Lionel Dumond

What's your dB IQ?

Thu, 01 Jun 2000 00:00:00 GMT

Hello, and welcome to the long awaited Part Two of "What's your dB IQ?" I would sincerely like to thank all who have written me since the publication of Part One to tell me that you found it informative and helpful. If you're jumping in fresh at this point, I heartily suggest you check out the original article first, as the material that follows here will build on the concepts introduced therein. Don't worry… we'll still be here when you get back!

Back in Part One I provided a detailed explanation of the term decibel, showed how the concept was derived, and presented examples of its use as a measurement of electrical power and voltage, as well as sound power, intensity, and pressure. I also wrote about standardized reference levels, and defined several decibel designations, such as dBspl, dBm, dBu, dBV, and so forth. There was also a brief exercise in conversion between different dB designations, presented as a discussion of nominal I/O levels between so-called "professional" and "consumer" gear.

As we press on undaunted, we'll delve a little further into the topic of the decibel as it pertains to acoustics and audio engineering. We'll do some dB calculations, develop a few handy rules-of-thumb (set off by borders, so they'll be obvious), and get into some detail about weighted SPL measurements and other attempts to quantify the "loudness" of sound, especially as humans perceive it.

If you'll recall, in the introduction to Part One we presented a few Decibel Brain-Teasers to test your knowledge of the topic. How many of you felt brave enough to tackle them then? For the ones that still remain unanswered, let's roll up our sleeves, sharpen our pencils, and do a little cipherin'!

Decibel Comparisons

The sound of your sweetie whispering sweet nothings into your ear is 20 dB. The sound of your mother-in-law yelling at you is 60 dB. How many times louder is your mother-in-law than your paramour?

The real question here is, of course, "How many times louder is 60 dB than 20 dB?" Since we're dealing with sound pressure levels here, the formula we've got to use is:

dBspl = 20 log (P / R)

where P is the sound pressure level in Pascals (Pa), and R is a reference value (in the case of dBspl, .00002 Pa). This is the formula we use to express the difference between two absolute pressure measurements in terms of decibels (remember, a decibel is always a difference between two measurements!)

Let's write down what we already know:

20 log (PM / R) = 60 dB
20 log (PS / R) = 20 dB

where PM is the sound pressure created by your mother-in-law and PS is the sound pressure created by your sweetie. Therefore;

20 log (PM / R) – 20 log (Ps / R) = 40 dB
20 log (PM / PS) = 40 dB
log (PM / PS) = 2 dB
(PM / PS) = 100

So, the answer is that your mother-in-law yelling at you is 100 times louder than the whispers of your lovely betrothed. (Please don't invite me over for Sunday dinner, okay?)

There are a couple of things of note here. First, since the reference measurement R is a constant, and ends up dividing out of the equation anyway, it doesn't really matter what the actual value of R is – in fact, it needn't have appeared in the calculation at all. Second, there was never any need to derive the actual values of PM and PS – the problem was to determine how many times greater PM was than PS, so determining the value of the ratio (PM / PS) was enough. Knowing this, we could have started out with the equation

20 log (PM/PS) = 40 dB

and arrived at the answer a bit quicker.

So, what does this tell us? Well, we know that if any two pressure measurements are 40 dB apart, that simply means that the larger one is 100 times greater than the smaller one. This hold true whether the two measurements in question are 60dB and 20dB, 50dB and 10dB, 110dB and 70dB, or whatever. Go ahead—try different values in the formula and see it for yourself!

Let's take our new found decibel comparison skills and see if we can't derive a few handy "rules-to-remember," shall we?

How many decibels louder is a sound if you crank it up to be twice as loud?

Easy! Let PA be the sound pressure level after you've cranked it up. Let PB be the original SPL. If (PA / PB) = 2, then

20 log (PA / PB) = 20 log (2) = 6.0206 dB

which is damn close to (a mathematical term, meaning "approximately") 6 dB. We can therefore say that, as a rule of thumb,

An increase of 6dB of a sound means "twice as loud."

This is a pretty handy thing to remember! If you are boosting a signal by 6 dB, that means you're doubling the amplitude. Conversely,

A decrease of 6 dB of a sound means "half as loud."

By the same method, we can also derive the following:

+ 1 dB means about a 12% increase in SPL
+ 3 dB means about a 40% increase in SPL
+ 6dB means about twice as loud (200% of the original SPL)
+ 12 dB means about four times as loud (400% of the original SPL)
- 1 dB means about 90% of the original SPL
- 3 dB means about 70% of the original SPL
- 6 dB means about half of the original SPL
- 12 dB means about one-quarter of the original SPL

In my opinion, these are values that a good engineer should know in his head. When you're turning knobs and making adjustments to a signal based on dB values, it's important to keep in mind what those values means in terms of actual sound levels.

Decibel Summation

There are two sound sources playing in a room: say, a drummer playing at 72 dB, and a guitar player playing at 66 dB. What is the loudness of the two playing together?

Remember, decibels are logarithmic units, and can't be simply added together to produce a meaningful result. If we added 72 dB and 66 dB in the example above, we'd come up with a figure of 138 dB—which is approximately the volume of a jet engine at one meter! Remember that the upper threshold of human hearing is about 120 dB. Common sense tells us that either our answer is wrong, or we're listening to one hell of a loud band.

There is a formula that allows us to combine decibels levels of different sounds from different sources in this manner. If the level of the first sound is represented as dB1, and the second sound, dB2:

dB of combined sound level = 10 log (10 ^dB1/10 + 10 ^dB2/10)

If we plug in the figures in the example, we'd have the following:

10 log (1066/10 + 1072/10) = about 73 dB.

There is something interesting worth noting about this equation: the result can never be higher than 3 dB above the higher of the two sound levels! In fact, if the two sound levels are the same (dB1 = dB2), the result will always be an increase of 3 dB. Also, the more the two readings differ, the smaller the increase in combined SPL will be, and of course the closer the combined result will be to the higher reading. There's nothing magic about any of this – it can all be deduced from the equation. It would be a good exercise to plug in a few numbers and verify these rules for yourself.

Let's summarize these general rules:

Combining two different sounds of the same level results in an increase of 3 dB.

The smaller the difference in level between two sounds, the greater the increase in their combined SPL will be (to a maximum of 3 dB).

The greater the difference in level between two sounds, the smaller the increase in their combined SPL will be.

But wait – haven't we already said that "twice as loud" means an increase of 6 dB? There appears to be a contradiction here, but really there isn't. When we talked before about a 6 dB boost in a sound being twice as loud, we were in fact talking about a single sound being boosted from it's previous level. The above formula holds true only when the two sounds being compared are different. It's a subtle but crucial distinction. The formula is useful, because when we talk about sound levels combining in this manner, we're almost always talking about different sounds. The fact is that if the two sounds are similar enough (i.e., phase correlated) the actual increase will be higher than predicted by the formula—and if the two sounds are the same, the increase in level can in fact be up to 6 dB.

You've scored some sweet tix to a Pearl Jam concert, where the SPL is 120 dB. As you're waiting in the interminably long line to get out of the parking lot after the show, you crank their latest CD on your car stereo at a level of 100 dB. How many car stereos played at that volume would it take to produce the same SPL as the concert?

Here we're combining not only two, but several, different sound sources to determine the combined result; however, the procedure is the same. Since each car stereo is playing at the same volume (100 dB), let's use the following equation:

120 dB = 10 log (n * 1010)

where n is the number of car stereos that would combine to produce a result of 120 dB. Solving for n produces a result of 100. That means it would take 100 car stereos playing together at 100 dB to produce a combined SPL of 120 dB!

Does this result surprise you? It shouldn't! From the rules of thumb we developed earlier, we know that two stereos playing together at 100 dB would be 103 dB. Applying the formula above, we find that adding a third stereo to the first two would produce a result of 104.8 dB (an increase of only 1.8 dB). Adding the fourth would make 106 dB, the fifth 107 dB, the sixth 107.8, and so on. You can see that each additional stereo makes less and less of an increase to the total result—which not only follows our rule, but should make intuitive sense based on your own everyday observations. When you yell at a small party, everyone hears you; when you yell at the same volume at a ballgame, you don't add much to the overall volume of the crowd!

How much louder is a 100-watt guitar amp than a 50-watt guitar amp?

Hey… after all this mathematical gymnastics, this one is a piece of cake! We know that watts are a measurement of electrical power. You'll recall the dB formula for power measurements is

Power dB = 10 log (P1 / P2)

where P1 and P2 are the two readings being compared. You can already probably figure out in your head that the answer to the problem is 3 dB.

10 log (100 / 50) = 3 dB.

Let's Join the Real World Now

So far, all the information we've presented about decibels as they apply to the behavior of sound has been purely theoretical. We've talked about the sound pressure level without respect to frequency, which is to say, we've been using an unweighted system of decibel measurement. However, studies done over the years—using real people listening to real sounds—have revealed that the perceived loudness of a sound depends on the frequency of that sound. In other words, the frequency response of the ear isn't flat at all, but is more sensitive to sounds in the middle of the audible spectrum than it is to low-frequency or high-frequency sounds. If you were to plot a frequency response graph of the typical human ear, you'd have a graph showing "3 dB down" cutoff frequencies at about 500 Hz on the low end, and 8 kHz on the high end. You'd also see a gentle yet pronounced "presence peak" at around 4 kHz, which indicates that the ear is most sensitive to sounds in that range.

Unfortunately, that's not the only fly in the ointment we have to deal with, because it's also true that the way in which the ear responds to various frequencies is dependent on the loudness of the sound itself. In other words, the ear's frequency response is flatter for loud sounds, and less flat when the sound is relatively soft.

Back in 1933, a couple of scientists named Fletcher and Munson did experiments on all of this stuff and developed a series of curves that very closely approximated the response of the human ear with respect to both frequency and loudness. These curves are sometimes referred to as the Fletcher-Munson loudness curve, or more commonly the equal response contour. (If you've worked with audio technology in any depth in the past, these are probably both terms you've already heard.)

These curves represent a plot of sound pressure levels of pure sine tones that are generally perceived as being equally loud. (They are, in fact, the inverse of the human ear's frequency response graph.) As an example, take a look at the sixth curve from the top. It shows that if a 1 kHz sine tone is played at 70 dB, a 70 Hz tone would have to be played at 80 dB to be perceived as equally loud; a 16 kHz tone would have to be played at 90 dB to be perceived as equally loud, and so forth. Note how the equal loudness contour is flatter for louder sounds and much less flat for softer sounds.

A few years after Fletcher and Munson published their results, acoustical engineers developed decibel weighting systems to compensate for the ear's imperfect response. Nearly all sound level meters incorporate electronic filters that reflect the now-standardized decibel weighting systems. The three most common weighting systems are A-weighting, B-weighting, and C-weighting, which attempt to compensate for soft, medium, and loud sounds respectively. By far, the most common designation you'll find is the A-weighting, designated as dbA or dB(A). The A-weighting system has, over time, been adopted by government agencies like the FAA and the EPA, and is used for nearly all legislative and other "official" purposes, such as noise control ordinances and the like. B-weighting is rarely used nowadays. C-weighting, which most closely corresponds to the unweighted measurement of SPL, is occasionally used for sound levels exceeding 100 dB.

The graph below shows the weighting curves used for most sound level meters.

Conclusion

There are several other factors which color the way in which humans perceive sound, all of which complicate to some degree any attempts to accurately express sound levels in terms of decibels. Frequency and loudness are not the whole story! These other factors include the sound's waveform, complexity, spectrum, harmonic content, duration, spatial properties, attack and decay characteristics, and the simultaneous presence of other sounds, to name a few. The study of how humans perceive, react to, and interact with music, noise, and sound is called psychoacoustics. Many in this field have tried to come up with systems to accurately quantify various aspects of sound perception, and the decibel is among the most important tools they use for this purpose.

As audio engineers, producers, technicians, and musicians, a clear understanding of the decibel is important to us as well, as we pursue our endeavors in the technology of sound. Again, I hope this Part Two of "What's your dB IQ" (along with Part One, of course) was helpful in helping to develop that understanding. If you have any questions, comments, kudos, beefs, or whatever… please feel free to write me at ldumond@prorec.com. I love getting the mail!

And once again… happy dBs!

Succeeding as an Intern

Thu, 01 Jun 2000 00:00:00 GMT

You've studied hard. You've paid your dues (or so you think). You've twiddled the knobs on an SSL 9000, used Pro Tools to slap together some fake radio spots, and actually participated in the quaint ritual of editing ¼' tape with a razor blade. But that's all history now. Yes sir… those weeks (or months, or years) spent at Big Al's Recording School and Storm Door Company are finally going to pay off! Bring on the gold records. Bring on the cover of Mix Magazine. Bring on the Grammy Awards! You are now ready to make records, break records, and basically set the Wide World of Audio Production on it's ear....

Gee, I wish things were really like this, but they're not. I hate to whiz in your Wheaties, but I'm here to tell you that life after recording school is not going to be an endless parade of stretch limos, CD-release parties, and royalty checks. If you've recently joined the ranks of recording school graduates being churned out by the thousands every year, congratulations--you are now qualified to take on a job with no future, no benefits, and best of all, no friggin' pay. Welcome to the Big Leagues, son.

Greenhorns Need Not Apply

If you're like most recent recording school graduates, you probably have little or no real-world experience in an occupation where experience is a highly valued commodity. Also, keep in mind that you are trying to compete in an industry in which over 80% of the participants are either self-employed or employed on a freelance basis, which means the other 20% are holding on to the jobs they have. You're faced with the classic Catch-22. How are you supposed to gain experience if you have no chance of breaking in without it?

This is where the internship comes in. You work for little or no pay for a specified period of time (usually, a few months at most) in some kind of production facility as a way getting a few notches in your belt. It's not glamorous, it's not always fun, and you certainly won't get rich. But, if you keep your ears and eyes open, you'll get a taste of day-to-day life in a real facility, learn some new gear, and hopefully make some valuable industry contacts along the way.

I know it may seem discouraging to realize that you've spent all that time and money in school just to qualify for a "lowly" internship--but the reality is that without some sort of prior training, landing an studio internship is nearly impossible. In the recording industry, even the competition for unpaid positions is pretty intense, and if you're completely wet behind the ears, you probably won't cut it. Every year, it seems there are more and more future engineers and producers chasing fewer and fewer available positions. Merely showing an "interest" or "aptitude" for the work isn't enough these days. Most studios don't have the time, patience, or resources to take on someone who has no background at all--even though they're not paying you--so that certificate or degree is still an important credential.

If your school is a good one, chances are they already have contacts with facilities that are willing to take on interns from time to time. These opportunities often quite limited and are thus usually reserved for the best and brightest in a given class, but don't be shy about jockeying for position if you think you've got a shot. Also, check into other placement resources that may be available, whether it's a list of studios that hire interns, a letter of recommendation, or help with your résumé. Ultimately, it's up to you to land the gig--but like any job, getting in is only the beginning. That first internship is your "foot in the door" to a career in the audio production industry, and how well (or poorly) you do may portend your future in The Biz.

In my many years in the industry, I've seen my share of interns come and go. Most were pretty good, some were excellent; others had no business being anywhere near a recording studio at all. (There was actually an "in-joke" at one place I worked, whereby those interns deemed particularly bad were bestowed with the "Golden Spatula Award"--no explanation necessary, I'm sure). From these and other experiences, I've worked up a list of pointers for all of you would-be studio rats out there, to help you survive—and hopefully, succeed in—a recording studio internship.

1. Treat it like a job.

Yes, I know there's no paycheck, but that doesn't mean you have the luxury of treating your internship like a summer vacation. Showing up late, leaving early, taking extended breaks, and taking days off whenever you feel like it are sure-fire ways to blow whatever chance you may have at success. Even though you're "only an intern," your boss and other personnel will be treating you and evaluating you as one of their own. You want them to be able to count on you as a colleague, and a big part of that means being there when you're supposed to. As Woody Allen once said, fifty percent of being a success is just showing up.

2. Put in the hours.

One thing is almost universally true--in any employment situation, regardless of the endeavor, clock-watchers rarely get ahead. I have personally been involved in situations as a freelance engineer where getting the call meant putting in 18 to 20-hour stretches for several days in a row! While you probably won't find yourself in such an extreme situation, you do need to realize that long and sometimes crazy hours are the nature of this business. Standing at the door with your coat on at 4:45 doesn't exactly convey enthusiasm for the work you're doing. Be prepared to put in the same hours as your co-workers do—even if it means postponing that hot date or missing the big game on TV. Establishing yourself as someone willing to go above and beyond the call of duty will go a long way toward garnering favor in your employer's eyes.

3. Do the crappy stuff.

One of the most important factors separating the good interns from the bad ones is how aware he or she is of his or her status. That isn't meant to sound demeaning; it's simply a fact of life for an intern. This is an industry in which you are expected to "pay your dues," and paying your dues isn't always fun.

The sooner you realize your place in the pecking order--that you are, in fact, the "low man on the totem pole"--the more successful you will be. You aren't there because you're a gifted producer or a hotshot engineer, but only by the grace of your employer. Even if you're unpaid, you are still consuming time, space, and resources, and you need to do whatever it takes to give back more than you're getting. It's up to you to justify your place in the scheme of things.

Hooked on the Bean

While you're running around making yourself helpful, don't forget one of the most important pieces of equipment in any recording studio – the coffee machine! Like the traditional water cooler in most office settings, the coffee machine is often the hub of activity in a recording studio. It's natural if you think about it – many of the musical types with whom you'll be working have thus far led sleep-deprived existences full of late-night sessions and even-later-night gigs, kept afloat with greasy-spoon rocket-fuel java. A few years of that, my friend, and you've got a monkey on your back with Starbucks on his breath. Yessiree… more likely than not, your colleagues will be irrevocably hooked on the bean. And pul-leeze, people… forget the decaf. We need that kick!

You may think I'm joking, but I'm not--if you want to be instantly valued as an important part of the team, learn to make decent coffee! Whether or not you personally like or drink coffee isn't the least bit important. Besides, the secret to making decent coffee isn't that hard. Just make sure to keep the machine impeccably, spotlessly clean (inside and out), always use cold water, don't let it sit too long, and learn how everybody likes it (usually, strong and black).

All of this reminds me of an intern we once had who was taking an audio production course, and the internship was required for his graduation. Unfortunately, this poor fellow didn't know a microphone from a doorstop or a tape reel from a pancake, but I was still sorry to see him go because, well… gosh darn it, the kid made a damn fine cup o' joe. (As I recall, I even wrote a pretty decent recommendation to his professor, too.)

Of course, every intern wants that shot behind the console, but you won't ever get that chance until you earn the trust and respect of your colleagues, while giving them the respect that they deserve, too. One of the best ways to do that is to consciously look for the things that nobody else likes to do and do them--willingly, enthusiastically, and without complaining. No task, no matter how menial, should ever be considered beneath you. And if you volunteer to do them before you are asked, you are way ahead of the game!

You've got to prove yourself in small but important ways every chance you get. If you stand by and watch the senior engineer dusting the console, you're making a big mistake. If the trash cans in the place are overflowing, you've already failed. If the floors need sweeping, the bathroom needs attention, or the office files are in serious disarray, you're missing the best opportunity you have to make yourself a useful part of the team! This is all the boring, menial stuff that needs to be done by somebody. Let that somebody be you! Show that you're willing to "pay your dues" with a smile on your face, and you'll get the opportunity you deserve.

4. Try not to get underfoot.

There is a very fine line between being "accessible" and being "in the way." Unfortunately, it's a balancing act many interns never quite learn to master.

When I'm supervising interns, my own personal pet peeve is the guy who is constantly asking, "Is there anything you want me to do?" Look, let me tell you something… if I had a job for you right now, you'd be doing it! And if I have to interrupt what I'm doing to find something to keep you busy, then you really aren't making my life any easier, are you?

Another frustrating situation I've run across is the "Puppy-Dog Intern." This guy is like my shadow. He's a constant companion. No matter where I go, he's there. If I stop walking in my tracks, he bumps into me. If I turn to get something, he's right there in the way. If I'm trying to trace a cable, more than likely he's standing on it!

Of course, it's only natural to want to contribute and be as productive as possible, but like any job, there are going to be busy times and there are going to be slow times. There are going to be times when your help will be needed and appreciated, and others when it's best to lay back and chill for a while. One good way to be valuable without being a pain is to be a "self-starter"--by being aware of things that need doing and to do suggest them yourself.

For example, I would much rather have someone come to me and say, "I noticed the cables in the back room are a mess. Shall I wind them properly and organize them for you?" than ask "Is there anything to do around here?" Another good rule-of-thumb to remember is "if you're not part of the action, be part of the woodwork." In other words, if you're not actively participating in something, it's usually best to hang back. For example, if there is a session underway, don't stand over the producer's shoulder, hover behind the engineer's chair, or kibitz about what's going on. Sometimes it's a good idea to be not seen and not heard!

5. Do it the way they want it done.

In recording school, you were presented with a great deal of theoretical knowledge, hopefully punctuated with a few hands-on examples. Remember though, that audio production is both a science and an art, and as such, there is considerable variation in the way things are done among different studios and people in the business. Often, this is just a matter of individual taste. Sometimes, it's a function of a person's background--for example, you'll often see "old school" guys from the days of analog and vinyl do things far differently than engineers coming up today. And sometimes, a person's methods may be unorthodox (or, let's face it, just plain wrong) because, well… I've always done it this way, dammit, and no one's gonna tell me any different!

Inevitably, there will be some task or procedure you will see being done, or be expected to perform yourself, in a manner diametrically opposed to that which you learned in the textbooks. It's okay to ask—politely—why they do it the way they do if you don't understand. After all, there is always more than one way to skin a cat, and some ways are better than others. However, by no means should you ever insist that their whole method is "ass-backward" and offer to "set them straight"—you're asking for trouble if you do! Whether it makes sense to you or not, it's usually best to make like Burger King and "do it their way." Who knows… you might just learn a thing or two they didn't teach you in school.

6. Don't grumble.

Granted, it's not always easy to maintain a positive attitude under what can seem like trying circumstances. I know that emptying ashtrays isn't what you had originally signed up for, but remember it's not forever. Just think of your internship as part of your continuing education in the record business, because that's what it is. Nobody ever mixed a gold record on his first day on the job. Everybody has to start out somewhere. And every day of experience you gain in the industry will build on the day before. Think of every day as the first day of the rest of your career in the audio industry. Stay upbeat. Stay focused. Don't criticize, condemn, or complain. Even if it seems you'll never get your turn at the wheel, be patient. Stick with it, and your day will come.

7. Borrow and read the manuals.

One of the great things about an internship is the exposure to new and different equipment. Take advantage of it! Find out where they keep the manuals, and read them thoroughly. Watch the guys who know what they're doing as they work the controls. Ask questions when and where appropriate. If they will let you, stay late or come in early and experiment with the gear. Not only will you gain a deeper understanding of the tools of the trade, you'll show the boss that you're eager to learn and contribute. And you'll get your hands on the buttons, for real, that much faster—after all, the more you know about the gear, the more they'll let you do stuff with it.

Remember that interning in a real working studio is a great opportunity! Read. Ask. Experiment. Learn. Absorb everything you can while you have the chance. Remember… the experience you gain with a given piece of gear may someday prove invaluable.

8. Make and maintain contacts.

Whether you realize it or not, you're in show business now… and in show biz, more than almost any other profession, it's often not what, but who you know that matters. It's simply a sad fact of life, that sometimes how smart or talented or knowledgeable you are doesn't always count as much as making the acquaintance of the right person.

Of course making friends, maintaining relationships, and networking among your peers is an important part of any successful career, and now is a great time to for you to get started. Try to meet as many folks in the industry as you can during your internship tenure. Don't be shy about letting them know you're looking to find permanent work in the industry. Ask them for guidance, suggestions, and advice on the best way to proceed from where you are to where you want to be. Take whatever steps you can to make a name for yourself, even if only among the local music crowd. Make sure to collect as many addresses and phone numbers as you can, and stay in touch with them in the future as best you can.

9. Get letters of recommendation afterward.

You've spent weeks or months working hard at a job without even being paid. Don't you think you deserve something to show for it? A letter of recommendation (or, at least the promise of a decent future reference) can be one of the most important things you take away from your internship. Oddly enough, it's been my experience that most interns never even think to ask for one.

It's best to bring this up with your supervisor well before the end of your stay--if you wait until the last minute, you'll almost certainly find yourself the recipient of a hastily jotted toss-off, if anything at all. You may want to ask one or two of your colleagues if they'd be willing to write on your behalf as well. Unless you've been a complete failure as an intern (in which case, you would likely have been booted some time ago) they will almost always agree. Remember that they were once where you are now!

As the time draws near, tactfully follow up on their progress. If they've agreed to furnish you with a letter but haven't yet put pen to paper, don't despair; just realize that this is quite likely not at the top of anyone's to-do list at the moment. If your last day is quickly approaching and still nothing, what may happen is that he or she will ask you to write a letter for them to sign. If they offer this as an option, great--grab it and run!

What's Your Story?

Ever done hard time as a recording studio intern? Have or do you own, or work in, a production facility that employs interns? Care to share your experiences?

Send your war stories, horror stories, and success stories to me at
ldumond@prorec.com
. I'll take the best ones and compile them into a future article.

Please be sure to include your real name and city. I'll withhold your name (as well as those of the other guilty parties) if you indicate that you don't want them published.

In Closing

I can remember, almost to the day, when I decided that life in the recording studio was for me. Despite the ringing ears, the relocations, and the sacrifice that it's taken, I'm glad I made that decision! I can't say I've never questioned the choice—I still do sometimes—but all in all, I wouldn't have traded it for the world. If you've made the same decision, just know it won't be easy. And if you do somehow manage to make a living in this crazy business, I hope you'll find as much fulfillment and satisfaction in it as I have.

I hope you've enjoyed this article. As always, if you have any questions, comments, complaints or good martini recipes, write me at

ldumond@prorec.com

. I love getting your letters!

What's your dB IQ?

Thu, 01 Apr 1999 00:00:00 GMT

Probably one of the most misunderstood concepts in all of audio science is that of the decibel -- what it means, and what it doesn't mean. This really isn't very surprising. Most books you'll find on audio production gloss over the subject, offering a rudimentary explanation at best. What's worse, you'll often see the term "dB" bandied about with little consideration given to the context in which it's being used -- are they talking dBu, dbV, dBm, dbVU, dBFS, dBspl, or what? And being a rather complex concept, the term doesn't lend itself well to a one or two sentence glossary-style entry. For example, if you look up "decibel" in a book, you might read something like "the smallest increment of sound which the human ear can detect" -- which only happens to be true in general, but certainly isn't an accurate definition by any means.

What's more, this lack of understanding isn't limited to beginners. There are folks I know who've worked in this field for a long time who aren't clear on the whole thing -- even though many of them think they are.

A keen grasp of all this decibel stuff is crucial to the understanding of just about every other audio concept there is. So my question to you is this: how well do you know your stuff? Think you've got it all down? Want to find out?

Okay, hotshot... it's pop quiz time. Sharpen a number-two pencil and let's go.

1. The sound of your sweetie whispering sweet nothings in your ear is 20 dB. The sound of your mother-in-law yelling at you is 60 dB. How many times louder is your mother-in-law than your paramour?

2. You've scored some sweet front-row tix to a Pearl Jam concert, where the SPL is 120 dB. As you're waiting in the interminably long line to get out of the parking lot after the show, you crank their latest CD on the car stereo at a level of 100 dB. How many car stereos played at the same volume would it take to produce the same SPL as the concert?

3. There are two sound sources in a room: say, a drummer playing at 72 dB, and a guitar player playing at 66 dB. What is the loudness of the two playing together?

3. What is the difference in dB between the nominal output levels of "consumer" gear (-10 dBV) and "professional" gear (+4 dBu)? (Hint: It's not 14 dB.)

4. What's the dynamic range of 16-bit digital audio in dB? Of 24-bit digital audio? Could you figure out the dynamic range of a 21-bit digital audio system?

5. How come you can go over 0 dB on an analog two-track, but can't go over 0 dB on a DAT?

6. How much louder is a 100 watt guitar amp than a 50 watt guitar amp?

If you can answer these questions with ease, then this article wasn't written for you. But, if you found yourself scratching your head and saying "huh?", read on. Beware the road signs that read "Caution -- Math Ahead!" I'm sorry about that, but it's truly not possible to gain a through understanding of dBs without a little cipherin' along the way. If you are one of those people whose eyes glaze over when faced with a pageful of equations, fear not. I'm right here beside you! Stick with it, reread it a couple of times if you have to, and before too long the light bulb will come on, I promise you.

Oh, and one more thing -- if you want to follow along with your own calculator, make sure you have a scientifical-type one that has a "log" function. Be careful not to get the "log" button (the function that computes the base-ten logarithm) confused with the "ln" button (the function that computes the natural logarithm, which is a kind of a log that marches to a whole different rhythm.)

And so, come with me now if you will, as we step through the looking glass into the Wonderful World of the Decibel -- where subtraction is division, rulers are longer on one end than the other, and nothing is as it seems...

Logarithmic Scales -- the Kind You Can't Play on a Piano

The term dB has meaning in all kinds of scientific measurement -- from sound, to electrical or mechanical power, to voltage, and so on. The decibel scale is an example of a logarithmic scale. Other examples of logarithmic scales used in scientific measurement are the Richter scale (used to denote the energy of earthquakes) and the pH scale (used to indicate the concentration of hydrogen ions in a solution).

Why do scientists use logarithmic scales? Well, one thing you have to remember is that scientists -- much like guitar players -- like things easy! And when you're dealing with a large range of numbers that have a bunch of zeros before or after the decimal point, using logarithms makes those numbers a whole lot easier to work with and compare to each other.

Let me give you an example. Let's say you're a scientist in the way-back old days. You're observing some phenomena dealing with sound. You've performed a few experiments, and have taken some measurements about the intensity of sound. You know that sound is the movement of energy, and that sound intensity is the amount of energy passing through a given area per unit of time:

Intensity = Energy / (Time * Area)

Since we know that the ratio Energy / Time is equal to Power (think hard... remember your high school physics?) we can say:

Intensity = Power / Area

And so, since sound intensity is the amount of sound power per unit of area, and we scientists measure power in watts, let's express our measurements in watts per square meter (W/m²), shall we? Okay, so far, so good....

Now suppose, if you will, that you've found that the smallest sound intensity that most people can hear is .000000000001 W/m². You also discovered that the intensity that makes people start to wince in pain is 1 W/m². Of course, you've taken a bunch of measurements in between as well, like .000792710162 and .000006288415. Just try conveniently comparing those numbers! Quick -- what's the difference between .000792710162 and .000006288415 ? Try figuring that one out in your head!

Let's see if we can make these numbers smaller and easier to work with. Instead of using these unwieldy numbers in their raw form, let's try taking the base-ten logarithm of these numbers, and working with those results instead. It just so happens that:

log ( .000792710162) = -3.1
log ( .000006288415) = -5.2

Now... doesn't that look a whole lot less daunting? We can easily see here that the difference in the logarithms of these wacky figures, is an easy-to-handle "2.1". Yes... but 2.1 what? What are we going to call this "difference" number? You suddenly come up with a brilliant idea -- you'll call it a Bel, after your boyhood hero, Alexander Graham Bell! (Alexander Graham Bell was your boyhood hero, wasn't he? Come on, work with me here, people...)

You take all this into your boss, hoping for a pat on the back and maybe even a big raise. He looks at it and says, "Hmmm... I like it. Good work. But I don't like that messy decimal point. Get rid of it!"

So, you decide to change your units a little bit, by multiplying both sides of your equations by 10.

10* log ( .000792710162) = -31
10* log ( .000006288415) = -52

But then, your difference isn't 2.1 any more, it's now 21. The units are now one tenth the size of the Bel you came up with before, so you decide to call this "new" difference unit a decibel, or dB for short. A decibel is one tenth of a Bel.

But gee... why stop there? Remember, we're lazy scientists here. Can't we somehow make this stuff even easier? We sure can, because we know that our friend the logarithm can do a cool trick -- it can turn subtraction into division!

10* log (x) - 10* log (y) = 10* log (x/y)

In other words, we don't have to take the logarithms of both numbers. We can simply derive the ratio of the two numbers, and only have to do one logarithm calculation, instead of two! (You have to remember, this was long before scientific calculators were invented -- scientists had to use slide rules to figure all this stuff out, so the fewer logarithm calculations, the better!)

10* log (.000792710162 / .000006288415) = 21 dB

So, now you know why scientists like to use logarithmic scales -- when you're dealing with a huge range of numbers, it just makes comparing things a whole lot easier!

The Decibel Difference -- Nothing Else Compares!

Cool. Now we have a convenient way of comparing two unwieldy numbers. But -- and this is a very crucial point to understand -- the number we always come up with is not an absolute value, but is a comparison between any two measurements (or, more precisely, the difference in their logarithms, times 10). Remember, we came up with this decibel thing to show the difference between two measurements. Therefore, it makes no sense unless you're dealing with two numbers, right?

Let's see if we can't refine our system a little more...

So far it's been a pretty hard day; so you decide to knock off work early and have yourself a couple of brewskis on the ol' laboratory expense account. As you're holding down a barstool at your favorite local pub, you find yourself starting to scribble on a napkin.

Suddenly, you shriek out loud, "Eureka! How about, instead of dividing two measurements we want to compare, that we just pick some reference number that will remain constant, and divide all our raw measurements by that number? The results would still be the same!" Here's what you scribbled on your napkin. Let's call the reference number "B."

10* log (x/B) - 10* log (y/B)
= 10* log [(x/B) / (y/B)]
= 10* log (x/y)
= difference in dB

By golly... it works! You ecstatically wave your hands in the air, jump off your stool, cram the napkin intp your pocket, and impulsively kiss the waitress as you skip out the door. Everyone in the place now thinks you're a nut.

Back at the lab, you're now faced with the task of picking a good reference number. Looking over your experiments, you remember that the smallest intensity you found most people could hear was 0.000000000001 W/m². How about using that as a reference? This way, any single measurement could be expressed in dB, as long as we always remember that we're comparing it to 0.000000000001 W/m² . You can see that your dB is still expressing a difference -- there are, in fact, still two numbers involved. It's just that the second number will just be an "implied" quantity of 0.000000000001 W/m² .

You anxiously check it out to see if it works:

10* log ( .000792710162 / 0.000000000001) = 89 dB
10* log ( .000006288415 / 0.000000000001) = 68 dB
89 dB - 68 dB = 21 dB

Man... are you a genius or what? You've invented the decibel! Good for you. Just don't go around getting a swelled head over it now.

======================

Do You Have the "Power" to Take the "Pressure?"

So, now you know the story of the decibel. To review:

The difference (in dB) between any two POWER measurements (let's call them "x" and "y") is

dB = 10* log (x/y)

There's a darn good reason I put that word "POWER" in all capital letters. And here, we come to another crucial point -- there is a distinct difference in sound power (watts), sound intensity (watts per square meter), and sound pressure (Pascal). Hold on, I'll explain what I mean in a minute. Pay close attention now; because this is a major source of confusion when it comes to decibels!

As long as we're talking sound power (in watts) or sound intensity (watts per square meter), the formula given above is perfectly good. However, when you commonly hear the term "dB" bandied about in terms of "loudness of sound," people usually aren't talking about sound power or sound intensity, but more accurately sound pressure levels, or SPL. After all, it's the pressure a sound exerts on our eardrums that determines how "loud" that sound is to us!

Sound power is measured in watts. Sound intensity, as we've already discovered, is the ratio of power over a given area, and is measured in watts per square meter (W/m²). Pressure, on the other hand, is a measure of force over a given area. Since force is expressed in Newtons (N), pressure can therefore be expressed as Newtons per square meter (N/m²). The more commonly used unit is the Pascal (Pa). (1 Pa is the same as 1N/m² .)

The relationship between intensity (I) and pressure (P) is defined as

I = P²/r

where r (the Greek letter rho) is "air impedance" -- a constant which is determined by atmospheric pressure, air temperature, and so on. For normal conditions, and at room temperatures, the value of r is very close to 400. Therefore, the threshold intensity of hearing we mentioned before -- .000000000001 W/m² -- correlates to a pressure reading of about .00002 Pa:

.000000000001 W/m² = (.00002 Pa)² / 400

At the other end of the scale, the threshold intensity of pain, given before as 1 W/m² , correlates to a pressure of about 20 Pa:

1 W/m² = (20 Pa)² / 400

So, in terms of sound pressure levels, the range of human hearing is about .00002 Pa to 20 Pa.

There's still one other thing to consider concerning the relationship between intensity and pressure. Notice that the intensity of a sound does not vary directly with pressure, but as the square of the pressure. Look at the formula again: I = P²/r. Think about what is happening here. When pressure doubles, intensity quadruples. When pressure quadruples, intensity goes up by a factor of sixteen, and so on. Unfortunately, that means for pressure calculations, our old formula, dB = 10* log (x/y), won't work any more!

Fortunately for us though, it should be too tough to "pressure" our equation into shape for our purposes!

dB = 10 * log (x/y), where "x" and "y" are measurements of sound intensity

Substituting the expression P²/r for these intensity measurements, we get

dBspl = 10 * log [ (P_x²/r) / (P_y²/r) ]
= 10 * log (P_x²/P_y²)
= 10 * log (P_x / P_y)²
= 20 * log (P_x/P_y)

This is almost the same as our formula for power, except this time we have to multiply by 20 instead of 10.

It is this term -- dBspl -- that is being referred to about 99% of the time when you hear or read the term "dB" concerning the loudness of a sound. This is in fact exactly what you are measuring when you use a sound level meter -- the pressure on the mic, referenced to .00002 Pa, and expressed in terms of dBspl. (Often, the dB reading shown on a meter is a weighted measurement; something we'll explore more deeply in Part II of this article.) The problem is that almost nobody attaches the little "spl" to the end, which tends to make things needlessly confusing! When it comes to loudness, just remember that "dB" is just kind of understood in practice to mean "dBspl."

The reference measurement against which SPL is measured is .00002 Pa, which is a good approximation of the lower threshold of hearing for most normal, healthy youths. That's the figure to use for the bottom of the fraction in the formula above.

dBspl = 20 * log (P_x/ .00002 Pa)

Since log 1 = 0, it doesn't then take a lot of head scratching to figure out what the dB level of the smallest audible sound is, does it?

20 * log (.00002 Pa / .00002 Pa) = 0 dB SPL

Pay attention to that last equation, because it points up a neat-o property of the dB: Because log (1) = 0, any measurement that is equal to the reference measurement is always "0 dB," regardless of the dB designation in question. This holds true for dBm, dBu, dBV, dBFS... it doesn't matter. You see this more clearly in a moment.

The loudest sound pressure we can tolerate is about 20 Pa, as we've already noted. Can you figure out how to express this in terms of dB? If you've been following closely so far, it should be no sweat:

20 log (20 Pa / .00002 Pa) = 120 dB

So, now you know that the dynamic range of human hearing is approximately 120 dB. Actually, I'm sure you already knew that. But, now, you know why.

Of course, just about every other sound you can hear is going to fall somewhere in the range between 0 dB and 120 dB. Just for fun, I've included a handy-dandy table of common sounds, and their approximate loudness in dB (remember, in this table, dB means dBspl!)
.

The Secret's in the Recipe... uh, I mean, Reference

To review so far:

1) The difference (in dB) between any two POWER measurements (let's call them "x" and "y") is

dB = 10* log (x/y)

2) The difference (in dB) between any two PRESSURE measurements (Let's call them Px and Py) is

dB = 20 * log (Px / Py).

Okay... all well and good so far. But you know as well as I do that there's more to this story. So, what you're probably asking right about now is, "what's the deal with all this dBu, dBv, dbV, dBm, dbVU, dBFS, and so on? Help me! I'm drowning in alphabet soup!"

Relax... once you understand what the term "decibel" really means, all the rest of this stuff is easy! Remember, the designation "dB" denotes a comparison between two numbers (of the same type, of course -- you can't use dB to directly compare, say, volts to watts). One of those numbers is usually a reference of some sort. That little bit on the end merely reminds us what's being measured. The reference level to which various measurements are compared have become standardized over the years, as I'll explain below.

dBm and dBVU

Now we've already discussed the derivation of dB as concerns power measurements. We talked about it in terms of sound power (measured in watts), but as we know, there are lots of other phenomena that can be discussed in terms of "power" -- like electricity, for example.

Back in the good ol' days, before LED ladders and fancy LCD displays and such, audio engineers relied on a device known back then (as it is today) as a VU meter. A VU meter is nothing very fancy; merely a cool little device consisting of a needle attached to a magnet, such that as the current in the magnet increases, the needle turns clockwise. "VU" stands for Volume Unit, a term developed during the early days of radio.
The problem with VU meters when they were first introduced was that all of them were different. That is, until the late 30s -- when a bunch on engineers sat down and decided to standardize the VU meter such that whenever the power in a circuit was one milliwatt (1 mW rms), the VU meter would read 0 dB, or, in other words, 0 dBm = 0 dBVU.

The little "m" in dBm stands for milliwatt. The dBm is a measurement of power, always referenced to 1 mW.

dBm = 10 * log (Power / 1mW)

Knowing this, it's pretty easy to express the power in a circuit in terms of dBm. You already know what happens when the power in the circuit is 1mW. This is the reference power level -- and when a measurement is equal to the reference level, the formula always yields zero, remember?

10 * log (1 mW / 1mW) = 10 * log (1) = 0 dBm

As we've said, this is the reading that would normally cause a VU meter to read 0 dBVU. If you see that your VU meter jumps from 0 dBm to +3 dBm, you know that the power in the circuit has doubled, because

10 * log (2 mW / 1mW) = 10 * log (2) = 3 dBm

What if the needle drops from 0 dBm to -6 dBm? What is that telling us?

10 * log (0.25 mW / 1 mW) = 10 * log (.25) = -6 dBm

What's happened in this case is that the power in the circuit has dropped to one-quarter of its previous value.

Now I would hate for you to think that every VU meter you see is graduated in dBm, for this is not the case! When other types of gear was invented, engineers decided to use VU meters for them as well, and calibrated the meters such that the measurement that was equal to the reference measurement would make the meter read 0 dBVU. Take the tape recorder for example. They put VU meters on tape recorders, but this time, instead of calibrating 0dBVU to an electrical power reference, they calibrated the meters to a magnetic power reference. Magnetic power, often referred to as flux, is expressed in Webers per meter (W/m). Since the magnetic flux of recording tape is pretty small (good thing, too -- I'd hate to have my fillings sucked out by a reel of tape!) we usually use nanoWebers per meter (nW/m). A nanoWeber is one-billionth of a Weber.

The VU meters on a tape recorder were set to read 0 dBVU at whatever the recommended recording level of the tape was. The first recording tape put out by Ampex sounded pretty good at 185 nW/m, so those old Ampex machines used 185 nW/m as the 0 dBVU reference on the VU meter. The cassette standard is 160 nW/m, so the VU meters on a cassette deck are calibrated such that a flux of 160 nW/m on the tape causes the VU meter to read 0 dBVU. The dB meters on modern professional analog decks are calibrated to match the much hotter recording levels possible with today's high-tech oxide formulations -- 250 nW/m and up. The spec sheet on a particular deck will usually tell you how many nW/m "0 dBVU" means on that deck's meters.

dBu (a.k.a. dBv)

Think back again to that high school physics class. We know from Watt's Law that there is a relationship between power and voltage, don't we? Watts Law states that

P = V² / R

where P is the power in Watts (W), V is the electrical potential in Volts (V), and R is the resistance (or, in this case, impedance) in Ohms (W).

If you remember from our discussion of dBm, the reference power used in that case was 1mW. As I said, this standard was developed in the 1930s -- and back then, the input impedance of every piece of audio gear made was 600 ohms, period. Tape decks, mixers, preamps, power amps, everything -- if it had an input, the resistance from the hot wire to the ground connector was 600 W. That's just the way things were back then!

So the question now becomes: What voltage does it take to generate 1 mW of power across a 600 W impedance?

P = V² / R
.001 W = V² / 600 W
V² = .001 W * 600 W
V = sqrt (.001 W * 600 W)
V = .775 Volts

We see that it takes about .775 V to generate 1 mW of power at 600 W. Even though the 600 W impedance "standard" went the way of the dinosaur long ago, the .775 V reference remains as the reference voltage for dBu.

There's one more very important detail to note here. Notice that power varies as the square of the voltage. This is the analogous situation to the dB derivation of sound pressure, given earlier. In fact, voltage can be (and often is) thought of as "electrical pressure." If we do the same substitution calculation for voltage as we did for SPL, we get the " times 20" formula

dBu = 20 * log (voltage / .775 V)

There's something you may be wondering at this point -- why the "u?" Well, it wasn't always that way. The term "dBu" was originally written as "dBv" (with a small "v"), and in fact you'll still see old-timers use "dBv" sometimes. The problem was that folks kept getting their "dBv" (small "v") confused with "dBV" (capital "V"), and we can't have that now, can we? So, "dBv" was changed to "dBu." Just remember that if you ever see the term "dBv" being used anywhere, it means "dBu," or that the person using it has fallen into the same small v-capital V confusion.

Speaking of which...

dbV

We've come a long way since the days when the impedance of gear was 600 W. On today's stuff, you're likely to encounter much higher impedances -- 10,000 W or more. At impedances that high, power dissipation is pretty low (in fact, we're talking microwatts here now) because impedance and power are inversely proportional (Watt's Law again: P = V² / R).

Remember that the dBu uses .775 V as a reference, which some engineers thought to be kind of awkward -- and because the 600 W standard had been abandoned, there was no special significance to that .775V reference level any more. So, a new standard was developed -- dBV -- that instead, uses a nice, round, 1V as a reference instead.

dBV = 20 * log (voltage / 1V)

You can see that dBu and dBV are very similar! They both compare voltages; it's just the reference levels that are different.

A side trip -- "Professional" vs. "Consumer" Levels

There has always been a lot of confusion about the whole issue of the nominal operating level of "professional" gear versus the nominal operating level of so-called "consumer" gear.

You may have heard that professional gear is "+ 4 dBu" and consumer gear is "- 10 dBV." Because only professionals used this stuff back when it was new (and expensive!) technology, and the older dBu designation was all they had to work with, the original designation of operating level (expressed in dBu) has stuck. By the time consumer audio products were introduced in a big way, the dBV has been invented, and so dBV was used for consumer gear. (Remember, they are both simply ways of comparing voltage levels -- nothing more. That + 4dBu is somehow inherently "better" than -10 dBV is a big, fat myth, kept alive by the somewhat arbitrary labels of "professional" and "consumer" attached to them.)

I'll just bet that a lot of you have glanced at this "+4 / -10" thing, and just assumed that the difference between the levels is 14 dB. But now we know better, don't we?

The reason the difference isn't 14 dB is because the reference levels between dBu and dBV are different! Remember, dBu is referenced to a voltage level of .775 V, and dBV is referenced to 1V. Armed with the knowledge you now possess, can you figure out what the true difference in operating level is, between + 4 dBu and -10 dBV?

+ 4 dBu = 20 * log (voltage / .775 V)
voltage = 1.228 Volts

- 10 dBV = 20 * log (voltage / 1V)
voltage = 0.3162 Volts

20 * log (1.228V / 0.3162V) = 11.79 dB

You can confirm this by doing a little experiment. Plug a piece of consumer gear with -10 dBV outputs into a piece of gear with + 4 dBu inputs. If they both have VU meters, calibrated such that 0dBVU on each piece of gear corresponds to it's nominal operating level, you'll find that 0dBVU on the consumer gear causes a reading of -11.79 dBVU on the pro gear's meters.

dBFS

We now come to a relatively new, yet all-important dB designation, the "dBFS," which stands for "decibels full scale" -- a kind of dB designation created especially for digital gear.

This one is a little strange because, unlike all the other dB varieties, the reference level isn't at the bottom, or somewhere near the middle, but at the very top possible measurement. This means that "0 dBFS" designates the highest possible level, and that all other measurements expressed in terms of dBFS will always be less than 0 dB -- in other words, a negative number. This is why, on digital gear using VU meters (where 0 dBVU means 0 dBFS), the "0" is at the top of the scale, and the meter can never read higher than that.

Let's take16 bit digital audio as an example. "The term "16-bit" means that the level of any sample can be stored as a 16-bit binary number (a binary number with 16 placeholders). As we know, the binary number system only has two digits, "0" and "1." Therefore, the highest possible 16 bit binary number is the number with all "1"s: 1111 1111 1111 1111 (binary). So the formula for dBFS in a 16 bit digital system is:

dBFS = 20 * log (sample level / 1111 1111 1111 1111)

It's easy to see why they say "you can't go over 0 dB in digital." That's because, at the highest possible sample level (which is the dB reference level, in the case of digital):

20 * log (1111 1111 1111 1111 / 1111 1111 1111 1111) = 0 dBFS

Also, using the same formula, we can easily figure out the dynamic range of a 16-bit system, because we know the smallest possible sample level (other than zero, of course) is 0000 0000 0000 0001.

20* log (0000 0000 0000 0001 / 1111 1111 1111 1111) = -96 dB

So, now you know why the meters on a 16-bit DAW usually read from 0 dB to -96 dB, when displayed at their highest resolution. By following the same logic, it's easy to figure out that the dynamic range of a 20-bit digital audio system is 120 dB, and that for 24-bit digital audio, it's 144 dB. I'll let you do the math if you want to. (Hint: on most calculators, it's easier to convert the binary numbers to decimal numbers first -- otherwise, you're likely to run out of digits!)

Coming in Part Two

This concludes our discussion on the mighty decibel, at least for now. In the second and final installment of this article, we'll talk about weighted decibel measurements, which are an attempt to more closely associate raw dBspl readings with the way humans actually hear sound. We'll figure out some handy rules-of-thumb concerning decibels, and pull off a few Stupid Decibel Math Tricks. I'll also help you solve the rest of those brain teasers at the beginning of this article. (You won't want to miss that, huh?) Of course, if you feel ambitious, there's no need to wait for me -- feel free to go ahead and get started on them now. The article you just read presents all the knowledge required!

I'll leave you with a handy Decibel Cheat Sheet. You might even want to print it out and hang it on your fridge!
images/Wcf7c6c7b89459.gif" width="576" height="108" alt="">
* binary 1111 1111 1111 1111 converted to decimal form = 65536
Until next time... happy dBs!

Lionel Dumond is the Managing Engineer at Sweetwater Sound Productions in Fort Wayne, Indiana, and is the Media and Mastering Editor here at ProRec. Unfortunately, he wasn't able to successfully explain the decibel system to the cop who wrote him the ticket for the cherry bomb mufflers on that ‘68 Mustang...

The Waves Native Power Pack II

Mon, 01 Mar 1999 00:00:00 GMT

The landscape of Marketing History is littered with ill-fated products, saddled by their well-meaning creators with monikers meant to cash in on the glittering reputation of their legendary namesakes. As we've all learned, however, sharing a name with a classic isn't necessarily an indication of repeated glory. Remember the Mustang II? The Exorcist II? Or, God forbid... New Coke?

The lesson learned here is that, if you're going to adopt a name synonymous with Greatness, that you'd better be able to walk your talk. The pressure to excel, to succeed, to live up to heightened expectations is tremendous. Many have tried. Many have failed. Few are up to the task. But every so often, these untested heroes with the oh-so-familiar names surprise, delight, and even astound us. It can be done. Just ask Ken Grffey, Jr.

Or, for that matter, ask the developers at Waves Ltd., who are calling their latest bundle the "Native Power Pack II." Whoa... wait just a minute here! Heck, we all know that this new bundle of plug-ins has got to be good -- after all, they're from Waves, arguably the best known and most highly regarded developer of audio plug-in effects in the world. But do they attain the rarefied "can't-live-without-it" status of the original NPP? Are they really that good? That's the thing I was determined to find out the minute I read the press release -- "Waves Announces the Native Power Pack II." "Indeed," I muttered skeptically. "We shall see about that." I could hardly wait to put these new plug-ins through their paces!

The Basics

The Native Power Pack II contains four plug-ins: The Renaissance Equalizer, the Renaissance Compressor, the MaxxBass, and the DeEsser. These plug-ins have been available for a while now in their Pro Tools/TDM versions, which require a proprietary DSP card to run. These new versions can run on the host processor of either a Mac or PC, hence the term "native."

These plug ins do not actually replace, but are meant instead to complement the C1 compressor, L1 Ultramaximizer, and Q10 Equalizer in the original Native Power Pack, a feat they accomplish quite nicely. While you'll find many similarities among these new plug-ins to the C1, L1, and Q10, the NPP II can pull off some stuff their venerable siblings never dreamed of. Once you load the NPP II into your system, you may find yourself relying on those original plug-ins a lot less in the future. I know I certainly have!

System requirements are fairly modest. For Macintosh, a Power Mac running MacOS 7.6.1 or higher, with a 603 CPU (120 MHz or faster) and at least 32 MB or RAM is required. For PC, a Windows 95/98, Pentium II (166 MHz or better) machine with at least 16 MB of RAM is required. Keep in mind that these are the minimum requirements, which means you'll have to have at least this much horsepower to get the software to load and run. If you plan on actually doing anything very useful with your audio, you'll probably need more juice that what I've described here. Full system requirements are available at the Waves website, http://www.waves.com . Another thing you'll want to keep in mind is that the Renaissance plug-ins use floating point architecture, so the FPU performance of the host processor is of utmost importance.

The four plug-ins in the NPP II are just that -- plug-ins, which means they require a host program to run. The Mac NPP II supports most popular Mac-based DAW and multitrack audio sequencer software, such as Cubase, Logic, Digital Performer, AudioSuite, and others. The PC plug-ins are all DirectX applications, which means they are supported by just about all the PC audio software out there, including Sound Forge, Cakewalk, WaveLab, Samplitude, and more. Again, if you're not sure if the host software you plan to use has the right stuff, check the website for more comprehensive information.

Sweet Anticipation!

Simply can't wait to try them out? Waves allows you to download a full version of the NPP II right from their website, which is what I did. You'll also need to download the small WaveKey utility which, when run, identifies your dongle and generates a data file which you email to Waves. Waves will then email you back data and instructions to authorize your dongle to run the program. I know, I know... all of this sounds like a major pain in the butt, but trust me, it's very easy. Once I emailed the required file to Waves, I had the necessary stuff the next day and was up and running in no time. The installation was flawless. Of course, you can always buy the retail version from your favorite dealer, which comes with a disc, a WaveKey (if you don't already have one), and printed documentation.

And now, on with the show!

The Renaissance EQ

The Waves Renaissance EQ (like it's predecessor the Q10) comes in a "full size" version (in the case of the Renaissance EQ, six bands of EQ) as well as smaller, more efficient versions (four-band and two-band, respectively), with the only difference between them being the number of available bands. If you don't need all the power of the six-band version, you can save some CPU horsepower by utilizing one of the smaller versions. In terms of sonic output, all of them sound the same.

You'll notice a familiar Q10-type interface here; an interactive response-graph with movable band markers, and precise numerical control of all parameters. But that, my friend, is pretty much where the similarities with the Q10 end.

The first thing you might notice is that there is no input trim control. That's because of the floating point architecture that these plug-ins use. With 32-bit floating point I/O and 48-bit internal processing, headroom is virtually unlimited (it's literally in the thousands of dB) so it's impossible to overload the input, regardless of the incoming signal level. (In case you;re wondering, Waves uses a proprietary dithering process to convert it's internal 48-bit data to 32 bits on output.) The other advantage gained by using such high resolution in processing and I/O is to minimize artifacts and to maintain accuracy of the finest details in the audio data. (In fact, both Renaissance plug-ins in the NPP II use this ultra-high resolution floating point architecture.)

The other major differences between this EQ plug-in and any others you may have used before won't be apparent until you actually start playing with it, and listening to the results. Unlike most parametric EQ plug ins, where all the bands are the same, each band of the Renaissance EQ is unique, optimized for doing a specific job in a specific part of the frequency spectrum. For example, Band 2 supports bell-type and low-shelf filters, but not high shelf or cut filters. This band/filter specialization greatly increases the resulting accuracy of the effect -- precise settings of bandwidth, range, and gain were not only possible, but actually audible as a real world result.

I'll bet I know what you're asking right about now: "with only six bands, and with each band being different, how can steep slopes on cut or shelf filters be achieved?" On most plug-ins, including the Q10 (and, in fact, with most analog parametric EQ hardware as well) filter slope is fixed, and sharper filters are created by "stacking" multiple filters within the same frequency band. But here is the very cool thing about the Renaissance EQ: the steepness of a cut or shelf filter can be controlled by using the Q control! This means that very sharp, and very precise, cut or shelf filtering can be created using only a single band on the Renaissance EQ. This is an extremely cool feature! I have found that I can create filters using a single band of Renaissance EQ that used to take four bands with the Q10.

The other really neat thing about the Renaissance EQ is that the bell filters are both asymmetrical and fully parametric at the same time. Asymmetrical means that at any given value of Q, a cut in gain has a narrower bandwidth than a gain boost. This is the way that many classic, vintage equalizers work; however, most of those (for example, old Pultecs and such) aren't fully parametric. The Renaissance EQ provides the best of both worlds -- amazingly realistic analog-type response, with exacting parametric control. Very sweet stuff!

Pros: Includes smaller two and four band versions to save CPU cycles. 32-bit float 1/0 and 48-bit processing sounds incredible. Extreme filter accuracy. Adjustable-slope on shelving filters. Asymmetric bell filters impart true-to-life analog-type characteristics. All in all, just about the finest software equalizer plug-in you'll find out there.

Cons: I'm thinking, I'm thinking...

Renaissance Compressor

Waves claims that the Renaissance Compressor "combines the technologies of Waves' C1 Compressor/Gate and the famed L1 Ultramaximizer." Describing it in this way doesn't nearly do it justice. While it lacks some of the detailed features of either of those plug-ins, it more than makes up for that by doing some amazing things neither the C1 nor the L1 can.

The Renaissance Compressor makes no attempt to copy the classic "gain-response" graph normally found on compressor plug-ins, opting instead for a simpler, cleaner, and (in my opinion) far more functional approach. Sliders are provided for the basic five compressor controls, with numeric entry available for each one, of course. In the middle is the gain reduction meter. When trying to set dynamics levels, this is the most important piece of feedback you can get -- it tells you how much compression or expansion is actually taking place -- so you want that meter to be clear, responsive, and easy to read, and this one is. The rectangle above the output meters is a limiting display. The amount of limiting that is actually taking place is indicated by its color -- it turns yellow whenever the limiter kicks in, followed by a brighter yellow for heavier limiting and red for very hard limiting.

When writing reviews, I'm not generally given to superlatives, but I really have to say that very first time I heard it, I was completely floored by this plug-in! Sure, it compresses, like the name says. But where this baby really shines is how uncannily it emulates the sound of just about any and every classic piece of dynamics-processing hardware there is. By careful setting of the parameters, this thing can be an LA2A, an 1176, or a 1960! It can act like a tube EQ or a solid state one; with snappy VCA-like response, or more forgiving opto-coupled behavior. The Opto setting really is gorgeous -- the release time actually gets a little slower as you ease up on the threshold, just like a honest-to-God piece of real vintage iron!

One more feature that deserves a mention is the Automatic Release Control (ARC). Let's say you're attempting to use a compressor as a dynamics "leveler," just gently smoothing out the transients in a mix. For a situation like this, you want to set a fairly long release time. This setup works great as long as the dynamics remain fairly consistent and predictable. But what if, all of a sudden, you get a really big fast peak coming at you? That long release time you've so carefully set is gonna mean trouble -- pump city! The common solution is to use two compressors in a chain -- one with a fast release to handle those quick peaks, and another to act as the leveler. But with the Renaissance Compressor's ARC control, this case is handled automatically. In fact, with the ARC and the built-in limiter, the Renaissance Compressor can act as three dynamics processors simultaneously -- something not even those multi-megabuck behemoths can do!

The only real drawbacks of the Renaissance Compressor is that is lacks a sidechain input, so it can't do split-mode, EQ-dependent compression like the C1. If it did, I don't think I'd ever have a reason to touch another compression plug-in ever.

Pros: Killer dead-on emulation of just about any vintage compressor ever made, including tube, solid state, opto-coupled, and VCA-controlled sounds and behaviors. ARC function makes it almost impossible to screw up. Functional, no-nonsense display. Built-in limiter. Hands-down the best software compressor I've ever had the pleasure of using... and trust me, I've heard them all.

Cons: No sidechain.

MaxxBass

Let's say you want to enhance and smooth out the low-end response of a mix, or to prepare material for playback on a specialized system, such as a kiosk, interactive exhibit, or small multimedia-type computer speakers. For a job like this, you might reach for an equalizer, or perhaps a sidechain compressor. Next time, you might give this handy little plug-in a try instead.

The MaxxBass is a versatile tool that purports to beef up and add punch to the low end of program material. However, unlike the more traditional approaches, it achieves the desired result by rolling off some of the existing low end and "replacing" it with higher-frequency harmonics that are related to the rolled-off portion of the signal. By doing this, all sorts of effects can be achieved -- including the preservation of low end that might exceed the physical limits of certain systems.

You'll notice several sections in this plug in. The critical parameters are Frequency, which is the crossover point below which the plug-in treats material as "bass". The other critical parameters are the sliders, which allow the user to attenuate the original bass, and to determine how much generated harmonic material is added. There are also settings which allow control over the nature of the harmonics generated, and to achieve compression of the added harmonic material.

The one thing I really wanted to test was Waves' claim that this plug-in can actually increase the perceived bass response of smaller systems, by "fooling" the ear into thinking it was hearing low frequencies that aren't really there. As the theory goes, when the listener hears related low-frequency harmonics, the ear and brain will fill in the missing low-end information -- information the playback system isn't actually capable of reproducing.

Luckily, I had a pair of Optimus multimedia-type computer speakers handy. I also ran the test through a Zenith clock radio equipped with an auxiliary input. I ran the stereo buss of my mixer into both of these systems. Then, for each one, I determined its low-end frequency response by monitoring just the "Bass" portion of the signal (MaxxBass allows you to do this easily) and set the Frequency accordingly. I then cut the bass slider all the way (thus completely removing the non-reproducible part of the signal) and started moving the MaxxBass slider up, listening closely to the results.

I was pretty impressed by what I heard. When I switched in the MaxxBass, I could somehow "magically" hear the stuff that I darn well knew was below the low-end response of the system! It was actually a bit weird at first -- sort of like the feeling you get when staring at an optical illusion, when you know what you're seeing is not real, but you "see" it anyway. In fact, that's a great way to describe what the MaxxBass did -- it created an "auditory" illusion, right there on those little 2-inch speakers. Wow!

This is simply NOT a trick you can pull off with any equalizer; in fact, there's no other plug-in I know of at all that does exactly what the MaxxBass does. With an EQ, you can roll off bass, but once you do, it's gone. Subsequently boosting the low end of the remaining signal isn't going to help much in that situation. The MaxxBass preserves frequency response by actually creating new sonic material that is precisely related to the rolled-off low-end. All I can say is, if you are producing audio for any specialized system -- television, radio broadcast, CD-ROM, computer games, public installations -- wherever you fear bass response may suffer, you need the MaxxBass!

Pros: Smoothes out and enhances low-end response. Extends perceived bass response well below the physical limitations of smaller systems. Easy to set up and operate.

Cons: Very easy (and sometimes, very tempting!) to "overdo" this effect.

DeEsser

The fourth plug-in included in the Waves NPP II is the DeEsser. This does exactly what you'd expect -- it attenuates sibilant sounds in a recording. It's basically a EQ-sensitive compressor, but it comes pre-equipped to handle this rather specialized task, with a hard knee, fast attack, high ratio, and optimal sidechain EQ bandwidth already set up for you.

Setting this one up is a lot quicker and easier than setting up a general-purpose multiband compressor to handle this job. By monitoring the Sidechain output, one can easily set the frequency (from 2kHz to 16 kHz) that best captures the "ess" sounds one wants to tame. The Threshold is then lowered to taste. That's it... brain-dead simple.

It also works in either Split or Wideband mode. Wideband mode is the behavior most of us are already familiar with, and the one would likely opt to use for a solo track. In Wideband mode, when the track contains material that falls within the selected sidechain band and which exceeds the threshold, the entire track is attenuated at that point. This is the old tried-and-true method, and when done right, it works splendidly.

But wait... what if you're attempting to control sibilance in, let's say, the vocal track of an entire mix? You certainly do not want the volume of the whole track being turned down at that point. Try it and see what happens. Serious pumpage! (In fact, with the hard knee behavior and high ratio of this effect, it's really easy to make a full spectrum track bob like a cork on the water. This isn't desirable in most cases... but it could lend itself to some really wild effects, too!) Fortunately, Split Mode takes care of this case. In Split Mode, only the portion of the meterial that falls within the sidechain bandwidth (and that exceeds the threshold, of course) is attenuated, not the whole spectrum. This is a real boon to mastering engineers, who have to deal with sibilant vocal tracks, out-of-control crash cymbals, and too-harsh string or horn sections all the time in the context of a full mix, with no opportunity to treat the individual tracks at that point. I gave the DeEsser a real workout by digging up some rock mixes I mastered about two years ago that I remembered had some really harsh sibilance, due to a hashy, overdriven mic pre. It worked so well on these mixes that I found myself thinking how many hours I could have saved if I had access to the DeEsser back then! The optimized setup and the ability to monitor the sidechain signal takes an incredible amount of guesswork out of this all-too-common audio task.

Pros: Optimized for sibilance control. Split Mode works great, and can be a real lifesaver in many full-mix situations. Ability to monitor sidechain makes for super-easy operation.

Cons: Fixed bandwidth on sidechain. Though I didn't run into any material where this was a problem, it's something to keep in mind.

The Big Finish

No review of a Waves product is complete without the obligatory mention of... well, you know, that... that... infernal... thing. Waves calls it a WaveKey, but in computer circles it's more commonly referred to as a "dongle," a little piece of hardware that connects to your parallel port. It's a copy-protection device, or more accurately, a "license" protection device, since you can't run any Waves plug-in without it, and you obviously can't attach it to but one computer at a time. You can actually copy the software all you want, but without the dongle, you can't run any of it.

Let me say now that I have been using Waves software for more than three years, and never, ever have I given the dongle a second thought (except once when I misplaced it during a move, but that's another story). It passes data right through, and has never interfered with, nor diminished the performance of, anything I have ever hooked to the same parallel port. Nor has any other Waves user I know ever reported trouble with their dongle. (Nope -- I'm not even gonna go there...) It's completely innocuous. Heck, it even comes in a nice neutral color (white), so it goes with everything!

In spite of that, I know there are those among you that have a moral objection to dongles -- not to mention key authorization disks, CD-key codes, or any other copy protection scheme of any kind whatsoever. I've noticed this a lot more among longtime PC users; Mac users are pretty used to dealing with things like this, but the freewheeling, do-it-yourself PC culture is only now starting to come face-to-face with the harsh realities of hard-core software protection, especially as more and more companies that started out developing for Mac (of which Waves is one) port their stuff over.

All I can say is; if this is you, then you're really missing out. If you pass up the NPP II just because of an objection to the necessary dongle, I feel sorry for you; the Waves stuff -- not just the NPP II stuff, but all of it -- is just too damn good to deprive yourself of just because you suffer from a case of Dongle Aversion. So do yourself a favor. Buy it, slap the dongle on your computer, and like me, you'll never think about it again. Trust me.

The Native Power Pack II is an amazing bundle of software plug-ins. The Renaissance Equalizer and Compressor are, hands down, the best of their kind money can buy; well worth the price of the NPP II all by themselves. The MaxxBass and the DeEsser could almost be thought of as freebies -- but they shine too; and as far as I know, there's no other plug-in out there that does what either of these does nearly as efficiently or as well. The Native Power Pack II truly does live up to its legendary name. Ease of use, an abundance of cool features, and above all, great sound is what the NPP II is all about!

--------------------------------------------------

Lionel Dumond is the Media and Mastering Editor of ProRec Webzine, and an engineer and manager of the production studios at Sweetwater Sound in Fort Wayne, IN. He's thinking of changing his name to "Bob Ludwig II", just for laughs...