Design Article
Loudspeakers: Objective evaluations - Part 1: Sound source radiation patterns
Floyd Toole
11/22/2010 7:41 AM EST
Figure 18.1a shows an ideal point source that, as a function of distance, experiences a rapid increase in the surface area over which the sound energy is distributed. Because the energy per unit area (sound intensity) is inversely proportional to the square of the distance from the source, this phenomenon has come to be called the inverse-square law. The sound level correspondingly falls rapidly, at a rate of -6 dB/dd (dd = double-distance). This happens only in the far field of the source.
Beranek (1986) suggests that the far field begins at a distance of 3 to 10 times the largest dimension of the sound source. At this distance, the source is small compared to the distance, and a second criterion is normally satisfied: distance2 = wavelength2/36
In the near field, as shown in Figure 18.1b, the sound level at any frequency is uncertain. Figure 18.1c shows estimated distances at which far-field conditions should prevail for a loudspeaker system and for its components. This would be the minimum distance at which a microphone should be placed for measurements and at which listeners should sit to have a predictable experience.
In a room, closely adjacent reflecting surfaces must be considered to be part of the source. This means that the far field for the combination (loudspeaker plus a very early reflection) can be very far away.
Diffusers behave as secondary sources of sound, and they can cover significant areas of room surfaces. Cox and D'Antonio (2004, p. 37) point out that listeners should be placed as far from scattering surfaces as possible, at least three wavelengths away. For devices that are effective to 300–500 Hz, this is a minimum distance of about 10 ft (3 m). As they realistically point out, "In some situations, this distance may have to be compromised."
So what is heard while standing close to the loudspeaker and its immediate environs can be very different from what is heard farther away, especially if one is moving around and by doing so enhancing the audibility of any near-field lobing or acoustical interference. Such effects are especially audible with stable broadband sounds like pink noise. Back in the listening area, sitting down, listening to music or movies, the audible result will be very different and much more pleasant.
FIGURE 18.1 (a) The classic illustration of spherical spreading, originating with a point source. In the far field, the sound level falls at a rate of -6 dB per double-distance. (b) A graphic illustration showing the disorderly near field and the predictable far field behavior of a source. (c) Estimates of the distances at which far-field conditions are established for a three-way loudspeaker system and for its components, singly and in combination.
|
|
In recording control rooms, it is common to place small loudspeakers on the meter bridge at the rear of the recording console. These are called near-field or close-field monitors because they are not far from the listeners. As shown in Figure 18.1c, the near field of a small two-way loudspeaker (the midrange and tweeter of the example system) extends to somewhere in the range 21 in. to almost 6 ft (0.53 to 1.8 m). Including the reflection from the console under the loudspeaker greatly extends that distance.
There is no doubt, then, that the recording engineer is listening in the acoustical near field, and that what is heard will depend on where the ears are located in distance, as well as laterally and in height. The propagating wavefront has not stabilized, and as a result this is not a desirable sound field in which to do precision listening, but as they say, perhaps it is "good enough for rock-and-roll."
|
Some of these far-field distances are much greater than the 1 m distance universally used for specifying loudspeaker sensitivity (e.g., 89 dB @ 2.83 v @ 1 m). There is no problem here because in the standards that specify the rituals of loudspeaker measurements, it is stated that the measurement should be made in the far field, whatever that may be, and then the sound level that would be expected from a point source at 1 m should be calculated.
For example, if a measurement is made at 2 m, 6 dB should be added to arrive at the sound level at the reference distance, even though 1 m may be within the near field of that particular loudspeaker. The 1 m standard distance is therefore a convenience, not a directive that a microphone should be placed at that distance. Many people have misunderstood the intent of the standard distance, including some major players in the loudspeaker business.
If it is necessary to make measurements within the near field, useful data can still be obtained by spatial averaging: making several measurements at the same distance but at several different angular orientations with respect to the loudspeaker and averaging them. This is another of those uncertainty principle situations. By spatial averaging we have a better idea of the true frequency response, but we don't know the axis to which it applies. If we measure at a single point within the near field, we know the axis precisely, but we don't have a good measure of the frequency response.
|
|
|
|
|
sharps_eng
11/23/2010 7:01 PM EST
Brands of loudspeakers are recognisable by their characteristics, in one case in my experience by their lack of charateristics - I once walked past a London shop doorway and for a second or two felt a bubble of clear sound round my head as I passed; sure enough, at the back of the store was a Tannoy 15in monitor pointing outward. Its outstanding phase-coherence provides a unique and (sadly) unusual experience. As the author points out here, however, room placement and loudspeaker format interact hugely in making the LS more or less sensitive to listener position and movement, and any improvement through this work will be welcomed by audiophiles everywhere.
Sign in to Reply
selinz
11/25/2010 1:46 PM EST
It's interesting that the whole phase linearity thing came and went in the late 70s. Recently the "big thing" is frequency response customization. Most high end recievers have Audyssey (you set specially calibrated microphones in the primary listening areas)MultiEQ room calibration or equivalent to take care of room characteristics, but generally at the cost of phase linearity. There's no denying that it sounds great, by my 1977 vintage phase linear speakers still have the effect of turning your head to see if it's real or not, particularly with brass and drums...
Sign in to Reply
Ralex
11/26/2010 6:55 AM EST
It's interesting how little emphasis it is about what sounds natural and good to the ear. Too many people are into technical mumbo jumbo and the results are too many products doing room correction sounding real bad. I investigated this during the 2000s and wrote about it in AudioXpress 2004 issue august, september and october. Basically, I found the the room messes up the direct sound, meaning that the only thing that can be corrected sounding good in the mid and high frequency domains is the sound coming from the loudspeaker. A corrected loudspeaker sounds great and very realistic, the same goes for a phase linear loudspeaker! The low end can sound great by removing offending peak resonances in a room. Any parameteric eq can do this. Ask any live sound engineer. Easy.
Sign in to Reply
WKetel
11/26/2010 4:25 PM EST
What is unfortunate is that for a large portion of our citizens it will not matter in a few years. The practice of playing sounds that I will not call music at very high levels has been demonstrated to cause permanent hearing damage, to the point where it will not matter what the phase relationship is for the sound reproduction system.,The very non-linear frequency and amplitude response of the hearing of these individuals will assure that nothing that they hear will be undistorted or sound good. Truly a tragedy, but truer a reality.
Sign in to Reply
kdboyce
11/26/2010 6:13 PM EST
WKetel has a good point, and if I may amplify (pun intended) it a bit.....
Audio professionals and audiophiles alike bemoan the fact that the vast majority of the sound buying public has no clue as to what good sound really sounds like. (Do you think I got the word sound in there enough? )
Now the market is flooded with so much MP3 material that the younger generation thinks this is what it should sound like. Couple that with so many other types of audio codecs (MP3Pro, OggVorbis, AAC, FLAC, to name a few), the digital audio world is more like alphabet soup than audio and are hard to distinguish between.
I was fortunate enough to work in a group with people who did know high quality and how to get it on the amplifier side of the equation. And they listened to this on really good speakers.
What would spell doom for good music would be when the sound recording engineers that mastering the recordings to take advantage of MP3. UGH!
As for me, I won't stick earbuds in my ears. There is no faster way to ruin your ears than earbuds and an MP3 player. Well maybe being in a performing Rock band is faster....
Sign in to Reply
K6TVM
12/1/2010 12:12 PM EST
Around the 1940's a group (Bell Labs?) did a study to find out what kind of sound the public actually likes. They compared in a blind test really good sound systems against really bad sources, like the telephones and table-model radios of the day - the five-tube AC/DC things. Turns out the most popular source was also the most restricted - the little AC/DC radios with the three-inch speakers! You wonder why they like MP3 and think cell phones sound okay?
I wouldn't worry too much about the public at large - they never did like good sound!
John Amos
Sign in to Reply
kdboyce
11/26/2010 6:19 PM EST
Additional comment:
Really good speakers and amplifiers make a BIG difference.
In Taipei, there was a restaurant run by an audiophile who had set aside part of his establishment as a kind of studio which could be seen by patrons. Also, huge vacuum tube amplifiers and a massive speaker stack was visible at the front. They had to be capable of several hundred watts per channel at a minimum.
This guy would play very high quality recordings over that system at a volume level that was quite low compared to what it could be. The sound that came out of that set up was glorious and was really appreciated by all. So much so that it was the quietest Chinese restaurant I had even been in. :-)
Sign in to Reply
sharps_eng
11/27/2010 4:29 PM EST
My Tannoy DL2000s have an annoying response 'bump' at 50Hz (an original feature) which I have always intended to iron out with acoustic treatment. Hasn't happened yet - (a) I'm an electronic engineer so in true style my hifi is only ever half-finished and (b) the room keeps getting remodelled by the C-in-C :-)
So by being flexible and easily repeated, audio analysis and pre-compensation can keep pace with changes in a room. This research should benefit the effectiveness of that DSP acoustic code in receivers.
Sign in to Reply
divide_by_zero
12/8/2010 6:12 PM EST
Sigfried Linkwitz (of Linkwitz-Reilly crossover fame) has a web site where he not only sells speaker systems, plans and kits, but also discusses all of the parts of the audio reproduction chain. This includes a discussion of what live, unamplified music in a good venue sounds like and how most listeners are accustomed to badly recorded, close-miked, over-produced recordings played back by systems designed for these inferior recordings. He makes some compelling arguments. I'll pass judgment when I complete my Pluto speakers. See LinkwitzLab (dot) com.
Sign in to Reply
Cookie Jar
2/4/2011 1:50 PM EST
Shame on loudspeaker engineers using pink noise to measure frequency response "to avoid troublesome resonances". Unfortunately, music consists of sustained frequencies which tend to excite every resonance in your transducers, enclosures and room. If you measure loudspeaker response with sine-waves, you get a very wild response curve indeed. That is in fact your real world response curve, but it's bad news for marketing.
There's a lot of resonating going on. You have the mass of the cone and its spring. The Q of the enclosure resonances are commonly over 20!
Don't forget the resonances of the room.
The solution? Make your room dimensions a multiple of the cube root of 2 to spread the resonances evenly and use an open back full range electrostatic speaker. Nirvana!
Sign in to Reply
Sasquatch884
3/24/2011 11:00 PM EDT
I worked as a recording engineer from 1965 to 1980, the last 12 years of that period at Columbia Records studios in Manhattan. I worked on pop, jazz, rock, blues, R&B and classical music, with artists as diverse as Barbra Streisand, Bill Evans, Edgar Winter, Junior Wells, J.J. Jackson, and the Budapest String Quartet. I never once gave a rat's you-know-what about speaker specs. Here was the test I used - I placed a high-quality condenser mic, a Neumann or Schoepps, a few feet away from an open grand piano, put my head near the mic and had somebody competent play the piano. I would then walk into the control room, open the fader for that mic, get the volume up to about what I had heard in the studio. If the piano sounded different, the speakers failed. If it sounded really close, the speakers passed. This is all without equalization - not on the mic channel and definitely not in the monitor system. The only speaker that ever passed that test - unequalized and in different control rooms - was the original Electro-Voice Sentry II. I started using them at independent studios and then had Columbia buy a couple of pairs when I worked there. I still own a pair. They are simple, classically-designed two-way bass-reflex speakers, with a stiff 15-inch woofer, a small-throat horn tweeter crossed over pretty high and a cabinet with the resonance of a wet brick.
Sign in to Reply
Nic Cohen
7/27/2012 4:06 AM EDT
I do believe that the quality of sound in relation to the majority of the general public is lost when it comes down to the brand and functionality of the device they choose to use, as the quality does not always correspond to the ""quality"" of the product.....
Nic
www.kdweb.co.uk
Sign in to Reply