6 >
Members
Login | Register
Mailing List