So if I want to have an even off-axis response from the midrange and tweeter, which match well depending on the angle, the baffle width should play a role. Ok.
In that case, isn't the width more crucial for the midrange because the off-axis response of the tweeter depends more on the waveguide than on the baffle width?
If I were to build different test enclosures with different widths, I would match the enclosure where the midrange driver goes well with the off-axis behaviour of the tweeter, wouldn't I?
Or do I orientate myself on the given conditions and rather choose a suitable crossover frequency? I hope my questions are somewhat understandable
In that case, isn't the width more crucial for the midrange because the off-axis response of the tweeter depends more on the waveguide than on the baffle width?
If I were to build different test enclosures with different widths, I would match the enclosure where the midrange driver goes well with the off-axis behaviour of the tweeter, wouldn't I?
Or do I orientate myself on the given conditions and rather choose a suitable crossover frequency? I hope my questions are somewhat understandable
There are different objectives why would you do something, it's complicated topic but here is something that hopefully helps you to figure out more.
Directivity in general is important if room reflections are important. The further away you listen and the poorer/lively your room acoustics is the more you hear room sound in general, which can be simplified to be average sound the speaker radiates to all directions for the significant bandwidth. Then there might be some specific angle which might benefit from some special directivity, like if you know exactly how your speakers need to be setup and know exactly what the angle is, then you could try and optimize.
When it comes to bass, there is about no directivity as wavelengths are longer than room. Highs usually beam at least a bit, since wavelength gets shorter than the tweeter transducer diameter. In between you have to somehow slide the directivity from 0db of bass to about 10db of high treble but none of this matters if you listen in anechoic space where all you hear is direct sound, you would have to mind only about the direct sound.
In a typical living room it likely matters though. If you have a jump in directivity about at the crossover, then the sound might feel weird, "dry" on highs and "wet" for lows, because there is more room sound for lows as it will radiate to all directions while treble is more focused toward where the speaker is pointed at. So you'd likely want some kind of smooth DI so that it wasn't so apparent where DI changes. Also, listening height and listening angle relative to speaker might be important to consider. On some speakers there is only about one good listening axis, while on some other there is big area (cone of sound) where sound stays uniform and balanced. Now also practical issues are in the mix, like where you can position the speakers and can you always sit on the spot or wander around etc.
I bet there is lot of people who don't hear such things, or don't mind about it. On the other hand I think there is more to be optimize but that's only after you have the speakers in your room and you've listened your self and hear how your room affects the sound. It's rather easy to come up with a system with almost any directivity in software, and to meet good guidelines found in studies what would be a good DI, but to optimize further than that it has got to be optimized for your room and for your liking, you must know what to optimize for.
Alright, what's the crossover region got to do with it? since around crossover two (or more) transducers emit same bandwidth and in case they do not colocate in physical space then there will be some interference to some directions. This means some directions there is destructive interference at some wavelength, some sound cancels, which will make DI peak near the crossover. Whether you hear it or not is another question, but what you can do is play with the interference by adjusting directivity of the two transducers by manipulating their relative location and physical construct around them, which all affect how much sound goes to what directions, also crossover plays it's role on it. DI gets smoother the less sound gets destroyed in interference and positioning in room and matters if any of it is audible anyway.
All of this can be experimented in a simulator like VituixCAD, rather easily and in real time basically. What is missing from the simulation is how does it actually sound like and how important it is considering any other objective you might have with your project. To determine how anything sounds like, make two speakers, listen both and choose which one works better for you. If you don't want to do this, make a speaker that seems most interesting to you, juggle it so that directivity is relatively smooth and enjoy 🙂 Have fun!
ps. forgot to mention, sound interacts with physical world by wavelength. So, a waveguide can control sound only whose wavelength is small enough compared to physical size and shape of the device. A baffle, or a box, is just another "waveguide" so if a baffle is bigger than the waveguide attached to it, the baffle affects sound to lower frequency than the waveguide.
Directivity in general is important if room reflections are important. The further away you listen and the poorer/lively your room acoustics is the more you hear room sound in general, which can be simplified to be average sound the speaker radiates to all directions for the significant bandwidth. Then there might be some specific angle which might benefit from some special directivity, like if you know exactly how your speakers need to be setup and know exactly what the angle is, then you could try and optimize.
When it comes to bass, there is about no directivity as wavelengths are longer than room. Highs usually beam at least a bit, since wavelength gets shorter than the tweeter transducer diameter. In between you have to somehow slide the directivity from 0db of bass to about 10db of high treble but none of this matters if you listen in anechoic space where all you hear is direct sound, you would have to mind only about the direct sound.
In a typical living room it likely matters though. If you have a jump in directivity about at the crossover, then the sound might feel weird, "dry" on highs and "wet" for lows, because there is more room sound for lows as it will radiate to all directions while treble is more focused toward where the speaker is pointed at. So you'd likely want some kind of smooth DI so that it wasn't so apparent where DI changes. Also, listening height and listening angle relative to speaker might be important to consider. On some speakers there is only about one good listening axis, while on some other there is big area (cone of sound) where sound stays uniform and balanced. Now also practical issues are in the mix, like where you can position the speakers and can you always sit on the spot or wander around etc.
I bet there is lot of people who don't hear such things, or don't mind about it. On the other hand I think there is more to be optimize but that's only after you have the speakers in your room and you've listened your self and hear how your room affects the sound. It's rather easy to come up with a system with almost any directivity in software, and to meet good guidelines found in studies what would be a good DI, but to optimize further than that it has got to be optimized for your room and for your liking, you must know what to optimize for.
Alright, what's the crossover region got to do with it? since around crossover two (or more) transducers emit same bandwidth and in case they do not colocate in physical space then there will be some interference to some directions. This means some directions there is destructive interference at some wavelength, some sound cancels, which will make DI peak near the crossover. Whether you hear it or not is another question, but what you can do is play with the interference by adjusting directivity of the two transducers by manipulating their relative location and physical construct around them, which all affect how much sound goes to what directions, also crossover plays it's role on it. DI gets smoother the less sound gets destroyed in interference and positioning in room and matters if any of it is audible anyway.
All of this can be experimented in a simulator like VituixCAD, rather easily and in real time basically. What is missing from the simulation is how does it actually sound like and how important it is considering any other objective you might have with your project. To determine how anything sounds like, make two speakers, listen both and choose which one works better for you. If you don't want to do this, make a speaker that seems most interesting to you, juggle it so that directivity is relatively smooth and enjoy 🙂 Have fun!
ps. forgot to mention, sound interacts with physical world by wavelength. So, a waveguide can control sound only whose wavelength is small enough compared to physical size and shape of the device. A baffle, or a box, is just another "waveguide" so if a baffle is bigger than the waveguide attached to it, the baffle affects sound to lower frequency than the waveguide.
Last edited:
Some quick responses for your questions 😀 I cannot be more specific, sorry.
But there is also vertical off-axis response, which affects DI, and might matter sound wise as well. c-c and other stuff play their role here.
If you want to optimize also the vertical off-axis, then you might want to play with varying baffle size.
Have fun!🙂
*) game is determined by you! set main goal for your project, be it good sound or good looks or just good time and learning experience, perhaps all of them or something else. Determine priority list and stick to it. If you are after for good sound, prepare for long journey because you must start from what you want to hear, then how your room affects it, then what kind of a speaker system would output sound in such way that being affected by room you get maximal pleasure with your brain processing it all 🙂 Although, making speaker kit or anything that you fancy could take you quite far already, first 90% of the trip and prepare you to start the next 90% 😀
Yeah it's fine
Yeah they would all be part of the same game*, also size and number of the transducers and their relative location to each other and what construct is around them and so on.
Even horizontal off-axis response is easy to achieve when the two transducers are similar in size, say a small tweeter but on a waveguide, and a woofer. Or two direct radiating transducers that are close to same size to get suitable crossover point where the woofer does not beam yet. Tweeter would have low DI up until higher frequency as it is smaller of the two. Waveguide is a way to get the tweeter appear bigger in this sense, get woofer and tweeter directivities match better as they are effectively more similar in size.
But there is also vertical off-axis response, which affects DI, and might matter sound wise as well. c-c and other stuff play their role here.
the problem is not width of the baffle but different size of the transducers. If both transducers are on same baffle, their "low frequency directivity" would be similar to each other. But, the bigger transducer would start beaming on lower frequency than the smaller transducer, which is the thing you should be thinking.
If you want to optimize also the vertical off-axis, then you might want to play with varying baffle size.
This is part of it, there is other things in a loudspeaker system than directivity, that needs to be fine for good sound. For example the crossover. Say that you chose 6" waveguide, 1" dome for it and a 6" woofer and put them on a miminal baffle. Now, their directivities would roughly match around some bandwidth about +/-6" wavelength and only concern is the crossover frequency, which would also be somewhere +/-6" wavelength. Question now is, does the tweeter manage with such crossover frequency? and does the woofer have good sound still, no significant breakup? And answer is, buy suitable drivers that do. Or change the plan if you already bought some drivers, make crossover higher or lower if the drivers can do, play with c-c to juggle the DI, change waveguide, do what ever for best compromise you set yourself for.
Have fun!🙂
*) game is determined by you! set main goal for your project, be it good sound or good looks or just good time and learning experience, perhaps all of them or something else. Determine priority list and stick to it. If you are after for good sound, prepare for long journey because you must start from what you want to hear, then how your room affects it, then what kind of a speaker system would output sound in such way that being affected by room you get maximal pleasure with your brain processing it all 🙂 Although, making speaker kit or anything that you fancy could take you quite far already, first 90% of the trip and prepare you to start the next 90% 😀
Last edited:
My first 3D printed Baffle. Took the new & fast printer about 11h ... will probably take 2 days for a serious speaker with big waveguide ...
But everything worked and is precise.
Btw - how to glue PLA to wood?
But everything worked and is precise.
Btw - how to glue PLA to wood?
I had quite good results with normal PVA wood glue (ponal express in my case). even when gluing PLA to PLA.
I would consider / measure:
- polar frequency responses of midrange in the enclosure (can be translated to directivity)
- polar frequency responses of tweeter in waveguide and in enclosure (same here: can be translated to directivity).
- crossover frequency
- crossover fillter steepness
if you have fixed enclosure and waveguide you can still change crossover frequency and filters to fine tune matching directivities.
again this is very VERY dependent on your crossover frequency and order.
I learned from mabat's thread and by experimenting with waveguides that it does not make sense to stop considering the "waveguide" just because the physical waveguide edge / because the waveguide meets the baffle.
for higher frequencies you can probably just neglect the influence of baffle width and baffle shape, but it has big influence for the lower frequencies.
here below an example of a very mediocre waveguide in a wide baffle and in a rounded over narrow baffle (0-15-30-45-60-75-90° horizontal).
please take it with a grain of salt, as the measurement parameters were not exactly 100 % the same!
I wasn't impressed with PVA glue on PLA, but it's not awful. Epoxy seemed to bond a lot better to wood than PVA to PLA.
Epoxy - that should work. But for my prototpe ... just use some wood glue sounds very convenient ;-)
A nice thing about gluing PLA or PETG with PVA is you can dissolve it with warm water to disassemble.
Here's a more recent measurement after filling the tiny gap around the waveguide and with a few extra filters applied:
I am really happy with the result. Radiation is still wide (that is what I wanted). But diffraction seems to be less pronounced than without the waveguide.
I assembled a prototype.
Enclosure parts are all anodized aluminum and the waveguide is 3D printed (SLA).
The tweeter is SB26CDC-C000-4, the woofer is SB15CAC30-4, and the passive radiator is SB13PFCR-00.
I call this speaker CA5 (Ceramic+Aluminum+5" woofer).
I will upload the measurements soon.
Here's a rendering versus a real picture comparison.
No, the pattern is more controlled with the small waveguide than without. If the waveguide can be wider the pattern can be more even, over a wider range, but fits better a 2 way concept than 3 way.
Yes, this is your waveguide.
The routing template was not used because the baffle was made by a CNC specialist.
I changed the shape of the waveguide as follows so that the head of the screw is not exposed.
I used 1 aluminum profile as bracing.
It was used for the convenience of assembly.
I don't know if there's no resonance even without bracing.
@306xx_KR
I was just curious how you get the finish of the waveguide so well done, but than I saw got it with a SLA printer.
That makes aaaaallll the difference.
Unfortunately not cheap to get it done by one of those print shops. 🙁
Since you used a PR with the same SD as the woofer (as well as roughly the same xmax), I also assume that this speaker isn't designed to go very loud?
I was just curious how you get the finish of the waveguide so well done, but than I saw got it with a SLA printer.
That makes aaaaallll the difference.
Unfortunately not cheap to get it done by one of those print shops. 🙁
Since you used a PR with the same SD as the woofer (as well as roughly the same xmax), I also assume that this speaker isn't designed to go very loud?