Hello all,
I'll be posting this on a few other Audio/DIY forums because I almost always come across the same argument that Frequency Response isn't the only factor when it comes to Speakers/Headphones but I don't have a good knowledge on what else matters.
For example, let's place three speakers in a Anechoic chamber (Removing room acoustics for sake of discussion) as well as assume that the three speakers chosen will all have a FR of 40hz to 20KHz +/-1.5 db. With these assumptions if I the listener heard all three speakers (Maybe one being the top of the line Genelec Studio monitor, another being lower end KRK speakers monitors and the last being the Orion (a well received DIY speakers) I think most would agree that each would sound different. Where I get confused is that 1.5db is very difficult for humans to distinguish between yet (and I may have this all wrong) the three systems would sound quite different even though the listner is hearing the same frequencies from each speaker... So from my perspective 8hz should sound the same no matter the source. This is where I'm looking for some education. I know there are numerous graphs posted alongside FR graphs and these are what I assume cause a sound difference even though the FR is for the most part identical.
I tried googling and reading through forums to find the answer yet I'm not finding out nearly as much as I'd hope. How does a speaker produce a flat FR yet contain distortion or other anomalies while maintaining a flat FR?
If any of you know some good material to read and can link it for further explanation that's just as good as a lengthy post.
Thanks for your time. I just want to be able to created a great listening environment and am interested in how to accomplish this without using FR as my only guideline.
Joel.
I'll be posting this on a few other Audio/DIY forums because I almost always come across the same argument that Frequency Response isn't the only factor when it comes to Speakers/Headphones but I don't have a good knowledge on what else matters.
For example, let's place three speakers in a Anechoic chamber (Removing room acoustics for sake of discussion) as well as assume that the three speakers chosen will all have a FR of 40hz to 20KHz +/-1.5 db. With these assumptions if I the listener heard all three speakers (Maybe one being the top of the line Genelec Studio monitor, another being lower end KRK speakers monitors and the last being the Orion (a well received DIY speakers) I think most would agree that each would sound different. Where I get confused is that 1.5db is very difficult for humans to distinguish between yet (and I may have this all wrong) the three systems would sound quite different even though the listner is hearing the same frequencies from each speaker... So from my perspective 8hz should sound the same no matter the source. This is where I'm looking for some education. I know there are numerous graphs posted alongside FR graphs and these are what I assume cause a sound difference even though the FR is for the most part identical.
I tried googling and reading through forums to find the answer yet I'm not finding out nearly as much as I'd hope. How does a speaker produce a flat FR yet contain distortion or other anomalies while maintaining a flat FR?
If any of you know some good material to read and can link it for further explanation that's just as good as a lengthy post.
Thanks for your time. I just want to be able to created a great listening environment and am interested in how to accomplish this without using FR as my only guideline.
Joel.
Even in an anechoic chamber, polar pattern is critical. "Frequency response" is a simple thing for boxes of gain, but much less well-defined for transducers.
if you want some reading material on this topic (and have the patience) you could check out this thread http://www.diyaudio.com/forums/multi-way/166411-measurements-when-what-how-why.html
I thnk you will find it covers your question and also shows a pretty good cross section of beliefs/opinions.
As Sy mentions polar pattern is a key (though I'm unsure why it is critical even in an anechoic chamber, unless because there is no such thing as a perfect one?), think of it not so much as what the direct sound arriving at your ears sounds like (anechoic freq response) but how all the other reflected sounds arriving at your ears mix together when they arrive at your ears. Depending on the polar pattern of each speaker the reflections will be different, and since our brain interprets these reflected sounds to work out spacial information they will have an effect on the ultimate sound we perceive.
Tony.
I thnk you will find it covers your question and also shows a pretty good cross section of beliefs/opinions.
As Sy mentions polar pattern is a key (though I'm unsure why it is critical even in an anechoic chamber, unless because there is no such thing as a perfect one?), think of it not so much as what the direct sound arriving at your ears sounds like (anechoic freq response) but how all the other reflected sounds arriving at your ears mix together when they arrive at your ears. Depending on the polar pattern of each speaker the reflections will be different, and since our brain interprets these reflected sounds to work out spacial information they will have an effect on the ultimate sound we perceive.
Tony.
The more I've measured and listened to loudspeakers the more I'd say polar response has got to be the most important thing. The better controlled/more constant and less rippled the PR the better. If the tonal quality isn't right, there's just no real sense moving on. The next things I'd put on that list is efficiency(I'm a green kind of guy ha ha) and that is really b/c I want the maximum amount of output possible before distortion/compression becomes audible. Once you get there, you probably won't care about anything else. If you still do... more power to you.
Check these out: ****DanTheMan's blog****: Review of Polar graphs
care to guess which of these sound most correct?
Dan
Check these out: ****DanTheMan's blog****: Review of Polar graphs
care to guess which of these sound most correct?
Dan
I have a belief that maintaining the time information and reproducing it
correctly is more important than frequency response (and polar response).
Our brain can adapt quickly to variations in frequency response, but it has
a hard time reconciling distortion in time. I still haven't thought about this
clearly, but in my mind, this has got to be it.
Why else would full range drivers have a certain appeal that multi-way
speakers find hard to replicate, even though the full range drivers have
typically terrible on and off axis response?
Example: if two notes are reproduced at slightly different frequencies than
the orginal, but correctly spaced one after the other in time, they would
sound the same. But if they are spaced even slightly differently, they
would not sound the same at all.
correctly is more important than frequency response (and polar response).
Our brain can adapt quickly to variations in frequency response, but it has
a hard time reconciling distortion in time. I still haven't thought about this
clearly, but in my mind, this has got to be it.
Why else would full range drivers have a certain appeal that multi-way
speakers find hard to replicate, even though the full range drivers have
typically terrible on and off axis response?
Example: if two notes are reproduced at slightly different frequencies than
the orginal, but correctly spaced one after the other in time, they would
sound the same. But if they are spaced even slightly differently, they
would not sound the same at all.
Last edited:
Harmonic distortion, particularly second harmonic distortion at low frequencies and high volumes is a feature of many speakers. 10% THD is not uncommon, mostly second and some third.
Of course even 10% may be barely audible, it's popular to decry THD as a measure of quality, but it is measurable and it'd still be nice to eliminate it...
w
Of course even 10% may be barely audible, it's popular to decry THD as a measure of quality, but it is measurable and it'd still be nice to eliminate it...
w
Loudspeakers project sound differently in every direction. They also continue to project sound after electrical stimulation has ceased.
The room also has directional and delay characteristics that interact with the loudspeakers.
Some loudspeakers are intended to be omnidirectional, even those speakers that do not intend omnidirectionality can be so in the bass frequencies. Some speakers emulate mostly a point source (such as dynamic systems with dome tweeter), others emulate line or planar sources.
I currently like adding omnidirectional ribbon supertweeter (crosses in at 16Khz) to my line-source-like Acoustat 1+1 electrostatics. What matters is the overall effect, and not whether one "model" (such as "the line source model") is realized perfectly.
The room also has directional and delay characteristics that interact with the loudspeakers.
Some loudspeakers are intended to be omnidirectional, even those speakers that do not intend omnidirectionality can be so in the bass frequencies. Some speakers emulate mostly a point source (such as dynamic systems with dome tweeter), others emulate line or planar sources.
I currently like adding omnidirectional ribbon supertweeter (crosses in at 16Khz) to my line-source-like Acoustat 1+1 electrostatics. What matters is the overall effect, and not whether one "model" (such as "the line source model") is realized perfectly.
Another reason to believe that the brain can tolerate large amounts of harmonics
not present in the original material, but even slight amounts of distortion in time makes
it sound unnatural. You immediately know that the sound is 'not real' or reproduced.
I believe we are chasing the wrong target when going after vanishingly low levels of
THD - we are not happy even after achieving it.
True, but they're more musical than most loudspeakers out there.
I'm talking strictly about enjoyment, about connecting with music.
Most single driver designs achieve this quite easily, given all their
shortcomings. Now, if the criteria were soundstage depth, response
variations, and such other terms we use to describe how great our
systems are - and I like these things too - then obviously the full
range drivers will fall short.
All I'm saying is, the goodness behind full range drivers needs to be
explored more, rather than going on about polar response and such
other things. To me, they sound right despite all their shortcomings.
And what they got going for themselves is time coherence more
than anything else.
I'm talking strictly about enjoyment, about connecting with music.
Most single driver designs achieve this quite easily, given all their
shortcomings. Now, if the criteria were soundstage depth, response
variations, and such other terms we use to describe how great our
systems are - and I like these things too - then obviously the full
range drivers will fall short.
All I'm saying is, the goodness behind full range drivers needs to be
explored more, rather than going on about polar response and such
other things. To me, they sound right despite all their shortcomings.
And what they got going for themselves is time coherence more
than anything else.
I wish to add that I agree on the importance of polar response.
I've heard the Altec VOT and it imaged like no other system.
This must be primarily down to the superb off axis response.
The importance of good polar response is proven. Time to move on.
I've heard the Altec VOT and it imaged like no other system.
This must be primarily down to the superb off axis response.
The importance of good polar response is proven. Time to move on.
Thank you all for your replies so far! As I've stated, I've posted this in a few other forums so I'm keeping up with each.
I've been doing additional reading (thanks for the links) and they're helping although some of the information is still foreign to me (i.e reading polar pattern graphs) so that'll take some additonal time.
@InnerSpace
Thanks for your resposne, the reason I'm taking out room acoustics for this discussion is because I'm more interested in learning why speakers X,Y,Z sound different apart from the room. If my room has terrible bass gain then that's addresing another problem outside of the speaker. Currently I'm trying to discover how other measurements affect sound outside of FR. Polar Patterns are one I'm becoming more familiar with for example. /end
I'm a 'wanna be' audiophile I guess you could say with an obsession with logistics. I want to make my KRKs perform the best that I can and I'm looking into different DIY room acoustic solutions to accomplish this.
In general I've seen numerous speakers for sale with varying reviews about how great or poor they are yet many have very similar response graphs. This is what sparked my curiosity. Some speakers are two way where others are WMTMW, this like this get me curious as to why they sound different than say a two-way system. My assumption at this point is that this is a result of Polar Pattern.
Overall, I'm becoming more convinced that it's very important to find speakers that have a Polar Pattern that's condusive to your listening situation (i.e my KRKs are only listened to from my computer chair so I require a limited dispersion whereas speakers for a HT room might want a wider patter since they'll be delivering sound to 6-8 people in varying seating positions. Also, it seems that a pair of 1000$ speakers in a well treated room will give great results compared to those 20 000$ speakers in a untreated room. Am I on the right path? It seems room acoustics play a much larger role than the speakers themselves.
Thanks again and I'll continue to read on.
Joel
I've been doing additional reading (thanks for the links) and they're helping although some of the information is still foreign to me (i.e reading polar pattern graphs) so that'll take some additonal time.
@InnerSpace
Thanks for your resposne, the reason I'm taking out room acoustics for this discussion is because I'm more interested in learning why speakers X,Y,Z sound different apart from the room. If my room has terrible bass gain then that's addresing another problem outside of the speaker. Currently I'm trying to discover how other measurements affect sound outside of FR. Polar Patterns are one I'm becoming more familiar with for example. /end
I'm a 'wanna be' audiophile I guess you could say with an obsession with logistics. I want to make my KRKs perform the best that I can and I'm looking into different DIY room acoustic solutions to accomplish this.
In general I've seen numerous speakers for sale with varying reviews about how great or poor they are yet many have very similar response graphs. This is what sparked my curiosity. Some speakers are two way where others are WMTMW, this like this get me curious as to why they sound different than say a two-way system. My assumption at this point is that this is a result of Polar Pattern.
Overall, I'm becoming more convinced that it's very important to find speakers that have a Polar Pattern that's condusive to your listening situation (i.e my KRKs are only listened to from my computer chair so I require a limited dispersion whereas speakers for a HT room might want a wider patter since they'll be delivering sound to 6-8 people in varying seating positions. Also, it seems that a pair of 1000$ speakers in a well treated room will give great results compared to those 20 000$ speakers in a untreated room. Am I on the right path? It seems room acoustics play a much larger role than the speakers themselves.
Thanks again and I'll continue to read on.
Joel
Your KRKs have a very broad pattern and you need it for near field.
Read this: ****DanTheMan's blog****: Psychoacoustics
this:
****DanTheMan's blog****: Above the modal dominated frequencies--300Hz and up and those polar graphs again.
Well controlled directivity is immensely important--until the sound waves get so big that the room is part of the source.
Dan
Read this: ****DanTheMan's blog****: Psychoacoustics
this:
****DanTheMan's blog****: Above the modal dominated frequencies--300Hz and up and those polar graphs again.
Well controlled directivity is immensely important--until the sound waves get so big that the room is part of the source.
Dan
I should add: the issue to me isn't wether we should have controlled directivity or not, it's wether it should be narrow, wide, or somewhere between the extremes and how do we get it to below 300Hz w/o getting huge or having a strong rear radiation.
Dan
Dan
Many would say that (except for very small rooms) the "modal dominated" region ends closer to 150 Hz. . . . which may account for why so many of the supposedly "constant directivity" horn+bloom-boxes don't cut it for "natrual" sound. This may be why properly equalized corner horns sound better than horn-box combos.
In addition to smooth on-axis (not necessarily flat, but no abrupt response shifts) and similarly smooth power response (implies, but does not require, a uniform polar response) I'd add "no resonances" (at any frequency, since they put energy where it doesn't belong in the time domain) and no audible IM or aliasing distortions (any distortion that adds "hash" to the sound, even in brief impulses).
A loudspeaker must present to my ear:
a reasonable reproduction of the original sound,
reflected (room) sound that suggests the same source/balance,
and no sound that leaves me wondering "what's that?"
You are definitely correct about resonance, but the "not cutting it" doesn't make a lot of sense compared to other loudspeaker types. I'll agree that the typical pattern tightening above 1k isn't optimal. I'd actually add that to my argument for wide dispersion--the change is less abrupt and it's also a point for Dipoles.
Usually you can see resonance in the polar graph, but not with time resolution obviously. The exception being for me is when the resonance is in the low frequencies where I can't measure FR accurately. I may try to do some near field in the LF to see what I can find as I just haven't done a whole lot of it. The impulse is the only way I know to get a look at it there and I can't read them well enough to say for certain exactly where the resonance exists. You'd have to know how exactly how the slope of the ripples relates to frequency. I know if they are steep, it's HF. Shallow, low Freq. But that's a broad brush.
My set up isn't very good though. Fairly accurate measurements using my system is really above 800Hz ish or so as I'm just not getting enough data points and my measurement is too close for bass. CSD isn't particularly useful for finding audible resonance and useless w/o the impulse graph. Wavelet.... much better.
My question would be for Deward as it always is--"what's the original?" I still say the recording is, not the event. 😉 Just jabbing at you--I know what you'll always think.
Dan
Usually you can see resonance in the polar graph, but not with time resolution obviously. The exception being for me is when the resonance is in the low frequencies where I can't measure FR accurately. I may try to do some near field in the LF to see what I can find as I just haven't done a whole lot of it. The impulse is the only way I know to get a look at it there and I can't read them well enough to say for certain exactly where the resonance exists. You'd have to know how exactly how the slope of the ripples relates to frequency. I know if they are steep, it's HF. Shallow, low Freq. But that's a broad brush.
My set up isn't very good though. Fairly accurate measurements using my system is really above 800Hz ish or so as I'm just not getting enough data points and my measurement is too close for bass. CSD isn't particularly useful for finding audible resonance and useless w/o the impulse graph. Wavelet.... much better.
My question would be for Deward as it always is--"what's the original?" I still say the recording is, not the event. 😉 Just jabbing at you--I know what you'll always think.
Dan
Last edited:
Is it important to "win" listening contest(s)
IMHO, it's about the connection to the music, not "winning" anything!
YMMV, of course!
Best, Don
We are talking about speaker measurements that correlate to a positive listener experience. I can connected to music over a telephone.
Dan
Dan
- Status
- Not open for further replies.