Stereo systems

Stereo systems 1

On the pages dealing with room acoustics and recording we concluded that the response of the room would directly affect and color the sound. The same applies to any loudspeaker. The frequency response of the speaker is of outmost importance. There are, of course, many other aspects that determine the performance of a speaker: distortion, dynamics, phase, low-level behavior and phase integration between the elements. However, frequency response is one of the most clearly audible and easily understood ingredients. (In real life it is, of course, a very complex question, but a basic understanding will make a big difference for you as you will see from the following paragraphs).

Remember that the direct sound from the speaker is of outmost importance and that it should be as straight as possible. No EQ in the world can save a bad frequency response, as the measurements of the speaker in a room will only measure the overall sound. More about this subject at bottom of this page. (The reality is more complicated and I need to point out that the frequency curve should in fact not be perfectly straight as regards to the direct sound from the speakers. This is due to system errors and has to do with the fact that all stereo systems – with the limitations of radiating sound with only two speakers – cannot present a sound image compatible with the concert hall. The adjustments needed for these system errors are within 1 dB. That is to say, given the limitations of most speakers on the market, it means that the demand to keep deviations within 1 dB does put stringent demands on nearly any system available! Please, keep this in mind while reading the following text).

A very common extra loudspeaker in control rooms, the small Yamaha NS-10 M (M as in "Monitor") has a particularly bad frequency response and bad phase integration between the two elements.

The phase between the two elements is about 120 degrees out in the area between 3-4 kHz. The result is that the frequency response in that particular area will be strongly affected if you normally listen on the central axis between the two elements, or alternately, as often is the case in a control room, under or over the axis. In addition, the sound will also be affected if the "monitor" is lying on its side or standing up. You can see on the picture that the worst case is a 16 dB difference in the response when listening above the central axis. The sound transmitted to the room is also very irregular and this will result in undefined reflection-patterns from the room. This in turn, will not convey the original recording environment. All of this is, of course, bad for the listening experience; certainly, with this speaker, you cannot get a clear perception at all of the true recorded sound and of the room it was recorded in. In this matter, it is not a "monitor" speaker. Note also the very resonant treble response! So much resonance is due to a very bad treble element.

If someone would claim that you only need a good amplifier to remedy this problem — I have frequently seen this argument in stereo magazines – I can only state that no amplifier can remedy such bad phase integration or such bad frequency response. A worst-case amplifier with a very bad cable (read expensive with high inductance and capacitance) will affect the amplifier's frequency response only a few dB up or down.

Certainly, a better way to go is to have a speaker and a room that both have as much neutral response as possible. It will take too much space here to discuss all kinds of factors that contribute to high quality sound. A good source of information (in Swedish) is the society LTS whose homepage is: http://www.lts.a.se/lts/startLts.html. You can find lots of information there concerning building cheap and electrically efficient cables, building inexpensive, but high-quality loudspeakers and many more things.

Note that often you will see claims that frequency response is not so important, as you can always acquire an equalization unit to change it and also that the room will always affect the speaker's sound. Actually today, high-end modern surround amplifiers contain a small automatic computer that will measure the frequency response and change it to a flatter version for just your room. But, remember this: any change of the frequency response will be of the overall measured sound in the room. The direct sound from the speaker will be even worse after the EQ than before the change. Even if the overall sound in the room is bettered and more flat. It's in real life only possible to adjust the response in the bass register with an EQ without any bad side effects.

A much better way of proceeding is to acquire a speaker that is as flat as possible and then try to remove or reduce the (bad) influences in your actual room. The goal, ultimately, is to extract a flat response from both the speaker and the room. Remember also that any EQ will change the phase of the sound and also the time between the transmission of sounds in different registers. An EQ will therefore always distort the original sound. And we really would like to avoid any kind of distortion!

Generally speaking there are only two types of people with an interest in HiFi:

1. Those interested in a sound that is either soft, musical, dramatic, cool or whatever their preferences are ...
    (Most HiFi magazines offer endless advice in these matters).

2. Those interested in how the CD (or any recording medium) itself sounds.
    (These will be individuals who want a system that is transparent and that will not add or subtract
    anything from the recording).

Page 2