Your questions make a lot of sense, and mirror those I've heard over many years. All valid points.
My reservations about the new drivers started when I had to replace the original tweeters some time in 2001 or thereabouts. One of the 9000's had developed a subtle but annoying twittering sound that appeared very occasionally - the sound of a rubbing voice coil. I tolerated this for about a year, then finally decided to replace the 9000 with the 9500 and re-trim the crossover as necessary.
The big problem with the original 9000 was the sticky dome was a magnet for the weird crumbly expanded-foam packing used by Scan-Speak. By the time it got from Denmark, to Madisound, then to the customer, the dome was covered with little white flecks, which required tedious removal with a tweezer. A cosmetic but annoying issue.
Unfortunately, rather than replace the worthless grade of expanded foam - the cheapest and most crumbly stuff I'd ever seen - Scan-Speak changed the tweeter! So the boxes still arrived half-wrecked, but it didn't stick to the dome anymore - the bits of foam just rattled around in the box.
The new-series tweeters measured a little better - maybe - but they had a subtle dryness, a sandy quality, that wasn't there in the original 9000. Admittedly, back in 1991, the other dome tweeters on the market were little short of dreadful, with the awful inverted-domes being the worst of all. At the time, the 9000 was a very easy choice, the only one that measured decently and also sounded quite good.
But now the 9000 is long gone, replaced by tweeters that are optimized to a different set of subjective qualities - that I don't much care for. The 9300 and 9500 are sorta-kinda like a 9000, but some of the beautiful sweetness and "rosiny" quality on violins is gone.
I haven't warmed to the sound of modern high-end speakers designed around the current Scan-Speak, Vifa, Seas, and Dynaudio products. They sound kind of over-analytical to me, and don't quite get the tonality right, compared to the more old-school things they were doing 10 years ago.
I can't quite put my finger on what's going on, except to guess the original designers have retired and new guys are doing the subjective balancing now. That's why I considern Skaaning interesting, since it was founded by the man behind Scan-Speak, and I'm sure has the same sort of subjective balancing as my favorite drivers. I've heard a rumor (but no more than that), that the old man had a hand in designing the Vifa 5.5", which is why it measured and sounded better than the larger-format Vifas which were superficially similar but actually sounded different.
I now know more about driver design than I did then, and something as seemingly inconsequential as glues for the voice-coil to the cone are actually extremely important to the overall sound of the driver, the speaker, and the entire hifi system. What makes this interesting is some of the best-sounding glues are quite volatile and toxic, and are now banned in the EU and many parts of the USA (like Los Angeles, home of JBL and Altec). This is part of the reason some drivers have become off-shored to China, to evade EU and Los Angeles VOC emissions requirements.
So little bitty things like choice of VC former and glues have a big effect on driver sonics. They do not sound the same - and can be changed in production without too much effect on the driver measurements, so driver vendors can and do make these kind of production changes.
But these unannounced running changes do no favors for the loudspeaker manufacturer, who then have to re-balance the crossover, or even change drivers. This is part of the reason for the Version 2 speakers we see in the market - a lot of time it happens just because the original part is no longer available, and the marketing department has to make a virtue of necessity.
My reservations about the new drivers started when I had to replace the original tweeters some time in 2001 or thereabouts. One of the 9000's had developed a subtle but annoying twittering sound that appeared very occasionally - the sound of a rubbing voice coil. I tolerated this for about a year, then finally decided to replace the 9000 with the 9500 and re-trim the crossover as necessary.
The big problem with the original 9000 was the sticky dome was a magnet for the weird crumbly expanded-foam packing used by Scan-Speak. By the time it got from Denmark, to Madisound, then to the customer, the dome was covered with little white flecks, which required tedious removal with a tweezer. A cosmetic but annoying issue.
Unfortunately, rather than replace the worthless grade of expanded foam - the cheapest and most crumbly stuff I'd ever seen - Scan-Speak changed the tweeter! So the boxes still arrived half-wrecked, but it didn't stick to the dome anymore - the bits of foam just rattled around in the box.
The new-series tweeters measured a little better - maybe - but they had a subtle dryness, a sandy quality, that wasn't there in the original 9000. Admittedly, back in 1991, the other dome tweeters on the market were little short of dreadful, with the awful inverted-domes being the worst of all. At the time, the 9000 was a very easy choice, the only one that measured decently and also sounded quite good.
But now the 9000 is long gone, replaced by tweeters that are optimized to a different set of subjective qualities - that I don't much care for. The 9300 and 9500 are sorta-kinda like a 9000, but some of the beautiful sweetness and "rosiny" quality on violins is gone.
I haven't warmed to the sound of modern high-end speakers designed around the current Scan-Speak, Vifa, Seas, and Dynaudio products. They sound kind of over-analytical to me, and don't quite get the tonality right, compared to the more old-school things they were doing 10 years ago.
I can't quite put my finger on what's going on, except to guess the original designers have retired and new guys are doing the subjective balancing now. That's why I considern Skaaning interesting, since it was founded by the man behind Scan-Speak, and I'm sure has the same sort of subjective balancing as my favorite drivers. I've heard a rumor (but no more than that), that the old man had a hand in designing the Vifa 5.5", which is why it measured and sounded better than the larger-format Vifas which were superficially similar but actually sounded different.
I now know more about driver design than I did then, and something as seemingly inconsequential as glues for the voice-coil to the cone are actually extremely important to the overall sound of the driver, the speaker, and the entire hifi system. What makes this interesting is some of the best-sounding glues are quite volatile and toxic, and are now banned in the EU and many parts of the USA (like Los Angeles, home of JBL and Altec). This is part of the reason some drivers have become off-shored to China, to evade EU and Los Angeles VOC emissions requirements.
So little bitty things like choice of VC former and glues have a big effect on driver sonics. They do not sound the same - and can be changed in production without too much effect on the driver measurements, so driver vendors can and do make these kind of production changes.
But these unannounced running changes do no favors for the loudspeaker manufacturer, who then have to re-balance the crossover, or even change drivers. This is part of the reason for the Version 2 speakers we see in the market - a lot of time it happens just because the original part is no longer available, and the marketing department has to make a virtue of necessity.
FlorianO said:
Btw. a rock and a hard place.... XO smack bang in the middle of the presence region or too high.
With me swearing up and down by the (claimed!) 300 degrees dispersion of the Gallo Acoustics CDT tweeter XOed @ 3k I would still go for the big 12NDA520 XOed 'bit below 2k vs. the above CD+horn XOed at 7k.
But that again....single driver btw. 400 Hz and 7k (beaming ?) with none (or minimal) overlap at HF...
Geez Lynn, do you _really_ have make it so hard?
Don't worry, it's not as bad as it sounds.
The problem can be approached on purely subjective basis - in advance. We're talking about a different character of sound here. Let's look at the 500 Hz to 2 kHz region. Do you want this handled by a horn or a cone? That's the real choice, when you come right down to it.
One path is a low 500 Hz crossover, similar to the big classic studio monitors with one or two 15-inch drivers and a seriously large horn and compression driver. That Big Mama studio sound, probably not the last word in sweetness, but mondo dynamics, 'cause that's what that crossover and driver lineup will give you. That'll be the basic character of the speaker, never mind the dipole stuff.
The 18Sound drivers of interest are the 15NMB420 and 15MB700 twins, which appear to have identical cones, based on the cone mass and the FR curves.
The other path is - hmm - more of a Bozak big-Fifties-speaker kind of sound, more vintage, a bit more cone coloration, but probably mellower in a full-size classic-radio kind of way. The horn is more of a HF-only driver, and is tasked with sonically matching the midbass driver. If the horn or midbass driver run out of steam first, it'll sound weird. This implies that both need lots of headroom - SPL for the midbass, a generously high crossover for the horn, and a compression driver with a sweet character when pushed hard.
Looking at the second path and 18Sound, I see two drivers that look promising: the previously-mentioned 12NDA520, or the 12ND710. Both have low-mass cones and are specifically designed for midbass use, which is certainly what we want. Both drivers also look like they'd benefit from a fairly narrow 2.2 kHz notch filter, which implies the acoustic crossover should be either at, or a little below, the 2.2 kHz frequency.
The HF section? Well, a ribbon is going to working hard to make it to 2.2 kHz. Very hard. 500 Hz isn't happening. That leaves compression drivers, either 1, 1.4, or 2-inch format. The smaller formats tend - and this is a generalization - to sound like they're working harder than the larger formats. I think the sound of the woofers is going to be more predictable than the HF drivers, which are frankly all over the place, and more different than the measurements show.
I feel pretty confident about ordering the assorted midbass drivers, but I have a feeling I'm going to be auditioning a lot of horns and compression drivers. I'll be stuck up a tree if I don't like any of them.
Lynn Olson said:
In todays market of hifi drivers I came to the same conclusion, the sound of the Audiotechnology drivers are most natural to my ears. With regard to distortion/ FR linearity/CSD they also perform very good. I think Skaaning made the best compromise off all involved factors.
It's maybe interesting to note that I also was after some higher efficiency OB system. In the end I settled for "only 92-93 dB". Main reason was that I used a custom Audiotechnology 23I driver for the mids (shorter coil, different surround and spider). For the highs I paired the mid with a old time favourite of mine the Philips RSQ8P planar. For the low frequencies I chose for a ripole configuration with 2 15" pro drivers (nothing fancy overhere). I mainly opted for the Ripole because of the smaller form-factor (only a 45*50*50 cm cube) compared to a stock OB and a slightly different dispersion.
Therefore, I would suggest not to abandon the mid + pro ribbon route. The offerings of Stage Accompany seem to me a feasible candidate for the tweeter (they bought the Philips patents and improved upon it). However, I've a Raal 140-15D at hand still to be mounted on my OB (or in one of my other projects) and do think that it's better sounding and measuring than my Philips or any other tweeter I've used (including the Mundorf AMT you considered). IMHO the Raal is doing the same as the Audiotechnologies are doing, getting the proper balance. So it all boils down to preference do you want high efficiency and slightly less sound quality (Stage Accompany) or do you want less SPL and higher sound quality (Raal)?
Lynn Olson said:
With the excuse of ignorance: this seems like pleasing the "chick with guitar" crowd too, no ?
LaMa said:
Well, if I read correctly Lynn's post above it all boils down on how does the RAAL make it in the 2.5k -- 5k octave. That's afterall half an octave above the recommended 1.6k XO point (whatever that means).
Maybe -- like Michael suggested above -- an FR, CSD and distortion in that range should help *hint*hint*
?
One option I didn't mention is the possibility of combining the BMS mid/high 2" compression driver with a RAAL or Stage Accompany ribbon, crossing at the natural rolloff point of 7 kHz provided by the BMS, and crossing the BMS at the low end at 1.4 kHz to the 18Sound 12" drivers. (I note the BMS is rolling off at the same frequency as old-school 2" phenolic drivers, which sound very good in their frequency range.)
This may seem a little complex, but all of the drivers - 18Sound 12-inch, BMS 2-inch mid/high, and the ribbon are then used very conservatively, well within their linear ranges. This is especially important for compression drivers and ribbons, which greatly benefit from crossovers that are even modestly higher than the "recommended" crossover points.
The 18Sound would hardly be taxed by a 1.4 kHz crossover, either, since the first peak of any note is at 2.2 kHz, where it can be addressed with a mild notch filter about 3 dB deep.
The rest of the dipole could use the Beyma 18" woofer mentioned earlier, chosen for a relatively low-mass cone and smooth, linear midrange response (unlike most 18's).
The general design principle is to use all drivers well within their linear regions, paying attention to drivers that are sensitive to excess excursion - compression drivers and ribbons in particular.
The alternative to the CD/ribbon scenario is a one or two-metre-high stretched-film ribbon, of the large-area BG type. I'm not really a fan of the line-radiator concept or sound, preferring point-source radiators.
This may seem a little complex, but all of the drivers - 18Sound 12-inch, BMS 2-inch mid/high, and the ribbon are then used very conservatively, well within their linear ranges. This is especially important for compression drivers and ribbons, which greatly benefit from crossovers that are even modestly higher than the "recommended" crossover points.
The 18Sound would hardly be taxed by a 1.4 kHz crossover, either, since the first peak of any note is at 2.2 kHz, where it can be addressed with a mild notch filter about 3 dB deep.
The rest of the dipole could use the Beyma 18" woofer mentioned earlier, chosen for a relatively low-mass cone and smooth, linear midrange response (unlike most 18's).
The general design principle is to use all drivers well within their linear regions, paying attention to drivers that are sensitive to excess excursion - compression drivers and ribbons in particular.
The alternative to the CD/ribbon scenario is a one or two-metre-high stretched-film ribbon, of the large-area BG type. I'm not really a fan of the line-radiator concept or sound, preferring point-source radiators.
dunno is it any novelty or not
http://www.goodsoundclub.com/Forums/ShowPost.aspx?PostID=3766
http://www.goodsoundclub.com/Forums/ShowPost.aspx?PostID=3766
Lynn Olson said:
Iinteresting.... I'm a big fan of the good old SS 8200. A terrific driver that gets the harmonic structure right. Really brings the music alive. As I understand it, this tweeter has been in the Scan Speak product line a long time.
Pity more people don't use it. Maybe they're afraid it won't have enough "air" or "sizzle". But its the bottom end of its range that makes it so attractive. Mates up very well with large cone woofers for a seamless sound. A great tweeter that is a true pleasure to listen too, and I listen every day.
Zen Mod said:
Well, I must confess I've been reading bits and pieces from Romy the Cat's site, since BudP mentioned his name as one of his East Coast clients.
The gotcha with reading anyone's subjective opinions, especially with horns, is that perception and musical preferences are very much individual. I'm not usually in agreement with the sonic tastes of most horn-o-philes, but then again, the mainstream stuff isn't for me either. So I usually take the subjective descriptions - from anyone - with a big grain of salt.
Not having heard Romy's system - and that's the only was I can assess anything - I don't know what he's aiming for sonically. Sonic preferences aside, though, Romy's got it right about the strong and weak points of compression drivers and ribbons. Ribbons get in trouble when asked to go too low, and compression drivers are not particularly wideband devices, functioning at their best over two or three octaves. The most common mistake in speaker design is taking drivers into high-distortion regions - they might measure OK, but a design that does that is going to sound strained and harsh with symphonic and choral material.
That's why I get annoyed with hifi demos featuring small jazz combos and minimally-miked blues performers. Musical preferences aside, demos like that emphasize a sense of pace and tone color, but don't stress the system with a dense, wideband spectrum, dynamic terracing, and hall sound. That's much harder to get right.
mige0 said:
To get things straight: I used the Raal 140-15D and the Mundorf AMT 2340 (the smallest one). Also I did not use the drivers in the same configuration (4th order LR on the Raal @2kHz. 2nd order electrical and 3rd order bessel on the AMT @ 4 kHz). At low XO frequencies the AMT just sounds strained, even metallic (reminded me of the cheap metal domes). The AMT was crossed to a Audiotechnology 15H and the Raal coupled to a AT 18H. Like I said completely different setups. However, I could compare both drivers as I had both in the same room on the same time.
Also, I'm just a diy enthusiast, I find it quite difficult to put the listening impressions on paper.
Both drivers are very neutral and transparant in presentation. I know a lot of people are saying this of their loudspeakers, but I mean really neutral. The ringradiator from Scanspeak, the Seas Millennium or the Morel Supreme 110 sound coloured in comparison (I used those drivers too).
However, the AMT sounded more clinical and less natural it sounded more like the ceramic tweeters, but less harsh. It seemed that the focus was on the details in the music, not on the music itself. Also the off axis response of the AMT dropped very early. The AMT somehow sounded to much hifi (very suitable for the current high end loudspeakers), but did not offer the natural presentation of the Raal.
On the other hand while the Raal does everything as good or better compared to every tweeter I've tried, it's main feature is that it can be used as low as the manufacturer is claiming, actually the loudspeaker is sounding better the more of the spectrum is covered by the Raal. It seems the balancing act of loudspeaker design, as mentioned by Lynn, is IMHO better accomplished in the Raal.
Measurement wise the Raal shows better horizontal dispersion when compared to the Mundorf. Both drivers have very low distortion (hence their very neutral presentation). Also the Mundorf I used has far lower efficiency, because it's a 4 ohm driver with ~90 dB W/m.
Ok, what do you think about this: BMS4592ND 2 way compresion driver, with 2" throat exit, 1.75" tweeter and 3.5" midrange diafragm.
More here: BMS coaxial
The idea is from here, until now I didn't even now of coaxial compresion drivers: DIYaudio thread
I think this could be a nice solution to all those talks on this thread and in the "Large midrange for OB" one, about 4-6" CD beeing less stressed in midrange than say 2" CD's but viceversa in the trebles.
Of course it's questionable if this 1.75" tweeter would better a RAAL ribbon... I'm quite sure not. But again.. this is dual concentric, which has some advantages, especialy with a 6Khz cross point where the offset of the drivers is more obvious.
Beeing a 2" throat, I guess we could find a 15" cone to mount on. The 18sound coaxes are just 1.4".
This kind of CD would offer a 3way dual concentric driver from 200Hz up to 20Khz... That's something...
I hope it makes any sense...
More here: BMS coaxial
The idea is from here, until now I didn't even now of coaxial compresion drivers: DIYaudio thread
I think this could be a nice solution to all those talks on this thread and in the "Large midrange for OB" one, about 4-6" CD beeing less stressed in midrange than say 2" CD's but viceversa in the trebles.
Of course it's questionable if this 1.75" tweeter would better a RAAL ribbon... I'm quite sure not. But again.. this is dual concentric, which has some advantages, especialy with a 6Khz cross point where the offset of the drivers is more obvious.
Beeing a 2" throat, I guess we could find a 15" cone to mount on. The 18sound coaxes are just 1.4".
This kind of CD would offer a 3way dual concentric driver from 200Hz up to 20Khz... That's something...
I hope it makes any sense...
Zen Mod said:
The RAAL ribbon website makes strong claims for their "zig-zag" cut thin metal front plate for tweeters which they call INOX, but the "water drop" tweeter ribbon in the GoodSoundClub uses a more traditional waveguide type front bezel. Any opinions on which is better technology for a ribbon?
Lynn has your design vison changed or just firmed up after 80 pages of discussion? Is a dipole 12" coaxial driver plus 12" midbass plus two 15" woofers still the design target?
Hi
@ LaMa thanks for your description, which gives a good picture about the character of these drivers to me!
I think about buying the Mundorf AMT myself but rather the 2530 or 2640 type – hence my strong interest in any subjective comparison. At least the possibility for dipol usage they offer is very appealing.
Greetings
Michael
@ LaMa thanks for your description, which gives a good picture about the character of these drivers to me!
I think about buying the Mundorf AMT myself but rather the 2530 or 2640 type – hence my strong interest in any subjective comparison. At least the possibility for dipol usage they offer is very appealing.
Greetings
Michael
The Story So Far
Well, I've gotten more detailed information on the coaxes. I'm sorry to report I haven't found any with FR as smooth as can be had with separate midbass drivers and HF systems. The difference isn't small, either - all of the coaxes appear to have 10 dB or worse peak-to-dip ratios in the 1 to 5 kHz region, both in the bass-driver and the integral horn responses. The combined curves conceal this, but the separate LF and HF curves reveal the truth.
I am not willing to design a 15-element passive crossover, resort to digital equalization, or simply ignore response deviations of this magnitude. In addition, when response deviations are especially severe, highly directional peaks start to appear, which do not lend themselves to any kind of equalization, since correcting a peak that appears in one location makes the response worse elsewhere, especially in the time domain.
That narrows things down to the two 18Sound 12" midbass drivers mentioned earlier, with either 1, 1.4, or 2" compression drivers and Tractrix, Le Cleac'h, or Oblate Spheroid horns, and/or ribbon drivers to either complement or replace the CD/horn HF drivers.
I really don't know what my tolerance levels for horn coloration will turn out to be. I'll probably audition the Radian vs BMS compression drivers. Don't think I'll track down any fifty-year-old Altec 515's or Vitavox S2's. That's not my style, no matter how good these rare, costly, and unobtanium items are reputed to sound.
By the time I start building the system a couple of months from now, the RAAL ribbon that's designed to go into studio monitors (100 dB/metre, 2 kHz crossover, real-world power-handling) may be on the market, providing an alternative to the CD/horn route.
As for the bass, that's not so critical. At the minimum, I see a single 15" 18Sound driver selected for low distortion and low coloration, and probably powered with its own amp, equalization, and active crossover - possibly by Rane prosound. A pair of 15's offers more opportunity for subtle equalization between the floor and off-the-floor mounted drivers, as well as a welcome decrease in distortion and increase in headroom.
I'll be following GregOH1's project with great interest - he's made some sensible, real-world choices, and I'm very curious how his project turns out.
LineSource said:
Well, I've gotten more detailed information on the coaxes. I'm sorry to report I haven't found any with FR as smooth as can be had with separate midbass drivers and HF systems. The difference isn't small, either - all of the coaxes appear to have 10 dB or worse peak-to-dip ratios in the 1 to 5 kHz region, both in the bass-driver and the integral horn responses. The combined curves conceal this, but the separate LF and HF curves reveal the truth.
I am not willing to design a 15-element passive crossover, resort to digital equalization, or simply ignore response deviations of this magnitude. In addition, when response deviations are especially severe, highly directional peaks start to appear, which do not lend themselves to any kind of equalization, since correcting a peak that appears in one location makes the response worse elsewhere, especially in the time domain.
That narrows things down to the two 18Sound 12" midbass drivers mentioned earlier, with either 1, 1.4, or 2" compression drivers and Tractrix, Le Cleac'h, or Oblate Spheroid horns, and/or ribbon drivers to either complement or replace the CD/horn HF drivers.
I really don't know what my tolerance levels for horn coloration will turn out to be. I'll probably audition the Radian vs BMS compression drivers. Don't think I'll track down any fifty-year-old Altec 515's or Vitavox S2's.
That's not my style, no matter how good these rare, costly, and unobtanium items are reputed to sound.By the time I start building the system a couple of months from now, the RAAL ribbon that's designed to go into studio monitors (100 dB/metre, 2 kHz crossover, real-world power-handling) may be on the market, providing an alternative to the CD/horn route.
As for the bass, that's not so critical. At the minimum, I see a single 15" 18Sound driver selected for low distortion and low coloration, and probably powered with its own amp, equalization, and active crossover - possibly by Rane prosound. A pair of 15's offers more opportunity for subtle equalization between the floor and off-the-floor mounted drivers, as well as a welcome decrease in distortion and increase in headroom.
I'll be following GregOH1's project with great interest - he's made some sensible, real-world choices, and I'm very curious how his project turns out.
There is Simulation and there is Reality
Hi
to get an idea about the accuracy of the simulation with the EDGE software I made a compact test baffle.
The baffle is 0.30 m x 0.46 m
The driver is a pretty old but well known, restored and modified Dynaudio 21W54
Specs can be found under
http://www.gattiweb.com/images/dynaudio/21w54_data.pdf
" http://www.gattiweb.com/images/dynaudio/21w54_data.pdf")
The measured FR (smoothed and unsmoothed) at zero distance without any EQ applied can be seen below. Also the CSD and Sonogram, which is basically the top view of the CSD with the SPL coded in colors.
In the sonogram the rather strange effect of a resonance following a dip can be observed quite clear. The decay of that 750 Hz resonance is not linear but more in steps.
The simulation with EDGE on axis of the woofer gives following results:
The real equalizing was performed with the help of the DCX 2496
low shelf 1 = +12dB at 30 Hz 6dB
low shelf 2 = +15dB at 49 Hz 6dB
Notch 1 Q = 1.6 -6dB at 544Hz
Notch 2 Q = 1.6 -8dB at 3530Hz
Please note that I made this setup for measurement purpose only. It is optimised for a flat FR at 1m.
Measuring at a distance of zero 0.25m 0.5m and 1m with EQ applied gives the following results:
What can be seen here is that that kind of simulation gives sort of idea about what's going on and is fairly close to the measured baffle effect at 550Hz to predict an appropriate EQing.
Greetings
Michael
Hi
to get an idea about the accuracy of the simulation with the EDGE software I made a compact test baffle.
The baffle is 0.30 m x 0.46 m
The driver is a pretty old but well known, restored and modified Dynaudio 21W54
Specs can be found under
http://www.gattiweb.com/images/dynaudio/21w54_data.pdf
" http://www.gattiweb.com/images/dynaudio/21w54_data.pdf")
An externally hosted image should be here but it was not working when we last tested it.
The measured FR (smoothed and unsmoothed) at zero distance without any EQ applied can be seen below. Also the CSD and Sonogram, which is basically the top view of the CSD with the SPL coded in colors.
In the sonogram the rather strange effect of a resonance following a dip can be observed quite clear. The decay of that 750 Hz resonance is not linear but more in steps.
An externally hosted image should be here but it was not working when we last tested it.
An externally hosted image should be here but it was not working when we last tested it.
An externally hosted image should be here but it was not working when we last tested it.
An externally hosted image should be here but it was not working when we last tested it.
The simulation with EDGE on axis of the woofer gives following results:
An externally hosted image should be here but it was not working when we last tested it.
The real equalizing was performed with the help of the DCX 2496
low shelf 1 = +12dB at 30 Hz 6dB
low shelf 2 = +15dB at 49 Hz 6dB
Notch 1 Q = 1.6 -6dB at 544Hz
Notch 2 Q = 1.6 -8dB at 3530Hz
Please note that I made this setup for measurement purpose only. It is optimised for a flat FR at 1m.
An externally hosted image should be here but it was not working when we last tested it.
An externally hosted image should be here but it was not working when we last tested it.
Measuring at a distance of zero 0.25m 0.5m and 1m with EQ applied gives the following results:
An externally hosted image should be here but it was not working when we last tested it.
What can be seen here is that that kind of simulation gives sort of idea about what's going on and is fairly close to the measured baffle effect at 550Hz to predict an appropriate EQing.
Greetings
Michael
Hi Michael,
That is excellent work, but from a steady sine POV only!
Baffle edge responses take time to develop, ie. steady sine responses become modified wrt initial waveform and excitation.
Thus calculated equalisation 'compensation' might provide a 'smooth' steady sine response, but it will affect the reproduction we listen to because all music and transients will be affected by the series 'compensation' filters inserted in line.
Note how the initial dip in the waterfall response is due to the 'storage' of energy which causes the subsequent resonant peak. If the sinusoidal drive for that subsequent peak is flattened to suit the baffle response, what will happen to the initial transient response ?
Answer - driver reproduction will not sound right.
My view of programs like the Edge is to use them to view the effect of these resonances etc, so that they can minimised at the baffle design stage with driver offsets and non-rectilinear shape.
I feel that distorting the electrical waveform in time, to match a baffle response in time can only cause notable amplitude distortion to music and transients.
We need the main driver and baffle to be physically better at outset, rather than attempt to try for electrical compensation afterwards.
Cheers .......... Graham.
That is excellent work, but from a steady sine POV only!
Baffle edge responses take time to develop, ie. steady sine responses become modified wrt initial waveform and excitation.
Thus calculated equalisation 'compensation' might provide a 'smooth' steady sine response, but it will affect the reproduction we listen to because all music and transients will be affected by the series 'compensation' filters inserted in line.
Note how the initial dip in the waterfall response is due to the 'storage' of energy which causes the subsequent resonant peak. If the sinusoidal drive for that subsequent peak is flattened to suit the baffle response, what will happen to the initial transient response ?
Answer - driver reproduction will not sound right.
My view of programs like the Edge is to use them to view the effect of these resonances etc, so that they can minimised at the baffle design stage with driver offsets and non-rectilinear shape.
I feel that distorting the electrical waveform in time, to match a baffle response in time can only cause notable amplitude distortion to music and transients.
We need the main driver and baffle to be physically better at outset, rather than attempt to try for electrical compensation afterwards.
Cheers .......... Graham.
Thanks Mige0, that's a nice set of measurements, and especially welcome to see direct comparison between real-world and the Edge simulation, and with a driver that's well-known to all of us.
When I realized that Edge has no floor reflection, I started using it with mirror-imaged baffles mounted with the full set of drivers. Then I started to get a little suspicious of distance effects, and starting measuring everything at 3, 12, and 100 meters, so I could see the plane wave developing, and not get lost in the clutter of near-field effects.
The optimum shape that evolved looked like the rudder of a jet airplane, with the widerange driver mounted in the upper corner towards the "leading edge" (if it were on an airplane). I haven't posted this shape, because I'm not really certain if Edge is very reliable.
One effect that was surprising was seeing all of the energy moving towards the floor and reflecting across the mirror-image, and then diffracting off the lower surface of the baffle on the return trip. As I moved the "trailing" surface of the rudder to prevent more and more energy from reflecting off the floor (the rudder was raked back more steeply), you could see Edge reflecting less energy as the floor reflection became more and more shaded by the "trailing" edge.
Looking at this made me realize that thick felt at the bottom of the frame (in the back) would be a good idea, to prevent this reflection from bouncing off the floor and re-radiating off the baffle edge. Treating the entire rear surface of the baffle with thick wool felt is also probably a good idea as well, to minimize the energy diffracting off the edges.
When I realized that Edge has no floor reflection, I started using it with mirror-imaged baffles mounted with the full set of drivers. Then I started to get a little suspicious of distance effects, and starting measuring everything at 3, 12, and 100 meters, so I could see the plane wave developing, and not get lost in the clutter of near-field effects.
The optimum shape that evolved looked like the rudder of a jet airplane, with the widerange driver mounted in the upper corner towards the "leading edge" (if it were on an airplane). I haven't posted this shape, because I'm not really certain if Edge is very reliable.
One effect that was surprising was seeing all of the energy moving towards the floor and reflecting across the mirror-image, and then diffracting off the lower surface of the baffle on the return trip. As I moved the "trailing" surface of the rudder to prevent more and more energy from reflecting off the floor (the rudder was raked back more steeply), you could see Edge reflecting less energy as the floor reflection became more and more shaded by the "trailing" edge.
Looking at this made me realize that thick felt at the bottom of the frame (in the back) would be a good idea, to prevent this reflection from bouncing off the floor and re-radiating off the baffle edge. Treating the entire rear surface of the baffle with thick wool felt is also probably a good idea as well, to minimize the energy diffracting off the edges.
Interesting, though I never got that kind of agreement with reality from the Edge. When I simulated my baffles with moderate foldback (workaround=simulated as if flat) it didn't work for me. It wasn't cavity resonance either, but the added complexity of the acoustic path I suppose.
Graham,
Graham,
Actually it won't affect transient response much IMO because a transient by definition contains large HF elements. Those won't be affected by any EQ in the LF. So, the transient's perception through precendence effect of its HF compnents will remain unaffected.