Wow Scott now you have really got me thinking. The artistic side of me is getting all warm and fuzzy. I can envision some very interesting designs with folding wing cabinets. Does the baffle extension need to project to the ground or is it symetrical to it's width.
Does this option model the famous Abbey to it's reputation or does TC still have a bit of magic hidden in his design?
Great writing by the way. I love the way you right in real sentences with complete thoughts good for you.
Keep up the good work
Does this option model the famous Abbey to it's reputation or does TC still have a bit of magic hidden in his design?
Great writing by the way. I love the way you right in real sentences with complete thoughts good for you.
Keep up the good work
I feel like a newbie. I have not built that many speakers in the 20 years I have been in the hobby. In fact I know that there are quite a few people who have used the worksheets to design more speaker systems then I have even thought about in the past couple of years (Scott, GM, Bob Brines, ..... ).
Right now the worksheets only handle rectangular full height baffles. The math needed to do different shapes is not that hard but I just have not made the second pass yet to include this feature. I want to upgrade everything to the current level (figuring 90% of users will be happy with that) before going back and adding some of the other options being discussed. I don't want to be in the perpetual upgrade loop and never make anything available.
'Haven't got onto the floor-bounce part yet guys: my apologies, I got side-tracked writing up some consecutive Thor models (see the 'normal' speaker page for the thread if you're interested). However, I've been looking at wide-baffles a bit more, and I can offer a few thoughts. Nothing overly new or profound, but there you go, just what I have observed from playing around in the new sheet.
Firstly, beware the erroneous assumption that a wide-baffle will automatically allow the complete removal of BSC circuitry from a design; it doesn't work that way. I was lucky with my prototypes some months ago in that their placement in the room, and my listening position / axis / distance from the drivers allowed me to achieve this. That doesn't necessarily mean that it would work in that way for you; like everything else, it depends on your particular situation / circumstances. Let me quickly state that I remain a huge fan of wide-baffle designs. I think they are brilliant for many reasons, and well-worth exploring, but be careful, and always go in with your eyes wide open, and a good idea of what it is you want to achieve. They might be appropriate for you. Equally however, they might not.
I've noticed some statements of late in a couple of different places on the net (no names: that might not be quite fair) suggesting that a wide baffle can reduce the diffraction to such a low frequency that it has become omni-directional, and therefore no circuit is required. That's anechoic by the way. Well, I buy that, but they seem to have forgotten to mention that in anechoic terms, that would nominally require a baffle some 1.65m across. it's basically a mis-interpretation of what it is we're attempting with this sort of design: namely a reduction, or complete removal of circitry in the all-critical mid-range. Now, a design, say, 26" wide should not need any at all over 200Hz. However, herein lies the flip-side of the coin.
Let us assume we have our speaker in an anechoic chamber, or free-space, removing the influence of the listening room. Now think compensation-circuit values. You can see where this is going can't you? Yep, compared to a narrow-baffle, we've pushed the -3db point much lower, out of the midrange (good) but that causes us another problem; namely as a general rule, the lower the frequency, the higher the value inductor we shall need to utilise in our passive compensation circuit (all you active people can keep rubbing your hands in glee and plotting ways and means of taking over the audiophile world with your new wide-speakers. Pity they dropped out of favour in the 1970s: I blame WAF myself). And that is a bit of a pain.
OK, so we shan't despair (never good), because remember this was a free-space situation, and I'm fairly sure none of us listens in a field or an anechoic chamber. In room, things look better once again. Assuming the cabinets are close to a rear wall (front baffle within, say, 24"), the wide-baffle types win out over the narrow-baffle types every time in the simulations I've run, needing on occasion, like the Lowther MLTL I mentioned in my last posts, dramatically less in the way of compensation as they appear to be behaving to a large extent like an infinite baffle (or considerably more than a narrow-baffle design is at any rate). Pulled out further into the room, a narrow baffle design usually appears to have the advantage in this regard. Swings and roundabouts as usual. This is, of course, purely focusing upon the frequency and baffle-step response; there are of course other benefits / drawbacks to both methods.
My own conclusions so far?
1) A wide-baffle design is more domestically suitable close to a rear wall due to its size, so that's helpful, given that they appear to provide their best in-room performance for the lowest level of correction in such a position.
2) They appear to have other advantages due to reflections etc which I have yet to explore with any thoroughness. I have a fair bit of reading and experimenting to do before I would be willing to pass comments, but I do know that most people who have tried them out tend to find their influence in this respect quite benign, and more realistic than the narrow-baffles common at present.
3) If you want to run a wide-baffle design using full-range drivers, or a 2-way setup, pulled out in your room, go for active compensation, not passive, unless you have some serious amplification, and even then, with large inductor values, you could well run into trouble. Passive compensation, when correctly applied has never damaged the sound in my experience, but then, I've never needed to use extremely large value inductors. Everything in moderation.
4) If you want to go multi-way (as in 3 way), wide-baffles could be the best thing since the pre-sliced bread days. No compensation required in the midrange at all, assuming that your chosen drivers don't require a notch filter, which is usually the bane of such setups. OK, the bass-drivers will need compensation, but that should be less of an issue for them than in a 2 way or single-driver setup.
Cheers for now
Scott
Firstly, beware the erroneous assumption that a wide-baffle will automatically allow the complete removal of BSC circuitry from a design; it doesn't work that way. I was lucky with my prototypes some months ago in that their placement in the room, and my listening position / axis / distance from the drivers allowed me to achieve this. That doesn't necessarily mean that it would work in that way for you; like everything else, it depends on your particular situation / circumstances. Let me quickly state that I remain a huge fan of wide-baffle designs. I think they are brilliant for many reasons, and well-worth exploring, but be careful, and always go in with your eyes wide open, and a good idea of what it is you want to achieve. They might be appropriate for you. Equally however, they might not.
I've noticed some statements of late in a couple of different places on the net (no names: that might not be quite fair) suggesting that a wide baffle can reduce the diffraction to such a low frequency that it has become omni-directional, and therefore no circuit is required. That's anechoic by the way. Well, I buy that, but they seem to have forgotten to mention that in anechoic terms, that would nominally require a baffle some 1.65m across. it's basically a mis-interpretation of what it is we're attempting with this sort of design: namely a reduction, or complete removal of circitry in the all-critical mid-range. Now, a design, say, 26" wide should not need any at all over 200Hz. However, herein lies the flip-side of the coin.
Let us assume we have our speaker in an anechoic chamber, or free-space, removing the influence of the listening room. Now think compensation-circuit values. You can see where this is going can't you? Yep, compared to a narrow-baffle, we've pushed the -3db point much lower, out of the midrange (good) but that causes us another problem; namely as a general rule, the lower the frequency, the higher the value inductor we shall need to utilise in our passive compensation circuit (all you active people can keep rubbing your hands in glee and plotting ways and means of taking over the audiophile world with your new wide-speakers. Pity they dropped out of favour in the 1970s: I blame WAF myself). And that is a bit of a pain.
OK, so we shan't despair (never good), because remember this was a free-space situation, and I'm fairly sure none of us listens in a field or an anechoic chamber. In room, things look better once again. Assuming the cabinets are close to a rear wall (front baffle within, say, 24"), the wide-baffle types win out over the narrow-baffle types every time in the simulations I've run, needing on occasion, like the Lowther MLTL I mentioned in my last posts, dramatically less in the way of compensation as they appear to be behaving to a large extent like an infinite baffle (or considerably more than a narrow-baffle design is at any rate). Pulled out further into the room, a narrow baffle design usually appears to have the advantage in this regard. Swings and roundabouts as usual. This is, of course, purely focusing upon the frequency and baffle-step response; there are of course other benefits / drawbacks to both methods.
My own conclusions so far?
1) A wide-baffle design is more domestically suitable close to a rear wall due to its size, so that's helpful, given that they appear to provide their best in-room performance for the lowest level of correction in such a position.
2) They appear to have other advantages due to reflections etc which I have yet to explore with any thoroughness. I have a fair bit of reading and experimenting to do before I would be willing to pass comments, but I do know that most people who have tried them out tend to find their influence in this respect quite benign, and more realistic than the narrow-baffles common at present.
3) If you want to run a wide-baffle design using full-range drivers, or a 2-way setup, pulled out in your room, go for active compensation, not passive, unless you have some serious amplification, and even then, with large inductor values, you could well run into trouble. Passive compensation, when correctly applied has never damaged the sound in my experience, but then, I've never needed to use extremely large value inductors. Everything in moderation.
4) If you want to go multi-way (as in 3 way), wide-baffles could be the best thing since the pre-sliced bread days. No compensation required in the midrange at all, assuming that your chosen drivers don't require a notch filter, which is usually the bane of such setups. OK, the bass-drivers will need compensation, but that should be less of an issue for them than in a 2 way or single-driver setup.
Cheers for now
Scott
Scottmoose,
I'm glad to read about your interest and positive experiences with wide-ish speakers placed close to the wall. I think on-wall flat panel TVs will create a large demand for up-against-the-wall speakers, and good CAD tools to model speaker placement, point sources vs. linesources, and optimizing cabinet shapes at the wall boundary would be very helpful. Maybe modest depth waveguides can create a more 3D soundstage for on-wall speakers. There appears to be a market waiting for new solutions.
from Sterophile Sonus Faber recommends near wall placement.
"The Stradivari's wide baffle produced a singular sonic picture. Instead of the more common narrow-baffle, low-diffraction sound, in which a speaker "disappears" to leave behind a ghostly apparition of a three-dimensional sound picture, the Stradivari presented a more weighty, unusually solid picture that seemed to be a three-dimensional curtain wrapped behind the baffles and extending well back into virtual space. While more conventional baffles have produced wider, more transparent soundstages and perhaps more focused and upfront images, none has delivered such a solid and physically believable three-dimensional soundstage in my room—aided, I'm sure, by the Stradivari's rich, palpable midrange."
I'm glad to read about your interest and positive experiences with wide-ish speakers placed close to the wall. I think on-wall flat panel TVs will create a large demand for up-against-the-wall speakers, and good CAD tools to model speaker placement, point sources vs. linesources, and optimizing cabinet shapes at the wall boundary would be very helpful. Maybe modest depth waveguides can create a more 3D soundstage for on-wall speakers. There appears to be a market waiting for new solutions.
from Sterophile Sonus Faber recommends near wall placement.
"The Stradivari's wide baffle produced a singular sonic picture. Instead of the more common narrow-baffle, low-diffraction sound, in which a speaker "disappears" to leave behind a ghostly apparition of a three-dimensional sound picture, the Stradivari presented a more weighty, unusually solid picture that seemed to be a three-dimensional curtain wrapped behind the baffles and extending well back into virtual space. While more conventional baffles have produced wider, more transparent soundstages and perhaps more focused and upfront images, none has delivered such a solid and physically believable three-dimensional soundstage in my room—aided, I'm sure, by the Stradivari's rich, palpable midrange."
Attachments
Now that is very interesting, as it seems to bear out what I'm finding from running the simulations in Martin's software, and my practical experiences with a pair of WB MLTL mules I knocked up a few months back. Thanks for letting us know. I've not read any reviews of these beasts (all £22,000UK of them I believe) -I think the only UK magazine that featured them was Hifi News, and I don't buy that very often. Looks like we're on the right lines anyway!
All the best
Scott
All the best
Scott
Don't underestimate Martin -you might be surprised at what he's already done / has in the pipeline. These updates are only the start.
Re different drivers and crossovers tending to have the greatest effect, I agree 100% -duff-drivers and crossovers will never exactly help the sound. However, my primary interest is in cabinets, and their effect on the sound; selecting drivers and engineering crossovers etc is out of my scope, at least here. And in this regard I find wide baffle designs interesting. Assuming close[ish] rear-wall placement, in all the simulations I've run so far they've exhibited the smoother, flatter response curve, which is surely of some worth. They have also achieved it with smaller values in a BSC circuit that's been pushed into a range where human hearing is somewhat less sensitive. Whilst I haven't heard any problems with a properly engineered BSC circuit, that's not exactly going to do any harm, and the smaller values could be of interest for those with lower-powered amplifiers. This is just regarding baffle-step of course -I haven't really considered the reflection etc issues yet. Not having heard the Stradivari, I can't comment on their sound, but reading the above extract, I think I know what the reviewer is driving at -my own test mules exhibited similar characteristics. It's interesting at any rate!
Best
Scott
Re different drivers and crossovers tending to have the greatest effect, I agree 100% -duff-drivers and crossovers will never exactly help the sound. However, my primary interest is in cabinets, and their effect on the sound; selecting drivers and engineering crossovers etc is out of my scope, at least here. And in this regard I find wide baffle designs interesting. Assuming close[ish] rear-wall placement, in all the simulations I've run so far they've exhibited the smoother, flatter response curve, which is surely of some worth. They have also achieved it with smaller values in a BSC circuit that's been pushed into a range where human hearing is somewhat less sensitive. Whilst I haven't heard any problems with a properly engineered BSC circuit, that's not exactly going to do any harm, and the smaller values could be of interest for those with lower-powered amplifiers. This is just regarding baffle-step of course -I haven't really considered the reflection etc issues yet. Not having heard the Stradivari, I can't comment on their sound, but reading the above extract, I think I know what the reviewer is driving at -my own test mules exhibited similar characteristics. It's interesting at any rate!
Best
Scott
Scottmoose said:
The effect of baffle on frequency response also depends on driver size and XO frequency due to beaming effects. Additionally, the 2D shape and the surface curvature also plays a role in the smoothness of the response. These cannot be realistically compensated for without losing fidelity. Most engineers that rely on the FR amplitude response too heavily don't realize this. I'm not a math wizard, but I think that the math used to generate FR from transient data is not reversible. Therefore the same FR response can result from different transient data, but the different data will sound differentl while showing the same FR response.
I take the point that the baffle size and shape should be determined by the driver you are using. That's why I design all of my cabinets from scratch, specifically tailored for the driver I am using or have been asked to use, rather than just coming up with a generic cabinet design, stuffing a random driver in and hoping for the best. Remember that I'm not asserting, or advocating these as finished designs here though. They're just interesting concepts created using the new sheet that I'm offering to stimulate debate (which I for one am thoroughly enjoying), creativity, and provide a starting point for exploration by other people.
As for the rest, I'm afraid I'm not sure I fully understand what you're suggesting. Are you suggesting that Martin's MathCad models are inaccurate because they only work at present assuming a flat baffle, with no curves, backward folds etc? Watch this space if that is the case -I believe these features will be emerging in due course. Or that the frequency response plots they generate is based purely upon the amplitude of the driver and port output, and does not account for anything else such as reflections, and their own effect upon it? I don't know enough of the math or parameters behind the sheets to comment I'm afraid -Martin -over to you! Or are you suggesting that WB designs cannot be tenable because you would need to correct for reflections from the baffle, and this would result in a loss of 'fidelity?'
If that's the case, I suspect we're going to agree to disagree. For a start, I don't know what 'fidelity', in hifi terms, is. It's too subjective a concept. But I wonder 1) why correcting for them would inevitably result in a loss of fidelity, and 2) why you would automatically want to 'correct' for them in the first place? Those reflections can actually be of some use in generating a more realistic acoustic image. Imaging might be a trifle more diffuse, but I've yet to go to a point-source concert. (says me -a single driver fan! Oh well! If you can't contradict yourself... ;-) OK, fair enough, I can't see WB designs being for everyone, but I am suggesting they are a valid design concept worthy of lots of exploration, because they could well provide major benefits for some people. So far, at this stage, I can see a whole lot of potential ready and waiting for the reaping, and I intend to carry on looking at them as and when I can.
Like now. Here's an idea.
One issue we have with a WB design is that, although we've shifted the baffle-step problem out of the mid-range, we still need to correct for it, if it's pulled out into the room. And that's a bit awkward, because the low-ish frequency probably means we're going to need a dirty great inductor. Fine for those with 200wpc SS amps, but decidedly awkward of you use anything under about 50wpc. This means that you're going to have to think of either active correction, or go for a different type of enclosure, right? Not necessarily. A mixture to share with you, gentlemen. Of particular interest to us full-range types.
Take 1 full-range unit of your favourite type for covering the mid-range and treble. The FE108ESigma for example. Mounted in the sealed upper half of a nice, WB enclosure (a little oiled burr-veneer perhaps, some nice brass or silver edging to set things off well and boost WAF so it becomes a nice, arty piece of furniture -you get the idea). Now add, in the internally separate lower half of the enclosure (vented, sealed, bandpass, aperiodic, whatever you like really) two of your favourite 8" bass units. Mounted magnet to magnet in bipolar configuration.
Just a thought.
Cheers
Scott
As for the rest, I'm afraid I'm not sure I fully understand what you're suggesting. Are you suggesting that Martin's MathCad models are inaccurate because they only work at present assuming a flat baffle, with no curves, backward folds etc? Watch this space if that is the case -I believe these features will be emerging in due course. Or that the frequency response plots they generate is based purely upon the amplitude of the driver and port output, and does not account for anything else such as reflections, and their own effect upon it? I don't know enough of the math or parameters behind the sheets to comment I'm afraid -Martin -over to you! Or are you suggesting that WB designs cannot be tenable because you would need to correct for reflections from the baffle, and this would result in a loss of 'fidelity?'
If that's the case, I suspect we're going to agree to disagree. For a start, I don't know what 'fidelity', in hifi terms, is. It's too subjective a concept. But I wonder 1) why correcting for them would inevitably result in a loss of fidelity, and 2) why you would automatically want to 'correct' for them in the first place? Those reflections can actually be of some use in generating a more realistic acoustic image. Imaging might be a trifle more diffuse, but I've yet to go to a point-source concert. (says me -a single driver fan! Oh well! If you can't contradict yourself... ;-) OK, fair enough, I can't see WB designs being for everyone, but I am suggesting they are a valid design concept worthy of lots of exploration, because they could well provide major benefits for some people. So far, at this stage, I can see a whole lot of potential ready and waiting for the reaping, and I intend to carry on looking at them as and when I can.
Like now. Here's an idea.
One issue we have with a WB design is that, although we've shifted the baffle-step problem out of the mid-range, we still need to correct for it, if it's pulled out into the room. And that's a bit awkward, because the low-ish frequency probably means we're going to need a dirty great inductor. Fine for those with 200wpc SS amps, but decidedly awkward of you use anything under about 50wpc. This means that you're going to have to think of either active correction, or go for a different type of enclosure, right? Not necessarily. A mixture to share with you, gentlemen. Of particular interest to us full-range types.
Take 1 full-range unit of your favourite type for covering the mid-range and treble. The FE108ESigma for example. Mounted in the sealed upper half of a nice, WB enclosure (a little oiled burr-veneer perhaps, some nice brass or silver edging to set things off well and boost WAF so it becomes a nice, arty piece of furniture -you get the idea). Now add, in the internally separate lower half of the enclosure (vented, sealed, bandpass, aperiodic, whatever you like really) two of your favourite 8" bass units. Mounted magnet to magnet in bipolar configuration.
Just a thought.
Cheers
Scott
Err, good point. My apologies.
If you want to go WB with a single, full-range driver, it'll need to be fairly close to a wall. Not that that is likely to be a problem for the better-half... ;-) Positioned with the front baffle around 24" away or less, you should only need light BSC values. The above plan is really for those who need to have the cabinets pulled out from the wall, but don't have the amps to cope with the big BSC circuits this would entail. As lots of us full-range types use subs to augment the bottom end anyway, I hoped this would provide an acceptable compromise, familiar to many, and would allow the use of our favourite smaller units like the FE108ESigma etc (Ha! I've talked myself out of it after all!
)
Re-reading my post, I probably shouldn't have bothered mentioning full-rangers, as it's equally applicable to multi-ways as well -force of habit. I seem to recall Mirage doing something similar to this, though I could be mistaken.
Cheers
Scott
If you want to go WB with a single, full-range driver, it'll need to be fairly close to a wall. Not that that is likely to be a problem for the better-half... ;-) Positioned with the front baffle around 24" away or less, you should only need light BSC values. The above plan is really for those who need to have the cabinets pulled out from the wall, but don't have the amps to cope with the big BSC circuits this would entail. As lots of us full-range types use subs to augment the bottom end anyway, I hoped this would provide an acceptable compromise, familiar to many, and would allow the use of our favourite smaller units like the FE108ESigma etc (Ha! I've talked myself out of it after all!
)Re-reading my post, I probably shouldn't have bothered mentioning full-rangers, as it's equally applicable to multi-ways as well -force of habit. I seem to recall Mirage doing something similar to this, though I could be mistaken.
Cheers
Scott
Scottmoose,
I think all modeling have some limitation. It's basically a tool to reduce risk but not for fine tweeking a design. That's why we always need testing. I think Martin has done a great job. But I'm just trying to bring some awareness of the limitations. There are two aspects of baffle design. One is the radiation space, and the other is the additional transient wave caused by sudden lose of support as the wave reaches the edge. The former is considered mostly, and thus the BSC. The latter normally is less considered, and cannot be acurately compensated for electrically.
To solve the later issue, the wide baffel design is ideal if the diffraction pulse is delayed more than 4ms after the original pulse reaches the listening position. Than means distance of the center of the driver to the first edge must be probably over 1.3 meters. Another approach is to design round baffle edges, design special baffle shape, or add felt around the drivers in order to spread the lose of baffle support over time. I believe Martins Worksheets may not handle the later approach.
To solve the radiation space issue, I think it's a more simple problem. But you also need to consider first without the room modes, so that you get the right transient wave front, but then there is a tradeoff between the room modes and maintaining the correct transient wave front energy. This is where both need to be considered separately and not just looking at the frequency response charts.
I'm sure there are going to be lots of questions because its realy hard to explain in short words.
I think all modeling have some limitation. It's basically a tool to reduce risk but not for fine tweeking a design. That's why we always need testing. I think Martin has done a great job. But I'm just trying to bring some awareness of the limitations. There are two aspects of baffle design. One is the radiation space, and the other is the additional transient wave caused by sudden lose of support as the wave reaches the edge. The former is considered mostly, and thus the BSC. The latter normally is less considered, and cannot be acurately compensated for electrically.
To solve the later issue, the wide baffel design is ideal if the diffraction pulse is delayed more than 4ms after the original pulse reaches the listening position. Than means distance of the center of the driver to the first edge must be probably over 1.3 meters. Another approach is to design round baffle edges, design special baffle shape, or add felt around the drivers in order to spread the lose of baffle support over time. I believe Martins Worksheets may not handle the later approach.
To solve the radiation space issue, I think it's a more simple problem. But you also need to consider first without the room modes, so that you get the right transient wave front, but then there is a tradeoff between the room modes and maintaining the correct transient wave front energy. This is where both need to be considered separately and not just looking at the frequency response charts.
I'm sure there are going to be lots of questions because its realy hard to explain in short words.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.