1 - 20 of 23 Posts

#### Sean Spamilton

·
##### Registered
Joined
·
1,419 Posts
Discussion Starter · ·
Hey there, I'm looking to further my basic working knowledge of speakers ... I understand frequency response to some extent, the low end being where the bass rolls off , but what does the extra bit, + or - X indicate? i.e. what's that last part mean? "Frequency response: 38Hz-20kHz, ±3dB" .. Thanks!

#### flickhtguru

·
##### Registered
Joined
·
2,495 Posts
Ok if you ever see a freq resp without a +/- X db it is a bogus or meaningless claim, because you have no idea at what level the speakers would put out that lower or upper freq. +/- 3db is the typical spec for speaker and sub frequency response. 1db is considered to be the least amount of volume change that the human ear can detect. So if a speaker has a spec of 38hz to 20khz +/- 3db then theoretically that speaker will play that whole frequency range pretty flat with a variation of no more than +3db higher than the average and no more than -3db less than the average. So basically it will only have a variation of 6db within that freq range.

Now keep in mind that your room has a ton to do with how a speaker sounds and what the speakers in room freq response may be. Say a speaker is claming 30db at -3db but Audyssey sets the xover of that speaker to 50db that means that the in room frequency response of that speaker had a -3db point of between 41 and 50 hz (assuming the AVR has 10hz increments for the xover.) Only way to truly tell how a speaker sounds is to take it home and set it up in your room.

#### Bill Fitzmaurice

·
##### Banned
Joined
·
15,491 Posts

Quote:
Originally Posted by Sean Spamilton  /t/1471983/could-someone-explain-what-3-or-3-means-when-it-comes-to-speaker-frequency#post_23295065

"Frequency response: 38Hz-20kHz, ±3dB" ..
That should mean that if the sensitivity was, for instance, 87dB/watt, that with one watt input the lowest sensitivity from 30Hz to 20kHz would be 84dB, the highest 90dB. I say 'should', as manufacturers have a way of skewing data. IME if there is no measured SPL chart to back up frequency response claims those claims should be taken with a truckload of salt.

#### commsysman

·
##### Registered
Joined
·
7,063 Posts
When you look at the frequency response curve of the speaker, its output will normally "roll off" or decrease at some frequency at the low and high end.

The most frequent way of specifying how much it rolls off is the -3 decibel point. This is because the -3 db point is where the speaker output is at HALF POWER relative to the flat part of the curve.

This is considered to be the point at which the speakers frequency response has dropped off enough to be noticeably less effective at reproducing.

#### RicardoJoa

·
##### Registered
Joined
·
638 Posts
It means that the FR falls within + or - 3 decibels. in other words, some frequency will hit +3 decibel over the average of 0 and some will hit -3 decibels. The range ends in your example, 38hz would mean -3 decibels at the low end, and -3 20khz at the high end.

#### Sean Spamilton

·
##### Registered
Joined
·
1,419 Posts
Discussion Starter · ·
[
Quote:
Originally Posted by Bill Fitzmaurice  /t/1471983/could-someone-explain-what-3-or-3-means-when-it-comes-to-speaker-frequency#post_23295310

That should mean that if the sensitivity was, for instance, 87dB/watt, that with one watt input the lowest sensitivity from 30Hz to 20kHz would be 84dB, the highest 90dB. I say 'should', as manufacturers have a way of skewing data. IME if there is no measured SPL chart to back up frequency response claims those claims should be taken with a truckload of salt.
Quote:
Originally Posted by flickhtguru  /t/1471983/could-someone-explain-what-3-or-3-means-when-it-comes-to-speaker-frequency#post_23295286

Ok if you ever see a freq resp without a +/- X db it is a bogus or meaningless claim, because you have no idea at what level the speakers would put out that lower or upper freq. +/- 3db is the typical spec for speaker and sub frequency response. 1db is considered to be the least amount of volume change that the human ear can detect. So if a speaker has a spec of 38hz to 20khz +/- 3db then theoretically that speaker will play that whole frequency range pretty flat with a variation of no more than +3db higher than the average and no more than -3db less than the average. So basically it will only have a variation of 6db within that freq range.

.

Hmm.. So I understand why a speaker that was rated at +3db would play on average between 0 to 3db difference, but why would it go from -3db to +3db with a 6db variation if it was rated this way? wouldn't that mean that a speaker rated at -3db would play anywhere from -6db to 0db less the average by that logic? And I assume a flat frequency response is something to be striving for with a speaker? Or wait, looking at my example again, it looks like it has both a + and a - on top of each other, perhaps this is where your getting the 6db variation from right?
Quote:
Originally Posted by flickhtguru  /t/1471983/could-someone-explain-what-3-or-3-means-when-it-comes-to-speaker-frequency#post_23295286

Ok if you ever see a freq resp without a +/- X db it is a bogus or meaningless claim, because you have no idea at what level the speakers would put out that lower or upper freq. +/- 3db is the typical spec for speaker and sub frequency response. 1db is considered to be the least amount of volume change that the human ear can detect. So if a speaker has a spec of 38hz to 20khz +/- 3db then theoretically that speaker will play that whole frequency range pretty flat with a variation of no more than +3db higher than the average and no more than -3db less than the average. So basically it will only have a variation of 6db within that freq range.

Now keep in mind that your room has a ton to do with how a speaker sounds and what the speakers in room freq response may be. Say a speaker is claming 30db at -3db but Audyssey sets the xover of that speaker to 50db that means that the in room frequency response of that speaker had a -3db point of between 41 and 50 hz (assuming the AVR has 10hz increments for the xover.) Only way to truly tell how a speaker sounds is to take it home and set it up in your room.

I get that room effects can reek havoc on frequency response - having been through that with the positioning of my new speakers last month. I assume that no matter what volume / how many watts are driving the speaker the frequency response will stay the same, meaning there will always be a variation of ( in this example ) + 3 to - 3.
Quote:
Originally Posted by commsysman  /t/1471983/could-someone-explain-what-3-or-3-means-when-it-comes-to-speaker-frequency#post_23295470

When you look at the frequency response curve of the speaker, its output will normally "roll off" or decrease at some frequency at the low and high end.

The most frequent way of specifying how much it rolls off is the -3 decibel point. This is because the -3 db point is where the speaker output is at HALF POWER relative to the flat part of the curve.

This is considered to be the point at which the speakers frequency response has dropped off enough to be noticeably less effective at reproducing.

Hmm, I think where I'm getting lost is I don't understand what the frequency " CURVE " is... beyond having scene examples of it in REW pictures hear on the forum, and I might as well be trying to read music with them, as I don't understand them.

Is the bottom line then a speaker with less variation in the frequency response a better speaker? I already understand one with better sensitivity is best, but one with say a variation of only +1 would be super flat and therefore amazing? also, would it be better if it was + 1 or -1 or + and 1 at the same time? ( lol, I bet this sounds completely ridiculous, I'm sorry! )

#### Bill Fitzmaurice

·
##### Banned
Joined
·
15,491 Posts

Quote:
Originally Posted by Sean Spamilton  /t/1471983/could-someone-explain-what-3-or-3-means-when-it-comes-to-speaker-frequency#post_23296248

[

Is the bottom line then a speaker with less variation in the frequency response a better speaker?
To some extent, as flat axial frequency response has been proven time and time again as what the ear most prefers. But there are other considerations, such as off-axis response, vertical pattern control, sensitivity, time alignment and transient response, all of which combine with axial response to give an overall result. It's possible for a speaker that really shines in those other areas to subjectively sound better than another speaker that's poor in those areas but has flatter axial response.

#### Sean Spamilton

·
##### Registered
Joined
·
1,419 Posts
Discussion Starter · ·

Quote:
Originally Posted by Bill Fitzmaurice  /t/1471983/could-someone-explain-what-3-or-3-means-when-it-comes-to-speaker-frequency#post_23296307

To some extent, as flat axial frequency response has been proven time and time again as what the ear most prefers. But there are other considerations, such as off-axis response, vertical pattern control, sensitivity, time alignment and transient response, all of which combine with axial response to give an overall result. It's possible for a speaker that really shines in those other areas to subjectively sound better than another speaker that's poor in those areas but has flatter axial response.

Ok, well that's somewhat straight forward. What about the Low Frequency cutoff? for instance Mine says 30hz at -10db .. does that mean that when playing down to 30hz ( not that I let mine get down that far, I cross them over to subs at 80 ) there will potentially be -10db less sound?

#### ack_bk

·
##### Registered
Joined
·
11,586 Posts

Quote:
Originally Posted by Bill Fitzmaurice  /t/1471983/could-someone-explain-what-3-or-3-means-when-it-comes-to-speaker-frequency#post_23296307

To some extent, as flat axial frequency response has been proven time and time again as what the ear most prefers. But there are other considerations, such as off-axis response, vertical pattern control, sensitivity, time alignment and transient response, all of which combine with axial response to give an overall result. It's possible for a speaker that really shines in those other areas to subjectively sound better than another speaker that's poor in those areas but has flatter axial response.

This post should be stickied

#### flickhtguru

·
##### Registered
Joined
·
2,495 Posts

Quote:
Originally Posted by Sean Spamilton  /t/1471983/could-someone-explain-what-3-or-3-means-when-it-comes-to-speaker-frequency#post_23296526

Ok, well that's somewhat straight forward. What about the Low Frequency cutoff? for instance Mine says 30hz at -10db .. does that mean that when playing down to 30hz ( not that I let mine get down that far, I cross them over to subs at 80 ) there will potentially be -10db less sound?

Yes that is correct...

As Bill stated a flat req response or a "flat" freq curve are what most strive for. However if you are boosting your subs and running them hot you are in essence not keeping a flat curve. Also if someone says they have major mid bass slam that means that in their room or with their settings the mid bass region is boosted higher than the rest of the mid range and bass.

#### Sean Spamilton

·
##### Registered
Joined
·
1,419 Posts
Discussion Starter · ·

Quote:
Originally Posted by flickhtguru  /t/1471983/could-someone-explain-what-3-or-3-means-when-it-comes-to-speaker-frequency#post_23296596

Yes that is correct...

As Bill stated a flat req response or a "flat" freq curve are what most strive for. However if you are boosting your subs and running them hot you are in essence not keeping a flat curve. Also if someone says they have major mid bass slam that means that in their room or with their settings the mid bass region is boosted higher than the rest of the mid range and bass.

I mean I don't let my Fronts get down that low, those are my tower specs, 35 to 23000hz at + / - 3, with a LF cutoff of 30hz @ -10db. I just happen to have the AVR cross them over at 80 to the sub. I don't think I'm running my sub " hot "
my sub is 29 - 120hz at (+-) 3.

·
##### Registered
Joined
·
354 Posts

Quote:
Originally Posted by Bill Fitzmaurice  /t/1471983/could-someone-explain-what-3-or-3-means-when-it-comes-to-speaker-frequency#post_23296307

To some extent, as flat axial frequency response has been proven time and time again as what the ear most prefers. But there are other considerations, such as off-axis response, vertical pattern control, sensitivity, time alignment and transient response, all of which combine with axial response to give an overall result. It's possible for a speaker that really shines in those other areas to subjectively sound better than another speaker that's poor in those areas but has flatter axial response.

Bill, what do you think of measurements like this:
Quote:
The Venere 2.5’s listening-window response (a five-point average of axial and +/–15-degree horizontal and vertical responses) measures +2.65/–8.17 decibels from 200 hertz to 10 kilohertz.
http://www.hometheater.com/content/sonus-faber-venere-25-speaker-system-ht-labs-measures

#### arnyk

·
##### Banned
Joined
·
14,420 Posts

Quote:

Quote:
Originally Posted by Bill Fitzmaurice  /t/1471983/could-someone-explain-what-3-or-3-means-when-it-comes-to-speaker-frequency#post_23296307

To some extent, as flat axial frequency response has been proven time and time again as what the ear most prefers. But there are other considerations, such as off-axis response, vertical pattern control, sensitivity, time alignment and transient response, all of which combine with axial response to give an overall result. It's possible for a speaker that really shines in those other areas to subjectively sound better than another speaker that's poor in those areas but has flatter axial response.

Bill, what do you think of measurements like this:
Quote:
The Venere 2.5’s listening-window response (a five-point average of axial and +/–15-degree horizontal and vertical responses) measures +2.65/–8.17 decibels from 200 hertz to 10 kilohertz.
http://www.hometheater.com/content/sonus-faber-venere-25-speaker-system-ht-labs-measures

Certainly nothing exceptional in terms of performance. The money must be going into the nifty Italian cabinetry. The dip around 8 Khz and the kick up following it is found by many inexperienced listeners to be euphonic. The dip might be perceived as a slight reduction in harshness while the rise at 10 KHz adds a little "air".

Nothing that should bust a delicate eardrum... just not maximally accurate.

#### Bill Fitzmaurice

·
##### Banned
Joined
·
15,491 Posts

Quote:
Originally Posted by arnyk  /t/1471983/could-someone-explain-what-3-or-3-means-when-it-comes-to-speaker-frequency#post_23297290

Certainly nothing exceptional in terms of performance.
To say the least. I'd call it mediocre.

#### RicardoJoa

·
##### Registered
Joined
·
638 Posts
One thing for sure if thats what they aim then thats what they get. It is about enjoying music, if the listener enjoys it, thats what matter.

#### Bill Fitzmaurice

·
##### Banned
Joined
·
15,491 Posts

Quote:
Originally Posted by RicardoJoa  /t/1471983/could-someone-explain-what-3-or-3-means-when-it-comes-to-speaker-frequency#post_23297813

One thing for sure if thats what they aim then thats what they get. It is about enjoying music, if the listener enjoys it, thats what matter.
Owners of speakers tend to think that they sound great. They must, otherwise they'd get rid of them. How many owners will say 'My speakers sound like crap and I paid way too much for them' ? Even Bose has a high rate of owner satisfaction.

#### RicardoJoa

·
##### Registered
Joined
·
638 Posts

Quote:
Originally Posted by Bill Fitzmaurice  /t/1471983/could-someone-explain-what-3-or-3-means-when-it-comes-to-speaker-frequency#post_23297923

Owners of speakers tend to think that they sound great. They must, otherwise they'd get rid of them. How many owners will say 'My speakers sound like crap and I paid way too much for them' ? Even Bose has a high rate of owner satisfaction.

They dont think. Most dont even bother with vistiting these forums, let alone "listen" to anechoic FR, when most system is affected by room. But if they do visit these forum they will get brain washed that flat frequency is the ultimate good sound. Sorry, how many flat frequecy speaker are out there? -/+0.0db is flat, anything is just colored.

#### BobL

·
##### Banned
Joined
·
3,781 Posts

Can you tell if someone you know has a cold with a stuffy nose or a sore throat?  Defnitely! Sometimes you can tell this even with people you don't know.  Can you tell this if you are both in a different room? Definitely!  Although, the room has a lot of affect on sound our brains still know when something doesn't sound right. The room is still "coloring" the sound but our brains recognize it and knows it does not sound correct.  There have been MANY studies on what people prefer and as Bill stated time and time again they choose speakers with a flat frequency response, especially in blind tests.  We can have a dead room or a reverberant room but our brains still seem to know what is natural.  Also, as Bill stated there are many other aspects of a speaker which determine the sound we hear and some of those attributes might be poor or cause poor interaction in the room.

The best of speakers might be within +/- 1 db frequency response and if we measured this compared to electronic equipment (amps, etc) an engineer would call that broken because it isn't +/- a very small amount like .001db.  But, as speakers go they would be excellent as long if their other qualities are good as well.  It is tough to find many speakers with +/- 3db.  As Floyd Toole stated (Paraphrasing).  There are many aspects of a speaker which affects its sound but if it doesn't start off with a flat frequency response why bother.  It is arguably one the biggest aspects to get right in speaker design.  He also states that off axis response is very important especially for room interaction.  So a speaker that measures decent but not great both on and off axis might sound better than a speaker that measures great on axis but has poor off axis response.

Now I think magazines do a disservice to the public with many of these reviews and their superlatives used.  With a speaker that measures like above their superlatives should be about fit and finish only.  However, there are some studies that showed speakers that look better sound better in sighted tests but didn't fare as well when the tests were blind.  So maybe if we just make good looking speakers everybody will be happy

Nobody is getting brain washed from the forums.  The research and studies have already proven what many people have stated, it isn't their opinions it comes from real data.  The reason speakers are measured anechoic or quasi-anechoic is because they need to get the speakers to replicate the source without voicing a speaker to a room's response.  Just as if you wanted to record your friend to give an accurate depiction of what he sounds like you wouldn't record him when he has a cold or sore throat.  Speakers should replicate the sound and not alter it.

#### RicardoJoa

·
##### Registered
Joined
·
638 Posts
The are thousand of speakers rated on - /+ 3 db that sounds vastly different from each other. I wont disagree to start with a flat on and off axis speaker, but the real question is, how many flat response speakers are out there? It is obvious that these sonus arent trying for the flat response, probably neither of its other speakers. They have their goal, and if the met their goal and people like it thats what matter. Nothing wrong to like to hear a colored pair of speaker.

As you mentioned, there are other aspect of speaker design, we just dont know their off axis response, the real matter is off axis at xover points, phase aligment at xover point as well. But like you said, you can great on axis but if off axis is inconssistent, it may not sound good too.

#### Bill Fitzmaurice

·
##### Banned
Joined
·
15,491 Posts

Quote:
Originally Posted by RicardoJoa  /t/1471983/could-someone-explain-what-3-or-3-means-when-it-comes-to-speaker-frequency#post_23303530

They have their goal, and if the met their goal and people like it thats what matter.
If their goal was to produce a sheep in wolf's clothing at a premium price, they succeeded. Not that +,-3dB is the holy grail of speaker design, especially with today's DSP options, but IMO the response of the tweeter they used should have earned it a place in the trash bin. I wouldn't settle for that response in a speaker that I designed selling for 1/10 the price of the SF. And when there are literally dozens of tweeters that retail for \$30 or less that are far superior to whatever SF uses there's no reason why I'd have to.

1 - 20 of 23 Posts