6 DIY eD A7S-650 kits with FP14K clone - Page 4

well then do 16!!! Then you could stop with subs, for like, forever

with his ported setup perhaps, but two sealed wont equal what you have now. Im running two sealed atm and im looking for more

Ricci now has his XXX18's in ~8cuft dual-opposed sealed cabs. If you have d2 XXX18's, bridge that 14k clone into one box and buy two more and another amp. And then double that. Yyyeeeaaahhhhh.

Actually it's 9 cubes not 8.


MK I guess I am confused as to what you are actually looking for? You should have plenty of bass system now to do whatever you want to do as deep as you need it to go. You did back in the day with the 8 ED 18's too. Are you unhappy with the way it sounds or something in the presentation? I'm all for trying out different things, but switching your system out every 6 months? Is it just for fun or are you still missing something?

We have an impostor on our hands.

MK, who stole your identity?


JSS

Isn't that air volume too small for dual 18's. I though the xxx needed at least that much per driver to get the most out of them.

Not when you got a ton of power, DSP and 8 of them

This.

That.



I dont need to get the most out of them due to what the other guys are saying. Im well below the limits at REF and past. Before I could push the drivers to the danger zone, xmax or more right at amp clipping in a much larger room. Now I put another driver in the same space the power per drive has dropped but total system power has increased. Sensitivity has increased and excursion is cut in half. At amp clipping the excursion is about 40mm max on peaks which is comfortable for these. So I gain 3-5dB extra headroom with less stress on the drivers. Yes the modeling looks bad, the qtc is really high and it is unrealistic to put enough power on it in an enclosure that small to use all of the stroke. Increasing the air space to 10cubes or more for each would be better technically. I dont have the space for it so compromises are made. IB is what these really want.

Got some measurements? What mic and measurement system are you currently using?

As I've suspected for quite a long time, our room gain profiles are quite similar. And, now that you're posting far more accurate measurements than many from the past, I can say it without being tarred and feathered.

The low end bump in your close mic is the measurement systems noise floor because the level is too low below 10 Hz. If you bump the level to 110dB at the top end, the roll off should flatten out. I also wonder if you're using the default Tukey window and whether or not you're using a SC calibration file, both of which will tend to skew the result down low.

As I've posted in the past, it's my belief that a loopback measurement of the SC does not necessarily (and in fact most times doesn't) reflect the SCs effect on the measurement because you take the analog input and loop it to the analog output. Since all electronics have DC blocking caps and they are almost always in the analog output stage, the loopback reflects that. But, when you take a measurement, you don't use the analog output. The ADC sends it in the digital realm to REW and the analog output is out of the loop. So, a SC correction file can add bump where it doesn't belong.

Also, back when, M Bentz razzed me about my ACO Pacific hardware being inaccurate below 30 Hz because he misinterpreted the dBA specs. But, during that discussion between the 2 of us, he brought up windowing. I studied the subject fairly intently as a result and have experimented exhaustively since then and found his point about Tukey below 20 Hz to be a valid one.

In any case, the graphs you posted are accurate enough to compare native vs at the LP in both our rooms, which I've done below. All green traces are yours and all gold traces are mine, overlaid on each other's graphs:



I believe that every room has a more similar than not room gain profile and that the 'pressure pot that begins at the longest dimension that can support a standing wave' theory is bunk.

You gain 2 advantages over me: 1) that your masonry walls and floor exhibit far less transmission loss and 2) you have no stairwell or windows and that gives you higher output with less displacement. I gain 1 advantage over you: my listening space in on a huge 27' x 27' wood structure floor system, supported in the center by a Lam-beam, that provides a bump at around 6 Hz. It's a pretty good trade and brings us fairly even, IMO and according to the posted evidence.

I wish you would download my SpecLab settings and get some SL graphs fired up. It would be interesting stuff.

Ricci,

What speakers are you using to keep up with all that woofage?

MK, his close mike is better than yours, just sayin you should do something about it...

Just kidding Bosso, and MK both yall have killer setups, Im working my way there slowly...

Beast, what you have is a killer setup. Don't sell yourself short. Anyway, if you want to solidify it's awesomeness, just double the subwoofage you have now. Bam! Easy.

Dave what do you mean about the tukey window? Ive looked into it somewhat as will so Im curious. I noticed little difference so long as the window period is of sufficient duration for the frequencies of interest which in this case would be really long.

It's about leakage, not length of time. Windowing is like subwoofer design... compromises to get the best result for the specific test. Single or multiple sine waves vs frequency response vs room resonance analysis, dynamic measurements, etc.

In most of the papers on the subject Tukey is rarely even mentioned. I just took the info I gathered and then experimented with the different choices, added the research and results and chose from there.

I was wondering if the CEA 2010 stuff addresses windowing? Seems to me it would be extremely important if what you're after is max output from a shaped tone burst.

Ping me if you would like some good reading material...

Dave,

I have done the same. Read quite a bit and do a lot of experimentation which is how I arrived where I am using what I do currently. (I don't really care which window is mentioned in which papers to be completely honest.) Most of the papers are more concerned with generalities and measuring response in the normal range >30Hz and or in some cases measuring static environmental noise or in noisey environments. Agreed on all accounts with what you said about compromises and having certain settings more applicable for different scenarios. I'm on the same page there. The type of window is only one variable among a great many. The window length is critical. You also have the size of the FFT blocks, the bit depth of the measurement, sampling rate, measurement noise floor,etc...I've got a few pretty good papers I dug up on the subject but I am by no means an expert on the theory behind it. I could always use more info. Can you link me in a pm?

What is it about the Tukey type window in particular that you have reservations about for LF measurements? I have not found noise or sideband leakage to be an issue in the VLF range myself provided that the measured signal is of sufficient strength. Frequencies below 10Hz are very long and take quite a while to propagate enough for a good measurement signal to be captured. They also tend to take much longer to die off as well.

Anyway we are discussing low bass measurements here. I use a 48khz sampling rate and I always use the longest duration sweeps available in the software that I am using and restrict the measurement bandwidth as close to the area of interest as possible. I can't remember whether I'm using 24bit depth or not. Anyway I always try to get enough measurement signal strength that the noise floor is not a concern. For measurements down to 10Hz such as outdoor I use a closing window length of 500ms. At that point you can switch the window type through most of the various popular windows and there will be negligible difference between any of them above 10Hz. Tukey 0.25 or 10ms are both close to rectangular so I typically use that. It won't matter which is selected really. If I am trying to look down into the ULF range I will set the right window to 1000ms or more which results in most of the popular windows having very little difference from each other again. It is fairly easy to look at the measured impulse and see the noise floor and where to be with the windows in order to capture the deepest bass. Since we are talking about measuring the deepest frequencies possible at the limits of our gear and systems it doesn't make a lot of sense to be limiting the resolution in those frequencies to me. As you increase the window size it will change the low frequency response appreciably. There is a minimum window length required to resolve each wavelength. 3Hz takes 333ms to complete one cycle. As the window duration is increased there is a point at which increasing it further will not change the measured results in any meaningful way. Where this occurs is a good indication that the measurement should be showing the deepest bass freq's with about as much resolution as can be expected. This usually lines up well with the impulse dropping down into the noise floor unless it is unusually high.

Again I am no expert in this area, this is just my layman's gut take on it. Maybe someone stronger on the subject can reply. What window types/lengths and other general settings do you use for your VLF measurements?

I believe that the window settings combined with too low of volume levels and exaggerated or misused cal files make it it difficult to discern what is really going on in most deep bass measurements that are posted. For example with the Omnimic when I see measurements below 20Hz they often look very smoothed, or low resolution to me but no one ever posts what windowing is being used on the measurement. These settings can easily cause major differences. MK care to share?
Edited by Ricci - 9/14/12 at 1:07pm

Hey Beast, you anywhere near the Greenville (S.C.) area??

"I was wondering if the CEA 2010 stuff addresses windowing? Seems to me it would be extremely important if what you're after is max output from a shaped tone burst."

the shaped tone burst in cea 2010 is a sine wave run through a hann window filter. as i understand it, since the pre-filtered signal is periodic and the hann filter forces the "tails" to zero, there is no reason to apply another window filter (unless you have some measurement problems, such as might be caused by reflections).

interestingly, it seems that the choice of 6.5 cycles was to capture 1/3 octave "slices", which makes sense in the context of having 3 measuring frequencies per octave. or put another way, the 100hz measurement in cea 2010 isn't really just 100hz, it is a blend of the the 1/3 octave centered around 100hz.

If a Hanning window is used, then of course there is no reason to apply another one. I was just curious which window they used.

Hanning (Hann) works.