http://www.mcmaster.com/#f13-felt/=wnjab6
Courtesy of our member badman. http://www.audioheritage.org/vbullet...l=1#post268321
I've used the 1/2" stuff in the past but the 476's are loaded with two of the foam pads instead of just the usual one.
http://www.mcmaster.com/#f13-felt/=wnjab6
Courtesy of our member badman. http://www.audioheritage.org/vbullet...l=1#post268321
I've used the 1/2" stuff in the past but the 476's are loaded with two of the foam pads instead of just the usual one.
If you want to stick with foam you can try these, readily available in Europe :
http://www.ebay.fr/itm/PANNELLI-FONO...item4aaf3b800a
http://www.ebay.fr/itm/PANNELLO-ADES...item4ae4110565
(with or without adhesive backing)
You can also try their melamine foam (basotect), which is supposed to be more efficient in absorbing waves...
I don't think the EQ was done automatically on the M2: this type of EQ is easy to do manually (I did something similar on some 2344), the hard part is to get a meaningful (set of) measurement(s).The settings for M2 kindly posted by 4313b leads me to think that JBL has a computer algorithm to calculate several different combinations of EQ/PEQ/phase/delay with an acceptable final curve. Then they listen to each different combination to verify that a given combination of DSP settings indeed sounds good. That is why I think the M2 is getting so fantastic reviews. They have spent the time needed, they have the HW needed, they have a large anechoic chamber and they have the skills to find the right HW / DSP combination.
This is where JBL's infrastructure and (LSR) measurement technique is invaluable: being able to automatically take several hundreds of measurements and deduce 30°/10° averaged curves as well as first reflexions and sound power all in once is great for evaluating and EQing a loudspeaker.
But now that we know that the M2 *is* well behaved the "only" thing that you would need is a 30°/10° averaged measurement to be used as your EQ target, and (manually) EQ that flat as JBL did for the D2+M2 waveguide.
Here is what I did for the 2344:
I mounted the 2344 on a PA loudspeaker pole so that the rotation angle was at the horn slot.
I then set the MLS length to something like 20sec, and slooooowly/smoooothly rotated the pole while taking the measurement, up to around 15°.
This gives a usable 30° horizontal average measurement (+/-15°, the horn being symmetric) that retains phase information (because the slot-mic distance is constant) and can be windowed to exclude reflexions.
I don't think the 10° vertical measurement is that important given the small angle (+/- 5°), but you can obtain it by rotating the horn on the pole so that the verticals become the horizontals... (you can also of course do the diagonals that way)
+/-15° horizontal averaged measurement (caution: the 2426J used was probably somewhat out of spec...) :
EQ curve (~15 manual EQ and filter points) :
final +/-15° horizontal averaged measurement (900Hz LR 36dB/oct target) :
Here is the on-axis response:
I don't have any single point off-axis response measurement to show (did not keep them, sorry).
If you open both the 30° averaged and the on-axis response in two tabs of your browser and alternatively switch from one to the other, you will see that the on-axis response has a small and low Q depression around 8kHz that is not there in the averaged measurement.
In fact the response is hotter 15° off-axis than on-axis at these frequencies, which seems to be a common trend in many biradial horns....
For reference, here is the on-axis response measurement of a 2426H on the 2344, taken from the brochure:
As you can see the depression around 8kHz is clearly visible here (and you can also guess that the 2426J I was using was probably not up to spec...)
Very nice work!
I'm not sure if you took this into account or not, but did you use a baffle? I have never taken measurements of the 2344, but most of the horns I have measured, measure differently when in a baffle than when freestanding. I'm not sure if a baffle the size of your final speaker would markedly change your results, but it might.
Widget
Hi POS,
So averaged and on-axis on the same graph.
so the differences are not so much, may be the response on the half of the off-axis response angle would better match, but the differences are not so much as can be expected, may be as 2344 horn is quite wide dispersion angle over +/- 45 deg off axis (horizontal)
regards
ivica
Hi,
Excellent data. I can try to do replicate that for the M2/2451Ti (without baffle) over the weekend if time permits. How far out did you measure? I will measure outdoors to use a long time window and long sweep.
I have some on and of axis measurements from last week. I will see if I can find them and post here. They are with a out of spec driver but gives a picture of what happens.
Do you have any suggestion for a European felt supplier as I would be preferred felt over foam? I have checked here but they only offer thin felt. If not, I will follow Badman’s link and order from the US.
Regarding DSP settings, I think you can get to the same final correction curve using different combinations of DSP settings as they interact with one another. Whether the different combinations sound different is a different story. The M2 settings are rendered good by many so I will try to keep as close to them as possible.
What I can say is that the same DSP parameter settings do sound different in different brand DSP units. I have not tested if they indeed measure different (with the same settings) and if corrected to the exact same output would sound the same. My guess has always been that they use slightly different algorithms to treat the sound so even if a PEQ is set to the same exact parameters they do not do exactly the same thing. -And that the implementation of the A/D converters is different. That is why I feed my BSS digital signal.
Maybe 4313b or someone else on this excellent forum can shed some light on how the JBL DSP setting process works.
Kind regards
//RoB
The solution to the problem changes the problem.
-And always remember that all of your equipment was made by the lowest bidder
Thanks Widget
There was no baffle but I figured it would not change the response that much: the butt cheeks are off the baffle anyway, and in a 4430/35 only the lower lips would have some sort of baffle support. Maybe I was wrong and as it could possibly change the response down low though.
This was more of a test, as obviously the driver I had at hand was not good. I'd love to get a proven 2426H/J to test though, and publish the EQ for anyone to use.
In this case this baffle thing would have to be addressed.
Another question is driver to driver variation: would a precise EQ be worth it with driver that can vary in their response from one (good) unit to another (good) as much as the 2426 can (even new ones). I think the D2 is much more consistent in this regard (similar to BMS driver maybe), as among the 3 I had, all measured almost exactly the same, even those that had distortion problems...
Obviously to make precise EQ usable you need a horn that has an even reponse over a wide angle *and* a driver that has a good unit to unit consistency...
Not sure about the distance. It was indoor (more of a test than anything serious...) and the mic was probably 1.5m away from the horn...
If you want to rotate the horn during measurement you have to use an MLS signal.
Maybe something like that?Do you have any suggestion for a European felt supplier as I would be preferred felt over foam? I have checked here but they only offer thin felt. If not, I will follow Badman’s link and order from the US.
http://www.audiophonics.fr/fr/traite...mm-p-5410.html
What thickness are you looking for?
Any minimum phase correction leading to the same amplitude curve should give you the exact same result, regarless of the actual number of EQ points and type used underneath.Regarding DSP settings, I think you can get to the same final correction curve using different combinations of DSP settings as they interact with one another. Whether the different combinations sound different is a different story. The M2 settings are rendered good by many so I will try to keep as close to them as possible.
There are many variations in EQ types and interpretations from one unit to another: constant Q vs proportional Q EQs, Fc position of a shelving filter, Bessel Fc interpretation, etc...What I can say is that the same DSP parameter settings do sound different in different brand DSP units. I have not tested if they indeed measure different (with the same settings) and if corrected to the exact same output would sound the same. My guess has always been that they use slightly different algorithms to treat the sound so even if a PEQ is set to the same exact parameters they do not do exactly the same thing. -And that the implementation of the A/D converters is different. That is why I feed my BSS digital signal.
This can really be a source of problems and a major cause of sonic differences when going from one unit to another.
This is something I tried to address in my M2 settings analysis document: https://docs.google.com/document/d/1...haring&rm=demo
The best way to avoid compatibility problems is to avoid shelvings and bessel filters when designing your EQ. Then you only have to deal with constant vs proportional Q... (easy enough)
Hi POS,
Very nice work about M2 equalization You have shown us in the mentioned link.
So if I have understood your work 'inverse' response would be M2 & D2430K FR response.
I have done that from your figures and attached here where the mentioned response of M2 with D2430K is shown as red line.
regards
ivica
Hi Ivica,
Yes more or less, but taking into account the passive network and 800Hz LR 36dB/oct acoustical high-pass target (flat otherwise, with no attenuation up high as shown in the LSR measurement set, but is it true?...), the listening window averaged response of the D2/M2 waveguide with no filter should look like that:
http://www.audioheritage.org/vbullet...l=1#post373835
Hi POS,
Very interesting behavior over 8kHz (neglecting about 10dB lower then mid frequency range)
If such response can be get from the measurements of the driver and horn combo, they behave as two drivers are in the combo, one VHF connected to CD horn and UHF driver with diffraction slot (such as 2405). Interesting....(almost unbelievable - for me)....
regards
ivica
There are currently 1 users browsing this thread. (0 members and 1 guests)