Common Digital Audio Myths

Like most people, I long believed that when it comes to digital audio, higher numbers meant better sound. Fortunately for us, that is simply not the case.

Myth 1: Higher Sample Rates Mean Better Sound

Audio sample rate—44.1, 48, 96 kHz, and so on—is often compared to frame rate in film or video. Everyone agrees that a video sequence at 30 frames per second is better defined than one at 12 frames per second.

With audio, however, the principle is completely different. Sample rate is tied to frequency range: to represent a frequency accurately, the sample rate must be at least twice that frequency. A sample rate of 44.1 kHz, the one used for CDs, can therefore reproduce frequencies up to 22.05 kHz accurately—that is, 44.1 divided by 2.

Now, the human ear perceives sounds from about 20 Hz to 20 kHz. That has been established for decades—and we are still looking for the freak of nature who can hear beyond that limit! And that is, of course, under ideal conditions. For a normal adult, the average upper limit is more likely around 16 or 17 kHz. So despite all the goodwill, talent, or gifts of the best sound engineers on the planet, they still have human ears—and human ears lose high-frequency perception with age.

Since the entire audible spectrum is already fully represented in a signal sampled at 44.1 kHz, higher sample rates add nothing audible or useful to the audio. They do not increase its fidelity or its quality. In fact, several studies even suggest that very high sample rates, such as 192 kHz, can be less accurate than lower ones.

Myth 2: Information Between Samples Is Lost

Another widespread belief is that digital audio does not faithfully represent what happens between samples, leaving gaps that prevent accurate waveform reproduction. Higher sample rates are often presented as a way to reduce this problem.

Once again, this is a myth. Even though sound is sampled at intervals, the position and content of the original signal remain intact. I will spare you the mathematical explanation, but if you are interested, Dan Lavry explains it in great detail in his article “Sampling Theory for Digital Audio.”

Put simply, despite the common stair-step representations of digital audio—the kind we see in every DAW—the waveform reconstructed from those samples is just as precise, continuous, and smooth as an analog signal.

Myth 3: Higher Bit Depth Means Higher Fidelity

To understand what bit depth really affects, we first need to look briefly at analog-to-digital conversion. Each sample is assigned a numerical value, and the conversion process creates a noise floor, similar in principle to the hiss produced by analog media such as tape. The distance between this noise floor and the highest level before clipping is what we call dynamic range.

Bit depth—8, 16, 24, or 32 bits—ultimately controls one thing: the available dynamic range. In other words, it determines the distance between the noise floor created by the conversion process and the highest level before clipping. As long as the signal stays within that available range, bit depth has no direct effect on the accuracy of the waveform itself.

Strictly speaking, an 8-bit waveform can be just as accurate as a 24-bit waveform, as long as it fits within the 8-bit dynamic range. That range is 48 dB, meaning everything between -48 dB and 0 dB can be reproduced accurately. At 16 bits—the bit depth of a standard CD—the range increases to 96 dB. The noise floor is so low that you would have to crank the volume all the way up to hear it.

Besides, if we compare the dynamic range of digital audio with that of analog media, we may be in for a shock. What is the bit depth of a good old audio cassette? 9 bits, if it is in perfect condition, but 6 bits is much more realistic. That is three times less than a standard CD. Yet hundreds of millions of them were sold! Our beloved vinyl record? 11 bits at its very best. The best studio tape machine? 13 bits, maybe 14 if you are willing to add saturation. (And adding saturation is the whole point of using a tape machine, isn’t it? Oh! You thought it was for high fidelity? Sorry!)

Myth 4: Set the Recording Level as Close to 0 as You Can

Once again, this is false. This method may make sense with analog equipment, but it is extremely risky in the digital world.

First, digital audio does not offer the same safety net as analog: saturation. In the analog domain, pushing a signal above 0 produces saturation, and the warmth and roundness it creates are part of what people love about analog gear.In digital audio, going above 0 produces clipping—a very unpleasant crackling distortion. In other words, going above 0 is a risk you want to avoid. Always.

As we have seen, digital audio offers a ridiculously wide dynamic range compared with analog. There is no reason to risk clipping the signal just to record “hotter.” At 24 bits, you have enough dynamic range to capture with equal accuracy a fly buzzing at the back of the room and a jet taking off inside your house!

There is another problem: many transients—the initial attack of a sound—are too fast to be displayed accurately by a meter. The actual level of a snare drum hit, for example, can easily exceed the reading shown on your meter by 10 dB.

That is why it is much safer to aim at least 12 to 15 dB below zero. At 24 bits, you can even go much lower without causing any problems. There is no need to “fill the bits” or “fill the meter,” as some people still say in the audio world. A signal recorded at -25 dB on your meter will still be more than 70 dB above the noise floor.

If you still have doubts, watch Monty Montgomery’s excellent video. He demonstrates several of the concepts discussed here using analog measuring equipment and a 20-year-old E-Magic interface!

Conclusion

The 44.1 kHz / 16-bit combination is an optimal format for reproducing audio in digital form. That said, a 48 kHz sample rate will certainly cover all your needs without overloading your hard drives. It has been a standard sample rate in digital audio production for years, and it is also commonly used in multichannel and immersive formats such as Dolby Atmos.

Along the same lines, when recording and mixing, it can be useful to take advantage of the insurance policy that 24-bit audio provides. Its low noise floor gives you more room to manage levels.

In short, 44.1 kHz / 16-bit is already enough to reproduce audio faithfully. For production, however, 48 kHz / 24-bit is the smart choice: it gives you excellent sound quality, plenty of headroom, reasonable file sizes, and full compatibility with modern digital workflows.