Preparing for REW+BFD

[HClarkx]

Fast response ...... I hope you are getting some work done ... I need to soon.

I didn't follow your 20, 40, 60, 80 ... are these the choices on the sub variable phase control?

This leads me to ask one more question, can REW provide phase or "distance" at each of an array of sample frequencies?

Bruce, I'm sure you are right that if an auto EQ room correction system falls short, it's going to be below ~100 Hz and more often than not, rooms will have severe problems down there.

In a bit of experimenting with sub locations with a higher WAF rating, I did find some with a range of +10 to -15 db under 100 Hz that would take some powerful EQ to solve. And, frequencies at the peaks and valleys were different at each seat and no EQ system can solve that.

Harrison

[Trekari]

I have an analog phase control on each sub, variable from 0 to 180.

It simply has some dashes to represent where you are on the phase 'scale.' So the first big dash is 20 degrees, the second is 40, the third 60, etc. In between the big dashes...aw, here:

http://av123forum.com/showpost.php?p...&postcount=141

[brucek]

This leads me to ask one more question, can REW provide phase or "distance" at each of an array of sample frequencies?

You simple take multiple REW sweep measures from 0-200Hz with the sub and mains playing, and adjust the phase of the sub(s) until the crossover area provides a smooth transition.

Or you wait for the next version of REW in beta testing right now that has an RTA feature and adjust the phase live with the RTA running.

I did find some with a range of +10 to -15 db under 100 Hz that would take some powerful EQ to solve. And, frequencies at the peaks and valleys were different at each seat and no EQ system can solve that.

It's not a perfect solution, but REW allows up to 8 sweeps to be averaged to produce a new response, that you then create your filters from for the BFD, and this gives a fairly good response over the measurement area.

brucek

[HClarkx]

Trekari,

It just occurred to me that the dip you are worrying about in the 90-95 Hz region is right about where a crossover might fall. What is your mains-subs crossover set at?

If there is any phase problem between your subs and mains the dip should be wider, but still one has to wonder.

[Trekari]

I had the xover on the fronts at 80Hz, and the sub crossover was set to 180 actually.

After my measurements of each front speaker/center on Friday I will set their xovers to where each makes sense and integrates with the subs the best.

[Wayne A. Pflughaupt]

HClarkx wrote:

I've seen your thoughts on this but haven't seen anything proving that the room decay is fully offset by the filter time response (i.e., that the two match).
Also, can you refer me to tests that confirm the time components added by the filters correct room decay characteristics?

I haven’t seen any comprehensive in-room testing, other than Ethan Winer’s EQ vs. Bass Traps extravaganza. The test has some problems, though, mainly that the sheer number of measurements presented is mind numbing, and the guy who set up the equalizer did a really bad job of it.

Ethan also did a simple evaluation of the effectiveness of the Audyssey MultEQ, which claims to reduce ringing. His conclusion was that it was a mixed bag as far as dealing with extended low frequency signal decay, aka “ringing” or “modal ringing.”

Since there seems to be a dearth of in-room time domain testing, other than these two, I’ve done some of my own (although I haven’t posted any results). After studying waterfalls ’til I’m bleary-eyed, I’ve found pretty much the same thing as Ethan, that actual in-room measurements show that it’s a mixed bag using an equalizer as a fix for extended low frequency decay issues.

I compared multiple waterfalls (as alluded to here, you won’t get a true picture from a single sweep) of equalized response using what I’ll call “modal filters,” that REW recommended to improve ringing (which is what REW does when you run the “Find Peaks” routine), to the “smoothing filters” I like to use (i.e., simply EQing out peaks and nulls with no regard for any effect on ringing). What was apparent was that the modal filters improved ringing for some peaks REW had identified, but for others there was no improvement at all.

Even then, the improvement waterfalls did show for modal filters was only apparent in the short-duration 300 ms window. When the window was lengthened to 600 or 1000 ms, any advantage modal filters showed over smoothed fully vanished. The overall trend was that modal filters had the effect of quickly attenuating decay in the short term (compared to smoothing filters), but only down to a certain level. In the long term, both sets of filters were showing comparable results (this despite the fact that most of the smoothing filters were set for boost!). Which is to say, there was no reduction in ringing.

That’s one thing I have noticed here at the Shack where people (including myself) have posted graphs showing that an EQ filter improved ringing: They’ve all been short-duration 300 ms windows, not long-duration.

You can clearly see this “fast attenuation but only to a certain level” effect with the first set of comparison graphs from Ethan’s Audyssey evaluation - note the region between 56 and 20 Hz.

Ethan and John both maintain that narrow filters have the greatest effect on time domain. That’s one thing I have yet to evaluate at length, but intend to.

Regards,
Wayne

[Trekari]

Well, it turns out that BJC was exceptionally slow in shipping my order.

Placed 2am Tuesday morning, is only supposed to ship out today.

So unless I get the desire to do manual measurements, I won't be moving my PC into the living room until (at BEST) Friday, possibly Saturday, depending on when they arrive.

I plan on doing sweeps of all 7 speakers individually with and without Audyssey, so those of you interested in the benefits of Audyssey will want to stay tuned for another sample.

Secretly I'm also hoping that my in-room non-Audyssey sub response is a bit more suited for a house curve than the relatively flat post-Audyssey graph. So I have my own reason for running the comparisons, hehe.

[HClarkx]

Wayne A. Pflughaupt wrote:

Ethan also did a simple evaluation of the effectiveness of the Audyssey MultEQ, which claims to reduce ringing. His conclusion was that it was a mixed bag as far as dealing with extended low frequency signal decay, aka “ringing” or “modal ringing.”

I recall the Ethan Winer monolog on his comparison of Audyssey and room treatments. Didn't Chris Kyriakakis of Audyssey take issue with his results and quite conclusively demonstrate that the Audyssey FIR filters do address decay? The Audyssey filters are FIR so it would seem they should.

It occurs to me that I said something stupid earlier about filters compensating room decay. One may want to adjust room decay but one doesn't want to eliminate it. Doing so would befuddle sensitive ears that expect a "real" room. I wonder though whether a filter that corrects a given peak or dip in frequency response will necessarily be altering the decay at those frequencies appropriately. I.e., is there a correlation between a dip in frequency response at a listening position and the room decay at that frequency. I'm thinking not. A reflection from a wall/ceiling near the sub can combine with the direct wave (as in Tekari's case) and cause quite a response dip, but that doesn't mean that there will be additional reflections that continue bouncing around the room at that frequency (that reflection may have come via the only solid surface in the room). If I'm right, any decay correction wouldn't necessarily be appropriate.

. . . . I’ve found pretty much the same thing as Ethan, that actual in-room measurements show that it’s a mixed bag using an equalizer as a fix for extended low frequency decay issues.

Seems there's an opportunity for some definitive comparisons of Behringer, Audyssey, and other approaches.

I wonder though, the ear must deal frequently with variations in decay. Decay may not be much of an issue if the ear/brain takes care of it. Unless maybe the decay problem is severe. I mean a boomy room is a problem, but beyond that maybe decay isn't much of an issue.

. . . . . What was apparent was that the modal filters improved ringing for some peaks REW had identified, but for others there was no improvement at all.

I think it's significant that you didn't find REW's recommendation to degrade decay. That would definitely be a negative result. At least if the improvement ranges from none to some, the results are in the right direction. But, fixing decay implies we know what decay is appropriate. Maybe we should to go london and measure the Philharmonic hall and then adjust our own decay to match it when we are listening to the London Philharmonic. Or maybe not, doesn't the recording already provide that effect? So what is the target? I think I've seen "desirable" HT decay times mentioned so presumably there or optimum ones (do they vary by music/movie type?). Can we actually hear the difference if the decay is in a credible range?

That’s one thing I have noticed here at the Shack where people (including myself) have posted graphs showing that an EQ filter improved ringing: They’ve all been short-duration 300 ms windows, not long-duration.

My guess is that if the ear/mind can be troubled by inappropriate (?) ringing, it's probably the longer duration ringing that is problematic.

You can clearly see this “fast attenuation but only to a certain level” effect with the first set of comparison graphs with Ethan’s Audyssey evaluation - note the region between 56 and 20 Hz.

I do see that. It would seem likely that the lower the frequency, the more difficult it is to control ringing. There's an awful lot of energy bouncing around to be controlled by a speaker cone. I mean after the bit of energy is sent out, additional energy has to be emitted to counter the ringing. Can a filter that only needs to make a small correction in frequency response at a certain listening position be expected to help much with a monster ringing problem at that frequency.

Ethan and John both maintain that narrow filters have the greatest effect on time domain. That’s one thing I have yet to evaluate at length, but intend to.

I look forward to that. I may download REW, connect my Audyssey mic, and see what happens when I turn my Sound EQ on and off.

Best,
Harrison

[Wayne A. Pflughaupt]

I think it's significant that you didn't find REW's recommendation to degrade decay.

Well, since REW only recommends cutting filters, there isn’t much chance that would happen! However, I did notice some sonic degradation with the modal filters (which were more numerous and narrower than I like to use). The uneven bass notes I complained about in the Hard Knee article returned.

Decay may not be much of an issue if the ear/brain takes care of it. Unless maybe the decay problem is severe. I mean a boomy room is a problem, but beyond that maybe decay isn't much of an issue.

To my ears it isn’t, at least not like long decay in the upper frequencies is.

Can we actually hear the difference if the decay is in a credible range?

Supposedly you can with bass traps (although I’ve never heard them).

Here are a few more graphs, this time from Ethan’s Traps vs. EQ face-off (the ones added after the fact). Again, the equalizer was not adjusted properly, which may account for why we don’t see the “fast attenuation” effect. But one thing's for sure, looking at the Traps graph, I don’t think anyone can say they got they get that kind of reduction of decay from equalization (although I don't think I want 17 traps in my living room!).

Regards,
Wayne

[HClarkx]

Convincing enough. Do I read these as running out to 518ms? I vaguely recall 600ms being an appropriate decay rate. Does that ring a bell?

The 18-50 Hz range looks to be several seconds with traps and a bit less with the EQ. Higher frequencies are brought down very quickly by the traps and only modestly by the EQ. I wonder if the more uniform decay rate across the 18-322 region provided by the EQ isn't preferable to the longer decay at the low end and very short decay above about 50 Hz in the traps option. Does Ethan discuss that? If my recollection of 0.6 seconds is right, the EQ looks better. But, maybe that .6 isn't right.

I think I'm lucky. My room is large, well broken up, lots of random absorbing surfaces, few opposing sizable parallel surfaces, etc. But, I'll check it with REW as soon as I master it. Right now it's telling me my "input" is too low and I can't get it up.

If Audyssey does as well with decay as it does with frequency, I should be in good shape. Below are my before and after sub plots averaged across 10 measurement locations in a three-seat bubble. Each seat tested individually looks about the same before and after, so even though these are averages for the 10 measurement locations, they are representative of each seat. Vertical divisions are 5 db. Horizontal range is 20Hz and up.

Before:

After:


Harrison

[Wayne A. Pflughaupt]

I wonder if the more uniform decay rate across the 18-322 region provided by the EQ isn't preferable to the longer decay at the low end and very short decay above about 50 Hz in the traps option. Does Ethan discuss that?

Don’t think so, but if I know Ethan, he’d prefer to have truncated decay all the way down to 0 Hz! One thing he did note, I think it was at an after-the-fact discussion thread at AVS, was that he didn’t use the correct traps for low frequencies – another sad shortcoming for what should have been a stupendous experiment. IIR most bass traps absorb to 70-80 Hz or so, but you can see the ones he was using were pretty much “done” by ~140 Hz.


Trekari, we’re not hijacking your thread here, just talking amongst ourselves here until you get back to us with your sweeps.

Regards,
Wayne

[HClarkx]

I've read that listening in a too-dead room can be disconcerting. I know being deep in a cave is an interesting experience. Presumably any room echos are on top of anything provided in the recording, so are "extra" but still expected by the ear/brain. This book provides some perspective. Note especially the top of page 26, though all of section 2.7 (starting on page 24) is interesting.

http://books.google.com/books?id=JIu...iTTF0#PPA27,M1

Harrison

[Trekari]

I don't know where to start with this kind of problem :'(



The above graph is both subs active, phase 160 on the Right, 0 on the Left. No mains.
The Right sub Xover is set to 100, the L Sub is set to 80

No Audyssey, No BFD Filters.

Now:



The above graph is both subs set to Phase 0, no mains, direct input to Sub-In on the amplififer, no BFD, no Audyssey.

[Trekari]

So...in order to remove that nasty dip at 81Hz, Ihad to move the sub up onto my couch. Does that mean it is my ceiling that is creating that null?

For instance, moving the R sub onto the top of the LSub produces:

[HClarkx]

The phase numbers are very strange ... is this plot the result of trying to optimize the curve by changing them? That shouldn't be necessary with your sub locations. If there is any benefit to a phase difference, that difference should be only a few degrees, maybe 5 or 10 at the most.

What does the plot look like with both subs set to 0 degrees?

Harrison

[Trekari]

The most recent 2 graphs in the previous post were Phase 0

[HClarkx]

You inserted a new plot while I was typing ......

Re the new plot ....

It doesn't look like you removed the notch, just moved it up 10 Hz.

If you get that dip with both subs set to 0 phase (assuming no wiring errors) and in their correct positions (near the mains in your case) and working together, then it's probably there with either sub working individually (though that's worth confirming) and you are going to have to fix it with the BFD. I.e., it's a room problem that isn't going to be fixed with any reasonable sub relocations (unless you have some other locations that you'd be happier with from an aesthetics standpoint ... in which case you might want to try those locations).

Harrison

[HClarkx]

Both phase 0 but with one on the sofa, right?

I don't think there's anything to be gained for you in moving the subs other than to desirable locations (next to the mains as before or further to the left and right).

I'd limit my testing to those positions.

Harrison

[HClarkx]

Limit your phase adjustments to a few degrees one way or the other, but only for fine tuning (I mean very fine tuning). At this point if the phase adjustment helps the notch, it's probably going to create problems elsewhere so is not a good solution to the notch (and surely won't help measurably in any event).

Harrison

[Trekari]

I suppose I should start with the basics.

Here is a plot of the LEFT subwoofer fed directly, Xover on the sub is 180, no mains, completely solo.



Now here is a graph of the R sub solo, no mains, Xover set to 180 also.



BOTH graphs show the subs in their aesthetically ideal position (where you've seen them in the photos) The only caveat is that the R sub is not on its spikes because I was moving stuff around and they make it harder

Here is a graph of both subs, Phase ZERO, playing together.

[HClarkx]

That's better.

Both subs together look better than either alone. That's the way it's supposed to be.

You do have that dip between 80 and 90. If audyssey doesn't take care of it, you can address it with BFD.

Hopefully Audyssey will minimize and smooth out the other minor deviations as well so that you can use the BFD just to create your house curve.

Please do try the bass tone control in your receiver. You may find it does an excellent job of beefing up the lows to your preference. If it's like mine, it will. Mine pretty much compensates for the low-volume ear-sensitivity problem all humans have.

Harrison

[Trekari]

Now, here is the LEFT channel included with both subs, Double-Bass in ON in my receiver, and the FRONT channels are set to full range.



Now let's turn Double-Bass off, leaving the front speakers as Full Range. NO change to the subs, still both set to Phase 0.