Your browser is out-of-date!

Update your browser to view this website correctly. Update my browser now


Spring Cleaning


It may be July where you are, but magazine lead times being what they are, the tulips are just coming out as I write this. Thus, I am compelled to do some spring cleaning, which in magazine-speak means filling up this space with odds and ends and follow-ups that I’ve been trying to work into a column but haven’t yet figured out how. Until now. So, a few scraps from my desk.

My April column about problems with Apple’s OS X drew a lot of response from readers, which was about evenly divided between “Tell it, brother!” and “Paul, you ignorant slut!” It also garnered a couple of solutions to the problems I presented. Steve Olenick told me about Cherry Picker, a clever piece of “donation-ware” that lets you edit and convert synthesizer patch name files and have them show up in Digital Performer, without all the mess and nail-biting involved in tweaking XML files. It will work with old OMS and FreeMIDI files, and let you audition patches in connected synths using MIDI sequences or loops. It’s a little fussy and doesn’t solve the problem of getting patch names into, say, Pro Tools, but it’s a step in the right direction.

Jeffrey Henning wrote, “To my knowledge, the Oxford 911 FireWire 400 bug was taken care of over a year ago with a firmware patch for the enclosures.” That’s true for some of the larger drive manufacturers such as LaCie, but my drive comes from a smaller company, AcomData, and there are no firmware updates for OS X — there are some for Windows, of course — on its Website ( I’ve written the company about this and have yet to receive an answer. Another reader said that there is a “912” firmware upgrade available for these drives, but the 912 firmware is only for FireWire 800 drives.

But there’s a whole other side to this Apple thing that I left out of the April column for risk of sounding like a bit of a crybaby. Now, however, the problem has been widely reported on in other publications. (For example, there was a great piece about this in the May issue of Wired called “Think Belligerent.”) What I’m talking about is how much Apple has shut itself off from the press just in the past couple of years.

Compared to most high-tech companies, Apple’s relationship with the people who write about it has always been open and supportive. Writers have been able to talk to and quote company engineers and programmers about problems and plans, and it has given readers the idea that this is a company that cares about, and respects the intelligence of, its customers. Just a little more than two years ago [February 2003], when I wrote my first Mix column about OS X, I had a long conversation with an engineering manager in Apple’s audio group, who offered many cogent thoughts on why things were the way they were (i.e., not very good) and optimistic expectations for the future, most of which came true.

But that engineer is now gone, and even though I still know people in that group (some of my former students, for example), I found that when I started researching April’s column, none of them would talk to me. Nowadays, all press inquiries, no matter whom they come from or what they’re about, have to go through the company’s PR office — and most of them apparently die there.

So I was never able to get any answers to my questions or even any information on how some of the problems I outlined in my column might be addressed in Tiger. Of course, this may be because none of the problems are addressed in Tiger: Of the “over 200 new features” the company is touting, the only one of any interest to audio pros is device aggregation, which means that you can have multiple audio interfaces active at the same time. A nice idea, but not if the machine can’t find any of them when it wakes up. Maybe they’ll fix that with the next OS release. But you didn’t hear it from me — or anyone else.

As I was working on my May column on how we listen, I wanted to include something about the deteriorating level of cabling in the consumer audio world, focusing on one particular parameter: metal-to-metal contact area between connectors. But I ran out of space.

Compare the predominant means of connecting components in the various parts of the audio industry. At the top are XLR connectors. Wonderful things: They lock together positively and reassuringly, offer three well-insulated current paths and provide strain relief for all three conductors. Each pin is 20.7mm long (all of these measurements are from Switchcraft’s Website,; there are small variations among different types and manufacturers) and 2.4 mm in diameter, making for a surface area (length × diameter × µ) of 156.1 mm2 per pin, or 468.3 mm2 total contact area.

Come down to the musical instrument and “prosumer” level. These are largely connected with ¼-inch plugs and jacks. These components lock pretty well, although they can usually be yanked apart given enough force, and only the most expensive provide sufficient strain relief that the cables can’t be pulled out of the connectors should you trip over them. Wiring schemes vary, but in general, the insulation between conductors is only as good as the cable itself and often not that good: Shields are generally pinched around inner conductors, and when you’re using balanced ¼-inch plugs, the terminals for the inner conductors are quite close together. The plastic inserts between the tip, ring and sleeve provide adequate insulation, but they can break if the plug gets abused. The ¼-inch plugs are 30.16mm long and 6.35 mm in diameter, so they have a maximum contact area of 602 mm2. Take away the insulators, which take up maybe 25 percent of the total in a TRS configuration, and you still have more than 450 mm2. Pretty similar to XLRs.

Go down to the “hi-fi” world and you’re dealing with RCA phono plugs. Nasty things to work with, prone to oxidation problems because of the way the plug’s shield wraps around the jack, and with little space to insulate the inner from the outer conductors, these beasties nevertheless work okay if they’re really well-constructed. Otherwise, they short out, the hot lead deteriorates and they snap off. But in terms of surface area, they’re not too bad: High-end versions of the plugs have a pin that’s 15.9mm long and 3.18 mm in diameter, giving a contact area of 159 mm2. The barrel is 6mm long, with an inner-facing diameter of 9.5 mm, providing 180 mm2. So the total surface area is 339 mm2, more or less, which is respectable.

But sales of all these connectors — if what I find myself buying at Radio Shack these days is any indication — pale in comparison to the new king of audio: the ⅛-inch miniature stereo phone plug. Once used only for Walkman headphones, these little suckers are everywhere now: CD players, iPods, MiniDisc recorders, “multimedia” speaker systems, airline seats, computer and soundcard inputs and outputs, and even some synthesizer and processing hardware. Every time I look around, I need another ⅛-inch cable or ⅛-inch-to-double-RCA adapter. I must own a dozen of the latter, but I keep running out of them.

Of course, a lot of that has to do with their high failure rate. The cable is absurdly flimsy: The shields comprise microscopic copper threads interwoven with fabric, and a good yank invariably breaks the copper at some spot that you can’t get to. The plugs are cheesy and plastic, and the conductors are way too close together. Should you happen to step on an end, you can just throw away the cable. Differing dimensional standards among manufacturers don’t get terribly in the way when you have room to wriggle — as you do with ¼-inch and RCA plugs — but when the plugs and jacks don’t quite line up with something this tiny, you get drop outs, buzzes and crosstalk. And contact area? Well, here’s the math: Overall length is 14.29 mm and the diameter is 3.175 mm, for a total of 142.5 mm2. Take off 20 percent for the insulators, and you have about 114 mm2, a little less than one-quarter the area of an XLR. And that’s for two channels, not one.

Doesn’t that make you feel better?

Every semester, I get a new crop of students who tell me how great Web audio sounds. They insist that this year’s compression algorithms are way better than last year’s, and at 392 kbps, they just can’t tell the difference between an MP3 and a CD. And every semester, I go onto various sites and listen to the latest uploads and don’t change my mind. Web audio has without doubt reached the level of the analog cassette: good enough for your car and maybe to have on in the background, but I wouldn’t want to listen to it on the big speakers at home for more than about 15 seconds at a time.

I recently put together a mix CD for a reunion party with 180 or so friends whom I hadn’t seen in 30 years. I had a great time digging up ’60s and ’70s music from various sources that I thought would be meaningful to this crowd. There was stuff I had on CDs, which sounded terrific when the disc was well-mastered and horrible when it wasn’t. There was stuff I only had on vinyl, so I fired up the old Technics turntable and digitized them, throwing in a little crackle filtering where needed and rolling off below 80 Hz to get rid of the rumble. But some of the music I wanted to include I didn’t have on disc and so I resorted to searching it out on the Web. When laid side-by-side with the other formats, even the highest-bit-rate MP3 and AAC versions I found paled in comparison. Neil Young’s “Cinnamon Girl” sounded far more anemic than spicy, while The Temptations’ “Ball of Confusion” put me more in mind of something you’d whack around with a tennis racquet.

It reminded me of a wonderful little test that engineer and educator John Monforte came up with when he was teaching at the University of Miami, which will show anyone, even with the tinnest of ears, what’s wrong with compressed audio. I’ve been using it on my students, and afterward, they don’t bother me about the subject again — until the next semester.

Here’s how it works. Take a well-recorded CD, one with a good, wide frequency spectrum and plenty of dynamic range. Rip it onto your computer as a mono linear PCM .AIFF or .WAV file. Make a copy of that file and “flip the phase” or, more accurately, invert the polarity. Take the inverted copy and merge it with the original. What’s the result? Nothing, of course.

Now take the original mono file and convert it into a lossless AAC file using your favorite audio editor. Save that and then convert it back to linear PCM. Invert the polarity and combine it with the original. You might see a couple of artifacts on the screen, but play the result and, again, you’ll hear nothing as lossless really does mean lossless.

Now take the first mono PCM file and convert it to an MP3 at the highest rate and best quality your software will allow. Save the MP3 and then bring it back into your audio software so that it up-converts it to PCM. Again, invert the polarity of this file and merge it with the original. Now listen, and you’ll hear a fairly robust, if somewhat subdued and frequency-limited, rendition of the original. But remember that this is the difference between the two signals: This is what the compression process takes out of the audio file. Case closed. Thanks, John.

And finally, about three years ago [March 2002], I did a column on my experience with tinnitus (it’s still under control, thank you) and age-associated hearing loss, and I was contacted by a company called EarQ Technologies (not to be confused with the hearing-aid company called “EarQ”). The company was selling an audiological self-testing system that could be used to custom-tailor a room’s frequency response to an individual’s own threshold response curve so that the result would be, from the individual’s perspective, “flat.” The company’s slogan was, “Fit your mix to your ears.” The system had received a favorable review in Mix and other plaudits from the musical and scientific communities, but these write-ups were strictly about how the product made it easy for people to monitor their hearing loss without having to go to an audiological clinic. They didn’t address the issue of whether it could fix one’s mixes.

I don’t want to beat a dead horse as the system is now out of production, but I do still get asked about it periodically. So I think it’s worth sharing what I’ve been able to find out about the concept. According to Harvard Medical School audiologist Dr. Chris Halpin, who gave me my hearing test and is a former musician himself, the EarQ system can’t work that way. “An elevated threshold for a tone does not imply abnormal loudness above the threshold for that frequency,” he writes. “In sensory hearing loss from age or exposure to noise, the apparent loudness does not shift with thresholds. You just lose the quiet tones.

“The loudness grows abnormally fast from threshold back to normal perception so that a loud sound, like 90dB SPL, for someone with normal hearing is the same loudness for someone with a threshold loss of as much as 60 dB. This is known as the ‘recruitment’ phenomenon and also happens in other sensory systems.

“To adjust a room or mix EQ based on thresholds is not a valid concept in sensory hearing loss. Instead, you should be thinking that when it sounds good to you, it will sound good to the audience.” But you already knew that.

While rummaging, Paul Lehrman also found an old box full of Mad magazines, so you may not be hearing from him for a while.