why tons of commercial speakers sound like crap

[doors666]

There's an interesting thread going on at techtalk forums. Some one has shown the freq response graphs of tons of commercial speakers and man, do they look like crap. There are tons of super expensive speakers listed there, e.g. B&W 800D, 805D, cm1, wilson audio duette, sophia2, puppy, vienna acoustics mahler, mozart, maggies, sonus faber etc. Whats going on here, these guys with millions in research budget cant do proper xo design or what? Some of these look like some random drivers thrown in, some of them can use reversing tweeter polarity. Do they intentionally sell crap and just dont care.

I know that flat freq response isnt everything, but still, it isnt all that hard to get that, when diyers with 1000$ measurement equipment can get there, why cant these guys?

 

[magma]

if everyones speakers had the same parameters they would sound pretty much the same wouldnt they ?

 

[doors666]

if everyones speakers had the same parameters they would sound pretty much the same wouldnt they ?

Not really, no. there are a lot of other important factors. power handling, power response, impedence, xo point, phase and not to forget the distortion profiles. I know flat response isnt everything, but it cant hurt to have that. I mean, why should some frequencies sound louder or quieter than others?

 

[acoustics]

A lot of commercial speakers out there in the market have a very less worth of drivers inside them , for example the mid segment rti speakers from polk audio use a $8 mid range and $5 tweeter and a crappy crossover , and their lsi series has a $ 25 worth of vifa tweeter which make them special and people are being charged extremely high just for the fancy looking enclosures .Now i dont mean that using expensive drivers make the speaker good , its just that people are being overcharged for the products in a rediculous way. Alot of branded speakers have poor implementations of good and expensive drivers and a lot of brands have cheap drivers as well as a bad implementation .
DIY is the best but one should have a good knowledge of each and everything b4 throwing their money in .
when a person buys a branded speaker he never cares to check the actual response curve because of two things one having the blind faith in the branded speakers and secondly he does not have access to the equipment to check the FR .

 

[reignofchaos]

Unless those speakers were tested in the same anechoic chamber with the same equipment, looking at those frequency response graphs makes no sense whatsoever.

Personally having heard both, I'd pick a Wilson (in that thread's worst list) over a Revel(in that thread's best list) any day of the week. There's way too many speakers out there that have a ruler flat impulse response but sound like crap.

 

[doors666]

A lot of commercial speakers out there in the market have a very less worth of drivers inside them , for example the mid segment rti speakers from polk audio use a $8 mid range and $5 tweeter and a crappy crossover , and their lsi series has a $ 25 worth of vifa tweeter which make them special and people are being charged extremely high just for the fancy looking enclosures .
DIY is the best but one should have a good knowledge of each and everything b4 throwing their money in .
when a person buys a branded speaker he never cares to check the actual response curve because of two things one having the blind faith in the branded speakers and secondly he does not have access to the equipment to check the FR .

I agree, but the speakers mentioned in the thread are mostly costing thousands of dollars, some of these brands are known to use real good quality drivers.

 

[keith_correa]

Measurements are one side of the story. Voicing them is another. Like someone says at the PE forum - "You do not listen to wiggles" - or something to that effect.

 

[Thad E Ginathom]

Don't know. People buy speakers that sound "good," or "right," not speakers that sound "correct". Which is, I suppose, saying that High-Fidelity is a myth.

Studio monitors are supposed to sound and measure flat. I think hifi speakers are designed to have a "voice." Like the voice: buy the speakers! Whatever the "wiggles" say!

 

[navin advani]

for example the mid segment rti speakers from polk audio use a $8 mid range and $5 tweeter and a crappy crossover , and their lsi series has a $ 25 worth of vifa tweeter which make them special and people are being charged extremely high just for the fancy looking enclosures ..

Agreed.

For example about 30 years ago a cousin of mine gave a Bose 301 for repair. I tracked the woofer and tweeter to the OEM and they cost $7 and $1.5 each. The Old Bose 301 had 2 cone tweeters. So total parts were $10 per speaker. $20 for a pair. The crossover was a bulb and a cap. I assume it cost Bose to mass manufacture the boxes for $20 (1982 $) per pair. Total cost of part $40. Price paid by my cousin $300 (1981-2). See pic.
http://forums.audioholics.com/forum...series-1-circa-1975-anyone-know-these-301.jpg

But.....a manufacturer also has to pay for packing, shipping, handling, insurance, overheads, distributor margins, dealer margins, etc... so they typical speaker sells for about 6 times it's direct cost and 10 times the cost of all it's Raw Materials. The high end speakers have larger spreads as they have fewer volume. So a Sonus Faber, Wilson Audio, or Vandersteen would typically sell for 15 times their Raw Materials (drivers, box, crossover, etc..). I wont get into brands like KEF, B&W, Revel, Focal, Theil as they make their own drivers (or get custom drivers from OEMs).

Unless those speakers were tested in the same anechoic chamber with the same equipment, looking at those frequency response graphs makes no sense whatsoever.

Personally having heard both, I'd pick a Wilson (in that thread's worst list) over a Revel(in that thread's best list) any day of the week. There's way too many speakers out there that have a ruler flat impulse response but sound like crap.

...and I loved the Salon but did not care that much about the Watt+Puppy+Thor even when shown by Dave Wilson himself. Maybe I expected too much.

Sometimes I just chalk up these difference as "Differnt Strokes for Different Folks".

 

[deba]

The extra money you pay for these expensive audio equipments is not for what it sounds like but the brand name it carries ........

 

[captrajesh]

For example about 30 years ago a cousin of mine gave a Bose 301 for repair. I tracked the woofer and tweeter to the OEM and they cost $7 and $1.5 each. The Old Bose 301 had 2 cone tweeters. So total parts were $10 per speaker. $20 for a pair. The crossover was a bulb and a cap. I assume it cost Bose to mass manufacture the boxes for $20 (1982 $) per pair. Total cost of part $40. Price paid by my cousin $300 (1981-2). See pic.
http://forums.audioholics.com/forum...series-1-circa-1975-anyone-know-these-301.jpg

There is a pair of these speakers available for sale in Chennai.

 

[Decadent_Spectre]

I have said quite often, personal taste. I have also repeated, people DON'T (usually) like a flat response. Is it accurate? No. Is it enjoyable? Yes. Take your pick. People can call it transparent all they want but it is not. There is reason to believe people prefer a certain amount of reflections/distortion and as anyone knows frequency response does not tell you the whole story. Usually people say even order distortion is more pleasant while odd order is not. Of course all this differs from person to person. People must be educated enough to understand what they want scientifically (ie measurements) and choose speakers accordingly, DSP necessary, to suit those subjective tastes rather than choosing brand names and assuming that throwing money at it will solve the problem.

Also what is the smoothing? Under what conditions were the measurements taken? Were they the same for each speaker? Was the equipment the same?

Pro audio for the win.

Audiophile garbage for the loose.

Simple as.

 

[maniac_2004]

could we have a list of speakers to avoid ..

and a list ( DIY / commercial ) which are VFM

will be very useful .

 

[murali]

Measurements are one side of the story. Voicing them is another. Like someone says at the PE forum - "You do not listen to wiggles" - or something to that effect.

You can have both right, the reason I stick to the Vandys. With the new Quatro now my main speakers, high pass filtering the main amplifier and equalising the bass side, the design philosophy which makes most sense, it gets the sound right and the measurements good enough. Even my older 2Ces got them right.

cheers.
murali

 

[doors666]

I have said quite often, personal taste. I have also repeated, people DON'T (usually) like a flat response. Is it accurate? No. Is it enjoyable? Yes. Take your pick. People can call it transparent all they want but it is not. There is reason to believe people prefer a certain amount of reflections/distortion and as anyone knows frequency response does not tell you the whole story. Usually people say even order distortion is more pleasant while odd order is not. Of course all this differs from person to person. People must be educated enough to understand what they want scientifically (ie measurements) and choose speakers accordingly, DSP necessary, to suit those subjective tastes rather than choosing brand names and assuming that throwing money at it will solve the problem.

Also what is the smoothing? Under what conditions were the measurements taken? Were they the same for each speaker? Was the equipment the same?

Pro audio for the win.

Audiophile garbage for the loose.

Simple as.

This is not about the harmonic distortion, but linear distortion. Most of the diy designers first shoot for a flat response in s/w and then they tweak and voice the speakers to taste, if you look at most of the diy speaker's responses, they are reasonably flat. The graphs posted in the thread are showing huge ups and downs. I know the fr does not tell the whole story, but its not that it doesnt tell you anything. I dont expect a speaker to have a ruler flat frq response, but i dont expect this bad a response either.

People have preferences about raising or lowering a particular range a little bit depending upon taste, but not to have 5-10 db ups and downs.

check out this thread, it has the reponse for B&W 602s2 (I own these), it looks relatively a lot flatter than the B&W plots for the latest 8 series.

Even I am confused a little, I dont expect these companies to make bad speakers, I mean some of these companies are legendary. I havent heard most of the speakers, so I cant really link the graphs to the sound quality.

 

[Decadent_Spectre]

As I said, listener preference. Flat isn't enjoyable. Flat is more of a clean canvas so you can paint as you prefer.

The human ear is very tolerant, there was a paper that discussed this, IIRC it said that dips were less noticeable than peaks and wider bandwidth peaks/nulls were more noticeable. I don't have the link on me as I read it long back but you may want to read AES papers by Floyd Toole.

I wager some of these companies intentionally put in a curve to target a certain market. That is their R&D, psychoacoustics.

The measurements posted are not verified. We don't know how/where they were measured. In room responses can actually look very similar to what was posted and often worse, particularly off axis.

Smoothness in the response is desirable but a flat response is not, tweaking to personal taste is absolute necessity. You must also account for ELC. Everyone's hearing is also different.

 

[Thad E Ginathom]

I wager some of these companies intentionally put in a curve to target a certain market. That is their R&D, psychoacoustics.

I wouldn't take that bet! I'm sure they do.

People will by what they are comfortable with and the sound that they like: the market will, of course, cater for them. Isn't that going to be true at all price levels?

Bare in mind that there is only a small subset of hifi buyers, those that regularly attend classical concerts, that know what they are comparing their music to anyway. Even those of us who regularly attend live music (classical carnatic, in my case) get it via the PA systems: it is nowhere near acoustic, even when the instruments are.

Speaker experts, especially our DIYers can answer this please... if you do actually set out to make a flat-frequency-response speaker, would it necessarily be expensive? OK, I know that some studio-monitor gear is quite pricey for a small speaker ... you'll probably tell me that frequency response is only a part of the picture anyway, but,,,

 

[Hari Iyer]

There's an interesting thread going on at techtalk forums. Some one has shown the freq response graphs of tons of commercial speakers and man, do they look like crap. There are tons of super expensive speakers listed there, e.g. B&W 800D, 805D, cm1, wilson audio duette, sophia2, puppy, vienna acoustics mahler, mozart, maggies, sonus faber etc. Whats going on here, these guys with millions in research budget cant do proper xo design or what? Some of these look like some random drivers thrown in, some of them can use reversing tweeter polarity. Do they intentionally sell crap and just dont care.

I know that flat freq response isnt everything, but still, it isnt all that hard to get that, when diyers with 1000$ measurement equipment can get there, why cant these guys?

When you check a cross-over simulator the following graphs are important. One of the driver mounted in free-air, the other of the filter and third of the speaker system. Taking example of the B&W Diamond 80S speaker there is drop of response at 600Hz by 3dB which can be for the baffle step because the response of the woofer goes down below 600Hz due to the step response which needs to be compensated in the cross-over. Again the response rises above 3KHz to 10KHz @ of 6dB which is because this might be a center speaker where a tighter mid-range is required. Also the response of the mid-range and tweeter in free-air has to be taken into account while considering the response of the speaker system as a whole. The speaker system (driver + cross-over) response when measured in anechoic room with small signal response in near field (1 watt/ 1 meter) will display the final response of the speaker system. By looking at the graphs it looks like a simulated response.
Cheers,

 

[doors666]

I wouldn't take that bet! I'm sure they do.

People will by what they are comfortable with and the sound that they like: the market will, of course, cater for them. Isn't that going to be true at all price levels?

Bare in mind that there is only a small subset of hifi buyers, those that regularly attend classical concerts, that know what they are comparing their music to anyway. Even those of us who regularly attend live music (classical carnatic, in my case) get it via the PA systems: it is nowhere near acoustic, even when the instruments are.

Speaker experts, especially our DIYers can answer this please... if you do actually set out to make a flat-frequency-response speaker, would it necessarily be expensive? OK, I know that some studio-monitor gear is quite pricey for a small speaker ... you'll probably tell me that frequency response is only a part of the picture anyway, but,,,

Nope its not expensive. I have built designs that have almost ruler flat response, it uses iron core inductors and electrolytic caps in the crossover (i didnt choose these, the designer used these and recommends these). the entire crossover will cost about 60-70$ for a pair.

 

[kaushik]

Thing is what you want really matters .feel costly high end audio is more hype less passion

that .1% .001% improvement line pulls many people .Its not wrong to pay ,just ask yourself is that you always wanted, or you are going because 'you should have .....series..speakers..with ... drivers' so be careful!

is it needed to know how ears work, what nerves operate their ..i dont think so