So I built a pair of Super Malcom corner horns. SEOS 18/BA750 with 4 ns6's, and subs hidden down there (better ones to come there sooner than later). They kinda started off as prototypes but outperform the LSR-32's (by a lot, credit to everyone involved and Bwaslo espeically for the malcom idea...) so they are in service now, next I need to buy a 6 ch amp (thinking parasound hca 806), and sell the hodge podge Im using now (bryston 4bst on the inside woofers, EV7100's for seos and outside...very strange combo I know). Anyway they are setup 4 way active, SEOS/Inside Woofers/Outside Woofers/Subs. XO point is 700hz for the 18/750's and inside woofers. Sounds amazing, never heard a sweet spot (sweet area?) like this before... So I know about power tapering the mini-line array, or how Bwaslo did it with ala .5 way... and have messed with it some (they are naturally now by virtue of higher gain amp on inside woofers), but does anyone know about delay shading? I know you can do it, and have tried it...seemingly getting good results with about 50 samples @44.1 (so about 1ms) delay to the outside woofers. But there are a lot of variables because it seems to affect the FR a bit, so that has to be adjusted for before the results can really be compared. Also dealing with the different placement I'm getting some sort of SBIR or floor bounce I wasn't getting before so I'm trying to track that down. So far just from messing and testing this seems to work fairly well, that's outside woofers down ~6dB delayed ~1ms (no .5 style shunting). The mic there is about 5ft from the right speaker and about 3 ft out of center to the right and the yellow overlay is left and green right, the house curve is intentional. I don't *think* the delay shading is responsible for the dip @ 200hz, I think it's SBIR or bounce from my desk. The mic is essentially on axis of the left speaker (yellow) and maybe 40 degrees off the right. Considering 200 is close to Schroeder in this room it's very good and I can maybe clean that up with part of my soon to come sub project...but I have no idea what I'm doing with the delay shading. Basically entering random values and testing. So does anyone know anything about delay shading? Any articles or resources I could read would be greatly appreciated. Thanks-
Well it certainly works- Here is 10 sample increments at probably 40 degrees. This is inside so some modal stuff going on but you can see that the shading is happening. It looks like around 70 samples @ 44.1 is about as high as it can go before it starts to mess with the FR. Seems the woofers are breaking in a bit more and will require running the measurement for ARC again. I found a parasound HCA 1206 which should be here thursday so by then I plan to do some more serious retuning when all the amp channels match. They are sounding amazing...
Shading and delaying will primarily alter the directivity of the woofer array. Given your crossover point, it shouldn't affect on-axis measurements except where the directivity changes affect how the speaker interacts with its near reflections. That might be the cause of the 200hz dip when you narrow directivity. You are also likely getting some pretty funky results with -6db and 1ms of shading. In fact, you would probably be better off not running the outside woofers. I would shoot for closer to 1-2db and more like .1-.2ms of shading. With that said, I think Bill's method is better for this very short array. I would actually just use the shading portion of his crossover. I know you want to use active DSP, but there is no benefit over the passive shading combined with DSP for all else. Ultimately, you are going to have to find a way to measure the directivity of these drivers while testing different delay and attenuation if you want to go that route. It will be rather difficult to predict what will happen so trial and error is your best bet IMO.
Shading and delaying will primarily alter the directivity of the woofer array. Given your crossover point, it shouldn't affect on-axis measurements except where the directivity changes affect how the speaker interacts with its near reflections. That might be the cause of the 200hz dip when you narrow directivity. Right, what your looking at in the first post is on axis of left vs off axis of right. The 2nd post is about 40 degrees off axis (left), stationary mic, the only changes are delay in samples. Sorry if that was confusing. FWIW the room is well damped and other than a glass block window on the right (source of 200hz dip I found out...duh), the largest reflective slat is only 22" wide, so ~650hz or so. The ISD gap from the 1st reflections is around 14ms, other than the window for that position (not in sweet spot). Long story short, the measurements of interest in both posts are off axis... You are also likely getting some pretty funky results with -6db and 1ms of shading. In fact, you would probably be better off not running the outside woofers. I would shoot for closer to 1-2db and more like .1-.2ms of shading. With that said, I think Bill's method is better for this very short array. I would actually just use the shading portion of his crossover. I know you want to use active DSP, but there is no benefit over the passive shading combined with DSP for all else. Do you have more info or articles I can read on delay shading? My test shows more like .8-1.5ms. And that makes sense to me given the speed of sound @1.1ft/ms. I can't really see how 10ths of a ms would be enough relative to a 26" line source. But I admit total ignorance on delay shading, I've tried to play with EV's LAPS II, but I can't make sense out of it yet. And sure, depending out how it all pans out I might do some sort of hybrid active/passive setup. DACs and Monitor controller channels I already had and were unused, amp channels can be bridged so not really a financial difference in my setup, but ultimately might as well do passive for the WG and inside woofers at the least. Why funky? Even with no parametric taper ~.8ms should not cause combing in the pass band: http://www.sengpielaudio.com/calculator-combfilter.htm Even mild combing at the top relative to tighter pattern at the bottom might be worth the trade off, also off-axis of the woofer is probably not half space at those freqs...or I'm probably missing something totally obvious. Ultimately, you are going to have to find a way to measure the directivity of these drivers while testing different delay and attenuation if you want to go that route. It will be rather difficult to predict what will happen so trial and error is your best bet IMO. Yeah I figure there are a few givens. It would seem the outside woofers should probably be down 6dB at the XO point, no? So really there are only 3 main options with variables between that will likely be small audibly. Basically outside woofers parametrically tapered like Bwaslo did, + delay shade to match WG's, outside woofers not tapered with power and delay shading, and a combo of say parametric -3dB @xo, power shade -3dB, delay shade to match. BSC and house curve can be handled by audyssey. Sorry for the novel...