AIFF vs MP3 and AAC (MP4)

DO MP3s AND AACs SUCK?

Click here to watch the video.

I didn’t run right out and buy an MP3 player when Apple first came out with them. After all, I knew the music wasn’t as good as the CD cuts. Heck, they tell you right in the specs that they use lossy compression. That means that they throw away part of the audio material. I’ve been recording for 38 years. I didn’t want anything to do with that inferior garbage.

But when my portable CD player broke a couple years ago, I borrowed an MP3 player and used iTunes to compress some Led Zeppelin songs. It sounded OK mostly, but I felt some of the audio was missing. I gave the player back. I didn’t realize until later that there are various “flavors” of MP3s. It turned out I was using a compression resolution recommended more for speech than hi-fi sound.

A few years passed and lust got the better of me when I laid my hands on the iPod Touch. Besides, after running into some sticky shed syndrome trouble with my old reel to reel tapes (see my post on The Secrets of Home Tape Baking), I was determined to transfer my collection of 1000 cassettes onto something more stable. Some form of space compression would certainly be required. Since I didn’t consider the material critical, I just started converting the tracks to the highest AAC setting available. But after processing a couple hundred, I got curious.

 

A HEARING TEST

I was getting ready to toss out the transferred cassettes and quite a few CDs, and decided I’d better make one more check before the source materials were gone for good. I imported an AIFF, MP3 and AAC (MP4) version of the same CD into Pro Tools and ran listening tests by soloing the stereo tracks. I was fully prepared to be disappointed, but I’ll be darned if I had no clue which was which. This wasn’t a blind test. I was the one clicking on the solo buttons and knew full well when I was playing the AAC tracks. To my ears, both the AAC and MP3s were indistinguishable from the original CD cuts.

Now I know there are some audio engineers out there who will argue with me on this. But before anyone gets all high and mighty and in my face, you might want to be sure you’ve tested yourself in private first. I say “in private” because Gizmondo.com recently ran some tests where they swapped Monster Cables with Coat Hangers for speaker wire. Guess what, the pros couldn’t tell the difference. Also, on a recent TV show I saw some audio engineers try to tell the difference between digital and analog recordings. They were right about half the time, exactly what you would expect if they were guessing. If you do decide to test yourself you might want to know that I used the highest conversions rate for AAC and MP3 encoding offered in iTunes—320 kbps. I used the automatic setting, not variable bit rates.

Also when testing, be sure you consider all the following factors: Some of what you hear comes down to the quality of the encoders, not just the bit rate. Don’t forget the devices themselves that play them back—a computer with a high quality sound card vs. one of the Apple players. Some material is more sensitive to audio artifacts when compressed. Even a slight difference in volume between the two audio tracks will likely make the louder one seem superior.

 

A VISUAL TEST

Next I began to wonder if maybe I might be experiencing age-related hearing loss that prevents me from hearing subtle differences. Audiologists say that around the age of 18 we begin to slowly lose some of our hearing cells. But the loss is so slow that most people don’t notice it until around the age of 50.

To see if any of that might apply to me, I used Photoshop to stack the waveforms of some AIFF, MP4 and MP3 audio on top of each other so I could compare them visually (Figure 1). I was surprised to find that the waveforms were nearly identical. I guess I expected them to look quite different, maybe with clear variations in the peaks and troughs and perhaps some cycles missing altogether.

 


Figure 1 — Circled areas represent visible differences between AIFF and MP4 waveforms of the same audio source material.

 

Even when the music changed to include more high frequency content, as with the lower frequency data, it was only slightly altered. And when you zoom out from the waveforms and view them as you would if they were playing in real time, the differences are virtually invisible. It’s the same as zooming in on a digital photo to see the pixels. But when you zoom back out, you see the whole picture, not the squares. So both visibly and audibly, the music is essentially, perceptibly identical.

Now to be fair, I am keeping some material in the AIFF format. That includes final mixes of my own projects, some sound effects CDs and about 100 of my very favorite CDs. But that’s it. Everything else goes MP4.

 

SAFE, NOT SORRY

If you still want to be safe after running your own listening tests, try using the Apple Lossless algorithm. It decompresses flawlessly—although I heard some engineers say it’s “pretty good.” I mean, come on guys. It’s lossless! Other lossless utilities include Monkey’s Audio, FLAC, Shorten, and Dolby TrueHD. They all provide files about 50-60% the size of the original.

Even at the highest quality of AAC, I can store about 5 times as many recordings in the same amount of space as AIFF files. That means almost 750 cuts on a DVD backup. Most of my library fits on one 500 GB hard drive (with 3 backups, of course). Now that’s convenience!

One thing you don’t want to do is put an AIFF audio file through multiple generations of compression and decompression. The results are as clearly audible as the color smear in third generation videotape. (You gotta be born before 1990 to know about those.) The lower the initial resolution, the more severe the audio deterioration.

 

JUST DO IT

In my opinion, the 320 kbps MP4 spec is not just good enough. It’s really, really good. Now, doesn’t that make you question why you’re eating up all your hard drive space recording at 96 kHz for your high end audio projects?

Leave a Reply

Your email address will not be published. Required fields are marked *