So you noticed it can do midrange and bass 👍 the other top performing midrange is a compression driver on waveguide/horn.... See how its a win win?where jbl cut the 15"
As for the 15" talk....This is a core topic (sd) of all drivers, and interesting to talk about here along Purifi, a top performing driver maker, and how Sd will affect the outcome of performance.
Well, discerning by the interesting conversation that has developed, along with keeping a good thread active, I'd say all is well....
If you have something interesting to say, that is more "on topic" or to your liking rather, go ahead and post it!
cut at 400hz is midrange for you ??
Purifi dont make pro speaker that work with horn please start a horn thread
I didn't say that wiggle is the problem, but vertical dispersion of a 2-way with LR4 xo. To be more precise, it's a problem measurement-wise, not necessarilly audible at spot. Sometimes it is compensated in on-axis response to get smoother PIR. If that is good design or a way to overcome a problem, who knows. Anyway, it is an issue that raises lots of questions and discussions during this millenium among speaker designers and reviewers.
Owl's face...
Midrange is considered down to about 120hz. Being able to crossover over lower, is also a desirable trait. Purifi woofers are of the quality level of pro woofers and low distortion so it should work out with proper implementation. Meaning it should have at least a good chance to keep up with the sound quality of a compression driver on a waveguide/horn, making it worthy of mating to one...cut at 400hz is midrange for you ??
Purifi dont make pro speaker that work with horn please start a horn thread
No I don't assume. Why are you constantly putting words in my mouth and nitpicking about irrelevant details? This happens in many other threads too. Isn't this just a discussion forum for hifi diy enthusiasts, not a scientific publication. Me and IMO most others JUST want to discuss here without double checking and peer controlling every sentence. Sorry.
My example of LR4 was JBL M2 and I showed vertical spectrogram by Erin's Audio Corner.
Princeton University Audio Lab have measured eg. Gedlee Nathan's vertical dispersion, and it shows typical 2-way owl's eyes. It has 10" woofer, and I don't know it's xo topology, but if I remember correctly it isn't pure bread LR4. I couldn't find bigger models' verticals.
One other good source for vertical measurements is ASR
My example of LR4 was JBL M2 and I showed vertical spectrogram by Erin's Audio Corner.
Princeton University Audio Lab have measured eg. Gedlee Nathan's vertical dispersion, and it shows typical 2-way owl's eyes. It has 10" woofer, and I don't know it's xo topology, but if I remember correctly it isn't pure bread LR4. I couldn't find bigger models' verticals.
One other good source for vertical measurements is ASR
Please do not minimise crossover type, nothing is more important in controlling DI wiggles, for a given physical configuration.
That something has gone wrong with the EQ in the M2 somewhere along the line because the directivity index is pretty much identical
Certainly the resolution is very high, the spikiness of the response that low in frequency doesn't really make sense though so I wouldn't see that as heart breaking truth but a strange result to investigate the cause of (which is obviously not practical in a review loudspeaker).
It makes sense to me because between 400Hz and 1600Khz is two musical octaves (400, 800, 1600Hz) but mathematical frequency span of 1200Hz (1600-400)
To get a 1/24 octave frequency resolution one needs a resolution of 1200 / 48 = 25Hz.
For 25 Hz resolution you need a window of 40ms.
To get a window of 40ms at 1m mic measurement distance, the reference axis has to be 7.5m above the ground.
So unless JBL/Harman were measuring with their speaker in an anechoic chamber some 7.5m (some 25 ft) from the nearest boundary eg. lifted their speaker 7m off the ground, they simply cannot get the resolution that the NFS gets...
To get a 1/24 octave frequency resolution one needs a resolution of 1200 / 48 = 25Hz.
For 25 Hz resolution you need a window of 40ms.
To get a window of 40ms at 1m mic measurement distance, the reference axis has to be 7.5m above the ground.
So unless JBL/Harman were measuring with their speaker in an anechoic chamber some 7.5m (some 25 ft) from the nearest boundary eg. lifted their speaker 7m off the ground, they simply cannot get the resolution that the NFS gets...
You misunderstand me the 250 to 400Hz region seems to have a slightly exaggerated spikiness that isn't quite the same in the 4367 where the woofers are very similar. The M2 was measured on it's side which is the only real change in measurement conditions. It had nothing to do with the resolution of NFS vs Harman's Anechoic chamber. I don't remember off hand what kind of resolution they can get but they usually smooth at 1/20.
I can't imagine that comparing graphs at 2Hz resolution is particularly useful, 1/20 oct is the minimum required by CTA2034, 1/48 would seem reasonable as otherwise you could be looking partly at rounding errors in the calculations, Vituix applies this by default.
Also at 300 to 400Hz the size of chamber seen in Harman images should be almost completely anechoic meaning that there is effectively no boundary and the distance to it doesn't really matter, at lower frequencies that would not be the case but that isn't the range where there is so much discrepancy.
I can't imagine that comparing graphs at 2Hz resolution is particularly useful, 1/20 oct is the minimum required by CTA2034, 1/48 would seem reasonable as otherwise you could be looking partly at rounding errors in the calculations, Vituix applies this by default.
Also at 300 to 400Hz the size of chamber seen in Harman images should be almost completely anechoic meaning that there is effectively no boundary and the distance to it doesn't really matter, at lower frequencies that would not be the case but that isn't the range where there is so much discrepancy.
By raggedness between the 250Hz to 400Hz you mean the peaks and dips?
"I noticed the resonance in the SPIN data and this made me curious what was going on. I first performed a near field measurement of the woofer with and without the grille. The results showed no difference in this region (meaning, the grille wasn’t the cause). So I measured the impedance of the woofer in the enclosure and then removed the woofer from the enclosure and measured the raw woofer impedance. As you can see, the enclosure is responsible for a resonance around 270Hz (I presume this is a standing wave as the frequency is roughly equivalent to the height of the enclosure). There is another resonance that is in the woofer itself centered around 400-500Hz; I presume this to be cone edge/surround resonance. This is responsible for the dip you see in the frequency response at this frequency."
[snip]
"There is a series of resonances in the midrange that looks concerning in the data. In fact, that is why I went through the trouble of removing the woofer from the enclosure and testing; so I could try to find where the fault was. As I mentioned earlier, I found a resonance in the enclosure at around 270Hz and another resonance in the woofer which was responsible for a 3dB dip from about 400-600Hz. Objectively, the dip doesn’t concern me as much the series of resonances in the 200-300Hz region. However, even though I knew these resonances were there, they were not audible to me in my listening tests. That could simply be too low in amplitude as they are only about +1.5dB above the mean in this region. It could be that the room is so dominant of the sound in this frequency. It could be a combination. While I am it, I’ll also note that JBL’s own specification sheet for the 2216Nd (link) shows the same woofer resonance. Additionally, the JBL 4367 (which uses the same woofer) has some interesting data in their white paper (link) which coincides with the 500Hz dip shown in my data."
Reference:
https://www.erinsaudiocorner.com/loudspeakers/jbl_m2/
Short of sending the speaker back to Harman, or reinforcing the cabinet the @augerpro way: https://www.somasonus.net/box-construction-methods we'll probably never know for 100% what going on below 600Hz. But removing box resonances, identified on a impedance trace is often a painfully slow process, not in the realm of any review.
But I stand by what I said regarding the resolution. Both the Canadian NRC and Harman's anechoic chambers are only good down to 100Hz and 60Hz respectively, which is a hint to their resolution.
"I noticed the resonance in the SPIN data and this made me curious what was going on. I first performed a near field measurement of the woofer with and without the grille. The results showed no difference in this region (meaning, the grille wasn’t the cause). So I measured the impedance of the woofer in the enclosure and then removed the woofer from the enclosure and measured the raw woofer impedance. As you can see, the enclosure is responsible for a resonance around 270Hz (I presume this is a standing wave as the frequency is roughly equivalent to the height of the enclosure). There is another resonance that is in the woofer itself centered around 400-500Hz; I presume this to be cone edge/surround resonance. This is responsible for the dip you see in the frequency response at this frequency."
[snip]
"There is a series of resonances in the midrange that looks concerning in the data. In fact, that is why I went through the trouble of removing the woofer from the enclosure and testing; so I could try to find where the fault was. As I mentioned earlier, I found a resonance in the enclosure at around 270Hz and another resonance in the woofer which was responsible for a 3dB dip from about 400-600Hz. Objectively, the dip doesn’t concern me as much the series of resonances in the 200-300Hz region. However, even though I knew these resonances were there, they were not audible to me in my listening tests. That could simply be too low in amplitude as they are only about +1.5dB above the mean in this region. It could be that the room is so dominant of the sound in this frequency. It could be a combination. While I am it, I’ll also note that JBL’s own specification sheet for the 2216Nd (link) shows the same woofer resonance. Additionally, the JBL 4367 (which uses the same woofer) has some interesting data in their white paper (link) which coincides with the 500Hz dip shown in my data."
Reference:
https://www.erinsaudiocorner.com/loudspeakers/jbl_m2/
Short of sending the speaker back to Harman, or reinforcing the cabinet the @augerpro way: https://www.somasonus.net/box-construction-methods we'll probably never know for 100% what going on below 600Hz. But removing box resonances, identified on a impedance trace is often a painfully slow process, not in the realm of any review.
But I stand by what I said regarding the resolution. Both the Canadian NRC and Harman's anechoic chambers are only good down to 100Hz and 60Hz respectively, which is a hint to their resolution.
Resolution is lost through time gating, it is my understanding that Anechoic chambers use a correction not a gate at the reflection point. If the measurement is not gated and there is no reflection at the frequency of interest there is nothing to reduce the resolution in the measurement method itself. We are a long way from exploring Purifi woofer's here though...
I dunno I think we're all on the same book, if not on the same chapter.
Scan-Speak has shown us their 38WE.
A Purifi 15" woofer will hopefully have a smoother flatter frequency response than a paper from JBL or Yamaha or any other established PA driver manufacturer using a paper/fibre cone.
How you determine that is through measurements. And that will be hard...
Luckily for camplo, Carsten Tingaard, who was responsible for those XXLS woofers from Peerless, so no doubt he will not leave a stone unturned and is working on bigger drivers behind the scenes at Purifi...
I, OTOH, have no desire for a 15" "midwoofer"
Scan-Speak has shown us their 38WE.
A Purifi 15" woofer will hopefully have a smoother flatter frequency response than a paper from JBL or Yamaha or any other established PA driver manufacturer using a paper/fibre cone.
How you determine that is through measurements. And that will be hard...
Luckily for camplo, Carsten Tingaard, who was responsible for those XXLS woofers from Peerless, so no doubt he will not leave a stone unturned and is working on bigger drivers behind the scenes at Purifi...
I, OTOH, have no desire for a 15" "midwoofer"
The advantage of increasing the Sd is that the excursion dependent distortion is reduced (plays louder for the same distortion or lower distortion for the same SPL). Moreover, the Sd modulation for the same excursion is typically lowered when Sd is increased (surround area is a smaller fraction of the Sd).
However, the scaling is not so clear for the current dependent distortion (current dependent Bl modulation aka FFM and magnetic hysteresis distortion).
Note also that the technologies we use offer a substantial reduction of distortion compared to conventional tech (eg FFM, Sd modulation, hysteresis and Bl(x) variation reduced by maybe 20-30dB). This means that a substantial increase of Sd is needed to offset this distortion advantage when using conventional tech.
The drawbacks of increasing Sd is less dispersion (beaming starts earlier), reduced upper bandwidth, larger and more expensive motor needed.
The next milestone for us at Purifi is a 10”. Hopefully we start prototyping this year - waiting for some tooled parts to arrive.
cheers
Lars
However, the scaling is not so clear for the current dependent distortion (current dependent Bl modulation aka FFM and magnetic hysteresis distortion).
Note also that the technologies we use offer a substantial reduction of distortion compared to conventional tech (eg FFM, Sd modulation, hysteresis and Bl(x) variation reduced by maybe 20-30dB). This means that a substantial increase of Sd is needed to offset this distortion advantage when using conventional tech.
The drawbacks of increasing Sd is less dispersion (beaming starts earlier), reduced upper bandwidth, larger and more expensive motor needed.
The next milestone for us at Purifi is a 10”. Hopefully we start prototyping this year - waiting for some tooled parts to arrive.
cheers
Lars
Not wanting to polemicize, but some of us might count this as an asset though.