I need opinions from some pretty smart people on this one, as it's just slightly past my pay grade:
The basic question concerns polar response graphs, and how to interpret them as to how the speaker will sound / react in an actual room.
It’s a good question on general principles, but also (of course) pertains to a specific quandary I have right now. Here are the details:
-----------------------
I just bought a used pair of EAW KF394’s.
They have a basic dispersion of 90º H x 45º V.
They are a 3-way box, using the BMS 4594ND coaxial for the mids & HF. This is attached to a 90x45 horn, but this horn’s outer dimensions are square. I don't yet know where the tens cross to the coaxial, but it surely must be around 400 - 500 Hz, as it typical with the 4594.
---------------------
For both ergonomic and aesthetic reasons, I am thinking about rotating the horn 90 degrees, and then mounting the cabinets sideways. In other words, the horn would remain 90º H x 45º V, but the tens would now be at the left & right, instead of the top & bottom.
Because the tens are simply vented, with no apparent dispersion control, I assumed that this would make no difference in the expected 100 - 500 Hz “low mids” range. - But then I checked the polar response graphs. I’m now a bit puzzled.
There is, to my surprise, a serious variance in frequency coverage in the (stock) vertical direction, from 100 to 500 Hz. (esp above 250 Hz)
## But I’m not even sure which way would sound best in an actual room, which is why I’m posting this query. I’m trying to decide if this “sideways box” orientation (again, with the horn remaining as it was) is a bad idea, or could possibly be an improvement. (Again, in an actual room, with real-world reflections.)
The response graphs can be seen here:
https://eaw.com/docs/1_Current_Products/KF/Spec_Sheets/KF394_SPECS_revA.pdf
I've also pared down the pertinent data to a single jpeg, so it's easier to see, but I don't know how to post it. (Any ideas?)
————————————————————————
So, what do you guys think?
Am I reading the graphs wrong?
If not, should I worry about this? 250 - 500 Hz is the “mud” octave, anyway. Maybe I’m better off with less on the sides, because that octave tends to build up in smallish rooms, anyway?
Not to mention, stock horizontal gets pretty tight above 600 Hz. Almost as tight as the HF. So really, the vertical orientation has a more even response above 200 Hz ! Why not make that the horizontal instead?
MAYBE EVEN PUT SOME 100-500HZ absorption at the bottom of the speakers, to keep that from hitting the floor?
All thoughts would be welcome & appreciated.
Thx.
The basic question concerns polar response graphs, and how to interpret them as to how the speaker will sound / react in an actual room.
It’s a good question on general principles, but also (of course) pertains to a specific quandary I have right now. Here are the details:
-----------------------
I just bought a used pair of EAW KF394’s.
They have a basic dispersion of 90º H x 45º V.
They are a 3-way box, using the BMS 4594ND coaxial for the mids & HF. This is attached to a 90x45 horn, but this horn’s outer dimensions are square. I don't yet know where the tens cross to the coaxial, but it surely must be around 400 - 500 Hz, as it typical with the 4594.
---------------------
For both ergonomic and aesthetic reasons, I am thinking about rotating the horn 90 degrees, and then mounting the cabinets sideways. In other words, the horn would remain 90º H x 45º V, but the tens would now be at the left & right, instead of the top & bottom.
Because the tens are simply vented, with no apparent dispersion control, I assumed that this would make no difference in the expected 100 - 500 Hz “low mids” range. - But then I checked the polar response graphs. I’m now a bit puzzled.
There is, to my surprise, a serious variance in frequency coverage in the (stock) vertical direction, from 100 to 500 Hz. (esp above 250 Hz)
## But I’m not even sure which way would sound best in an actual room, which is why I’m posting this query. I’m trying to decide if this “sideways box” orientation (again, with the horn remaining as it was) is a bad idea, or could possibly be an improvement. (Again, in an actual room, with real-world reflections.)
The response graphs can be seen here:
https://eaw.com/docs/1_Current_Products/KF/Spec_Sheets/KF394_SPECS_revA.pdf
I've also pared down the pertinent data to a single jpeg, so it's easier to see, but I don't know how to post it. (Any ideas?)
————————————————————————
So, what do you guys think?
Am I reading the graphs wrong?
If not, should I worry about this? 250 - 500 Hz is the “mud” octave, anyway. Maybe I’m better off with less on the sides, because that octave tends to build up in smallish rooms, anyway?
Not to mention, stock horizontal gets pretty tight above 600 Hz. Almost as tight as the HF. So really, the vertical orientation has a more even response above 200 Hz ! Why not make that the horizontal instead?
MAYBE EVEN PUT SOME 100-500HZ absorption at the bottom of the speakers, to keep that from hitting the floor?
All thoughts would be welcome & appreciated.
Thx.
Last edited:
They are still in shipping, so I can't. But even if I had them, I'd have to disconnect the tens in order to do guess at the xover point. That's a bit of work.
FWIW, many designs cross the 4594 at 400 Hz. (They are spec'ed to 300 Hz.)
- But regardless, that doesn't even matter. The actual crossover point does not affect the answers to my questions at all. Only the polar response does.
-----------------
As for "just trying it." No thanks. Turning one horn and rigging some kind of pole or truss mount would be do-able, but then I'd have to go do A/B tests in at least a dozen different rooms before I'd be comfortable with my own personal opinion. Not going to happen.
And no, I'm not overthinking it.
Like I wrote in the opening post, I'm interested in what should work best in real-world conditions. That's why I'm looking for real-world opinions, from guys that have dealt with this before.
Last edited:
To measure xover....just don't feed the tens when in biamp mode....
As far as crossover freq...that is horn control dependent, which means it takes a real big horn to go much below 600Hz..
But....Specs seem to indicate the hf/VHF kicks in a little below 500Hz, which makes sense because the 10's can't keep up with the CD, and no risk of over excursion of the CD down low.
By just try it, I mean just turn the speaker on its side
...
As far as crossover freq...that is horn control dependent, which means it takes a real big horn to go much below 600Hz..
But....Specs seem to indicate the hf/VHF kicks in a little below 500Hz, which makes sense because the 10's can't keep up with the CD, and no risk of over excursion of the CD down low.
By just try it, I mean just turn the speaker on its side
...
Mark, thanks for that clarification. - Very helpful.
Well, no, because that switches the M-H dispersion, which is exactly what I WON'T be doing.
T
By just try it, I mean just turn the speaker on its side
...
Well, no, because that switches the M-H dispersion, which is exactly what I WON'T be doing.
Yes. I got the passive version, not the self-powered "NT" version.
The focusing software (the Gunness processing that EAW calls "greybox" ) comes with both the NT internal amps, and either of their two processors, which are the UX3600 and the very expensive UX8800.
----------------------------------------------
In case anyone is interested in this general line from EAW, they are phasing out ALL of the "NT" speakers in their lineup, and currently only making the passive versions. The lovely KF394 will continue for a while, but not the NT version.
I believe this is because everything is changing to all-digital connections.
It's becoming standard now for high end plate amps and touring rack amps to have direct AES inputs, plus of course digital mixers are now well-established.
(I myself use a DAW, running on Windows, but it's the same idea.)
EAW is about to release a new high end processor that also has AES in & out: The UX48. It should be available first quarter of 2019.
I'd LOVE to get one, but sadly I'm told it will be at least as expensive as the current UX8800, which means well-north of $4K.
As for amps, I have two spare Lab Gruppen Fp6400's that I can use for the biamped KF394's
- so no worries there ! (unless I someday get a UX48.) The nice thing about using a high-end processor is that you can get by with really basic amps. No extra DSP needed. The EAW processors even do voltage-based limiting, based on the specific amps you're using.
The recommended active crossover appears to be just under 500Hz.
With regards to the original question, it looks like mounting the cabinet horizontally (and rotating the HF horn to keep is 90x40) would give you a very consistent beamwidth in the horizontal plane. That's generally a Good Thing.
However, your vertical beamwidth would end up with a lot of variation. Bad Thing.
Doing as you suggest would mean you're bouncing less lower-midrange off the walls, but more towards the ceiling and floor.
I'd leave them as they are.
Chris
With regards to the original question, it looks like mounting the cabinet horizontally (and rotating the HF horn to keep is 90x40) would give you a very consistent beamwidth in the horizontal plane. That's generally a Good Thing.
However, your vertical beamwidth would end up with a lot of variation. Bad Thing.
Doing as you suggest would mean you're bouncing less lower-midrange off the walls, but more towards the ceiling and floor.
I'd leave them as they are.
Chris
Hey Cableaddict, just in case you haven't found pricing you like on the UX processors, last time I looked around for them, the 8800 was under $2900, and the 3600 under $1200.
(Which kinda begs the question why would any one ever get an 8800...when two 3600's are less..??)
Congrats on the 394's btw, enjoy!
(Which kinda begs the question why would any one ever get an 8800...when two 3600's are less..??)
Congrats on the 394's btw, enjoy!
I think you're probably right, if you mean crossover frequencies are deliberately set apart from each other, if favor of better overall pattern control at perhaps an expense to on-axis response.
Most all of the EAW legacy settings show spread apart x-overs..
Here's a thread I initiated on another forum when tuning a 695z...
I remember asking about the xover spread somewhere in the thread...
EAW legacy settings
1) The 10" are a spaced pair of offset horns with porting. They are around a wavelength apart at 600 Hz, the spacing results in the vertical pattern control variations, the horizontal pattern gradually and evenly narrows with increasing frequency.
2)The stock KF394 configuration keeps some of the low mid "mud" off the floor and ceiling, where standing waves tend to make those frequencies more problematic.
In most rooms, the horizontal response is more important than vertical, since the audience is dispersed wider horizontally compared to vertically.
Rotating the cabinet will make the low mid response less even across the audience, and put more low mid energy on the floor and ceiling.
3) Above 630 Hz, most of the control exhibited in the polar response graphs are due to the HF horn.
Your aesthetic and ergonomic considerations weigh against a relatively small low mid dispersion difference, but rotating the HF horn will result in a compromised, negative sonic effect in a frequency range you seem to be critical of.
Attachments
Thanks, Art. This is very helpful.
A very good point about low-mid standing waves, which Chris also mentioned & which I should have realized myself. Whatever the actual comb filtering, those effects will be much more prominent floor-to-ceiling than wall-to-wall. (unless the ceiling is REALLY absorbent.)
- And believe it or not, I actually understand your first paragraph! I have been reading today about MTM designs, and how critical the two outer speakers distance is for pattern response, in conjunction with the crossover point & slope. (My head hurts....)
I'm guess that they are using a 2nd order x-over at 500 Hz, because that does in fact tighten the response within the x-over region, so would certainly have some effect down to at least 250 Hz.
Would you agree, or am I understanding this wrong?
Last edited:
Wow - Where did you see a UX3600 for $1200? I haven't order one yet.
(Obviously, the prices are starting to drop because of the upcoming UX48.)
And yeah, I'm REALLY looking forward to using these speakers! They will be here tomorrow. First test will only be full range, without Gunness, but it should still be pretty sweet.
I don't understand what you mean by "tighten the response within the x-over region".
Second order is only a 12dB per octave slope, the processed response of the KF394 shows a 500Hz fourth order (24dB per octave) response for both the LF and Hf sections. The unprocessed response of both HF and LF sections are different, requiring different filters to arrive at the intended acoustical response.
Attachments
AH, that's how they're doing it.
I understand now. Thanks.
---------
FWIW, what I meant by "tighten the response within the x-over region" has to do with the way the polar response narrows, in an MTM system, with a 2nd order x-over. - I had just been reading about this stuff the entire day.
But I guess, had they done that, it would have tightened in both vertical and horizontal.
Last edited:
You were right, Mark. Slightly less than $1200. I ordered one today. Should have it within a few weeks.
In the meantime, I managed to get things sounding surprising great using the DSP in my computer mixer. (DAW)
I used the EAW recommend HPF protection, plus a third-order cross to the subs at 100 Hz. (I'm not brave enough to do gigs at 80 Hz without the processor's protection!) and then I looked at the full-range frequency response graphs and noodled around for a bit, until my ears were happy.
Did a DJ gig today for a young & hip crowd, and I was pretty much blown away.
OMG, these boxes are nice, even without the extra stuff the processor will be doing. Clarity for days, less overall volume needed, and even less feedback from my mic.
Plenty of solid low-mids way out in the room, too, though of course I'm not crossing at 80 hz yet. That will be interesting.
I think this is a really good design for DIY guys to consider. Again, I'm really liking them right now, and I don't yet have the processor, so no digital phase tricks happening, nor any Gunness focusing. If anyone wants pics of the internals, let me know. The way those tens are mounted is pretty trick. - Art explained it well above. They are also mounted on an angle, (is that the "offset" part?) facing into sort of a small "horn" area. I dunno what to call it, though you guys probably know.
Last edited:
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.