This is to mate with a sub bass horn for music and home theatre. My goal is to achieve a peak SPL of 110 db @ 3m for the mains and 120 db @ 3m for the sub. The sub can easily do 130 db at this distance, so I have quite a lot of headroom. This means 120 db for the mains at 1m, no mean feat!
The sound I'm looking for is dynamic, open, transparent. I want to reach live sound levels with a sense of ease. I prefer a relaxed sound, not too detailed, as I've found drivers such as Seas Excell are brilliant with well recorded material, but brutal with most pop and rock music.
I intend to do an active 3 way, probably with something like this:
W W MTM W W
1. Tweeter
This is what I find most difficult to choose. A CD in waveguide would do it, but I'm not sure I can pull it off with the same kind of sound I get from a good hifi dome. Essentially I want that sound on a larger scale.
Suggestions anyone?
2. Mid
I'd prefer a 6.5" or 8" driver like the Focal 7k from 300 - 2.4k. Looks are an issue here. I'd like to avoid something that looks like a typical pro driver. I like the look of the B&C PS21 (similar to Adire Extremis with inverted dustcap), but it looks like it has some uncontrolled breakup before the intended crossover region.
I'd like to find a good driver that is cheaper than the Focal but does the same job. Suggestions?
3. Woofer
Looking for a 12" driver with low distortion, reasonable excursion (7mm), decent power handling and a smooth extended response allowing a high crossover point. Currently the Peerless SLS seems the most cost effective driver I can find. Any other suggestions anyone?
It's more critical that the mid and tweeter reach the SPL target, but with the woofers, I can add on more later.
Why choose OB for a high SPL application? I have found from my prototype with existing drivers, that not that much eq is required. In fact, based on this I expect to be able to get down to 80 Hz easily without eq boost. For serious listening, I may lower the xo point to the horn sub to 40 Hz.
The sound I'm looking for is dynamic, open, transparent. I want to reach live sound levels with a sense of ease. I prefer a relaxed sound, not too detailed, as I've found drivers such as Seas Excell are brilliant with well recorded material, but brutal with most pop and rock music.
I intend to do an active 3 way, probably with something like this:
W W MTM W W
1. Tweeter
This is what I find most difficult to choose. A CD in waveguide would do it, but I'm not sure I can pull it off with the same kind of sound I get from a good hifi dome. Essentially I want that sound on a larger scale.
Suggestions anyone?
2. Mid
I'd prefer a 6.5" or 8" driver like the Focal 7k from 300 - 2.4k. Looks are an issue here. I'd like to avoid something that looks like a typical pro driver. I like the look of the B&C PS21 (similar to Adire Extremis with inverted dustcap), but it looks like it has some uncontrolled breakup before the intended crossover region.
I'd like to find a good driver that is cheaper than the Focal but does the same job. Suggestions?
3. Woofer
Looking for a 12" driver with low distortion, reasonable excursion (7mm), decent power handling and a smooth extended response allowing a high crossover point. Currently the Peerless SLS seems the most cost effective driver I can find. Any other suggestions anyone?
It's more critical that the mid and tweeter reach the SPL target, but with the woofers, I can add on more later.
Why choose OB for a high SPL application? I have found from my prototype with existing drivers, that not that much eq is required. In fact, based on this I expect to be able to get down to 80 Hz easily without eq boost. For serious listening, I may lower the xo point to the horn sub to 40 Hz.
Hi Paul
Have a look at Magnetar's OB speakers. I think that they would be what you are looking for. Here are two threads in which he discussed his system:
WMTMW Open Baffles
Big open baffle project
Enjoy,
Deon
Have a look at Magnetar's OB speakers. I think that they would be what you are looking for. Here are two threads in which he discussed his system:
WMTMW Open Baffles
Big open baffle project
Enjoy,
Deon
paulspencer said:
Suggestions:
IMO you are thinking to conventionally for this requirement.
Think either a conventional driver line array.. OR,
..your point source compression driver/high eff. cone setup - which likely will not be MTM'esq (strictly speaking).
Since enough has been said on the line array (in other posts), I'll focus on the "pro-sound" design.
#1 - Getting "Hi-Fi" sound from a pro setup requires broad horizontal dispersion (for a horn) up to about 7 kHz (..and the higher the better). Most 1 inch dome tweeters have about 160 degrees of dispersion up to 4 kHz. Furthermore the spl off-axis isn't that strong for these drivers up to about 7 kHz.
Now then, you just won't find a good horn that does this, BUT anything over about 100 degrees should be decent. So here is where you start your system design. Not with a driver, but rather with a horn!
IMO there is an excellent maker of horns at what is a relativly modest price from DDS. What you want to look at here is the polar plot and how "constant" the dispersion pattern is higher in freq.. Note that the smaller "circles" on the polar plot indicate the very highest freq.s for the spec'ed passband. Again, we are really concerned with HORIZONTAL dispersion, NOT vertical.
http://www.ddshorns.com/catalog.php?page=products
So then going through their list picks up these with greater than 100 degrees horizontal coverage:
http://www.ddshorns.com/catalog.php?page=CFD1110Pro
http://www.ddshorns.com/catalog.php?page=CFD14110Pro
http://www.ddshorns.com/catalog.php?page=CFD15110Pro
http://www.ddshorns.com/catalog.php?page=CFD2110Pro
Notice that NONE are really an actual 110 degrees of dispersion constant. They are ALL down about 6 db at even their lowest freq.s at the 110 marker. (..thats OK, functionally that gets us just about 100 degrees).
The next step then is deciding what type of Compression driver to use AND asking yourself how low/steep are you going to use the driver and how high do you want the driver to go (..bearing in mind the increasing loss of sound pressure off-axis at higher freq.s)? (..the questions are inter-related.)
For instance IF you go lower in freq. (often less than 1 kHz) then you'll likely want a 2 inch driver. IF you go higher in freq. then it will prob. be 1 inch driver. A 1.5 inch driver basically provides you with something "in-between".
Generally:
The lower in freq. you use a compression driver - the cleaner and more detailed the sound (provided you use an appropriate/steep filter).. Additionally, the lower in freq. you use a smaller diameter compression driver - the greater the distortion. The higher in freq. the driver operates the greater the dispersion pattern and the more stable the imaging AND the greater the apparent subjective sense of "air".
To add to this consider tonal balance and diaphram/suspension material, AND driver "break-in".
IF a driver has a considerably greater loss in spl off-axis at higher freq.s - THEN the sound will be a little "duller" ***UNLESS*** you listen to the thing O N - A X I S. IF you listen ON-AXIS then the thing will likely be painfully bright. To compensate you can either "to-in" or "toe-out" the driver when listening (and each has their own advantages and disadvantages).
Driver material-wise.. Titanium is genererally the "brightest" sounding material, but it can often sound a little "closed-in" due to the transition between diaphram and surround. At the other end of the spectrum is coated paper/phenolic which can sound a little "dull". (..the most costly - Be, can be more detailed than any other diaphram material, but usually has less "impact" at lower freq.s due to its reduced mass.) Aluminum and Mylar seem to provide a middle-ground. ANY type will likely sound "closed-in" without a *SERIOUS* amount of "break-in" time. ANY suspension on a compression driver needs to be "fatigued" to loosen them up and achieve a more spacious sound. Then you can compound the issue with magnetic field integrity, (basically hysteris), of the magnet and its material. The best is typically provided from a good field coil, the worst from a ceramic ferrite. Next to a field coil would be Alinco, then Neo (usually). I think this has more to do with processing and crystal structure on these permanant magnets more than the material itself (..but its a guess). Additionally, the higher the field strength typically the cleaner the sound.
On last thing to comment on with driver selection is the use of an additional driver as a super tweeter. Here you can extend the off-axis disperion by utlizing a smaller driver. The thing you need to be carefull about here is the off-axis dispersion at the crossover for the compression driver and how much power vs. distortion the super tweeter can handle (..i.e. the lower you crossover the super tweeter the more trouble it gets "in").
OK..
With all of that my suggestion is the following bearing in mind all of the above factors AND cost:
1. Go as low as you can with the horn: More detail = better sound and this isn't a case of more "fatiguing" sound. To me then the horn to select is clear =
http://www.ddshorns.com/catalog.php?page=CFD2110Pro
2. Now you need a good 2 inch driver for it that doesn't cost a fortune (..though you don't want to "skimp" here either). I want to stay away from titanium - to bright for your tastes. You also want it to have good off-axis performance to higher freq.s. Additionally, you'll want a fairly flat impeadance curve. My suggestion is:
http://www.bmspro.info/index.php?show=item&usbid=10281&id=54367
Mylar diaphram, decent impeadance, low "fs", decent dispersion, strong magnet, reasonable cost. (and Brett here has commented on them favorably.) You will of course need LOTs of "break-in" for them.
Note that there are other possibilites here - most notably "vintage" drivers. But I chose to stay with current production drivers for the selection.
Now you COULD go with the Co-axial versions for your super tweeter bandwidth.. BUT it won't really "gel" with the dispersion pattern and will have relativly high losses off-axis - so I wouldn't recomend it for this wide-dispersion design. So that brings us to the super tweeter.
For the super tweeter I want something that pretty well matches the dispersion of the Horn/Mid at the crossover freq. and to at least 11 kHz. I also want something that goes to about 4.5 kHz with a LR 4th order crossover. Current production there is one clear choice: The Beyma CP25.
http://profesional.beyma.com/ENGLISH/producto.php
***** There is one aspect of this design that should be clear - notice that the 100 degree horizontal dispersion pattern is VERY similar to the frontal "lobe" of a dipole. I.E. this design will come close to a *true* constant horizontal directivity for most of the operating passband, unlike virtually every loudspeaker *****
http://www.linkwitzlab.com/rooms.htm
I may continue with this latter for suggestions on the dipole mid/bass drivers.
Scott,
I thought you were a big fan of the 1.4 or was it 1.5 kHz crossover point? Or do I have you confused with something someone else posted, maybe Thylanter, back in the "percieve" thread.... Anyways, just got back from RMAF in Denver and I was really impressed with a couple of systems using the DDS ENG-1 90 waveguide, and the best in my opinion was one using the 1 inch BMS 4552 or something like that.
Paul,
I would look real closely at the combination above, shelved down just a bit up to around 12-16k and then left alone so as to get as much top end out of it as possible and bring the upper-mid / lower tweeter response down closer to a typical cone SPL. Not much choice in my opinion for a single cone that will give explosive dynamics in an open baffle without looking at the pro drivers. The implementation I heard that was very dynamic and yet still smooth and imaged like crazy was from Emerald Physics, and his system coupled the above CD/horn with a pair of Ciare 8" per side, and then had a separate bass panel system with Eminence 15s. If I was gong with a single cone mid below the waveguide, I would look at the Volt BM2500.4 and see how it would hold up on an open baffle run down to 200Hz, and then cross to a subwoofer system, and yes, the Peerless XXLS still seem like the best way to go from all reports, especially if you want to avoid aluminum drivers.
Personally, I am thinking about a computer based crossover and DRC so I am looking at the same kind of set up you are thinking about but will probably cross from whatever mid I end up with to a sealed sub between 100 and 200 hz. After listening to around 5 or 6 open baffle mid drivers (Rick Craig at Selah Audio was showing a great new system with Accuton mid on open baffle and what I think was a sealed sub below and ribbon tweet above) I am convinced that even if I can control room nodes reasonably well with DSP rather than brute force multiple drivers in a dipole sub configuration, I still want that dipole midrange sound, it just sounds so right to me I don't think I will ever be able to put a mid or midbase driver in a box (unless it was a really good dome, and they all seem to be self-contained anyways so the box is just there to hold em in place and give them a baffle as needed..).
I thought you were a big fan of the 1.4 or was it 1.5 kHz crossover point? Or do I have you confused with something someone else posted, maybe Thylanter, back in the "percieve" thread.... Anyways, just got back from RMAF in Denver and I was really impressed with a couple of systems using the DDS ENG-1 90 waveguide, and the best in my opinion was one using the 1 inch BMS 4552 or something like that.
Paul,
I would look real closely at the combination above, shelved down just a bit up to around 12-16k and then left alone so as to get as much top end out of it as possible and bring the upper-mid / lower tweeter response down closer to a typical cone SPL. Not much choice in my opinion for a single cone that will give explosive dynamics in an open baffle without looking at the pro drivers. The implementation I heard that was very dynamic and yet still smooth and imaged like crazy was from Emerald Physics, and his system coupled the above CD/horn with a pair of Ciare 8" per side, and then had a separate bass panel system with Eminence 15s. If I was gong with a single cone mid below the waveguide, I would look at the Volt BM2500.4 and see how it would hold up on an open baffle run down to 200Hz, and then cross to a subwoofer system, and yes, the Peerless XXLS still seem like the best way to go from all reports, especially if you want to avoid aluminum drivers.
Personally, I am thinking about a computer based crossover and DRC so I am looking at the same kind of set up you are thinking about but will probably cross from whatever mid I end up with to a sealed sub between 100 and 200 hz. After listening to around 5 or 6 open baffle mid drivers (Rick Craig at Selah Audio was showing a great new system with Accuton mid on open baffle and what I think was a sealed sub below and ribbon tweet above) I am convinced that even if I can control room nodes reasonably well with DSP rather than brute force multiple drivers in a dipole sub configuration, I still want that dipole midrange sound, it just sounds so right to me I don't think I will ever be able to put a mid or midbase driver in a box (unless it was a really good dome, and they all seem to be self-contained anyways so the box is just there to hold em in place and give them a baffle as needed..).
Greggo said:
It all depends on the application.
Thy was really big on that, BUT so am I in most circumstances - the thing thats really different is the use of a horn, and that calls for re-thinking of the design.
Its funny you should mention liking the 1 inch BMS driver in particular.. One of the distributors for the US (who has heard all BMS compression drivers), ALSO liked that one the best overall - but it IS more limited in its application. Note that it is the 4540ND, and costs a reasonable $100 US. The waveguide also only costs $70 US.
Notice however that the waveguide is not a constant directivity design:
http://www.ddshorns.com/catalog.php?page=ENG190Pro
..and that poses some serious limitations with respect to imaging and loudspeaker placement.. and also tonal balance.
A better design (at least to 4 kHz) would be the Supravox Pavillion 1:
http://www.supravox.fr/anglais/haut_parleurs/pav_1_pouce.htm
It goes a little lower in freq. and has a MUCH wider dispersion character. THAT and the BMS 4540ND might work very well with a normal sealed or bass reflex midbass.. AND a super tweeter like the Fostex FT96H:
http://www.madisound.com/cgi-bin/index.cgi?cart_id=4012702.8199&pid=349
(the only problem there is that the fostex doesn't go *quite* as low as needed for the design for full dispersion character - even with a good LR4th order crossover).
With the DDS waveguide or other horn/waveguides you get a very difficult if not impossible job in combining them to open baffle speakers due to the totally different in room power response. Tried lots of them with my open baffle speakers and only the very undeep shallow waveguides seem to match. I'm experimenting at the moment with the BMS 4540nd drivers in different waveguides. For better power response and great match to open baffle sound you can also have a look at the Mundorf AMT tweeters.
Compression drivers can sound as good and most of times even better than lot's of dome tweeters out there and at the levels you want to play with them they will surely sound a lot better.
Compression drivers can sound as good and most of times even better than lot's of dome tweeters out there and at the levels you want to play with them they will surely sound a lot better.
Sjef said:
A Waveguide and a Horn are not the same with regard to dispersion.
What I described above *should* be comparable in the forward plane (give or take a db at 45 degrees off-axis). This would NOT be the case with waveguide. (..and note this same problem has a corelary when combining front loaded horns to normal omnipolar midbass designs - i.e. you'll find LOTs of people with a very similar response to yours for a rather different design.) Its also why I mentioned the waveguide is more suited to an omnipolar midbass design.
Now there will be a deviation in power response due to the reverse phase (and additional room reflections) of the dipole driver. All thats required then is some attenuation.
Another thing of concern is that the dipole driver is in fact operating as a dipole at the crossover freq. so that the forward plane is a *true* CD design. This can be modeled effectivly in the Edge (freeware):
http://www.tolvan.com/edge/
..and note that while I really didn't go "into it" - there are several reasons for having a lower crossover for the horn here that are related to these issues.
Hi Paul,
I like the way you specify your design goals. Starting with what you are looking to achieve and then selecting the most appropriate technologies and techniques to achieve it.
However, I wonder if some of your assumptions are correct. 110dB peak SPL at the listening position isn't as difficult to achieve as you might be thinking. Remember that you have 2 speakers contributing to the sound field, you are listening within a reverberant environment, music has a high crest factor and also that the spectra of recorded music reduces amplitude as frequency increases.
110dB peak at the listening position of 3m is equal to 120dB peak at 1m in an anechoic environment. For a typically semi-reverberant listening room, 3m approximately corresponds to the critical distance where the reverberant sound field makes a roughly equal contribution to the direct sound. There's more detail about this on my web page:
http://www.aeronet.com.au/Acoustics.htm
So 120dB peak drops to 117dB peak.
2 speakers further drops the requirement to 111dB peak.
Music typically has a crest factor (ratio of peak to RMS) of 6dB or more. It depends on the kind of music of course. Heavily compressed music might only have a crest factor of 2-3dB, whilst classical music might achieve 20dB or more, but for typical contemporary music 6dB is a good figure to work with. Average SPL now drops to 105dB per speaker.
Considering that the spectra of contemporary music falls at approximately 2dB/oct above 100Hz means that the mid needs to output a little less than the broadband average, the tweeter less still, and the woofer proportionally more. Of course the crossover frequencies play an important part in determining how the energy is split between the drivers and the exact type of music will also be a factor. Assuming crossover frequencies of about 300Hz & 2kHz will mean that the woofer will need to produce something like 106dB average SPL, the mid about 104dB SPL and the tweeter about 98dB SPL.
So now the problem is a more manageable. 98dB is easily within the reach of a 1" dome tweeter. And the midrange and woofer only need have moderate efficiency. Of course in the above analysis I have assumed a 6dB crest factor which implies the amplifiers be capable of 6dB more than that required for the maximum average SPL.
Cheers, Ralph
I like the way you specify your design goals. Starting with what you are looking to achieve and then selecting the most appropriate technologies and techniques to achieve it.
However, I wonder if some of your assumptions are correct. 110dB peak SPL at the listening position isn't as difficult to achieve as you might be thinking. Remember that you have 2 speakers contributing to the sound field, you are listening within a reverberant environment, music has a high crest factor and also that the spectra of recorded music reduces amplitude as frequency increases.
110dB peak at the listening position of 3m is equal to 120dB peak at 1m in an anechoic environment. For a typically semi-reverberant listening room, 3m approximately corresponds to the critical distance where the reverberant sound field makes a roughly equal contribution to the direct sound. There's more detail about this on my web page:
http://www.aeronet.com.au/Acoustics.htm
So 120dB peak drops to 117dB peak.
2 speakers further drops the requirement to 111dB peak.
Music typically has a crest factor (ratio of peak to RMS) of 6dB or more. It depends on the kind of music of course. Heavily compressed music might only have a crest factor of 2-3dB, whilst classical music might achieve 20dB or more, but for typical contemporary music 6dB is a good figure to work with. Average SPL now drops to 105dB per speaker.
Considering that the spectra of contemporary music falls at approximately 2dB/oct above 100Hz means that the mid needs to output a little less than the broadband average, the tweeter less still, and the woofer proportionally more. Of course the crossover frequencies play an important part in determining how the energy is split between the drivers and the exact type of music will also be a factor. Assuming crossover frequencies of about 300Hz & 2kHz will mean that the woofer will need to produce something like 106dB average SPL, the mid about 104dB SPL and the tweeter about 98dB SPL.
So now the problem is a more manageable. 98dB is easily within the reach of a 1" dome tweeter. And the midrange and woofer only need have moderate efficiency. Of course in the above analysis I have assumed a 6dB crest factor which implies the amplifiers be capable of 6dB more than that required for the maximum average SPL.
Cheers, Ralph
RMAF hi-jack, then back to scheduled programming!
Well guys, I think the reasoning expressed in support of your suggestions are all very sound (no pun intended), but I have to wonder about a couple of things that have been well supported in other threads (just going on memory here, nothing specific) that, along with my recent listening experiences, have me leaving the door open on waveguides, and specifically the DDS waveguide.
1) Do we really want consistent horizontal dispersion from top to bottom? Or do our listening rooms, on average, perform better when we have a smooth transition in the midrange to upper mid that narrows things slightly which may lead to more stable imaging? Especially when the drivers cross in front of the listener, possibly making for a larger sweetspot to be shared with others. Much like the reinforcement of lower frequencies due to "room gain" must be considered, should we also consider the growing audibility of reflections as they rise in frequency. This is, IMHO, fairly well in line with the designs of Ted Jordan, and also in line with the others who are building speakers/kits with pro sound drivers and horns, looking at horns with roughly 90 degree covereage and then using 10 or 12 inch cone drivers that start to narrow in dispersion due to beaming, matching the 90 degree point of the cone with the cross-over point as much as possible. In the speakers from both Emerald Physics and Audio Kinesis (Duke, who posts on AA once in a while...), I think that is exactly what the designers did and the results sounded very fine to me. And I am spoiled with the normal Hi-Fi smoothness and 3D image more than the lively dynamics of those pro-sound converts among us (Class A SS and AudioPhysic Tempo IIs are my current rig, slowly to be replaced by DIY efforts over the next year or two...), so rightly or wrongly, I consider myself a tough test for anything other than the well established cones/domes of hi end home audio.
2) If you are focused on the home hi-fi environment, then aren't waveguides versus horns a very reasonable test that everyone should perform for themselves? I think Earle Geddes is defiinitely on to something, and I think the DDS product is a nice compromise between the extremes on waveguide design and implmentation. One of the primary reasons that I invested in a trip to the Rocky Mountain Audio Fest last week was to hear all the options for mid-tweet design, cones/domes, cones/ribbons, cones/waveguides, cones/horns, horns/horns, etc.... (as well as line arrays, but hey, let's not go there!) So having done so, I should state that my own bias here has definitely drifted away from horns and towards ribbons and wageguides (and I have really been reading a lot about horns lately, and wanted to like them, but didn't). I listened carefully to all the horn designs at the show and for some reason, as much as I wanted to get into them, none of them impressed me except for Bruce Edgars Titans, and even then I still liked the above aforementioned a little better and also some killer alternatives such as the Merlins and the Ebens, not to mention Wilson Beneshe (spelling?) who also made great sound at the show, while I am blabbering on the topic, let me also say that SoundSmith, AudioCraftsmen/FacetAudio, BlueCircle/SelahAudio, and Stello/FJ (all amps/spkrs) fill out my own list of best sound of the show (I disqualify the Wilson Max2s, MBLs, and other stratosphere systems because they just seem stupidly self-indulgent to me and makes little more if any appreciable music than many other less impressive efforts). And Shin, if you are reading this I think you would have the same reaction to the Manger that I had, which is I loved it, and I think "overly polite" is no longer a fair criticism with the new version, and I am very very very tempted to just build my DIY reference system around that driver, but at the end of the day, for the few times that I want to push the limits SPL wise, I don't think they will take me where I want to go. If I was a sound engineer doing a lot of jazz/classical work, they would be my nearfield monitor, no doubt about it. (OK, OK, end of rant on RMAF...)
3) I think a nice, dynamic, pro sound 10 inch cone, designed for mid-range duty, could be a good candidate if the response is smooth enough for mid-field home listening, as the natural move towards some beaming would blend well at a low crossover point of 1200-1500 hz, providing a very smooth, albeit changing, power response that narrows slightly as one moves through the mid band, which I am still open to believing may be a good thing rather than a bad thing.
4) I heard a few systems that brought in super tweeters well above 5khz, didn't like any of them (and I am convinced lobing was at least part of the problem), so I am now personally convinced that you either cross right below or right above the areas that most seem to agree are most sensitive to our ears, and that would be, again my own bias here on the exact numbers, 1.8-3.8 khz. Every time I hear something crossed just at or below, or just at or above (but not too far in either direction) these two boundaries they tend to sound better than competing designs that don't, with the Merlins being a good exception (if you spend nearly a decade perfecting something and are as passionate about music as Bobby Palkovic, you will probably find a way to make your 2.2khz crossover point just about invisible).
5) I will admit, and this is why despite my long rant here, I am very open and interested in the other design advice in this thread, that I am very skeptical of cross-overs in the 200-800 Hz range because this is supposedly (cant' remember the reference, but it was a very credible one, and I am convinced it explains why I have not been impressed with a bunch of different 3-way designs) a range that many are very sensitive to phase, and transient time/phase perfection is a very tough thing to achieve, so I stick to my limited knitting for now, but admit to being curious to hear some speakers designed to cross in that range, especially things like a TAD horn from 500hz up and TAD Woofer handling everything below, and then all kinds of variations of that theme employing sub woofers and super tweeters as well, and specifically what Scott has recommended in his posts above. I think I want to do exactly what I have already posted about, but I would hate to do it without hearing or better yet experiementing myself with some of the things that Scott has suggested here...
6) Though I am open to the "pros" expressed in posts above, not sure I buy into all the "cons", as the wave-guide to dipole system I listened to gave my audiophysics a run for the money in soundstage, smoothness, and fatigue-free listening, and then surely blew them out of the water when it came to micro/macro dynamics, pace/timing, realistic timbres and overall liveliness of the sound.
7) Personally, I wish I had the time/money to try everything mentioned in this thread so far...
8) And finally, Paul, I do agree with another poster that your SPL goal might be overkill, as would be 4 x 12 inch woofers (if you are crossing to a sub from there..). I think you would get all the SPL you would ever want out of a good CD tweeter/mid and a really solid 8-10inch cone (if not pushed above 1500hz due to thermal limits, and much below 150hz due to mechanical limits), and then need only a pair of 10, 12, or 15 inch woofers, especially if you are crossing over to a bass horn sub, or any kind of sub at 40hz or higher. I would bet that there are a few 10s and a lot of 12s that would give you everything you would ever want down to 60hz and everything you would really need down to 40hz.
Well guys, I think the reasoning expressed in support of your suggestions are all very sound (no pun intended), but I have to wonder about a couple of things that have been well supported in other threads (just going on memory here, nothing specific) that, along with my recent listening experiences, have me leaving the door open on waveguides, and specifically the DDS waveguide.
1) Do we really want consistent horizontal dispersion from top to bottom? Or do our listening rooms, on average, perform better when we have a smooth transition in the midrange to upper mid that narrows things slightly which may lead to more stable imaging? Especially when the drivers cross in front of the listener, possibly making for a larger sweetspot to be shared with others. Much like the reinforcement of lower frequencies due to "room gain" must be considered, should we also consider the growing audibility of reflections as they rise in frequency. This is, IMHO, fairly well in line with the designs of Ted Jordan, and also in line with the others who are building speakers/kits with pro sound drivers and horns, looking at horns with roughly 90 degree covereage and then using 10 or 12 inch cone drivers that start to narrow in dispersion due to beaming, matching the 90 degree point of the cone with the cross-over point as much as possible. In the speakers from both Emerald Physics and Audio Kinesis (Duke, who posts on AA once in a while...), I think that is exactly what the designers did and the results sounded very fine to me. And I am spoiled with the normal Hi-Fi smoothness and 3D image more than the lively dynamics of those pro-sound converts among us (Class A SS and AudioPhysic Tempo IIs are my current rig, slowly to be replaced by DIY efforts over the next year or two...), so rightly or wrongly, I consider myself a tough test for anything other than the well established cones/domes of hi end home audio.
2) If you are focused on the home hi-fi environment, then aren't waveguides versus horns a very reasonable test that everyone should perform for themselves? I think Earle Geddes is defiinitely on to something, and I think the DDS product is a nice compromise between the extremes on waveguide design and implmentation. One of the primary reasons that I invested in a trip to the Rocky Mountain Audio Fest last week was to hear all the options for mid-tweet design, cones/domes, cones/ribbons, cones/waveguides, cones/horns, horns/horns, etc.... (as well as line arrays, but hey, let's not go there!) So having done so, I should state that my own bias here has definitely drifted away from horns and towards ribbons and wageguides (and I have really been reading a lot about horns lately, and wanted to like them, but didn't). I listened carefully to all the horn designs at the show and for some reason, as much as I wanted to get into them, none of them impressed me except for Bruce Edgars Titans, and even then I still liked the above aforementioned a little better and also some killer alternatives such as the Merlins and the Ebens, not to mention Wilson Beneshe (spelling?) who also made great sound at the show, while I am blabbering on the topic, let me also say that SoundSmith, AudioCraftsmen/FacetAudio, BlueCircle/SelahAudio, and Stello/FJ (all amps/spkrs) fill out my own list of best sound of the show (I disqualify the Wilson Max2s, MBLs, and other stratosphere systems because they just seem stupidly self-indulgent to me and makes little more if any appreciable music than many other less impressive efforts). And Shin, if you are reading this I think you would have the same reaction to the Manger that I had, which is I loved it, and I think "overly polite" is no longer a fair criticism with the new version, and I am very very very tempted to just build my DIY reference system around that driver, but at the end of the day, for the few times that I want to push the limits SPL wise, I don't think they will take me where I want to go. If I was a sound engineer doing a lot of jazz/classical work, they would be my nearfield monitor, no doubt about it. (OK, OK, end of rant on RMAF...)
3) I think a nice, dynamic, pro sound 10 inch cone, designed for mid-range duty, could be a good candidate if the response is smooth enough for mid-field home listening, as the natural move towards some beaming would blend well at a low crossover point of 1200-1500 hz, providing a very smooth, albeit changing, power response that narrows slightly as one moves through the mid band, which I am still open to believing may be a good thing rather than a bad thing.
4) I heard a few systems that brought in super tweeters well above 5khz, didn't like any of them (and I am convinced lobing was at least part of the problem), so I am now personally convinced that you either cross right below or right above the areas that most seem to agree are most sensitive to our ears, and that would be, again my own bias here on the exact numbers, 1.8-3.8 khz. Every time I hear something crossed just at or below, or just at or above (but not too far in either direction) these two boundaries they tend to sound better than competing designs that don't, with the Merlins being a good exception (if you spend nearly a decade perfecting something and are as passionate about music as Bobby Palkovic, you will probably find a way to make your 2.2khz crossover point just about invisible).
5) I will admit, and this is why despite my long rant here, I am very open and interested in the other design advice in this thread, that I am very skeptical of cross-overs in the 200-800 Hz range because this is supposedly (cant' remember the reference, but it was a very credible one, and I am convinced it explains why I have not been impressed with a bunch of different 3-way designs) a range that many are very sensitive to phase, and transient time/phase perfection is a very tough thing to achieve, so I stick to my limited knitting for now, but admit to being curious to hear some speakers designed to cross in that range, especially things like a TAD horn from 500hz up and TAD Woofer handling everything below, and then all kinds of variations of that theme employing sub woofers and super tweeters as well, and specifically what Scott has recommended in his posts above. I think I want to do exactly what I have already posted about, but I would hate to do it without hearing or better yet experiementing myself with some of the things that Scott has suggested here...
6) Though I am open to the "pros" expressed in posts above, not sure I buy into all the "cons", as the wave-guide to dipole system I listened to gave my audiophysics a run for the money in soundstage, smoothness, and fatigue-free listening, and then surely blew them out of the water when it came to micro/macro dynamics, pace/timing, realistic timbres and overall liveliness of the sound.
7) Personally, I wish I had the time/money to try everything mentioned in this thread so far...
8) And finally, Paul, I do agree with another poster that your SPL goal might be overkill, as would be 4 x 12 inch woofers (if you are crossing to a sub from there..). I think you would get all the SPL you would ever want out of a good CD tweeter/mid and a really solid 8-10inch cone (if not pushed above 1500hz due to thermal limits, and much below 150hz due to mechanical limits), and then need only a pair of 10, 12, or 15 inch woofers, especially if you are crossing over to a bass horn sub, or any kind of sub at 40hz or higher. I would bet that there are a few 10s and a lot of 12s that would give you everything you would ever want down to 60hz and everything you would really need down to 40hz.
Hey Ralph - good explination. But some of your numbers don't fit at all with with what I've found. Were did you get them?
My experience has been that standard CD mastering has an average level 18dB below peak. Classical is typicaly -20 to -22dB. A lot of recordings these days are hot, like -12dB, or even -10dB for Metal. But -2 or -3dB? Yikes! Never seen that.
Also. My measurements of spectral content seems to show that most of the engergy is in the mids (no surpise) and that the highs fall off pretty fast - faster than pink noise. Also not much stuff below 50Hz.
I used to have a program that would give a spectrum analysis of an entire sound file. But I can't find it now!
So anyway, I was wondering were your numbers came from, as they are so different from mine.
My experience has been that standard CD mastering has an average level 18dB below peak. Classical is typicaly -20 to -22dB. A lot of recordings these days are hot, like -12dB, or even -10dB for Metal. But -2 or -3dB? Yikes! Never seen that.
Also. My measurements of spectral content seems to show that most of the engergy is in the mids (no surpise) and that the highs fall off pretty fast - faster than pink noise. Also not much stuff below 50Hz.
I used to have a program that would give a spectrum analysis of an entire sound file. But I can't find it now!
So anyway, I was wondering were your numbers came from, as they are so different from mine.
Re: RMAF hi-jack, then back to scheduled programming!
Good questions Greggo..
here are some answers:
1. Probably. At higher freq.s do not pay much attention to the effects generated by the room (..unless you have the loudspeaker very close to a sidewall). This isn't to say that room treatment can't improve (considerably even) sound quality, but just that it isn't nearly as great a factor as you would think (..or most would have you believe). At lower freq.s however, especially in the midbass region it can be VERY beneficial (..and only partially, and to a lesser extent, having to do with room modal deviations).
As to imaging - the more consistent the dispersion at higher freq.s, the greater the image stability. Imaging (directional cues) are largely an artifact of spl changes between 2-11 kHz (and mostly 2-7 kHz). You can ALTER this behaviour a'la Jordan/Geddes/"toe-in" (..or in opposite my way/"toe-out"), but you will pay some "penalties" for doing so. Particularly with full bandwidth tonal balance, but also with either imaging (relativly constrained width for the Jordan/Geddes method), or listening position vs. image "skew" (my way). Furthermore, to do it properly requires careful control of the dispersion vs. spl vs. freq. AND making sure that the listener axis is "flat".
2.The Geddes method IS effectivly a (modern) waveguide version of Ted Jordan's method. The major difference is that Ted's version is an even wider dispersion design for most of the bandwidth.
I'd bet that most of the problems you had with the horns were related to the usual: narrow horizontal dispersion, poor crossover power response summation, and poor control of linear distortion added by the horn. From this I'm guessing that imaging was more forward and more closely "aligned" with either "phantom center" or the respective loudspeaker (Lor R), and also that the tonal balance was either subjectivly bright or dull.
3.Cone driver beaming is in effect "horn"/waveguide loading - but a VERY poor version of it. It gives a more "large format planar" sound to imaging cues - more 2d "washed-out" sounding, or less "palpable"/3d. There are also difference in perceived "mass character", and as a result subjective dynamics vs. clarity.
4.That can be a real problem with super tweeters (and most implementations of them).. Virtually always because:
1. the filters are not "steep" enough.
2. the vertical dispersion of both drivers is not "narrow" enough.
3. phase integration is poor.
4. most "beam" horizontally directly in their operating passband.
5.SPL is increasingly important above 1.5 kHz while phase is below 1.5 kHz (..though note that they don't "scale" the same). (..of course phase directly effects spl with any network-"split" design, but thats not what I'm referring to here.) Now with a good active crossover (particularly a digital one if available), and the ability to physically alter driver "off-set" in any DIY design - phase really shouldn't be an issue. IF it was passive - well then that would be *completely* different (..and little more than an exercise in futility with such a design).
Now I'm *not* saying you "should" do a design as I've suggested, rather that the basic design is technically less compromised than you'll find for most other designs when considering dispersion. This is particularly true when you consider that it is *really* easy to mess this up (..either with dipoles and any "horn/waveguide" loading as Sjef has found, or with "horn/waveguide" loading and standard omnipole radiators). (..in fact at least twice a month you'll see the latter problem "crop-up" on the High-Eff. board of the audioasylum - the standard solution is to modestly "front horn" load the omnipole in an effort to match dispersion character at crossover, OR recomend a larger driver to functionally do something similar ..a'la Geddes.)
6. I'd bet that the dispersion of the waveguide/dipole was similar in dispersion at crossover. In otherwords it is UNLIKELY that the polar response of the dipole driver was behaving like a pure dipole, (-3db at 45 degrees off-axis), at crossover - presumably because of baffle size vs driver size and position on baffle. Again, this is something you can easily model in the Edge freeware I've previously mentioned. (..and its fast, fun, easy, and overall quite accurate.)
..........................................................
anyway.. just some thoughts.
Good questions Greggo..
here are some answers:
1. Probably. At higher freq.s do not pay much attention to the effects generated by the room (..unless you have the loudspeaker very close to a sidewall). This isn't to say that room treatment can't improve (considerably even) sound quality, but just that it isn't nearly as great a factor as you would think (..or most would have you believe). At lower freq.s however, especially in the midbass region it can be VERY beneficial (..and only partially, and to a lesser extent, having to do with room modal deviations).
As to imaging - the more consistent the dispersion at higher freq.s, the greater the image stability. Imaging (directional cues) are largely an artifact of spl changes between 2-11 kHz (and mostly 2-7 kHz). You can ALTER this behaviour a'la Jordan/Geddes/"toe-in" (..or in opposite my way/"toe-out"), but you will pay some "penalties" for doing so. Particularly with full bandwidth tonal balance, but also with either imaging (relativly constrained width for the Jordan/Geddes method), or listening position vs. image "skew" (my way). Furthermore, to do it properly requires careful control of the dispersion vs. spl vs. freq. AND making sure that the listener axis is "flat".
2.The Geddes method IS effectivly a (modern) waveguide version of Ted Jordan's method. The major difference is that Ted's version is an even wider dispersion design for most of the bandwidth.
I'd bet that most of the problems you had with the horns were related to the usual: narrow horizontal dispersion, poor crossover power response summation, and poor control of linear distortion added by the horn. From this I'm guessing that imaging was more forward and more closely "aligned" with either "phantom center" or the respective loudspeaker (Lor R), and also that the tonal balance was either subjectivly bright or dull.
3.Cone driver beaming is in effect "horn"/waveguide loading - but a VERY poor version of it. It gives a more "large format planar" sound to imaging cues - more 2d "washed-out" sounding, or less "palpable"/3d. There are also difference in perceived "mass character", and as a result subjective dynamics vs. clarity.
4.That can be a real problem with super tweeters (and most implementations of them).. Virtually always because:
1. the filters are not "steep" enough.
2. the vertical dispersion of both drivers is not "narrow" enough.
3. phase integration is poor.
4. most "beam" horizontally directly in their operating passband.
5.SPL is increasingly important above 1.5 kHz while phase is below 1.5 kHz (..though note that they don't "scale" the same). (..of course phase directly effects spl with any network-"split" design, but thats not what I'm referring to here.) Now with a good active crossover (particularly a digital one if available), and the ability to physically alter driver "off-set" in any DIY design - phase really shouldn't be an issue. IF it was passive - well then that would be *completely* different (..and little more than an exercise in futility with such a design).
Now I'm *not* saying you "should" do a design as I've suggested, rather that the basic design is technically less compromised than you'll find for most other designs when considering dispersion. This is particularly true when you consider that it is *really* easy to mess this up (..either with dipoles and any "horn/waveguide" loading as Sjef has found, or with "horn/waveguide" loading and standard omnipole radiators). (..in fact at least twice a month you'll see the latter problem "crop-up" on the High-Eff. board of the audioasylum - the standard solution is to modestly "front horn" load the omnipole in an effort to match dispersion character at crossover, OR recomend a larger driver to functionally do something similar ..a'la Geddes.)
6. I'd bet that the dispersion of the waveguide/dipole was similar in dispersion at crossover. In otherwords it is UNLIKELY that the polar response of the dipole driver was behaving like a pure dipole, (-3db at 45 degrees off-axis), at crossover - presumably because of baffle size vs driver size and position on baffle. Again, this is something you can easily model in the Edge freeware I've previously mentioned. (..and its fast, fun, easy, and overall quite accurate.)
..........................................................
anyway.. just some thoughts.
Thanks for all the responses. This is all I have time for now ...
AJ:
I will happily have drivers shipped from overseas if I think I can get better drivers, or more suitable, or better value. I'm quite fussy with drivers - not happy with something local if I think I can get something I like from overseas. I bought my subs that way.
Ralph:
I have simulated my own speakers, and measured them at the listening position with ultracurve. There is a big gap between how much SPL I thought I would get and what I really achieve. Then when I turn it down so they have some more headroom, I get even less.
More later ... *runs out the door*
AJ:
I will happily have drivers shipped from overseas if I think I can get better drivers, or more suitable, or better value. I'm quite fussy with drivers - not happy with something local if I think I can get something I like from overseas. I bought my subs that way.
Ralph:
I have simulated my own speakers, and measured them at the listening position with ultracurve. There is a big gap between how much SPL I thought I would get and what I really achieve. Then when I turn it down so they have some more headroom, I get even less.
More later ... *runs out the door*
AJinFLA said:
Waveguide and dipole - no. Waveguide/horn and omnipole - yes. For the effects of dispersion character and crossover they are effectivly identical problems - the added complication though is rear radiation of the dipole and that can be handled effectivly by attenuation.
Dipole and baffle (specifically in the midrange) for the dispersion character - yes (though most definitly not with the nearly limitless options with driver placement on the baffle, nor complex baffle shapes). (..this is from memory though vs. the edge, its been a while since I've done measurements.)
For instance if you take a nominal 6 1/2 inch driver and measure it without baffle - spl starts taking a "nose-dive" right around 1 kHz (..actually I think it was slightly above 1 kHz for my 166ESR if I remember correctly). I've extended baffle width/length with things like cardboard (and thicker foam board) to see what the effects were. I did not try altering the shape from a rectangle flat baffle though. Again though, its all from memory - so my ability to declare its actual accuracy is far from reliable.. BUT the thing is pretty sophisticated and others like MJK haven't seen anything to complain about yet.
Hmm, had forgotten about your loudspeakers (good segway)..
For your design (..If my memory is correct, a beauty from what I recall), the ring tweeter and the waveguide are effectivly summing with the dispersion character of the upper portion of the seas midbass's passband - and even that is "wider" in dispersion than a dipole (likely up to 1.5 kHz).
In fact (now that I'm actually thinking about it), that is a better solution (dispersion-wise) than SL's omnipole + presence dip IMO, because far off-axis that 8 inch driver will still be dispersion limited in comparison to the tweeter (and the 4th order LR is down -6db at crosspoint - so it isn't helping the midbass out much). At lower freq.s its likely calculated more along the lines of the Orion for dipole radiation..No?
Ralph, perhaps I have not correctly described my target. When I say peak SPL, I'm referring to the peak I would measure with Behringer Ultracurve DEQ2496.
This is what I have now:
Vifa P17 6.5" midbass and Vifa D25AG 1" aluminium dome - MTM in a TL box driven by an amp with around 50w RMS with a listening position 3m away. My subs are dual AE speakers AV12 12" drivers with 23mm xmax driven by Behringer EP2500 pro amp putting 650w into each channel. Sealed 70L box.
I use ultracurve to eq the on axis response. There is no HP on the mains, but the eq reduces the overlap with the sub, hence dramatically reducing excursion demands on the mains. Hence their output is no longer excursion limited.
In WinISD with 40w input, I get 107db @ 1m
What I can measure doesn't go much past 90 db. I'm pushing to get much more, and I prefer to have some headroom.
What I'd ideally like to achieve is 20 db more output than I have currently with this setup!
If that results in compromises I'm not happy with, I can settle for 10 - 15 db more.
If I were to try to do this with a dome tweeter, I think I would probably need a waveguide, fairly high power handling and probably more than one tweeter.
I notice you have an interesting article on your site on waveguides. I have turned up a few on the lathe, and made attempts at measuring them - not really done right but I didn't seem to get close to the gain that is supposed to be possible. In many cases I lost as much as I gained. I've looked at Peavey and Geddes material on waveguides, also Zaph's experiment, and tried to create a waveguide of similar shape.
Greg,
I have to say I'm extremely jealous of what you have been able to hear!
The 4 x 12" woofers - keep in mind they are in open baffle! Four is for displacement. They seem like a good balance. They have good xmax, a very smooth response, and a low distortion motor. They are very good value here and a pro driver with a shorting ring is likely to cost 3x as much. Hence four of this driver is similar in cost to say a good large pro 15" driver with a shorting ring. The pro driver would probably have higher inductance and compromises relating to design for high power handling. The 12" would probably handle 100 - 200 each so 400 - 800 total, although it's probably not needed as xmax is likely to be the limit. Ideally I'd like to get dipole bass as low as I could, but I'd settle for 80 Hz for movies.
I'm still open to other drivers, but for now the Peerless SLS seems a very good choice for open baffle midbass crossing to a mid driver. The Peerless XLS costs 2.5x as much for 1.8x the excursion and a limited top end.
One thing that I have often wondered about: can large mid drivers achieve the same transparency as smaller hifi mids? I know there are strong views on both sides. I'd love to be able to hear good examples to both to be able to evaluate with my ears. Surely in making a larger mid driver there are compromises made. For me a big question is which route to take:
1. 3 way with a 6.5" high efficiency midrange (my initial intention)
2. 2 or 3 way with large pro driver - say 10 - 15"
Given that it's OB, and I don't want it too wide, I think it will have to be 3 way so that I can make the baffle effectively wider for the midbass woofers.
This is what I have now:
Vifa P17 6.5" midbass and Vifa D25AG 1" aluminium dome - MTM in a TL box driven by an amp with around 50w RMS with a listening position 3m away. My subs are dual AE speakers AV12 12" drivers with 23mm xmax driven by Behringer EP2500 pro amp putting 650w into each channel. Sealed 70L box.
I use ultracurve to eq the on axis response. There is no HP on the mains, but the eq reduces the overlap with the sub, hence dramatically reducing excursion demands on the mains. Hence their output is no longer excursion limited.
In WinISD with 40w input, I get 107db @ 1m
What I can measure doesn't go much past 90 db. I'm pushing to get much more, and I prefer to have some headroom.
What I'd ideally like to achieve is 20 db more output than I have currently with this setup!
If that results in compromises I'm not happy with, I can settle for 10 - 15 db more.
If I were to try to do this with a dome tweeter, I think I would probably need a waveguide, fairly high power handling and probably more than one tweeter.
I notice you have an interesting article on your site on waveguides. I have turned up a few on the lathe, and made attempts at measuring them - not really done right but I didn't seem to get close to the gain that is supposed to be possible. In many cases I lost as much as I gained. I've looked at Peavey and Geddes material on waveguides, also Zaph's experiment, and tried to create a waveguide of similar shape.
Greg,
I have to say I'm extremely jealous of what you have been able to hear!
The 4 x 12" woofers - keep in mind they are in open baffle! Four is for displacement. They seem like a good balance. They have good xmax, a very smooth response, and a low distortion motor. They are very good value here and a pro driver with a shorting ring is likely to cost 3x as much. Hence four of this driver is similar in cost to say a good large pro 15" driver with a shorting ring. The pro driver would probably have higher inductance and compromises relating to design for high power handling. The 12" would probably handle 100 - 200 each so 400 - 800 total, although it's probably not needed as xmax is likely to be the limit. Ideally I'd like to get dipole bass as low as I could, but I'd settle for 80 Hz for movies.
I'm still open to other drivers, but for now the Peerless SLS seems a very good choice for open baffle midbass crossing to a mid driver. The Peerless XLS costs 2.5x as much for 1.8x the excursion and a limited top end.
One thing that I have often wondered about: can large mid drivers achieve the same transparency as smaller hifi mids? I know there are strong views on both sides. I'd love to be able to hear good examples to both to be able to evaluate with my ears. Surely in making a larger mid driver there are compromises made. For me a big question is which route to take:
1. 3 way with a 6.5" high efficiency midrange (my initial intention)
2. 2 or 3 way with large pro driver - say 10 - 15"
Given that it's OB, and I don't want it too wide, I think it will have to be 3 way so that I can make the baffle effectively wider for the midbass woofers.
Hi Paul,
Thanks for the extra info.
Something doesn't add up here. Either you have a serious problem with the way your system is set up, or else there is some issue with your measuring equipment or technique.
The P17 for example has a raw sensitivity of about 87dB/w/m. A stereo pair of them driven with 50W should therefore be capable of something like 100dB at 3m.
Given that you're not excursion limited, with adequate amplifier power they should be able to get you another 6dB or more. Say around 106dB at 3m.
Cheers, Ralph
Thanks for the extra info.
Something doesn't add up here. Either you have a serious problem with the way your system is set up, or else there is some issue with your measuring equipment or technique.
The P17 for example has a raw sensitivity of about 87dB/w/m. A stereo pair of them driven with 50W should therefore be capable of something like 100dB at 3m.
Given that you're not excursion limited, with adequate amplifier power they should be able to get you another 6dB or more. Say around 106dB at 3m.
Cheers, Ralph
Hi Panomaniac,
Please take my comments regarding crest factor with lots of salt. The intent was to point out the relatively low average power level of music.
I agree that 2-3dB is underestimating the crest factor. 2-3dB represents a pure sine wave! But I was mostly intersted in a worst case analysis that could reasonably be designed to. As you suggest, in reality, even less average SPL is required by music than my analysis suggests.
Your comments regarding the power level dropping by more than 3dB/octave (pink noise) are interesting. I have gone through the process of measuring it for myself, and I found about a 1.5 - 2dB/octave drop. My measurements were made with a variety of popular music styles. I'll have to see if I can find my experiments and confirm this.
I agree that there's not much below 50Hz for most music.
Cheers, Ralph
Please take my comments regarding crest factor with lots of salt. The intent was to point out the relatively low average power level of music.
I agree that 2-3dB is underestimating the crest factor. 2-3dB represents a pure sine wave! But I was mostly intersted in a worst case analysis that could reasonably be designed to. As you suggest, in reality, even less average SPL is required by music than my analysis suggests.
Your comments regarding the power level dropping by more than 3dB/octave (pink noise) are interesting. I have gone through the process of measuring it for myself, and I found about a 1.5 - 2dB/octave drop. My measurements were made with a variety of popular music styles. I'll have to see if I can find my experiments and confirm this.
I agree that there's not much below 50Hz for most music.
Cheers, Ralph
Allowing for BSC, 1w1m = 88 db
1w3m = 79 db
10w=89 db
40w=95 db
Power compression will be probably 3 - 5 db. I had a look at power compression for Beyma pro drivers, the only company I can think of that shows power compression apart from JBL. The Beyma drivers I looked at typically had 2db compression around 40w rising up to 5 db at their nominal power rating. Since the woofer has a nom
This isn't really much of a suprise to me.
There is nothing wrong with my system. If our numbers don't agree, it's about the numbers and how they are worked out.
Now I'll admit I didn't use an SPL meter as such, but my diy mic via Ultracurve. But the sensitivity of the unit was entered into ultracurve settings so the absolute SPL levels should be accurate. I haven't compared it to an SPL meter.
I doubt that two pairs of speakers in stereo is going to add 6db to the output, not unless they are right next to each other. I'm also not sure the room is going to contribute 3db to the output.
We may talk about crest factors, but what I'm interested in is peak SPL, not the peak that can be sustained for a millisecond, but the peak that I can measure on ultracurve. Perhaps another measure of peak SPL would register 10 db louder or more due to a shorter time window.
I don't know if this clears anything up or not.
1w3m = 79 db
10w=89 db
40w=95 db
Power compression will be probably 3 - 5 db. I had a look at power compression for Beyma pro drivers, the only company I can think of that shows power compression apart from JBL. The Beyma drivers I looked at typically had 2db compression around 40w rising up to 5 db at their nominal power rating. Since the woofer has a nom
This isn't really much of a suprise to me.
There is nothing wrong with my system. If our numbers don't agree, it's about the numbers and how they are worked out.
Now I'll admit I didn't use an SPL meter as such, but my diy mic via Ultracurve. But the sensitivity of the unit was entered into ultracurve settings so the absolute SPL levels should be accurate. I haven't compared it to an SPL meter.
I doubt that two pairs of speakers in stereo is going to add 6db to the output, not unless they are right next to each other. I'm also not sure the room is going to contribute 3db to the output.
We may talk about crest factors, but what I'm interested in is peak SPL, not the peak that can be sustained for a millisecond, but the peak that I can measure on ultracurve. Perhaps another measure of peak SPL would register 10 db louder or more due to a shorter time window.
I don't know if this clears anything up or not.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.