Hi Jack, looks like it's coming along very well !
I'm not at all surprised that a logical step by step implementation of minimum phase EQs is working well.
One thing I think the shelving filters accomplish, is allowing our auto-EQ generators to work within acceptable magnitude-adjusting bounds.
I mean, we see in REW a warning when the auto-EQ doesn't like the level we chose. And my FIR generator, FirD, gives warnings when EQ requires adjustment that clips output at the design level chosen.
I think the substantial downward frequency response rollofff of the arrays almost demands shelving to be in place before our programs can make reasonable adjustments...
I firmly believe in tuning sub and array separately. Driver-by-driver tuning is almost my religion haha...and then combining the drivers/sections with linear phase crossovers, and it is my religion super haha.
Hey, the notch at 60Hz on your graphs looks so deep and sharp, I'm wondering if there isn't a polarity inversion between sub and array?
Again, looking like really nice progress
Can you make any comments yet, about the sound vs your synergy?
100% agree re' shelving and auto-EQ. I was surprised how well auto eq worked once I applied the shelving. Going to try DRC for the auto-eq shortly, build some more panels, and get a couple more moving blankets to drape the walls.
re' sub: I agree there also, just being lazy but also thinking ahead to a final room arrangement. In the garage, I can use my corner woofers up by the garage door and place the arrays well away from the walls and door in front of them. Then time alignment will mean something. In the living room, its going to be more of a multi-sub thing with the arrays tight to walls, as I'm trying to emulate now.
But 60 hz is a mode or boundary null. XO is at 80 Hz and I can reverse polarity there for a sharp null. From the step, I'd say my delay could be off by a cycle or half.
There is no high pass on the array (yet), just a low pass on the sub as I wanted as much overlap as I could get. OTOH, THD starts to rise below 200 Hz on my array so dynamics might benefit from a high pass. still a lot of things to try.
re' sub: I agree there also, just being lazy but also thinking ahead to a final room arrangement. In the garage, I can use my corner woofers up by the garage door and place the arrays well away from the walls and door in front of them. Then time alignment will mean something. In the living room, its going to be more of a multi-sub thing with the arrays tight to walls, as I'm trying to emulate now.
But 60 hz is a mode or boundary null. XO is at 80 Hz and I can reverse polarity there for a sharp null. From the step, I'd say my delay could be off by a cycle or half.
There is no high pass on the array (yet), just a low pass on the sub as I wanted as much overlap as I could get. OTOH, THD starts to rise below 200 Hz on my array so dynamics might benefit from a high pass. still a lot of things to try.
I built 2 more 2" panels and tried them out with the array both against the garage door and out in the room.
For the against the door case, the array was resting on a 2" panel resting in turn against 4" girders on garage door.
The responses with no sub and only shelf filters, FDW5:
In both positions, I had some near side wall boundary nulls that my 2" panels were ineffective against (not surprisingly). Against the door, there was some ripple 300Hz ++ which was due to the hurricane braces on the door holding the panel away from the door surface, mitigated by using multiple panels across the door. It turned out I didn't need the blankets on the door shown in the picture. I regret not taking a picture of the against the door case.
From a measurement perspective it seems like a tossup between the two positions, acoustically likely another story, which likely can be changed with ambience speakers.
I'm going to need to add a base to my array before I go much further with it out in the room, I already knocked it over once!
For the against the door case, the array was resting on a 2" panel resting in turn against 4" girders on garage door.
The responses with no sub and only shelf filters, FDW5:
In both positions, I had some near side wall boundary nulls that my 2" panels were ineffective against (not surprisingly). Against the door, there was some ripple 300Hz ++ which was due to the hurricane braces on the door holding the panel away from the door surface, mitigated by using multiple panels across the door. It turned out I didn't need the blankets on the door shown in the picture. I regret not taking a picture of the against the door case.
From a measurement perspective it seems like a tossup between the two positions, acoustically likely another story, which likely can be changed with ambience speakers.
I'm going to need to add a base to my array before I go much further with it out in the room, I already knocked it over once!
Attachments
+1 yes!!!! Which drivers were you thinking of using?.........Oh not a bass array but a base. You should use a bass array for the bass though
.I've looked at bass arrays. A decent one suitable for music but not quite HT would use 16 Faital Pro 5FE120s; a pretty small driver but its happy in a small volume so is compatible with my SB65array. If box size weren't an issue, this 6.5" peerless woofer has a lot more displacement and costs have as much -SDS-160F25PR01-08
but I don't think I'll need an array with up to 4 15" Acoustic Elegance woofers on call, and 2 12" on standby.
but I don't think I'll need an array with up to 4 15" Acoustic Elegance woofers on call, and 2 12" on standby.
I'll be very interested to see what you end up doing for bass support....whether you go with a column or not given your bank of 15"s.
Sometimes I feel like the real value in the floor-to-ceiling line array concept is for down low...where theory and practice and requisite center-to-center spacing are appear to be in harmony.
I worked on bass support and pretty pleased with the measurement at least when viewed at FDW5:
It definitely sounds good, glad I decided to ignore that 200 Hz null.
How I got there:
I moved the array back by the garage door, integrated the 15" corner woofer into the system and then lifted the front sub up onto the countertop when I realized the 70ish Hz mode was a ceiling reflection. Happy that didn't cost me my back.
EQ wise I retained the shelf filters in the miniDSP, went directly to DRC with the single measurement, then fine tuned the result with just 3 PEQs. Very streamlined process, much less work than manual eq and certainly good enough for this stage.
I got a surprise when I opened my default FDW5 window wider.
The 200 Hz first reflection null no longer dominated and a narrower null near 100 Hz appeared!?. Looks like I'm going to need thicker panels....
Murphy placed that first reflection point right on my window!
It definitely sounds good, glad I decided to ignore that 200 Hz null.
How I got there:
I moved the array back by the garage door, integrated the 15" corner woofer into the system and then lifted the front sub up onto the countertop when I realized the 70ish Hz mode was a ceiling reflection. Happy that didn't cost me my back.
EQ wise I retained the shelf filters in the miniDSP, went directly to DRC with the single measurement, then fine tuned the result with just 3 PEQs. Very streamlined process, much less work than manual eq and certainly good enough for this stage.
I got a surprise when I opened my default FDW5 window wider.
The 200 Hz first reflection null no longer dominated and a narrower null near 100 Hz appeared!?. Looks like I'm going to need thicker panels....
Murphy placed that first reflection point right on my window!
Attachments
Yes it does... is it there when you measure the sub by itself or the array by itself?
You might get away with some shifting of energy.
If it is pretty consistent off axis some DRC excess phase correction might help too.
Upload a raw IR and I'll apply my cocktail of correction to see what that does.
First you'd need to see what it does under different angles, can it be fixed with DSP or should it be left alone...
You might get away with some shifting of energy.
If it is pretty consistent off axis some DRC excess phase correction might help too.
Upload a raw IR and I'll apply my cocktail of correction to see what that does.
First you'd need to see what it does under different angles, can it be fixed with DSP or should it be left alone...
Its there all around. I measured in 2' circle around my primary LP and saw artifacts at that frequency in all positions but changing with distance from array and angle. I think my first step is garage door insulation; perhaps 3.5" thick denim or fiberglass over entire door if I can decide what to cover it with. Its the array, not the sub which XOs at 100 hz
Here is a zip of the .wav I passed to DRC
Here is a zip of the .wav I passed to DRC
The original input:
Here's a quick prediction after the DRC step(s):
A link to both steps within REW:
measurements.zip - Google Drive
Now all that would be needed is to apply some EQ and a target.
The frequency response in a 6 cycle view:
And a longer gated window of 25ms with Var smoothing (Which I'd use for EQ/target):
After applying a 25 Hz second order roll off and some quick EQ the 25 ms gated plot looks like this (Var smoothing):
See attached FIR filter, converted back to 96000. All my operations were run at 44100. Light EQ applied (all that I'd use), all that is missing now is a target (house) curve...
Here's a quick prediction after the DRC step(s):
A link to both steps within REW:
measurements.zip - Google Drive
Now all that would be needed is to apply some EQ and a target.
The frequency response in a 6 cycle view:
And a longer gated window of 25ms with Var smoothing (Which I'd use for EQ/target):
After applying a 25 Hz second order roll off and some quick EQ the 25 ms gated plot looks like this (Var smoothing):
See attached FIR filter, converted back to 96000. All my operations were run at 44100. Light EQ applied (all that I'd use), all that is missing now is a target (house) curve...
Attachments
Last edited:
With a target applied:
See attached FIR filter. A couple of warnings though, I see a couple of steep correction peaks at 110 Hz, 400 Hz, 16000 Hz and 17800 Hz. Handle with care. You could shave those down if this sounds decent at all. Same goes for the previous correction FIR file, as all I did with this one is add a target curve.
See attached FIR filter. A couple of warnings though, I see a couple of steep correction peaks at 110 Hz, 400 Hz, 16000 Hz and 17800 Hz. Handle with care
. You could shave those down if this sounds decent at all. Same goes for the previous correction FIR file, as all I did with this one is add a target curve.Attachments
Last edited: