Hey Conrad.
I ran those REW sweeps you requested. I did a measurement for each of the 6 subs by themselves, and also 2 additional measurements for each of the 2 sub pairs. They all have the acoustic timing reference, but when I was setting up the sweep parameters in the dropdown box you mentioned I only had options for L and R channels (no C or surrounds, so I'm assuming I need to change a setting or something in REW), so I used the main R speaker as the timing reference source in the base layer.
In the Dropbox file I shared (link below) I also included a measurement from 20-18,000 Hz with 1/24th smoothing applied to make it more readable. I was just wondering if anything in the higher frequencies jumped out at you as an issue, specifically in the 500-8,000 Hz range.
I did have a dumb moment though when I was adjusting the sweep's volume in REW... It's usually set at the default -30dBFS in REW and my AVR set to -10dB. I must have had a brain block because when I decided to change it in REW to -20dBFS I typed it incorrectly and it reverted to 0dBFS instead and for the life of me I couldn't get it to change (until I figured out how dumb I was being after the fact). So while I was having that fail I just decided change the AVR volume down to -30dB instead. So I said all that just to say that the overall volume for that 20-18,000 Hz sweep was a little bit different compared to the sweeps that I ran for the subs.
In your previous post you had said, "At the end of all this you'll have two virtual subs: sub one will be the four main subs, sub two will be the pair of VNF subs." I may be getting ahead of myself in this process, but if you could elaborate a little on doing the VNF subs separately... How would I go about doing that in the minidsp and REW? Or would I be utilizing the 2 sub outs on the AVR, each going in to a separate minidsp? I was just a lil fuzzy on the specifics of it.
Shared with Dropbox
www.dropbox.com
Thanks for this, very useful. I've attached my version here. You can get a nice response with no EQ, just using time alignment.
This is just using time alignment, all using one minidsp:
Here's what I did (are you ready!?)
- FL and FR subs, as we know, integrate well so I just used the combined response you already measured.
- I tried the FL and FR with the center but I couldn't get it to work.
- Instead I added the rear sub with a 1.5ms delay on the rear. This actually gives a really nice response. I could have stopped there but as you have six subs let's integrate six subs
- Adding in the center gives you a sharp null at about 106Hz but that likely won't be audible. As there's no timing change you can play about with having that sub or not, just watch the levels. More on that later.
With that combined four sub response I EQd it, exported the EQ'd response and re-imported it so that we now have a measurement called (F + R) + C-EQ -- (Fronts, Rear, Center, with EQ)
- For VNF we know we want to handle things differently. Your VNF response is pretty poor, it's rolls off hard at 40Hz. Those subs really aren't adding anything but it's fine, we can EQ them flat and you'll get all that chest slam from 40-100Hz in the back.
- The combined response of the two VNF subs can't be improved with timing, so I just EQd it quite hard and got a nice rising response. I exported and re-imported the EQd VNF subs as a new measurement.
Combining those responses (four subs, plus the two VNF) with a bit of timing gives this:
Orange is the four main subs, blu is the VNF subs, green is all combined.
So the VNF and the main subs have different EQ and different responses but they don't interfere with each other.
In your previous post you had said, "At the end of all this you'll have two virtual subs: sub one will be the four main subs, sub two will be the pair of VNF subs." I may be getting ahead of myself in this process, but if you could elaborate a little on doing the VNF subs separately... How would I go about doing that in the minidsp and REW? Or would I be utilizing the 2 sub outs on the AVR, each going in to a separate minidsp? I was just a lil fuzzy on the specifics of it.
Here's the chain:
AVR output 1 - minidsp input 1 - minidsp output 1 - Y splitter - front left and front right subs
AVR output 1 - minidsp input 1 - minidsp output 2 - rear sub (1.5ms delay)
AVR output 1 - minidsp input 1 - minidsp output 3 - center sub
AVR output 2 - minidsp input 1 - minidsp output 4 - Y splitter - VNF Subs (7ms delay)
So the front L and R subs are a "virtual" sub on one minidsp output and the VNF subs are also a "virtual" sub that take up one minidsp output. I think this is what you already had to be honest.
It also gives you independent trims in the AVR for the farfield subs vs the VNF subs. Changing the level of the VNF subs doesn't have much of an effect on the overall response so play about until you're happy.
I've attached the EQ for outputs 1, 2, 3 (FF Subs) and output 4 (VNF subs)
It's worth going step by step to ensure that what you get at each stage matches the predictions in REW.
As for levels, if you turn that center sub off the same EQ should still be good but the 106Hz null will disappear. The level will drop by a dB or two though, so you might need to trim adjust.
I'll be interested to see if, once this is applied, you still need to run the same levels. You're running your subs 15dB hot which is absolutely nuts. I'm wondering if it's not SPL you're craving but ULF. Your response is rolling off at 30Hz which is pretty high.
After all that, as a complete curveball, you might be better off selling the six and getting a pair of ported 18" monsters from the likes of JTR, HSU or PSA. If you put them in the right place they'll blow you away! Smaller driver subs are only ever going to be able to play so low, so loud. More drivers doesn't give you a lower response. It'll give you more output with less distortion and a smoother response with minimal EQ, but if you want a response that hits the teens you need bigger drivers.
Being in the US I'm guessing you're on concrete, which means to get the room wobbling you need ported. You have lots of options. Just a thought.
Here's the MDAT I worked in:
Integration - Conrad.mdat
I had a look at your full range response too, nothing in there to worry about except maybe that spike at 200Hz. Other than that I wouldn't EQ. Is that post Audyssey?