The reason I wanted to use the auto-eq/mirror filter because in looking at polar charts, it appeared that normalize might be emphasizing the vertical sidelobes in the HF. Looking at them now, I don't see much difference in the sidelobes but I do see a more intense and uniform mainlobe, compared to the manually equed mainlobe.
Yes, normalize is equivalent but learning how to create mirror filter was worth the time. I'm not sure what my workflow will be when I have a real array to tune but that might be part of it, at least for a quick start.
Yes, normalize is equivalent but learning how to create mirror filter was worth the time. I'm not sure what my workflow will be when I have a real array to tune but that might be part of it, at least for a quick start.
Attachments
Good stuff, RonDRC-FIR should be able to improve the top you get.
Shading might be another way and I keep forgetting that you guys don't need the bottom end to go low. I would have only used shading on the top end, to keep the bottom end go as low as possible.
I've often thought/dreamed about a synergy like horn in the middle of a line array of woofers. I think you can appreciate the why of that idea.
I see what you mean about the FDW options in DRC-FIR.
Probably the biggest tuning decision in FirD centers on the same issue...how much correction to apply, and when in the process. It's all done with usual partial octave smoothing choices...works well...but i can see DRC offers more to play with.
Yes, freq shading for HF/VHF only...no need lower down where line theory is holding together.
It could be that simple amplitude shading might help too (as opposed to the freq shading just mentioned) .
One thing that was stressed several times in the 3-day Smaart class i recently took, is how much it helps to vary the amplitude of speakers that have lobing issues when coupling.
A little bit of difference in amplitude goes further than i would have expected in reducing the depth of combing, without costing that much in beneficial coupling.
And a big yes to the idea of a synergy in the middle of a line !!
Don't laugh at me, but I'm already wondering if I could put my coax cd and horn into a line of six 12" drivers.
I'd need the 12" line drivers to cover up to about 700Hz, and of course as low as possible. Been looking at cheap 12"s this morning
Would appreciate thoughts on shading implementation for a 16 driver CBT array. I would prefer not to use 8 or 16 channels of amplification, even 4 channels would not be ideal as this will actually be a 2 way array similar to the CBT36. And the ultimate goal is for a multichannel system... so 8 or 16 or 32 channels per speaker times 7+ speakers...
Well, hoping there is a more practical way. Would be willing to dedicated 2 channels for the 16 woofer column. Using a mix of otherwise identical 4 and 8 ohm drivers I can work out shading to follow the Legendre curve pretty closely with as many as 8 steps sparingly using a power resistor here or there. But maybe that isn't needed.
How would you guys approach passive shading of a 16 driver column? I have debated whether to treat it as a 16 driver arc (top driver down 40dB) or an 18 driver arc with top two dropped. Is there a straightforward wiring scheme for groups that I am overlooking?
The tweeter column will use 48 drivers, 3 for each woofer, and i assume that once I solve the woofer column scheme I can duplicate it fairly closely using 3 tweeter groups.
Thanks.
Well, hoping there is a more practical way. Would be willing to dedicated 2 channels for the 16 woofer column. Using a mix of otherwise identical 4 and 8 ohm drivers I can work out shading to follow the Legendre curve pretty closely with as many as 8 steps sparingly using a power resistor here or there. But maybe that isn't needed.
How would you guys approach passive shading of a 16 driver column? I have debated whether to treat it as a 16 driver arc (top driver down 40dB) or an 18 driver arc with top two dropped. Is there a straightforward wiring scheme for groups that I am overlooking?
The tweeter column will use 48 drivers, 3 for each woofer, and i assume that once I solve the woofer column scheme I can duplicate it fairly closely using 3 tweeter groups.
Thanks.
I posted a spreadsheet earlier in this thread showing several options for 24 driver shadings. The methodology can certainly be adapted for fewer drivers. Chances are you already have your own.
Typically the first two groups can be done w/o resistors; just by series paralleling differently per Keele papers and Griffin thread, then resistor dividers for remaining groups.
I would do 16 drivers as 8,4,2,2.
I would start from a larger array - 18 or 19 - and truncate down to 16 where the weights get small.
With reasonably efficient drivers the R's are no big deal except for the inconvenience of having to put them somewhere, which is less inconvenient and costly than the multiple amplifier alternative. Yes there is a slight frequency dependence of weight due to driver Z variation but its small if you use both a series R plus another R across the drivers to swamp their Z.
How many groups do you need? Jim Griffin was satisfied with 3. Keele has done 5. I doubt you need to do more. Vituix can model the weights and show you the resulting vertical directivity, allowing you to optimize/compromise sensibly.
Typically the first two groups can be done w/o resistors; just by series paralleling differently per Keele papers and Griffin thread, then resistor dividers for remaining groups.
I would do 16 drivers as 8,4,2,2.
I would start from a larger array - 18 or 19 - and truncate down to 16 where the weights get small.
With reasonably efficient drivers the R's are no big deal except for the inconvenience of having to put them somewhere, which is less inconvenient and costly than the multiple amplifier alternative. Yes there is a slight frequency dependence of weight due to driver Z variation but its small if you use both a series R plus another R across the drivers to swamp their Z.
How many groups do you need? Jim Griffin was satisfied with 3. Keele has done 5. I doubt you need to do more. Vituix can model the weights and show you the resulting vertical directivity, allowing you to optimize/compromise sensibly.
I read the thread and saw the spreadsheet mentioned but didn't actually see it. Will go back and look for it. Thanks.
Downloading vituix now to start learning.
I may use two amp channels for the woofer column simply for power requirements but not certain of that.
I think I will truncate the array. -40dB seems a waste. Even the next driver at -24dB is doubtful whether it contributes substantially to the steering. 18 truncated to 16 seems most reasonable.
Although my current xsim 8 segment curves do look pretty.
Thanks again.
Downloading vituix now to start learning.
I may use two amp channels for the woofer column simply for power requirements but not certain of that.
I think I will truncate the array. -40dB seems a waste. Even the next driver at -24dB is doubtful whether it contributes substantially to the steering. 18 truncated to 16 seems most reasonable.
Although my current xsim 8 segment curves do look pretty.
Thanks again.
Here is an update on my spreadsheet modified for 16 drivers.
I added the refinement of scaling the first group to exactly 0 db.
Truncating from 17 instead of 18 worked out better in that the second group came out to -3.57 db which is what you get using the Keele/Griffin series/parallel connections of two groups with 2n and n drivers respectively.
I added the refinement of scaling the first group to exactly 0 db.
Truncating from 17 instead of 18 worked out better in that the second group came out to -3.57 db which is what you get using the Keele/Griffin series/parallel connections of two groups with 2n and n drivers respectively.
Attachments
Thanks... I had started a very similar spreadsheet, glad to know I am on track. Played with a lot of variations, think I will settle on something a little different... 4, 2, 4, 2, 2, 2.
That is because I plan to split the column into two amp channels vertically to ease wiring complexity and minimize power burned through resistors. Also because I happen to have amplifier channels to spare but they are only modest power (~200w/4r). This split balances power between the two channels reasonably well and gives a pretty easy schematic when dropping gain on upper half -3dB.
That is because I plan to split the column into two amp channels vertically to ease wiring complexity and minimize power burned through resistors. Also because I happen to have amplifier channels to spare but they are only modest power (~200w/4r). This split balances power between the two channels reasonably well and gives a pretty easy schematic when dropping gain on upper half -3dB.
Hi all, Here's a few straight line vs cbt observations.....
Please take them as purely preliminary as it's surely too early to have any hard conclusions...
I've spent most of my time measuring/tuning/then listening to the straight arrays. This has been freestanding outdoors, and in room corners as well as freestanding out in the room a bit. Different rooms too.
I keep coming back to EQ correction that takes the speaker flat outdoors, and then uses different degrees of low shelving depending on whether in a corner, and or which room. I little parametric help can be needed in the corners, but other wise that's it for now. Sounds good.
Any attempt to tune to a narrow area indoors hasn't yet been as good as just doing the above.
For simple listening comparisons to the cbt, I use the same EQ settings that were made for the straight line. (This has been in mono, as I only have two driver baffles.)
First, comparing the cbt freestanding out in the room to the straight in a corner.....
As expected the corner straight has significantly stronger lower end. I could almost live with its bottom end, when I'm listening at low to moderate levels.
But I couldn't live with bottom end of the cbt. Without the corner loading the little TC9's just don't have the displacement imo. That's just not solvable with EQ.
The table kinda reverses comparing from mid range up. I've found the cbt easier to keep open sounding. It's cool how easy it adopts to different room positions. I should stop and say the only tuning variable in these comparisons is the amount of low shelving and it's freq point/Q.
(I'm also guessing the corner interaction with the straight isn't fully sorted out, which helps make the cbt comparatively easier....
One thing that's really cool imo is the difference in where the image comes from.
On the straight line, as many have said, it follows ear-level height, moving up and down the line.
With the cbt, it stays in the center of the arc, even when I stand right in front of it looking down on the arc.
It has a focal point! Jim, if you or any other cbt owners read this, do you experience the same imaging phenom?
I do wish the focal center was a little higher though...feels a little weird to have something so tall image only 3 ft off the ground.
Today, I've converted the cbt to a second straight unit, and have begun some stereo listening. Happy to say the baffle moved over easier than hoped for. I was sweating it a little, because now matter how careful I tried to be, getting the 16 baffle mounting bolts to line up with t-nuts and inserts in both boxes was tricky. Well, at least one side interchanges well
Gonna listen for a while, and integrate some subs.
Then will try two cbt's in stereo and with subs...
After taking a break to just listen for a while, I'll get back on the measurement trail...
Please take them as purely preliminary as it's surely too early to have any hard conclusions...
I've spent most of my time measuring/tuning/then listening to the straight arrays. This has been freestanding outdoors, and in room corners as well as freestanding out in the room a bit. Different rooms too.
I keep coming back to EQ correction that takes the speaker flat outdoors, and then uses different degrees of low shelving depending on whether in a corner, and or which room. I little parametric help can be needed in the corners, but other wise that's it for now. Sounds good.
Any attempt to tune to a narrow area indoors hasn't yet been as good as just doing the above.
For simple listening comparisons to the cbt, I use the same EQ settings that were made for the straight line. (This has been in mono, as I only have two driver baffles.)
First, comparing the cbt freestanding out in the room to the straight in a corner.....
As expected the corner straight has significantly stronger lower end. I could almost live with its bottom end, when I'm listening at low to moderate levels.
But I couldn't live with bottom end of the cbt. Without the corner loading the little TC9's just don't have the displacement imo. That's just not solvable with EQ.
The table kinda reverses comparing from mid range up. I've found the cbt easier to keep open sounding. It's cool how easy it adopts to different room positions. I should stop and say the only tuning variable in these comparisons is the amount of low shelving and it's freq point/Q.
(I'm also guessing the corner interaction with the straight isn't fully sorted out, which helps make the cbt comparatively easier....
One thing that's really cool imo is the difference in where the image comes from.
On the straight line, as many have said, it follows ear-level height, moving up and down the line.
With the cbt, it stays in the center of the arc, even when I stand right in front of it looking down on the arc.
It has a focal point! Jim, if you or any other cbt owners read this, do you experience the same imaging phenom?
I do wish the focal center was a little higher though...feels a little weird to have something so tall image only 3 ft off the ground.
Today, I've converted the cbt to a second straight unit, and have begun some stereo listening. Happy to say the baffle moved over easier than hoped for. I was sweating it a little, because now matter how careful I tried to be, getting the 16 baffle mounting bolts to line up with t-nuts and inserts in both boxes was tricky. Well, at least one side interchanges well
Gonna listen for a while, and integrate some subs.
Then will try two cbt's in stereo and with subs...
After taking a break to just listen for a while, I'll get back on the measurement trail...
Interesting observations. I'm interested in CBT for a mixed home theater/music room application. Some of the attributes seem a home run for theater... output/dynamics, smooth vertical response over a range of heights to cover multiple rows, -3dB per doubling distance to maintain better balance over multiple rows etc. ie many of the good characteristics of line arrays.
Other attributes are good for theater and music... smooth horizontal polar response, imaging.
But that perceived imaging height worries me. Ideally for theater the front stage speakers match screen center or close to it. Having audio center too far below screen center may be an issue. And for surrounds generally these are elevated a bit above listener level to reduce attenuation by seats/people behind. A CBT isn't designed to be elevated. I had hoped that the good vertical coverage would mitigate that somewhat.
I'm building a pair to experiment with. Will just have to see/hear for myself.
Other attributes are good for theater and music... smooth horizontal polar response, imaging.
But that perceived imaging height worries me. Ideally for theater the front stage speakers match screen center or close to it. Having audio center too far below screen center may be an issue. And for surrounds generally these are elevated a bit above listener level to reduce attenuation by seats/people behind. A CBT isn't designed to be elevated. I had hoped that the good vertical coverage would mitigate that somewhat.
I'm building a pair to experiment with. Will just have to see/hear for myself.
I'll have to say I haven't heard any image variation issues with my CBT arrays in my listening room. I'm 6'3" and I can walk around the room or sit and listen and hear excellent sound as I move. Plus I can bend and listen near the arrays and the imaging does not change.
Don Keele has done extensive tests which measure his prototype CBT units in-room and compare to a point source speaker. He was interested in near/far coverage, vertical coverage, ceiling coverage, horizontal coverage and the like. Somewhere on his website he has a vintage (2004) photo of himself and Floyd Toole standing on stools while listening to illustrate the sound coverage.
Unless the array is a full-up unit with proper design/shading and such I would not jump to conclusions that a CBT will have an image problem.
RHosch is welcome to hear my CBTs and decide about some of the design issues that he is weighing.
Jim
Don Keele has done extensive tests which measure his prototype CBT units in-room and compare to a point source speaker. He was interested in near/far coverage, vertical coverage, ceiling coverage, horizontal coverage and the like. Somewhere on his website he has a vintage (2004) photo of himself and Floyd Toole standing on stools while listening to illustrate the sound coverage.
Unless the array is a full-up unit with proper design/shading and such I would not jump to conclusions that a CBT will have an image problem.
RHosch is welcome to hear my CBTs and decide about some of the design issues that he is weighing.
Jim
Last edited:
Hi Jim, sounds like a misunderstanding...
I didn't mean to give the impression that I hear image variation issues with the cbt.
Really, I meant to give the opposite impression....that the image stays fixed in the center of the array's height, no matter what ear level height is.
I find that remarkable, and meant to be asking you and others if you experience the same phenomenon?
I think/hope I have the right design shading in place,... based on the Keele's papers, your recommendation, and nc535's spreadsheet... which all appear to be closely in sync .
Using 6 groups of 4 drivers each, I have the following shading in place:
0, -0.5, -1.6, -3.4, -6.2, -10.9.
Look ok?
Like you, I also find the cbt very consistent sounding as i move around the room...excellent really. Can't wait to get to stereo...step at a time ..
The qualm I expressed was simply that the image's focus height was lower that what I'd expect, looking at a speaker over 6 ft tall.
Not sure that's a problem at all
Although I can see RHorsch's home theatre concern, of matching the sound imaging height with screen height.
Looking forward to how you'd describe your image height...thx!
I didn't mean to give the impression that I hear image variation issues with the cbt.
Really, I meant to give the opposite impression....that the image stays fixed in the center of the array's height, no matter what ear level height is.
I find that remarkable, and meant to be asking you and others if you experience the same phenomenon?
I think/hope I have the right design shading in place,... based on the Keele's papers, your recommendation, and nc535's spreadsheet... which all appear to be closely in sync .
Using 6 groups of 4 drivers each, I have the following shading in place:
0, -0.5, -1.6, -3.4, -6.2, -10.9.
Look ok?
Like you, I also find the cbt very consistent sounding as i move around the room...excellent really. Can't wait to get to stereo...step at a time ..
The qualm I expressed was simply that the image's focus height was lower that what I'd expect, looking at a speaker over 6 ft tall.
Not sure that's a problem at all
Although I can see RHorsch's home theatre concern, of matching the sound imaging height with screen height.
Looking forward to how you'd describe your image height...thx!
Don't discount psycho-acoustic effects. I know the dialogue in my TV always seems to come from the speaker's lips even though my center channel is below the speaker. So groundplane CBT might also be convincing.
But have you considered straight floor to ceiling array? If you equalize it flat on a vertical angle corresponding to that of your seating riser, it should work extremely well.
But have you considered straight floor to ceiling array? If you equalize it flat on a vertical angle corresponding to that of your seating riser, it should work extremely well.
I did consider straight arrays and have not ruled that out as an option. I have played with line arrays in the past, would like to evaluate the CBT concept. There are two concerns I have about straight arrays in my room... For one, ceiling is going to have slopes and I worry about that interaction. Floor to ceiling array assumes a ground plane on ceiling just as floor, but maybe it wouldn't be such a big problem. The other is that I like the horizontal response of the CBT better as it helps self correct. A line array isn't much different than the a point source box of same width and driver complement that makes up the array. Aside from that, I've always found straight arrays to have a little "larger than life" sense to them, which is great for some things but not for everything. Curious to hear the CBT myself to compare.
For theater application need more LF output than the epique can muster, probably more than the CBT36 as well, and not fond of the tweeter used in the 36. So I'll build a sister to the 36 with a little more volume displacement on low end and little higher quality tweeter for the top end.
For theater application need more LF output than the epique can muster, probably more than the CBT36 as well, and not fond of the tweeter used in the 36. So I'll build a sister to the 36 with a little more volume displacement on low end and little higher quality tweeter for the top end.
^I like the way you rationalize your experience and plans! Prototyping is fun and it really gives you insight that is invaluable. But that is not easy and cheap with arrays, only dirty!
Line arrays are not popular here in Scandinavia, I haven't heard any full-size ones. But as option to cbt, I might try how a dsp-controlled frequency-shaded vertical line array would work. Make it sort of WMTMW by setting gradual low-pass so that only the center of the line goes fullrange. That should make it virtually point-source and focuse eg. speech in the middle-height.
Line arrays are not popular here in Scandinavia, I haven't heard any full-size ones. But as option to cbt, I might try how a dsp-controlled frequency-shaded vertical line array would work. Make it sort of WMTMW by setting gradual low-pass so that only the center of the line goes fullrange. That should make it virtually point-source and focuse eg. speech in the middle-height.
Last edited:
hey ya wesayso
I've been reading the DRC-FIR documentation quite a bit...I'm thinking that's what your post was referring to ?
The documentation is very well written, impressively informative.
The more i study it, the more i understand what any good FIR generating program is doing.
I've found in comparison to the FIR generator I'm using, FirDesigner (FirD), is that as you build the FIR filter parameters, FirD let's you see how well the FIR filter is working towards matching target goals.
FirD has traces that continually respond to every setting to show deviation from ideal response.
It also generates a 'real time' wavelet ..a FIR spectrograph... that continually shows the effect of the FIR filter parameters as the file is devised.
With regards to the neat FDW applied filtering in DRC-FIR, at first I thought that offered more control than FirD's fractional octave smoothing.
But I now think ultimately FDW windows and fractional octave smoothing, are both trying to do the same thing...give us the ability to decide how much correction to apply.
The cool thing about FirD, is that you can split the frequency domain into as many bands as you want, and apply different levels of fractional octave smoothing to each.
You can choose the frequency bands based on the error from ideal, that the program shows.
So you don't have to live with an FDW throughout the whole spectrum. Maybe you can split the frequency spectrum in DRC_FIR ?..I haven't got that deep yet..
Between spectrum splitting, and being able to observe real time error, as well as on-the-fly-wavelets, ....it seems at this point that FirD is at least equal if not a better mousetrap than DRC-FIR.
Plus...a biggie... I know FirD and get super results...so I kinda can't see investing in DRC..not yet anyway.
My guess is that the great results you have achieved may have more to do with passionate attention to room acoustics in combination to learning your FIR program very well, and not so much the program itself.
I feel I've held the same passion, only i don't care about rooms any more.
You might be right about about me, my room and how I use DRC-FIR. In relation to that, I've spend most of the time to optimize the settings specifically for use with the straight arrays I was dealing with.
In your case, I've said it before, indoors, you'd be better off with the CBT. Without addressing the draw backs of straight arrays indoors (the parallel walls) the CBT has more advantages due to it's positional independence. The way it blends within the room, if you're giving it enough space, you simply can't get the same result from a straight array without addressing the room.
Outside the straight array turns into a finite array, for me only inside use makes it interesting enough to devote any time to it. I don't know what a CBT does outside, I figure one of the draw backs is that the shading that eats away from it's SPL potential.
A note about it's low imaging, with a proper target for your DSP you can get around that. Balance is everything for both tonality and imaging. Well, to be more exact: balance and timing.
In your case, I've said it before, indoors, you'd be better off with the CBT. Without addressing the draw backs of straight arrays indoors (the parallel walls) the CBT has more advantages due to it's positional independence. The way it blends within the room, if you're giving it enough space, you simply can't get the same result from a straight array without addressing the room.
Outside the straight array turns into a finite array, for me only inside use makes it interesting enough to devote any time to it. I don't know what a CBT does outside, I figure one of the draw backs is that the shading that eats away from it's SPL potential.
A note about it's low imaging, with a proper target for your DSP you can get around that. Balance is everything for both tonality and imaging. Well, to be more exact: balance and timing.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.