A Note on Equalization by Benjamin      Back to Music      September, 2004 In order to fully enjoy music, it is essential that the quality of the sound be excellent. Not only must the music be well played, but it must also be well recorded and played back in a pleasing way. We may not have much control over the recording, but there is something we can do to make the music sound much better when we play it back and that is to adjust the equalization. It is my strong opinion that with reasonably good equipment we can make the music sound great without spending a ridiculous amount of money, by taking the following into account. Equalization (or EQ) means to use electronics to boost or diminish the various frequencies in the music being played. We cannot adjust each of the frequencies, since they are infinite in number, but we can adjust different frequency bands by using a parametric equalizer. In the old days, we simply had a treble and bass control button (i.e. two bands), but this could be quite effective as will become clear in the following. There seems to be a misconception that sufficiently good equipment should need no EQ, since the goal is to reproduce the music as 'accurately' as possible. At most, some slight EQ might adjust for the acoustics of the room. This misconception is false, since I am quite convinced from years of experience that many recordings - especially classical recordings - sound flat and dull until the EQ is significantly adjusted. Either that or there is something seriously wrong with my hearing, but I don't think so because I play the classical guitar and that sounds fine. In a nutshell, the bass and treble need to be boosted by about 2-3 dB for the bass and by up to 6 dB for the treble. The following graph shows how I adjust the equalization in my iTunes music player when I play music at my computer, which is where I listen to most of my music. Audiophiles may immediately object that a computer cannot provide decent music, but this is untrue if one uses good external speakers. I have the Harmon Kardon Soundsticks hooked up to my Apple computer, along with a subwoofer. It sounds great for most music, better than I have heard on much more expensive equipment in some music stores, provided I don't play it too loud. For loud music, I use a normal stereo system, or, more likely, in order not to bother the neighbors, I blast the music on my rather excellent car stereo. The EQ for the car stereo is similar to what I will now describe for my iTunes player. A similar EQ would also apply to any stereo system, for most recordings. (Some of the newer recordings seem to have this EQ built into the recording process, thus confirming my philosophy. In this case, a flat EQ is preferable.) The graph shows each of the 10 bands in my iTunes parametric equalizer. (The bottom two bass bands are scrunched together on this plot.) For reference, the values of the equalizer are also presented in this table: Note that the midrange (about 500-1000 Hz) is unchanged, while the bass below 500 Hz is gradually raised to 3 dB at the lowest frequencies and the treble above 1000 Hz is gradually raised to 6 dB. These values can be adjusted to taste; the important point is to give a significant boost. (In my car, the separate subwoofer control actually pumps the bass quite a bit more at the lowest frequencies.) Boosting the bass and treble can also be done with the simple bass and treble buttons found on most audio amplifiers. I have always turned these buttons up since as long as I can remember, on every stereo. However, a parametric EQ evidently gives us more control over the exact shape of the equalization. The effect of this EQ is to make the bass sound rich and solid and the treble to sound spacious, airy and crystalline. This combination is what makes music sound great! I think that many of the 'audiophiles' who still believe that the old analog vinyl LPs had more 'air' are simply listening to CDs with the treble suppressed. On the other hand, those who think CDs sound 'harsh' are probably listening with EQ that is boosted too much in the 2000-4000 Hz range. This 'knee' in the curve is critical, and you must experiment until you get it right. I also think that a decent medium-priced loudspeaker is about as good as a much more expensive one, and any apparent difference is again due to EQ (i.e. different speakers emphasize different frequencies). Since a decent speaker has distortion that is almost certainly inaudible at a reasonable volume, what else could the difference be? One might discuss the radiation patterns of the speakers and the phase relationships between them, but I think these are relatively minor issues, unless you are trying to fill a large room with sound. I have never heard a music system for a large audience that sounded really good. Most of the rock concerts I have been to sounded terrible. But if it's just you sitting in front of your speakers in your living room, you don't need to spend a ridiculous amount of money. Notice that music stores may let you switch between different speakers during the demo, all allegedly coming from the same amplifier, but how many of them let you adjust the EQ (or even have an equalizer in the chain)? (At the same time, I'll admit that when I buy audio equipment, I err on the side of spending more money, just to be sure. Most of my skepticism is directed at the stratospheric prices of so-called 'audiophile' equipment. Where can one honestly test this stuff anyway?) A few final comments. I cannot detect any difference between MP3s recorded at 192 kbps (or above) and the original CD. At least I couldn't on a few CDs I recently tested. So I don't think that 'digital distortion' is really an issue, if it ever was. That is, if I can't even hear an improvement going from the 'compressed' MP3 digital format to the 'full' CD digital recording, then why would analog be any better (unless it were a matter of EQ)? By the way, this also indicates that those who think that CD is better than MP3 are probably just being fooled by an EQ difference, unless their ears are truly more sensitive than mine. I doubt it. And I doubt that the new SACD format is really intrinsically any better. If a given SACD disc sounds better, it is probably just the EQ again, or maybe it is better recorded. Since SACD is proprietary, to hell with it. I have bought virtually all of my huge music collection with my own money, but I love the flexibility of MP3 for many perfectly legal reasons. And I hate having that flexibility denied. It's like having to get a stranger's permission before driving your own car. I don't even like buying Apple's proprietary AAC format, though I do on rare recordings if I have to (and I know how to make MP3 out of that). The SACD advocates boast that SACD goes up to 40000 Hz, unlike the 'mere' 20000 Hz of standard CD. The simple truth is that we really cannot hear above 20000 Hz unless we are dogs. Those who think otherwise should play with a software tone generator. Most problems with recordings are the fault of the engineer. The classical recordings tend to be pedantically recorded with flat EQ, which makes the music sound flat, while the popular music people, going by their instincts, long ago learned how to 'enrich' music with an EQ similar to what is described here. Forget about the theoretical 'authenticity' of the music; what matters is how it actually 'sounds'. Duh! However, one distortion problem does often arise with my iTunes (but never with my car stereo): clipping. This is when the music exceeds the maximum value that the digital numbers can hold (all 1's). This can happen due to amplification somewhere in the digital chain, including with the EQ. When this happens, I simply shift the curve down to where the maximum boost is 0 dB. That is, I subtract 6 dB from all the values indicated on the graph. Then no clipping should occur if the music was properly recorded. However the volume may then be too soft, so when I need to (and can do so without distortion) I shift the curve back up by some amount. What matters for the EQ is the relative values of the various boosts (i.e. the shape of the curve) and not whether it is shifted up or down. Shifting it simply amounts to a volume change, but just changing the volume won't fix the problem if the clipping occurs before the music reaches the amplifier (e.g. right after the EQ). A final problem is not to drive your amp too hard (i.e. turn the volume up too much). You can't blame that on the EQ. In conclusion, let me emphasize that nothing is more important than the quality of the recording. Examples of superb recordings would be the Enigma best hits album Love, Sensuality and Devotion or the B-Tribe albums mentioned on my music page. On my middle-of-the-road music systems, these have a rich, sparkling, and luxurious sound that puts many classical music recordings to shame. I think that many classical recording engineers simply don't have a good ear for a gorgeous sound. They probably strive for 'accuracy' according to their instruments rather than just listening. Even recent classical music recordings can disappoint, by sounding flat, or dull or a bit harsh (especially the strings). The acoustics are crucial. The sound must be smooth and spacious but not muddy. This can be achieved through natural or artificial reverberation, but crystalline clarity must always be preserved and the atmospherics must not be excessive unless the music calls for it. A church may or may not be the best place to record. Some good recordings have come out of smaller companies like Chandos, for example the Margalit Schumann piano concerto. The quality of the sound on recordings from large companies like Deutsche Grammophon can vary from poor to excellent, which again shows how much depends on the ear of the actual recording engineer. Breaking News: I've just discovered that the phase associated with a given frequency can important to the presence, sparkle, spatial definition and overall realism of the sound. This is in addition to the amplitude of the frequency, controlled by the EQ. Apparently the relationship between the phases can be distorted during the recording process. There are hardware devices to correct this and also software plugins for the computer. I will have to investigate this. May or may not be aurally significant...