Quote:
Originally Posted by bigshot
The things that many audiophiles take for granted, like a flat frequency response, are the things that *count*. Think about it... When a set of speakers is rated to +/- 3db, can you hear what 3db sounds like? Yes. You can hear it, I can hear it, my 85 year old mother can hear it. And +/- 3db is REALLY GOOD SPECS for speakers.
If you run a signal through a silver cable and compare the output on the other end to the output of copper zip cord, it isn't going to be anywhere near as great a difference as that. In fact, the frequency response problem is going to be easily several orders of magnitude greater.
Next, pull out the specs for your system. Look at the specs for your CD player and amp. Compare those to the specs for your speakers... Again... not even in the same ballpark.
Now, look at your room. Is it acoustically perfect? Is the sound that comes out of your speakers exactly the same as the sound you hear in the room? I bet there are very few here who can answer yes to that. What's the point of spending a boatload of money on a sound system only to put it in a room that throws the frequency response off as much as 20db?
So, why are people worrying about minute differences in cables and distortion levels in CDs that don't even come close to the levels in their speakers, when they could be spending their energy correcting problems that are clearly audible... with a little bit of simple equalization?
|
Flat frequency response: What does that mean with speakers? Well, a measuring curve as close as possible to a straight line, you might say. Indeed. And with a bit of effort that's even achievable, still with the premise: «as close as possible», i.e., something reminding rather on a straight line than a jagged curve. And how can it be achieved? Well, let's say by carefully choosing the drivers (low-resonant membranes) and carefully designing the crossover network taking care of phase and the driver's frequency band. And let's assume decent room acoustics and speaker placement. But don't forget one essential precondition: a smoothing function for the measurement. When you measure a real-life frequency response, you'll get something like this:
See the red curve! The white curve stands for the smoothed frequency response. Why smoothed at all? Well, the unsmoothed curve includes all room reflections causing interferences. And note that this red curve represents the sound waves finally reaching your eardrums. Anybody want to try to smooth these jags with an equalizer?
A real-world equalizer can only smooth an already smoothed frequency-response measurement, but not a real-world frequency response.
Even if we're a bit generous and let the jags live, there's still the problem of direct and reflected sound. What's the goal of your equalizing? Do you want an even direct-sound frequency response (actually a legitimate goal, considering the sonic colors of the music instruments you may want to preserve) or do you want to have an even over-all frequency response, no matter if it consists of an uncontrollable and inhomogeneous mixture of direct and reflected sound? Not an easy decision! Because both approaches will inevitably lead to flawed listening results. The reflections actually shouldn't be there at all. They are an artificial coloration of the original recording. But at the same time the recording is in some way designed to be played in a room, maybe even an average living room, or a prepared listening room, certainly not in an anechoic chamber. A certain amount of reflections from different directions than the speakers themselves are wanted to simulate a real-world acoustic, not a free field. Imagine the sound of a cymbal in a very dry room with lots of carpet on the walls. That's not how a cymbal sounds in its natural environment -- a cymbal has an almost ball-shaped dispersion up to highest frequencies and consequently causes a lot of reflected sound from all directions.
But there's another problem: Particularly tweeters have problems with sound dispersion. There are a few exotic designs with nearly ball-shaped dispersion (some of them with other flaws resulting from this design goal), but common tweeters have quite arbitrary dispersion characteristics: the higher the frequency, the less dispersion. This feature makes for a distinct sonic characteristic. And of course the dispersion characteristics of the various tweeters vary in a wide range.
Then there are horn speakers. They pretend to «control» directivity, but in fact they just introduce a more or less arbitrary directivity with a lot of incoherence throughout the frequency spectrum. And they add coloration of a special kind to the original sound: inner reflections inside of the horns. The typical hollowness and (sometimes) squawkiness with strings so appreciated by classical lovers... They measure as massive decay in waterfall, but not necessarily in frequency-response plots. These colorations are not always that obvious in well designed horn systems, but they are there and audible with experienced ears.
And there are the crossovers: depending on the measuring angle, (frequency-wise) overlapping drivers will cause massive FR irregularities up to almost complete cancellations. This phenomenon can't be avoided, unless you're into fullrange speakers (...with their own massive problems, also such with sound dispersion), so all multiway systems suffer from this. It means individual radiation patterns of different speaker designs with individual, direction- and frequency-dependent degrees of direct and reflected sound which aren't included in the on-axis frequency response.
So if you equalize a speaker to a straight line, what does this mean? That it sounds exactly like another speaker measuring as a straight line? Far from that! The opposite is true. Different designs measuring virtually the same on axis will most likely sound radically different. And now keep in mind that we only have focussed on frequency-response aspects and have left other important criteria such as harmonic distortion and transient response apart. I don't have to tell you that they can have an equal degree of impact on the sound, although generally speaking frequency response is the most important criterion.
Headphones? There is no linear frequency response with headphones. The HRTF (head-related transfer function) is necessary to understand headphone measurements. And it has also to be considered that the scientifically elaborated HRTF is just a coarse scheme; in reality every individual has an individual HRTF.
And now the surprise: Despite the miserable response of sound transducers with every criterion -- compared to electronics components --, we can nevertheless hear sonic finesses in the latter and even in cables. It's hard to explain, but as most people have experienced, the great majority of audiophiles anyway, it's true. To the skeptics, and Steve the Bigshot in particular: At least you're able to distinguish the voices of Madonna and Kylie Minogue, aren't you? I'm not necessarily joking, because given the measurable flaws of sound transducers, that's not self-evident at all.
Quote:
| I know what causes upgraditis... It's a result of people not having a clearly defined goal for their sound system. They get whatever component sounds impressive to them in theory, and they figure it will work all by itself, regardless of whatever else they put with it. It doesn't matter if the sound is *correct* as long as it's *different*. This results in a colored sound presentation. After a few months of "listening to music with rose colored glasses" they tire of the coloration and randomly try another color... and another... and another. |
I know you know a lot of things, but wisdom is to know how little one knows.
Quote:
| ...But you won't get anywhere if you ignore the most basic concepts. |
That's certainly true. But you have a very limited perspective when you cling to basic concepts such as your frequency response. (And that only sound transducers matter.)
