Shopping Cart
Your Cart is Empty
Quantity:
Subtotal
Taxes
Shipping
Total
There was an error with PayPalClick here to try again
CelebrateThank you for your business!You should be receiving an order confirmation from Paypal shortly.Exit Shopping Cart

John Sciacca Writes...

Features, Reviews and a Blog by John Sciacca

Random Thoughts (Blog)

Random Thoughts (Blog)

JAES: I don't care WHAT you say! I say it's BETTER!

Posted on September 7, 2010 at 3:50 PM

(I’m warning you all up front; this is going to be a good bit drier and more technical-ey than my typical posts. So, if a normal blog post of mine is like a warm, sweet and gooey Cinnabon covered in icing that hurts your teeth so good, this is going to be like a piece of unleavened bread. With no salt. So get some red wine and brie ready, cause you’ve been warned.)

 

A former Sound & Vision (back in the good ole ampersand days before the move to the “+” symbol we all so know, fear and respect) editor and all around audio know-it-all (but in a good way; not that kind of smarmy, I-know-more-than-you so pipe down and listen while I pontificate on quantization theory and bit-depth resolution the way that some industry “experts” try and throttle the fun out of life with), Michael Riggs sent me an e-mail the other day.

 

He asked me about Vizio TVs. Which, was actually pretty straight-forward and boring and wouldn’t make an interesting story at all.

 

But THEN he sent me another e-mail which was quite fascinating. It was a reprint of an engineering report from the J. Audio Engineering Society (Vol. 55, No. 9, 2007 September if you want to keep your bibliographies up to date.) This journal uses terms like "Simulated Annealing Optimization" and "Parametric Spring Reverberation Effect." You know, a little light reading for when you've just GOT to take a break from Stephen Hawking's latest page turner. It also referenced the Boston Audio Society, who actually picked up one of my stories. (The fact that they picked up one of my posts is entirely all the more shocking after reading this engineering white paper which reads quite a bit like an actual engineering white paper, and not at all like my story which used the word “sucky” in the title. I guess even the high-brow, blue bloods like to stoop down for a chuckle with us commoners every now and again.) 


The article’s title was “Audibility of a CD-Standard A/D/A Loop Inserted into High-Resolution Audio Playback*” and basically put forth the following assertion:

Claims both published and anecdotal are regularly made for audibly superior sound quality for two-channel audio encoded with longer word lengths and/or at higher sampling rates than the 16-bit/44.1-kHz CD standard. The authors report on a series of double-blind tests comparing the analog output of high-resolution players playing high-resolution recordings with the same signal passed through a 16-bit/44.1-kHz ‘bottleneck.’ The tests were conducted for over a year using different systems and a variety of subjects. The test results show that the CD-quality A/D/A loop was undetectable at normal-to-loud listening levels, by any of the subjects, on any of the playback systems. The noise of the CD-quality loop was audible only at very elevated levels.

 

To boil this down, Mssrs. Meyer and Moran, the authors of this potential audio heresy, are claiming that CD audio sounds as good as we need, and that there is absolutely no audible difference between CD quality audio and high-resolution (which is routinely 24-bit/96 kHz) audio. Why should you care? Well, you probably shouldn’t. You probably can’t even play back these files anyhow. In fact, you’re probably wondering if you ever did run across a high-res file how you would convert it into a lossy MP3 so you could fit more of it on your iPod.

 

The more important question we all should be asking here is, “Why should I care?” (I, as in me, not “I” as in you are reading this to yourself in the first person.) Again, this is John’s blog, so of course that is always the ultimate query.

 

I’ll tell you why. I now feel like my ears have been lying to me recently. I have trusted these skin flaps on the sides of my head, and when they say, “Hey! Someone is yelling at you!” I turn my head. When they say, “Hey! That speaker in the theater is blown!” I complain. And when they say, “Hey! We like these high-res audio tracks! They are like a soothing balm to the white noise which is your life.” So if Meyer and Moran say there’s no difference, all I can wonder is…WTF?! Ears, so help me if you’ve been lying to me, I’m going to go back to destroying you with Q-tips and in-ear headphones!

 

Recently I’ve been listening to a ton of high-resolution audio tracks on the Sooloos music server that I’m reviewing. (You can read some more of my high-res love in here.) I’ve downloaded several albums at the largesse of Mr. David Chesky over at HDtracks.com. I’m enjoying Diana Krall, a really cool, hip jazzy duo called Dave’s True Story, Peter Gabriel, some Coltrane, and my old standby Rebecca Pidgeon. (I was REALLY hoping to get some great comments from David Chesky on Rebecca Pidgeon. He has worked with her several times, so I tried to prod him into a really juicy e-mail by sending him this: “It appears that you have met and worked rather closely with Rebecca Pidgeon. Can you share any insights on what she is like? She seems like a really interesting person. I loved her on the TV show The Unit...she was so icy as the wife of the colonel plotting her own rise through government. Plus, she's beautiful! Would love to hear your thoughts on working with her and how she is and about her personality.” Mr. Chesky is clearly a man who is not a fan of the jibber-jabber, and – like his audio recordings – keeps noise levels to an absolute minimum. His reply was simply, “A very nice and polite person. Works hard and is no nonsense. Very dedicated to her craft.”  I now feel like I know Rebecca well enough to call her Becs and then ask her over to listen to some music -- high-res, of course -- and drink her favorite beverage, which I’m picturing is a glass of red wine.)

 

Now, out of 554 trials, the high-res file was only correctly identified 276 times, or 49.82% of the time. Now, I don’t really give two hoots or a holler what the vaunted JAES says, I knows what I knows and what I knows is that they sound better. Now is this because I’ve installed Bionic Hearing and I’m hearing things that only dogs, LA Class hunter-killer class sub sonar men and certain breeds of Andalusian cave bats can? (Sadly, no. Living the life of Lee Majors super-hearing is still beyond my grasp.)

 

Clearly, JAES just polled a bunch of slack-jawed yokels who were trading their time listenin’ to some musics in exchange for a free cookie and glass of Tang. But, no! JAES sucker-punched this line of logic by saying that “audiophiles and/or working recording-studio engineers” – who made up 467 of the trials – scored only marginally better; scoring a barely better than the I-can-hears- better-when-I-picks-the-corn pone-outta-ma-ears with a 52.7% success rate.  (More surprisingly to me is that females sucked le ass on the test, scoring a horrible 37.5%.)

 

So have my ears betrayed me? Not exactly. The report – which from herein shall be forever referred to as The Report – has got a brother’s back here that I’m not totally buying into the 24-bit Kool-Aid. (Which is actually a good bit more expensive than regular 16-bit Kool-Aid.) The Report says that despite the fact that our ears are little better than chunks of lead at hearing the higher resolution, “Virtually all of the SACD and DVD-A recordings sounded better than most CDs—sometimes much better.” Why? Because “engineers and producers are being given the freedom to produce recordings that sound as good as they can make them, without having to compress or equalize the signal to suit lesser systems and casual listening conditions. These recordings seem to have been made with great care and manifest affection, by engineers trying to please themselves and their peers.”

 

So, forget that it comes on an SACD or DVD-Audio (actually, it’s probably a good idea to forget about those formats due to the fact that they are mostly defunct anyhow. We’ll have to wait like 20 years for them to make an LP like resurgence along with Beta.) or – way more likely – FLAC. The point is that these albums DO sound better. Whether it’s because of the mythical extra bits and resolution or because the engineer worked his fingers to the nub to suck every last droplet of audio-ness out of them, who cares? Better is better. And more importantly, The Report says that women don’t hear as well as men. So when Dana says she tells me something and I say, “I didn’t hear you,” that means she probably didn’t hear herself either! Vindication!

 

If you want a further education as to why all that extra resolution is unnecessary and possibly even contributing to the destruction of the Ozone layer, illegally harvesting whale blubber for cosmetics companies and making Baby Jesus cry, here is Michael Rigg’s excellent e-mail to me on the subject:

 

Audio is funny that way. Of course, all of Chesky's stuff sounds great because they care so much about the sound and know what they're doing. But theoretically, 44.1-kHz/16-bit PCM should be completely adequate, if not by a large margin. This is not widely understood largely because guys like me did such a crappy job of explaining digital audio when it first came along. A higher sampling rate just extends the bandwidth farther beyond the range of human hearing. (Sampling is often described as chopping up the signal, but it really effectively adds a lot of ultrasonic energy, as witness all those sharp corners in the explanatory diagrams. If you take the output from the sample-and-hold circuit at the front of an ADC and run it through a steep low-pass filter with a corner at half the sampling frequency you will get back exactly the original input waveform, with some phase shift that can easily be undone if desired.) Quantization, on the other hand, is not totally lossless, but the effect is to create a very small uncertainty in the level of each sample. This is really no different from what more traditional noise sources do -- different mechanism, same result. The net is that a well-executed, properly dithered digital channel is not readily distinguishable from an extraordinarily clean and precise analog channel.

 

Higher sampling rates and longer words don't hurt (except in storage requirement, which isn't much of an issue anymore), but they don't help, either. (Actually, the whole 24-bit thing is kind of a sham, since a true 24-bit resolution converter would have to run in liquid nitrogen or something to get the thermal noise down. Last time I saw a data sheet for a high-performance converter the spec'd dynamic range was about 18- or 19-bit equivalent; they may be pushing 20-bit now, and 22-bit is about theoretical max at room temperature. I don't know what anyone would do with real 24-bit dynamic range anyway. If you set the softest sound at the bottom of the range and put the volume control to where you could just hear it, you'd be instantly and permanently deaf when a sound at the top of the range came through. That's what I want in my living room!) So I wouldn't care except that the idea keeps getting pounded home that CD is a compromised, inferior medium when in fact it's the highest-fidelity mass-market audio distribution format we've ever had -- and as good as we'd ever need, in fact, at least for mono or stereo material. We just need the folks who make them to care about the quality of the masters they produce for the format. Then maybe we wouldn't have people running around claiming that vinyl (which is notably compromised with respect to real or potential fidelity) is a superior format. Or maybe we would. This is audio, after all.

 

So much for not writing a long email. Anyhow, one story: Probably about 12 years ago, when I was at Audio, David Chesky invited us over to their place, which was just a few blocks away, to listen to some 96/24 material they'd recorded. The neat thing was that a dCs rate converter was in the chain, so we could instantaneously dial the sampling rate down to 88.2, 48, or 44.1 kHz and back and change the bit depth in 1-bit increments. There was one passage in one recording where I sometimes thought I could hear a very small change in the cymbal sound when the sampling rate got down to 44.1. With the bit depth, we had to get down to 13 bits before I could hear any change.


Shocking? Well, 13 bits is still good for about 80 dB of dynamic range, which is a lot -- more, in fact, than we could dream of when I first got hooked on this stuff 40 years ago and more than all but a small proportion of music recordings actually deliver. Although I wouldn't want to give up the extra 3 bits now that we have them, some perspective is in order.

 

Better days,

Michael

Categories: September 2010, Music, Guest Blog

Post a Comment

Oops!

Oops, you forgot something.

Oops!

The words you entered did not match the given text. Please try again.

You must be a member to comment on this page. Sign In or Register

1 Comment

Reply Robert Bienstock
1:13 PM on December 10, 2010 
I am a huge fan of high-resolution music and I really do think that it makes a difference, but I also somewhat agree with your conclusions. IMO, I really do believe that for nearly every kind of music, a 44.1kHz sampling rate is sufficient provided that both the record and playback side are actually giving you that resolution. But I feel very differently about the word length (bit depth). Even though bit depth corresponds to dynamic range in the analog domain, I think that it is a mistake to thin that the effect on music of increasing the bit depth does nothing more than increase the dynamic range. This is particularly true with most recorded music that has a rather limited dynamic range. What increasing the bit depth does is to allow the softest possible sounds to be swallowed up by the ambient noise in the listening environment before they get swallowed up by the noise floor in the medium. Thus when I listen to music that has a higher bit depth, the only real difference I hear, subjectively, is that the reverb tails seem longer and all sounds appear to decay naturally, as opposed to decaying to a certain level and then abruptly disappearing entirely. And it doesn't take a lot of bits to make this difference. I think that even a change from 16 bits to 18 bits helps.