ORIGINAL: Thomas Campitelli
but mixes that sound good quiet always sound better loud, while the reverse is emphatically not true.
You know a great deal more than I, but I have a question as to whether a quiet mix will always sound good loud. How does this statement hold up in the face of Fletcher-Munson curves? Many thanks.
Whoa, this is a big question.
Technically, I should have said "almost always," but in my opinion, it's basically always true (at least for conventional music recordings intended for popular consumption).
As to your question, I should first say that Messrs Fletcher and Munson discovered a lot more than that infamous curve when it comes to how sound is percieved at different volumes, and a lot of research has happened since then. People can have a tendency to read an awful lot into one little chart, and to draw conclusions that aren't necessarily true based on incomplete information, kind of like the blind man who gets hold of an elephant's trunk and decides that an elephant is very like a snake.
For that reason, I suggest NOT relying overly much on technical information unless you really know something about acoustics (like, you have read and understood an actual scientific acoustics textbook or two). Trust what you hear and what you experience rather than what you think you know about the science. Web forums, magazine articles, and famous record producers are chock full of half-baked theories based on an incomplete understanding of psycho-acoustical principles.
So my advice to you is, when you hear a piece of advice like "try mixing at lower volume," rather than try to pick it apart and logically deduce the truth of it, instead try doing it, and see if you don't get subjectively better results. If you do, then great, you learned a new trick. If not, then no harm, no foul-- you learned at least one more thing that WON'T help you.
All that said, here is something I posted once before on this topic that touches on some of the technical stuff:
Autist recommended monitoring, while mixing, at 83-85dB SPL. I suggested that this was an unnecessarily and unusually high volume to monitor at, and suggested listening at very quiet volumes and only occasionally turning up the monitors to check high-volume and low-end playback. Here are the arguments on either side, as best as I can represent them:
85dB SPL is widely recognized as the sound pressure level (SPL) at which human hearing is most accurate. 85dB SPL is the intensity that a standard alarm clock is supposed to be from two feet away-- it is about as loud as a noisy restaurant, or rush-hour city traffic, and is often viewed as a sort of "cutoff point" above which hearing damage can occur with regular or long-term exposure. Most human conversation happens at about 60-75dB SPL, or about 1/4 to 1/2 as loud as 85dB. Movie theaters and movie soundtracks are usually calibrated to deliver an average level of about 85 dB.
Human hearing is not "flat," it's sensitivity depends on sound pressure level, and certain frequencies (low bass, for instance) are almost impossible to hear unless they are turned up fairly loud. Human ears have a sort of built-in EQ that exaggerates the mids and upper midrange the strict "presence range" from about 4-6kHz-- the frequencies of babies crying or consonents in speech--the human ear also more slightly exaggerates a broader range called the midrange (roughly 1,000-8,000 cycles per second, or 1-8kHz).
So human hearing can be thought to have three "tiers," of perceptibility-- the presence range, which is clearly audible, even to people with hearing damage, or even if it is very quiet, or masked by other noise; the somewhat broader "midrange" which encompasses the range of sounds that people genrally can identify, in most circumstances, and where most of the "melody" of music occurs, and the "full range" of human hearing, which extends from very slow-moving waves that are more felt than heard, to the very high frequencies that sound more like "air" than sound.
To get back to the question of what level one should mix at, the meat of the question is psycho-acoustical in nature, and requires a psycho-acoustical primer before any of it makes sense.
If you ever have an opportunity to spend some time in an anechoic chamber (a totally "dead," totally "soundproof" space), it is a weird, disorienting, and somewhat frightening experience. You will typically be walking out on a wire mesh "bridge" in a spherical room covered entirely with wedges of acoustical foam that appear to go into infinity (usually they are about 6 feet/2 meters deep). From the instant you walk in, any sound you produce will seem to instantly disappear inside your throat. After a few minutes, your heartbeat and breathing will start to seem impossibly loud, and if you have any trace of a cold or allergies or if you smoke, the mucus in your lungs and throat will rattle like a flimsy cage in your chest, still falling on a dead world, totally inside you. After a few more minutes, you will start to become acutely aware of your blood moving through your veins in deep, swishy pumps. Any digestion that hapens to be occuring inside you wil seem impossibly, humiliatingly loud. If you stay in the room for more than half an hour or so, and your hearing is not damaged, you will start to hear a steady, soft pattering, like quiet rain. This is the sound of silence--it is literally the sound of randomly-moving air molecules bouncing off your eardrums. It is an other-wordly experience, and not a pleasant one-- being in an anechoic chamber is almost universally regarded as disorienting, uncomfortable, and sometimes nauseating.
A far more common experience is being exposed to loud sounds. Anyone who has ever been to a loud rock concert is familliar with the experience of not being able to hear well for the next day or two. It's like walking from bright sunlight into the house, except your ears take longer to adjust than your eyes do. A few minutes after walking indoors in midsummer, you can see clearly, but it can often take a day or two after exposure to a few hours of loud music before you can clearly hear normal speech in real-world background noise.
More to the point, perhaps, is the fact that "quieter" recordings sound almost universally "lower quality" than "louder" sounds. Speaker salesmen have known this for years, as have admen-- TV commercials are almost invariably loud and heavily compressed. Decades of research has shown that, given the same piece of audio, people not only prefer, but better remember a louder version.
Another true fact is that human hearing is not "flat," like speaker response, not even close. Frequency response of human hearing improves the louder that sounds hit the ear, moreover, frequency and harmony perception actually CHANGES at different volumes-- this is important-- Fletcher and Munson demonstrated that two notes change from sounding out-of-tune, to in-tune, to out-of tune once again as you change the volume of the two notes-- this is for real. We are all familiar with the notion of tuning to A 440, and we take it on faith that that this tuning will work across all frequency spectrums and all instruments, but it doesn't. You can take two slightly out-of-tune instruments playing different octaves, and they will come to sound in tune at certain volumes. Similarly, if you take two "in-tune" instruments playing different octaves, and change the volume, they will sound out-of-tune.
This is for real, and has to do with the way that people percieve different frequency ranges differently depending on their volume-- extremely high-pitched or low-pitched sounds sound dangerous and scary if they're loud, and merely annoying if they're low. And their relationship to midrange sounds changes as they change in volume.
Somewhere in the middle of the volume range is where most people actually listen to recorded music-- in the car, in a bar, as background music while you're washing the dishes or waiting for the dentist, at a party while you're trying to have a conversation, at a cookout with a boombox in the window, in general on crappy playback systems under vastly subpar listening conditions, and so on... the remarkable thing about human hearing is that people hear loud music against a lot of background noise much the same way they hear quiet music against silence-- in both cases, people hear the midrange and presence range more than anything else. The only real-world circumstances under which people have "flat" frequency respone is in dedicated "listening" environments, where background noise is kept to a minimum, and where the music volume is medium-loud (about 85dB, to be precise).
So given all that, 85dB would seem like the perfect SPL level to work at, right? Well, sort of. That's the level that movie mixers and mastering engineers work at, but movie mixers are mixing "already mixed" songs, dialogue, and sound effects, and mastering engineers are obviously dealing with "pre-mixed" material. Actual professional recording and mixing engineers are little different, in my experience. For one thing, constant, day-long SPL levels of 85dB are right on the cusp of workplace safety standards-- constant levels of 85dB might not damage your hearing, but they definitely qualify as "bright sunlight" audio levels-- they make it harder to evaluate quieter sounds, and "dim" one's hearing. Moreover, monitoring at 85dB might give the engineer the most "accurate" perspective on the recorded material, but it does not necessarily present an accurate picture of what the listener will hear-- for an obvious example, listen to some audiophile recordings of medieval choral music, such as the Tallis Scholars' album "Spem in Alium"-- at conversation-level volumes, this masterpiece recording sounds out-of-tune.
The holy grail of commercial mixing engineers who have to sell records for a living is to create great-sounding mixes that sound just as good in all different kinds playback systems and listening environments. Unfortunately, this goal is impossible. There is no way to create a mix that is equally well-suited to a Jeep, a boombox, a noisy bar, a $400 home theater system, and a $400,000 audiophile listening room. There is no way to create a mix that sounds the same in a noisy car or bar as it does in a movie theater or a quiet listening room.
So the question is, where do you draw the line between mixing for the real world and the ideal scenario? In my experience, professional mix engineers who are mixing for popular consumption listen at all different levels, on many different playback systems, and tend to focus on making a mix that sounds good quiet, knowing that mixes that sound good quiet will have a tendency to translate well on a variety of playback systems at all volumes, both aesthetically and tonally, as pitch perception tends to be more forgiving at higher volumes, and more discriminating at lower volume.
If I were making dedicated audiophile recordings that I expected to be sold to and listenend to by people with excellent playback systems, I would do it differently.
My $.02, anyway.
post edited by yep - 2007/01/08 14:28:13