Thanks for that, your post helped me to find the one song that my son was looking for. A crazy song that was imaging like nothing else.
It just had to be Yello, and it was... that song was: Planet Dada (the flamboyant version).
...
Takla Makan has things going on that I don't particularly hear but definitely can feel (in my floor boards and couch)
A quick frequency analysis about halfway the song shows there's plenty going on
The main vocal is in the middle, nicely behind the plane of the speakers.. then they add an overdub that starts in the middle and spreads out left AND right....
Later on they have some keyboards that are bounced left and right really quick
It's a fun song, and yes, like anything from Yello, it's quirky. Thanks for the heads up, I hadn't heard this one... it's on their The Eye album.
I need to go caress the Blender now.
You are not alone
When I have built my new room to listen to speakers in, I will build some new speakers to listen to. I still seem to be stuck in the 'driver collection' phase.
For my Modified CBT24 project (24 full range drivers per side) I used SB65s for the arrays. In a sealed box a single
SB65 is 3 dB down in the range of 130 to 140 Hz. I use my Denon receiver as the active crossover between the arrays and subs at 130 Hz. Very good results for my application.
SB65 is 3 dB down in the range of 130 to 140 Hz. I use my Denon receiver as the active crossover between the arrays and subs at 130 Hz. Very good results for my application.
Just for you!!Am I alone in this or does the action on the forum seem a bit slow lately.
Always expecting to see nice new things, but there isn't that much action lately.
No new array projects on their way?
https://www.diyaudio.com/community/threads/spherical-bending-wave-corner-lines.404702/
//
@wesayso: I started to think about making a speaker with multiple small full range drivers after hearing a DIY speaker made by an acquaintance of mine. He was inspired by the Tekton design of using an array of tweeters for the midrange. However my acquaintance thought that it made more sense to use small full range drivers so if I recall correctly he has an array of 14 Mark Audio fullrangers that surround a tweeter. It is the only box speaker I have ever heard that has the transparency and purity in the mid range of an ESL or big ribbon. There is no sense of comb filtering and my guess is that the sound field is so chaotic that it kind of levels everything out, maybe a bit like distributed bass?
I have no doubt that this purity is due to the fact that each driver is barely working, just like your line arrays. So my question for you is, do you experience that same sense of transparency and purity but across the full frequency range ?
I have no doubt that this purity is due to the fact that each driver is barely working, just like your line arrays. So my question for you is, do you experience that same sense of transparency and purity but across the full frequency range ?
two very different drivers.
I like the TC9 for its balanced sound, but does not extend as high as the SB25.
The SB65 is a monster at reproducing strings, woodwinds and percussion.
It's the midrange that's super clear, but no doubt there are better solutions or options for an amazing top end or bottom end.
The coherence is wonderful though, with that clear midrange and after some DSP tweaks I get a great and dynamic sound.
I'ts lively and exiting, especially if you have the room working with the arrays. Some room response isn't a bad thing, but timing is everything.
The comparison to how distributed bass works I've made myself a couple of times. Actually it is a process of averaging. When using a single driver, it's arriving at the listeners ear, but we also hear the room. The driver will reflect off of walls, floor and ceiling and this combination is what we hear eventually. An array of drivers all have slightly different positions with regards to floor and ceiling, though the parallel planes are somewhat at the same distance. It is on these parallel planes that I put up absorbing panels,, at first reflection points. You end up with a direct sound that dominates and the reflections get averaged out, much like in a moving mic measurement or even an averaged multi position measurement. There's always some catch, the top end has very small wavelengths and comb filtering does exist. How we hear it isn't as detrimental as many people would want you to believe, but I'm not going to say it isn't there.
Having a crossover between an array and a single tweeter has other compromises. It isn't that ideal either. I chose to filter my array, to reduce the effect of combing. See the one post recap here: https://www.diyaudio.com/community/...range-line-array.242171/page-350#post-6752259
On the subject of averaging, lets look at the difference between a single driver and a filtered array at the listening spot with floor and ceiling reflections included (no side wall reflections in this one):
The orange line is the in-room prediction
If you ask me, that's where that magic in the midrange comes from...
The coherence is wonderful though, with that clear midrange and after some DSP tweaks I get a great and dynamic sound.
I'ts lively and exiting, especially if you have the room working with the arrays. Some room response isn't a bad thing, but timing is everything.
The comparison to how distributed bass works I've made myself a couple of times. Actually it is a process of averaging. When using a single driver, it's arriving at the listeners ear, but we also hear the room. The driver will reflect off of walls, floor and ceiling and this combination is what we hear eventually. An array of drivers all have slightly different positions with regards to floor and ceiling, though the parallel planes are somewhat at the same distance. It is on these parallel planes that I put up absorbing panels,, at first reflection points. You end up with a direct sound that dominates and the reflections get averaged out, much like in a moving mic measurement or even an averaged multi position measurement. There's always some catch, the top end has very small wavelengths and comb filtering does exist. How we hear it isn't as detrimental as many people would want you to believe, but I'm not going to say it isn't there.
Having a crossover between an array and a single tweeter has other compromises. It isn't that ideal either. I chose to filter my array, to reduce the effect of combing. See the one post recap here: https://www.diyaudio.com/community/...range-line-array.242171/page-350#post-6752259
On the subject of averaging, lets look at the difference between a single driver and a filtered array at the listening spot with floor and ceiling reflections included (no side wall reflections in this one):
The orange line is the in-room prediction
If you ask me, that's where that magic in the midrange comes from...
CBT array is an excellent solution for a finite array. It also has the benefit of reducing the effects of reflections in both the horizontal + vertical plane. So it is a viable option, but the type of shading used eats away some of it's dynamic output potential.
Straight array can be an excellent solution if it's a taller array, preferably from (almost) floor to ceiling. While the CBT papers show big differences in performance, they did not include the floor reflection for the straight array, nor was it a tall array (at 1 m), they did include the floor reflection for the CBT. The taller array does have even coverage in seated as well as standing position as our simulations have shown. You'll need EQ to get it to work, but do get quite a lot of output potential in return.
My current array is somewhere in-between both solutions. I tried to optimize the straight array to combat some of it's draw-backs, but in that process it starts to have more similarities to a shaded array or an expanding array. I wanted to keep the output potential of the straight array while focusing on vertical beam forming. What that does is make it function even more independent from floor and ceiling than the unfiltered straight array can manage. I did not loose any output potential and actually gained some on the top end due to the reduction of comb filter effects (a 5 dB gain on the top end).
If you have the room (a CBT takes up more floor space) and do not plan to use room treatment, a CBT can be an excellent choice.
If you want to get more output potential and don't mind treating parallel first reflection surfaces and don't mind using DSP/EQ to make it work, the straight array can be your solution.
In the end, you'll have to make up your own mind. Look up @nc535 thread and see how many ground we covered exploring all kinds of filtering for arrays including CBT. You can even join the fun with that starter model posted by Jack (nc535).
I'm just typing my opinion here, you have to make up your own mind. When I started to look at arrays I've looked into focused arrays, straight arrays, shaded arrays etc. trying to figure out what each could do for me in my room. I even considered using a tweeter in the middle, but never felt I needed it.
Straight array can be an excellent solution if it's a taller array, preferably from (almost) floor to ceiling. While the CBT papers show big differences in performance, they did not include the floor reflection for the straight array, nor was it a tall array (at 1 m), they did include the floor reflection for the CBT. The taller array does have even coverage in seated as well as standing position as our simulations have shown. You'll need EQ to get it to work, but do get quite a lot of output potential in return.
My current array is somewhere in-between both solutions. I tried to optimize the straight array to combat some of it's draw-backs, but in that process it starts to have more similarities to a shaded array or an expanding array. I wanted to keep the output potential of the straight array while focusing on vertical beam forming. What that does is make it function even more independent from floor and ceiling than the unfiltered straight array can manage. I did not loose any output potential and actually gained some on the top end due to the reduction of comb filter effects (a 5 dB gain on the top end).
If you have the room (a CBT takes up more floor space) and do not plan to use room treatment, a CBT can be an excellent choice.
If you want to get more output potential and don't mind treating parallel first reflection surfaces and don't mind using DSP/EQ to make it work, the straight array can be your solution.
In the end, you'll have to make up your own mind
. Look up @nc535 thread and see how many ground we covered exploring all kinds of filtering for arrays including CBT. You can even join the fun with that starter model posted by Jack (nc535).I'm just typing my opinion here, you have to make up your own mind. When I started to look at arrays I've looked into focused arrays, straight arrays, shaded arrays etc. trying to figure out what each could do for me in my room. I even considered using a tweeter in the middle, but never felt I needed it.
Maybe I need to point to yet another type of array. An array that I believe is a very smart solution that controls its output and makes use of the potential of the drivers used. The fractal array from @bbutterfield. Kind of taking the CBT into another dimension. I do fear the small vertical window though. Which is why I never made an effort to start using a single tweeter of any kind.
Hello to all the followers of this thread,
Without reading 400 pages of posts, I was wondering how the 25 drivers are wired and
what is the net-sum impedance?
With an odd number of drivers, I don't see how simple series/parallel wiring can work.
Do all the drivers receive identical signal/power?
CHEERS
Without reading 400 pages of posts, I was wondering how the 25 drivers are wired and
what is the net-sum impedance?
With an odd number of drivers, I don't see how simple series/parallel wiring can work.
Do all the drivers receive identical signal/power?
CHEERS
Series parallel wiring 5 groups of 5 drivers, returns the combined impedance back to that of a single driver.
https://www.diyaudio.com/community/...ver-full-range-line-array.242171/post-7319711
https://www.diyaudio.com/community/...ver-full-range-line-array.242171/post-7319711
( I wonder where my brain was )
One or two drivers going faulty (not open) and you've got quite the quiz.
I guess quality & headroom is the key.
OMG
One or two drivers going faulty (not open) and you've got quite the quiz.
I guess quality & headroom is the key.
I remember experimenting/simulating both focused and CBT arrays myself, based upon my TC9 driver size:
A focused array:
Driver configuration:
And the CBT array:
Driver configuration:
Comparing to my straight but shaded array:
Driver configuration:
These were quick and raw tests, well, except for the straight array as I invested quite a bit of time in that one. Don't look/stare at the frequency curve, pretend they are all straight, that's a matter of re-doing the EQ. Just look at the vertical coverage. Of special interest is the red DI line in the left middle graphs. All of them do pretty good in that regard. Mostly due to driver size (3.5" driver directivity + enclosure shape)
I ran these as kind of a sanity test, to decide if the filtered array I came up with was something to go for or not.
Clear to see the superiority of the focused array, if you can live with a super narrow vertical sweet-spot. And I do mean super narrow, almost laser beam like.
The CBT array as simulated here shows the widest vertical bundling. But it can't avoid comb filtering on the top end with a driver the size of a Vifa TC9 FD18-08. As I started with this driver size and did not consider starting over. Originally I chose this driver due to it's capability to do the whole spectrum, from the bottom low octaves (or at least as low as possible) to the top end (to the best of it's capabilities). The original idea was (and this idea remained) to have a (true) full range array.
The sim does need further explaining. My interest was to keep all 25 drivers to help out in the bottom end, so I used the same shading as was used in my straight array. This was done to keep the bottom end SPL capability as high as I could. If I were ever going for a CBT array, I'd still use this type of shading over the more conventional driver (group) shading as proposed in the CBT papers. That last shading option does not provide the bottom end SPL capacity I wanted to preserve. My shading is largely done as (high) shelves, so for the top end, it should not end up being a big difference to shading the entire driver (group). The biggest difference is that all drivers do contribute to help out in the bottom end with this frequency dependent shading. Vituixcad can sim this type of shading, and in the case of straight arrays gets very close to real world results, so I expect no difference when simulating arced arrays.
Last, but in my opinion not least, the shaded array I went with, as said, a chosen compromise of driver size and bandwidth. If one would use bottom end reinforcements like sub woofers, a smaller driver provides less of a compromise on the top end. Be aware that single subs may offer different output levels/distance compared to an array for mid and top frequencies, but within a room this is largely manageable as the room has quite a bit of influence (my bass gets louder when I lean back into the couch, even though most output would be generated by my stereo pair of subs).
So depending on one's own set of compromises, there is something to say for all types of arrays mentioned above, my preference for an array that only has to do ~150 Hz and up would be different from what my original goals were. My current arrays still help out with room modes, even though they are less potent with the 10F drivers than they were using the Vifa TC9. I shelve them down a bit and use less EQ on the bottom end. They do just enough to use them as part of a kind of multi-sub solution down to about 60 Hz where the subs can handle it on their own. I'm only using them where they get lots of help from being close to the back or side wall, so they don't need to do any hard work. In other words, their excursion per driver in the bottom end is negligible.(*)
To complete this post, I'll add the unshaded 25 driver array with the same settings as the above. After all, that's the one I could live with for at least 7 years without remorse.
Straight and unshaded array:
And it's driver layout (identical to the filtered array):
What this proofs to me is that you need the smallest driver you can get away with for the widest bandwidth you're aiming for.
This holds true for the CBT as well, size does matter. For my original goals, this meant a driver the size of the Vifa TC9 FD18-08.
With other priorities, it might have been something else.
One thing all of these arrays have in common: they can work well with (and within) the room. Determining what to go with would be a weighing of the set of compromises that are acceptable.
No room treatment, plenty of space, the CBT would be a very good choice due to it's nature to average out reflections off of both horizontal as well as vertical planes.
Room treatment on first reflection points opens up room for a straight array or straight shaded array. Be aware of what the lack of having this room effect does to your listening though. I did go with ambience speakers for a reason.
None of these speakers avoids the room boundaries, not even the focused array. There's plenty of energy going everywhere.
In the above plots, floor and ceiling effects where included, at a reduced level of -12 dB. No side wall or back wall effects were included.
If those were added, the CBT would show it's nature of easily dealing with those and would look better than a straight array. Which is why I make the distinction of what you are willing to do, treatment or no treatment. (you're still free to treat a room using CBT speakers)
Aside from these types of speakers, my interest would also go to coaxial solutions, like a well done coaxial driver or the very compact MTM type of speakers in a horn, in other words the Unity/Synergy type of stuff. For a small vertical coverage even a WMTMW could do. My preference is for seated use as well as standing, which only leaves the former solutions on the table.
(*) I've never seen the 10f drivers move, I did see the TC9 work out on songs like "Lullaby" from "A Perfect Circle".
P.S. I wanted to type all this up to clear up my point of view. If I were to start over, based on my own set of priorities, I'd probably go for a straight array again, but now having a pair of subs would allow me to use a smaller driver. I'd go for that solution based on space savings and having the option for easy absorbent panel placement hidden from view. A different room or different circumstances might alter that choice. The most important consideration for me is: what can one get that works with or within your room.
A focused array:
Driver configuration:
And the CBT array:
Driver configuration:
Comparing to my straight but shaded array:
Driver configuration:
These were quick and raw tests, well, except for the straight array as I invested quite a bit of time in that one
. Don't look/stare at the frequency curve, pretend they are all straight, that's a matter of re-doing the EQ. Just look at the vertical coverage. Of special interest is the red DI line in the left middle graphs. All of them do pretty good in that regard. Mostly due to driver size (3.5" driver directivity + enclosure shape)I ran these as kind of a sanity test, to decide if the filtered array I came up with was something to go for or not.
Clear to see the superiority of the focused array, if you can live with a super narrow vertical sweet-spot. And I do mean super narrow, almost laser beam like.
The CBT array as simulated here shows the widest vertical bundling. But it can't avoid comb filtering on the top end with a driver the size of a Vifa TC9 FD18-08. As I started with this driver size and did not consider starting over. Originally I chose this driver due to it's capability to do the whole spectrum, from the bottom low octaves (or at least as low as possible) to the top end (to the best of it's capabilities). The original idea was (and this idea remained) to have a (true) full range array.
The sim does need further explaining. My interest was to keep all 25 drivers to help out in the bottom end, so I used the same shading as was used in my straight array. This was done to keep the bottom end SPL capability as high as I could. If I were ever going for a CBT array, I'd still use this type of shading over the more conventional driver (group) shading as proposed in the CBT papers. That last shading option does not provide the bottom end SPL capacity I wanted to preserve. My shading is largely done as (high) shelves, so for the top end, it should not end up being a big difference to shading the entire driver (group). The biggest difference is that all drivers do contribute to help out in the bottom end with this frequency dependent shading. Vituixcad can sim this type of shading, and in the case of straight arrays gets very close to real world results, so I expect no difference when simulating arced arrays.
Last, but in my opinion not least, the shaded array I went with, as said, a chosen compromise of driver size and bandwidth. If one would use bottom end reinforcements like sub woofers, a smaller driver provides less of a compromise on the top end. Be aware that single subs may offer different output levels/distance compared to an array for mid and top frequencies, but within a room this is largely manageable as the room has quite a bit of influence (my bass gets louder when I lean back into the couch, even though most output would be generated by my stereo pair of subs).
So depending on one's own set of compromises, there is something to say for all types of arrays mentioned above, my preference for an array that only has to do ~150 Hz and up would be different from what my original goals were. My current arrays still help out with room modes, even though they are less potent with the 10F drivers than they were using the Vifa TC9. I shelve them down a bit and use less EQ on the bottom end. They do just enough to use them as part of a kind of multi-sub solution down to about 60 Hz where the subs can handle it on their own. I'm only using them where they get lots of help from being close to the back or side wall, so they don't need to do any hard work. In other words, their excursion per driver in the bottom end is negligible.(*)
To complete this post, I'll add the unshaded 25 driver array with the same settings as the above. After all, that's the one I could live with for at least 7 years without remorse.
Straight and unshaded array:
And it's driver layout (identical to the filtered array):
What this proofs to me is that you need the smallest driver you can get away with for the widest bandwidth you're aiming for.
This holds true for the CBT as well, size does matter. For my original goals, this meant a driver the size of the Vifa TC9 FD18-08.
With other priorities, it might have been something else.
One thing all of these arrays have in common: they can work well with (and within) the room. Determining what to go with would be a weighing of the set of compromises that are acceptable.
No room treatment, plenty of space, the CBT would be a very good choice due to it's nature to average out reflections off of both horizontal as well as vertical planes.
Room treatment on first reflection points opens up room for a straight array or straight shaded array. Be aware of what the lack of having this room effect does to your listening though. I did go with ambience speakers for a reason.
None of these speakers avoids the room boundaries, not even the focused array. There's plenty of energy going everywhere.
In the above plots, floor and ceiling effects where included, at a reduced level of -12 dB. No side wall or back wall effects were included.
If those were added, the CBT would show it's nature of easily dealing with those and would look better than a straight array. Which is why I make the distinction of what you are willing to do, treatment or no treatment. (you're still free to treat a room using CBT speakers
)Aside from these types of speakers, my interest would also go to coaxial solutions, like a well done coaxial driver or the very compact MTM type of speakers in a horn, in other words the Unity/Synergy type of stuff. For a small vertical coverage even a WMTMW could do. My preference is for seated use as well as standing, which only leaves the former solutions on the table.
(*) I've never seen the 10f drivers move, I did see the TC9 work out on songs like "Lullaby" from "A Perfect Circle".
P.S. I wanted to type all this up to clear up my point of view. If I were to start over, based on my own set of priorities, I'd probably go for a straight array again, but now having a pair of subs would allow me to use a smaller driver. I'd go for that solution based on space savings and having the option for easy absorbent panel placement hidden from view. A different room or different circumstances might alter that choice. The most important consideration for me is: what can one get that works with or within your room.
Last edited:
As has happened before, I've completely dropped the anti-cross talk from my system again. I did however change mid-side EQ (mid actually) to reflect the same effects it has at the ear. I recorded (to disk) the mid EQ by itself and with the anti-crosstalk actuated. Then removed the anti crosstalk and EQ-ed the mid to meet both graphs about half way.
The reason? The anti crosstalk (by itself) is very position dependent, ideally it requires one to sit in the exact same position and not move too much. In the combination with mid/side EQ this is less of an issue, but it still is more prominent than using (mid/side) EQ only.
I tried it recently as I was exaggerating everything settings wise to get a grip of how much I could move around the effects of staging/imaging.
I've done that many times before and yet still revisit these experiments if I have changed something, like the ambience settings or something similar. I did notice the effect the anti-crosstalk had under these circumstances (like the phantom center strongly following head movements). Though I've also had a lot of fun playing several tracks during these experiments.
Some tracks truly come to life with exaggerated settings like these, they make an unbelievable rendering that's drawing you deep into the music. Other tracks become harder to listen to or even plain unlistenable. But it's a sure way to get an idea of what is possible and that's why I do it from time to time. This time around I came to a similar conclusion as I've reached quite some time ago; even though anti-crosstalk based on a delayed counter signal works, it is just too position dependent (without additional tricks like tracking cameras etc.) and if I don't overdo the mid-side EQ, it gets me close to the same results. Mid-side EQ remains tricky business, it's quite easy to overdo it, where it doesn't work as intended anymore. It needs to be subtle enough not to notice it (*). A tweak like that won't work at all in the presence of early reflections. The reflections actually hide any crosstalk that's there (and in a stereo setup with two speakers, it's always there, guaranteed
).(*)= Of course it will be noticeable or else I wouldn't bother to use it
. It must have a natural feel, it should not be like a gimmick. The biggest effect is to clear up the phantom center, create better intelligibility. The other (side EQ) part gets you a better tonal balance between phantom center and sides. As said, it works in the absence of early reflections.
Last edited:
A couple of weeks later, I'm still enjoying what I'm hearing. In fact, I've had a lot of fun watching my whole family enjoying the music (and asking for more). We've played the whole "Into the Electric Castle" album (Ayreon, 1998 first pressing, not the remaster) from start to finish.
I keep enjoying that album which is heavily influenced by the 80's rock sound, no wonder as it's maker was a guitar player in an 80's rock band (Vengeance). But it also features Fish (ex Marillion) as one of the singers and Thijs van Leer (of Focus fame, 1969-1978).
I keep enjoying that album which is heavily influenced by the 80's rock sound, no wonder as it's maker was a guitar player in an 80's rock band (Vengeance). But it also features Fish (ex Marillion) as one of the singers and Thijs van Leer (of Focus fame, 1969-1978).
Last edited: