About Audio Standards

Last month I promised that, once again, I’d put my money where my mouth was. After sharing with you some concerns and criticisms of my approach to the issue of audio resolution voiced by several readers, I promised I’d share with you some of the audio standards I think are useful. Here goes.

Standards For Our Beloved End-User

Our beloved customers are the last stop on the resolution express, and any standards that aren’t cognizant of the conditions under which they view our product just aren’t relevant.

That said, we all too often take the position that our beloved end-users, Joe and Jane Six-Pack, have ears of tin, a TV they got on sale at Crazy Jerry’s Junky Appliances, and they live downstairs from Mel’s SlimFast Rock’n Roll Weight Room ‘n Aerobic Dance Spa. Hey, we figure Joey and Janey are lucky if they can pick out one word in ten in all the din.

That stereotype isn’t the whole truth, of course. End-users observe our work over a wide range of gear in a wide range of environments, and the best gear and environments are not only very good indeed, but they are getting better. We take aim at the middle or lower end of that big ‘ol end-user range at our own risk. However, if we really want to work to high standards, we’d better look at the high end of the end-user range (Louis et Louise Lafitte, s’il vous plait?).

And what might be the resolution values at such a high-falutin’ point on the range? Figure a 90 dB acoustic dynamic range (20-110 dB SPL). Figure 10 Hz. (a modern subwoofer, y’know) to 20 kHz. bandwidth. Figure low-distortion, low jitter playback gear. Figure five matched full-range channels in addition to a really decent subwoofer.

Maybe THAT’s the real playback standard we should try to deliver recordings for. There may be systems that do better than that, but they can’t be MUCH better. If you can make recordings that sound great over such a system, they are probably going to sound great most places (except for the occasional odd-ball recording that ONLY sounds good on bad systems). So what technical audio standards do WE need to get there?

Acoustical Standards

We need studios and sound stages that have a noise floor below NC-25 WHILE THEY ARE IN OPERATION (NC-25 is a noise floor tilted downward from 53 dB SPL at 63 Hz. that is at approximately 27 dB SPL at 1 kHz. and 20 dB SPL at 8 kHz.). This means that the noise emitted from cameras, dollies, monitors, amps, lights, HVAC, etc. must be kept to levels that are essentially inaudible.

IF we are going to be able to fully HEAR the recordings we make under these conditions, we need control rooms that have a similar operational noise floor. This is a TOUGH standard to meet, because it means silencing ALL of the mechanical gear (including fans) that live in a control room. Without everything isolated in an adjacent machine room, you just can’t hardly do it.

Further, there is a geometric requirement for the control room: it must have excellent lateral symmetry (including furniture), plus a well-maintained and viable median plane that is shared by the loudspeakers, listeners AND the room. Finally, the room and the loudspeakers must be configured to minimize low-frequency modal problems.

Resolution and Bandwidth Standards

We’d like our analog audio window to exceed the acoustic audio range, so it’d be nice to have 100 dB dynamic range (-80 to +20 dBV) and audio bandwidth from 10 Hz. to 20 kHz. This is trivial for the electronics, but close to or at the limits of technology for microphones and loudspeakers.

Interestingly, our 16-bit, 44.1 kHz. digital audio standard just about lines up with this window. There’s a strong argument to be made that 20 bits (120 dB dynamic range) is actually useful and I’ve speculated in my book, Total Recording, that an IDEAL system might need a resolution of 21 bits (126 dB).

At present, we have no significant evidence that frequencies higher than 20 kHz. are useful. On the low-frequency end of things, 10 Hz. is easy for electronics, but difficult for loudspeakers.

Coupled with these physical standards for resolution, we need to maintain linearity (very low distortion) across our ranges of amplitude and frequency. Further, we’ve found that similar resolution ranges for short-term time errors (jitter, wow and flutter) are also necessary. And, as Yeager eloquently pointed out, we may need greater resolution in some of our production tools, simply to head off problems with those particular tools.

Production Standards, Our Peers and Us

In production, the maintenance of such standards, through meticulous working process, documentation and system/facility maintenance, is essential. It is not sufficient to THINK we have high standards, or simply to have bought and paid for hi-rez gear – we need to be able to PROVE the active and on-going existence of those standards. Because audio and video production is a team sport, we need to include a sufficient proof-of-performance record with our work so that each person in the production path can verify the quality of the work that has gone before as well as his or her work for future users. This is in addition to proof-of-performance data that (hopefully!) we periodically collect and archive as part of the maintenance operation.

This means test tones and related source data recordings plus documentation at a performance level that just isn’t practiced very well right now. In general, the nature and relationship of various digital, analog and acoustical signals and signal levels is not well understood, documented or communicated by production people. We have little sense of the on-going erosion, the accumulation of errors that happens as an inevitable consequence of our kind of recording work. We have little sense of how to even measure such erosion and errors, let alone how to minimize it. Even when we do, we are usually under way too much deadline pressure to be able to do the kind of meticulous production-path maintenance that is at the center of truly high-quality work.

What Standards AREN’T Good For

At the same time, we’ve got to keep the obsession in perspective. The audio production chain, like any other, is only as strong as its weakest link. We cannot certify the high quality and resolution of our work by asserting that we use a 24-bit/96 kHz. digital recorder. The use of hi-rez buzzwords buys us nothing, except to suggest that there may be ONE place in the chain where we didn’t add too much error to the process. Only when we begin to specify SYSTEM minimums can we begin to claim those standards as meaningful and relevant to our work.

The Meaning Of It All

I particularly like what Bob Dixon said. “When the drive for excellence is combined with common sense, it is the most practical action of all. The issue is focus. A focus on excellence will chip away at [flaws in the] final product presented to the public. How good does it have to be? It isn't just the question of 16 bits being ‘good enough’. Our focus should be on ‘how good can we make it?’” Thanks, Bob.

Standards are about setting the minimum “goodness” that we, as professionals, will accept. Having high standards means having a “minimum” goodness that approaches the “maximum” goodness we can obtain. It ain’t easy. High quality hardly ever is.

Thanks for listening.

And that just about covers it . . .
Thanks . . .