My high school band director used to say, "what you do in practice you will do in performance." I must say, though I was something of a cut-up in band, this statement stuck, and over the last 30 years I have learned that it is quite true: If you spend time goofing around in practice, using bad form (since you're not actually 'performing' at the moment), practicing without a metronome, etc, the repetition of these actions will literally compete with your better skills and likely negatively affect your overall performance level. For this reason I have counseled bands, when preparing for a recording, to practice with a click -- literally don't play a note without a click going -- even if they have no intention of actually tracking with a click during the sessions. Why? Because the practice and repetition of this will improve their groove and make them play more like a single unit rather than a group of individuals, even when they play without a click.
This idea holds true for recording engineers as well. If you're going to be mixing records then I consider it helpful, indeed necessary even, to do your critical music listening in a consistent environment, ideally with consistent monitors and at similar levels that you use when mixing. Without such familiarity and consistency with your monitor system you may be surprised by how your killer mixes translate when played back on different systems.
Way back in 1937 a couple of fellows by the names of Harvey Fletcher and Wilden A Munson published what they called their "equal-loudness curves,' which were graphic curves showing human hearing sensitivity to pure tones at frequencies throughout the audio spectrum. These curves were more recently revised to what is now called the ISO 226:2003. Here is a graphic of both the original Fletcher-Munson curves and the ISO 226 standards.
So what does this show? Well, the blue lines are the Fletcher-Munson curves from 1937 and the red one are from ISO 226:2003. They differ a bit here and there, but overall are quite similar, and show us something fascinating: human beings (ALL human beings, musicians, mix engineers and music consumers alike) are more sensitive to some frequencies than others, and the perceived frequency relationships change with actual playback amplitude.
SPL (in dB) is given for the vertical axis and frequencies are given for the horizontal. The curves show the actual amplitude of frequencies throughout the spectrum needed to give the listener perceived equal level as given in phons (the unit of loudness level in pure tones... google it if you want to know more :)
So this graph shows that for a perceived equal loudness of 20 phons a 20Hz tone needs to be hitting 90dB of actual amplitude compared to only 20dB for 1kHz, 15dB for 4kHz, and about 35dB for 10kHz. That's for the perception of 20 phon for each frequency, which is very quiet. At 100 phon the 20Hz tone needs to be 130dB to match the perceived loudness of 1kHz at 100dB (ouch!!), 95dB for 4kHz, and 105dB for 10kHz.
At this point you're no doubt yawning, "interesting, Joel, but what is your point?"
Well, my point for bringing this up is this... a brief look at the graph and you can see that we humans are less sensitive to highs and lows in the audio spectrum, meaning they are harder for us to hear (compared to midrange tones) at softer levels. So when you're mixing you are much more likely to compensate for these ranges by using EQ to make them louder. When you monitor loudly, on the other hand, you will naturally perceive more of these extreme ranges and therefore create a flatter, less enhanced balance of frequencies.
This is why your mixing will benefit from doing any critical listening in a consistent environment with consistent playback system at consistent levels. I regularly listen to music in my mix room through the same ProAc Studio 100's that I mix with and at a similar level. Even when I'm listening to music, even for enjoyment, I'm constantly training my ear to my mix environment and playback system. So when I make decisions during a mix I am confident I know how those choices will translate when played back on multiple systems.
"What you do in practice you will do in performance."
For this reason also I avoid, at all costs, consumer EQ on car stereos and home entertainment systems (in fact I've never really invested in a 'good' home stereo, since I don't play music on it for anything other than some background music on occasion, usually for social gatherings). I am always a bit shocked when I get in someone's car to hear that they have there stereo with the 'treble' and 'bass' cranked, leaving all the mids comparatively buried. While this is all fine and good for consumers, I strongly encourage music professionals to adjust to listening to music flat through quality systems, so they can better trust their ears when production time comes. You won't be mixing through cheap, consumer, Baxandall EQ, so don't train your ears to expect it.
One other thing to take from all my ranting: whether you monitor loudly or quietly, in an excellent acoustic environment or not, using an expensive playback system or a pair of Chinese cheapies sold by the palette at your local music store.... get to know your system in its environment. Countless fantastic records have been tracked and mixed on Yamaha NS10's (!!) which are anything but accurate, full-range speakers. Yet they were everywhere for a couple of decades, and people got to know (and trust) them well. They were a known commodity that were in great supply, and they remain studio legends to this day (though I personally can't stand them... but that's just me).
Again, what you do in practice you will do in performance. So get used to listening to music the way you will be mixing it, so you can confidently make the right mix decisions to get the mix you're after, time after time.