WHY DIGITAL STINKS, PART 01: USER ERROR
By George Petersen
The old analog vs digital argument is some 30 years old and not getting any younger. I’m not going to bring up that old feud. There are good and bad digital systems, and there are good and bad analog systems.
But often one of the reasons that digital sounds bad stems from user error. Many of us spent years honing our analog recording skills, finding out how to work with the (many) eccentricities of putting electrical signals onto magnetic tape. We knew when to hit it hard, when to ease back, when to bias, when to overbias–you name it, someone tried it. And the results were often quite remarkable.
Then that nasty digital demon arrived. Without years of knowing the inside tricks, we tried treating digital like analog. And we all learned the lesson about digital overloads and that awful, nasty noise digital circuits exude when clipped. Fearing the evil digital dragon, many people backed off their levels, avoided the overs and started making recordings that were well below the dreaded 0dB mark. Besides, with no tape hiss, any recording could just be brought up to 0dB later with a little normalization—the tracks were free of overload distortion, so everybody’s happy, right? Well, not quite, because the resulting sound was often lifeless, sterile sounding and lacked the punch we so craved. When you start recording at -4 or -6 or even -10 dB just to avoid overload distortion, your available dynamic range suffers, and that 16-bit system you had starts putting out 12- or 14-bit recordings. To make matters worse, manufacturers started building digital gear with clip lights that kicked in around -4dB, so as users avoided that -4dB point, the situation became even more ridiculous. Fortunately, 20- and 24-bit digital systems arrived just in time for us to make some 16- and 18-bit recordings.
Unfortunately, even today, there still is no industry standard for what constitutes a digital clip. How many consecutive samples have to break that -0dB mark before an overload is indicated–or even heard by the listener? (One? Two? Four? Six?) I’m not implying that your waveforms should resemble square waves or the great mesas and plateaus of the American West, but the effect also depends on what kind of source you’re recording. Everybody knows that digital overloads can sound raspy and harsh, so while tracking, if a snare hit (an instrument that’s supposed to sound raspy anyway) has SOME clipping, is that really a bad thing?
Meters and those flashing clip lights are provided as a guide. Finding out what your system is capable of simply comes down to LISTENING. Take a little time, experiment with your DAW or reverb and determine the limits of your rig. As anyone who bent a few analog VU meter needles can tell you, sometimes rules were meant to be broken.
When not working on Mix stuff, George Petersen records and performs with the SF Bay Area-based rock band ARIEL. Check ‘em out at www.jenpet.com.
___________________________________
CLICK ON THE “COMMENTS” BUG BELOW TO READ/LEAVE COMMENTS…
November 29th, 2006 @ 3:45 pm
Indeed it is crazy. I had gone to recording school in late 80’s and learned one thing. Left the industry, became a software engineer and recently came back to do my own recordings at home. Working with digital proved interesting. My first recordings were totally lifeless, flat, sounded like I had recorded them on a 4 track tascam from 20 years ago. Then did a little reading and got myself some good rules to work with my hardware. I work with a Rosetta 800 at home as my one and only audio interface, my preamps are Focusrite ISA220’s. SO my hardware is not the problem. My mics are Neuman and Rode. All go into Logic. My rules are simple:
1- Get a level on the ISA from the Mic using the ISA’s VU meter. I understand what this level means
I let it hit 0 there, and if sometimes it pops barely over it, I’m OK with that, it rarely does though.
2- After the VU reads OK, adjust the sends on the ISA so I hit -4 to -3 (first yellow) on peaks. If by any chance I happen to nail a note that is hotter I usually don’t go higher than -2 to -1, never 0.
3- Don’t compress on the way in unless I absolutely HAVE to. I much rather work with the full dynamic range and let the 24 bit and huge digital headroom work its magic, then compress on mixdown if I need to for the right reasons.
4- (This is ONLY used as a last resort if I really can’t get the singer or the source instrument to sit tight) On the Rosetta 800 there is a soft limit button which does an incredible job of allowing to flirt with 0 without going over, and it sounds very very musical. Now most folks who work at home won’t buy an Apogee, so the first 3 rules still apply.
To me these are simple rules that I follow and I get recordings that are full of life, body, dynamics, and warmth, that are very malleable if they need to be on mixdown. I think in general we tend to over complicate the process with digital when it doesn’t need to be.
Dunno if this seems like a novice approach, it seems to work wonders for me
Thanks for a great blog.
R
December 31st, 2007 @ 8:58 am
Cool.