How do we measure fidelity? Of course, we all like to think we know it when we hear it, right? And some of us…my self included…like to validate our ears with some science. I like to grab files and do the spectral analysis, check the file for clipping and observe the “characteristics” of the graphs (way too many files seem to exhibit things that are acoustically impossible).
I’ve routinely focused on frequency response and dynamic range as the two most important aspects to sound. Using Pulse Code Modulation, sample rate and word length establish the ability of a recording/reproduction system to capture and recreate the frequencies and amplitudes associated with a selection of music. But is there more?
Yes. There is phase, which results from timing variations as components of a sound (or an aggregate sound) reflect off walls and surfaces prior to reaching your ears or a microphone. Phase is a critical aspect of sound and does “color” or affect our perception of a sound source but it is not in the same league as the previous parameters. That’s why some speakers have “time aligned” drivers and why you can inform your A/V Receiver about how far the speakers in your 5.1 surround system are from the center location.
So how about “depth recreation, dimensionality, inner detail, and lack of electronic grain”, which I quote from an article over at Audiophile Review? Steven Stone’s article on “What’s Good Enough” tries to convince “good enoughers” that there are other elements to fidelity that “are all difficult, if not impossible, to quantify”. This in defense of the merits of high-resolution audio.
The gist of the piece is to open up a list of audiophile traits that exist outside of normal acoustics…the science of sound. As a believer in the benefits of high-resolution audio, I should be right there with him…but I’m not. Not for the reasons that he and plenty of others espouse.
He states, “The most important aspect of higher definition music (above 44.1/16) is its increased inner detail and low-level definition compared to the same material at 44.1/16”. I was asked about this notion of “increased inner detail and low-level definition” by a reader earlier this week. My response was that it’s audiophile self-delusion or inaccurate use of terminology. I’m especially dubious when the graphic that Steven used to accompany the statement is flawed (not that I haven’t similar illustrations myself to demonstrate the notion of word lengths and their effect on dynamics. Thanks to John Siau of Benchmark for setting me straight).
Here’s the diagram showing the effects of “low and high bit depth” on fidelity from Audiophile Review:
Figure 1 – A common way to represent sampling using “stairstep” graphics to represent each sample. [Click to enlarge]
Here’s the graphic remade showing the momentary values of the samples. These samples are NOT steps but rather points along a curve that is rebuilt during the digital to analog conversion process. If the job is done well, there are no discrete values or “stairsteps” in the output.
Figure 2 – Another and more accurate representation of sampling using low and high bit depths (shorter and longer word lengths). [Click to enlarge]
We get what we get from the sampling systems (either DSD or PCM). They capture sound and music and recreate it using converters that have specifications for sample rate and word length. The resultant audio doesn’t contain attributes like, “depth recreation, dimensionality, inner detail, and lack of electronic grain”. There is good fidelity to the original or not. Of course, there are lots of approaches to recording involving microphone placement, choice of mikes etc. These do influence the quality of sound but no engineer I know says, ” If I place a microphone here or use this piece of equipment, I’m going to reduce electronic graininess”.
I admit it’s tough to write about the qualities of a wonderful recording. I try to avoid making statements like those but I do understand we have to go there sometimes. But these are expressions that try to relate an emotional or other intangible feeling…they are not acoustic qualities. And yes, they are not measureable because they aren’t science.
Stick with acoustic realities and you’ll be far better off.