Low radiates full-space or placed against back-wall half space.
mid half-space
High quarter-space.
So low energy is spread over a much bigger area as mid/high and looses on distance pressure compared to mid high. So the response mutates on bigger distance due polar behavior.
And high frequency's suffer more comb-filter effect and are there for also reduced compare to mid, due their shorter wave length.
So I tune low and high a bit louder to compensate the effects.
mid half-space
High quarter-space.
So low energy is spread over a much bigger area as mid/high and looses on distance pressure compared to mid high. So the response mutates on bigger distance due polar behavior.
And high frequency's suffer more comb-filter effect and are there for also reduced compare to mid, due their shorter wave length.
So I tune low and high a bit louder to compensate the effects.
Last edited:
The effect isn't linear with distance so how do you want to control such behavior?
That is why do not solve the problem.
And then the shape of the room your place the speakers in.
For high frequencies one can use directivity to control room influence but for low not frequencies.
That is why do not solve the problem.
And then the shape of the room your place the speakers in.
For high frequencies one can use directivity to control room influence but for low not frequencies.
Ok... here's a curve trying to explain what I'm trying to say. Here's my version of the OB speaker inspired by NP, standard Audio Nirvana 8" crossed to a Eminence 15A at about 150 Hz.
(It's an old curve, I have since been able to remove the dip at 150 Hz by adding rear wings)
The AN driver has two resonances at 3 kHz, 4 kHz, and a large response peak at 12 kHz. All these are very audible and I've tried to address them in different ways. They can be cured to some extent, but they never go away fully.
Now, I've also measured a B&W 603 and it measures ruler flat from 100 Hz to 15 kHz. When I listen to the B&W, I can easily hear how colored the Audio Nirvanas are. But hey, I would still take them over the B&Ws. But that's just me. I don't know how my preference fits into Olive, Toole, and Co.,but I'm sure there are many out there who prefer something similar.
One thing I like about the ANs is their high sensitivity. It somehow makes them sound more lifelike. This makes me think dynamic range has got to be one of the important characteristics of a loudspeaker that aspires to sound 'real'.
Also, they are coherent, for the lack of a better word, when compared to not just the 603s, but also some other high-end systems I've heard.
The B&Ws on the other hand have superior directivity, compared to the ANs. This manifests itself in the center image forming way behind the speakers as compared to the AN where the image moves forward. I believe this is because the ANs have less information going into the room at high frequencies which makes my brain work harder and in the end it cannot locate exactly where the center image is.
So, what are my conclusions? I think a speaker to sound 'real' must:
1. be point source or atleast single source over a large portion of the midrange
2. be capable of a large dynamic range
3. have high sensitivity
4. have flat FR
5. Good polar response
Hope this makes more sense.
(It's an old curve, I have since been able to remove the dip at 150 Hz by adding rear wings)
The AN driver has two resonances at 3 kHz, 4 kHz, and a large response peak at 12 kHz. All these are very audible and I've tried to address them in different ways. They can be cured to some extent, but they never go away fully.
Now, I've also measured a B&W 603 and it measures ruler flat from 100 Hz to 15 kHz. When I listen to the B&W, I can easily hear how colored the Audio Nirvanas are. But hey, I would still take them over the B&Ws. But that's just me. I don't know how my preference fits into Olive, Toole, and Co.,but I'm sure there are many out there who prefer something similar.
One thing I like about the ANs is their high sensitivity. It somehow makes them sound more lifelike. This makes me think dynamic range has got to be one of the important characteristics of a loudspeaker that aspires to sound 'real'.
Also, they are coherent, for the lack of a better word, when compared to not just the 603s, but also some other high-end systems I've heard.
The B&Ws on the other hand have superior directivity, compared to the ANs. This manifests itself in the center image forming way behind the speakers as compared to the AN where the image moves forward. I believe this is because the ANs have less information going into the room at high frequencies which makes my brain work harder and in the end it cannot locate exactly where the center image is.
So, what are my conclusions? I think a speaker to sound 'real' must:
1. be point source or atleast single source over a large portion of the midrange
2. be capable of a large dynamic range
3. have high sensitivity
4. have flat FR
5. Good polar response
Hope this makes more sense.
Attachments
Last edited:
I have been searching for years for a quote i THINK i read by PWK on frequency response. I believe he said something like "flat frequency response does not sound real" but i DO NOT have the exact quote and may be remembering it incorrectly. Can someone confirm? Maybe i imagined it?
Neat speakers ra7... i agree with your statements. funnily enough, i have a pair of old B&W too i take out from time to time (very rarely) to hear how my diy projects compare to something considered good.
Neat speakers ra7... i agree with your statements. funnily enough, i have a pair of old B&W too i take out from time to time (very rarely) to hear how my diy projects compare to something considered good.
Factor Analysis is a very good example of "fools rush in where angels fear to tread." I don't know if they used that technique and they are certainly not fools, indeed the opposite. But when it comes to "common causality", doing a FA is sure way to fool yourself.
Anybody with a stats package can do a FA and generally get something out the other end. I bet Harmon would find tube amps get rave reviews in their testing and would correlate with the other factors too.
But I generally agree with Speaker Dave's point (wouldn't I be dumb to disagree with this man!!).
Lipschitz and Vanderkoy often use true experimental methods. In the case of listening panels, you'd introduce some kind of frequency irregularity and see how the ratings changed. Every method has its own special drawbacks and best to use a bunch of methods.
Anybody with a stats package can do a FA and generally get something out the other end. I bet Harmon would find tube amps get rave reviews in their testing and would correlate with the other factors too.
But I generally agree with Speaker Dave's point (wouldn't I be dumb to disagree with this man!!).
Lipschitz and Vanderkoy often use true experimental methods. In the case of listening panels, you'd introduce some kind of frequency irregularity and see how the ratings changed. Every method has its own special drawbacks and best to use a bunch of methods.
Yeah... I have a pair of B&Ws on hand too.. and I use them too, to get a reality check from time to time. They serve this purpose quite well, don't they?
dewardh:
Hmmm... perhaps I didn't use the right words there. I do not know if there is a connection or not. It appears, that just by virtue of being more sensitive, the AN sounds more life like.
The ANs do sound terrible as you turn up the wick, so in a way they are limited in dyanmic range, but then I don't listen as loud and so may not need the ability to hit clean 105 db peaks, even though it may be desirable. Just by being as sensitive as they are gives me enough dynamic range.
My next project will be close to 100 db/W... or atleast I'm hoping so.
dewardh:
Hmmm... perhaps I didn't use the right words there. I do not know if there is a connection or not. It appears, that just by virtue of being more sensitive, the AN sounds more life like.
The ANs do sound terrible as you turn up the wick, so in a way they are limited in dyanmic range, but then I don't listen as loud and so may not need the ability to hit clean 105 db peaks, even though it may be desirable. Just by being as sensitive as they are gives me enough dynamic range.
My next project will be close to 100 db/W... or atleast I'm hoping so.
NO and yes , the best all round balance will come from equally weighted factors. Favoring one over the other will give you a speaker that is specific in its reproduction instead of being neutral.
Would that be the practice of medicine ? ... 😛
I find the numbers acceptable, they do correlate somewhat in real world observations.
This I think is funny, because we have numbers who display as the THD of equipment.
But it doesn't displays how we perceive them. Thus How the numbers are weighted to our perception. Like Earl Geddes did try to prove in his white papers.
And how, pray tell, do you sense what is radiating in other directions when in an anechoic chamber? By definition you can only hear the response from the particular listening axis you are on.
A couple of interesting tech papers (Salmi is one) talk about good speakers sounding the same in an anechoic chamber but sounding different in a room due to their particular polar patterns (due to d.i. vs. frequency).
David S.
Oops, response to early post, thought I was on the last page. Still....?
Last edited:
Your head is typically not clamped. So you're not restricted to a single axis- especially true with stereo.
Yes, I'm thinking of medical research where every drug and procedure must be put through exhaustive double-blind studies to prove that things really work with no placebo effect, no anechdotal evidence allowed.
This, as opposed to the dietary suplement folks who can claim anything and use marketing to convince the gullible that any vitamin, extract or suplement will have miraculous curative value.
Hmmm, which one sounds like the audio business??
David S.
But if you are going to compare two loudspeakers in an anechoic chamber you will have to define (set) the listening axis. Or are we saying "speaker A sounds better than speaker B when I move all around them"?
Surely, from a fixed vantage point only the response on that axis can be heard. If the test can't control exact listening position then near on-axis response changes may have some importance, but that is not the same as saying that speakers may be rank ordered by good and bad polar curves, in an anechoic chamber.
David S.
There's research on the effects of head clamping- that greatly inhibits localization. You don't have to get up and move around.
The head wouldn't have to move much to be comparable with wavelengths applicable with localisation, but I wonder how much greater the room itself would contribute to the noticeable difference compared with the program material differences over that small region of head movement.
I also doubt polar issues would be significant over this region, but I know I can detect room modes well into the midrange so I suspect that an anechoic chamber would significantly reduce the noticeable effect of head clamping.
Hey Dave,
If you have full access to those papers I'd be extremely interested in reading them as they touch very much on my OP.
I've been keeping up with the thread despite that it's weened off track (which is fine) but I seem to agree with you and the findings on the paper you referenced since it makes the most sense.
I have a hard time believing that speakers can sound different if both are flat and room modes are taken out of the equation. What sound is being produced on an axis I'm not listening to shouldn't be and important factor if i'm listening to an axis that is coming through as 'flat'.
I guess this is why major recording studios, audio engineers have high end studio monitors and extremely well designed room with a lot of money invested in room treatment as opposed to Tube preamps, McIntosh amplifiers and 30 000$ audiophile speakers.
Jr,
- Status
- Not open for further replies.