BFD equalisation for the digitally crippled.

[russ.will]

I know there are certain limitations about posting graphs on this forum, but I should explain that all my computers were bought with a view to photo processing. Room equalisation wasn't even on the radar when I bought them, so I have to make the best of what I have and throw in a large dollop of patience.

If anybody on here is running Mac iBooks and REW and can interface to the midi input on the BFD, I'd be well interested to hear fom you.

Some backgound. I've chosen not to do a house curve just yet, because 1) flat is tricky enough for me at the moment, because 2) I don't have REW or computer with the necessary soundcard and 3) because my head hurts. I also seem to have ended up with a 6dB lift down to 25Hz albeit by accident rather than design.

A house curve is on the cards and I'm looking into ways of doing it, but I'm a bit worried about broad filter boosting frequencies below 20Hz because the Monolith doesn't have a high pass filter to protect it, like the PB-10 for instance. I suppose I could do a broad boost and then apply a massive narrow cut centered on 16Hz, or I could follow Waynes advice with a broad, deep cut centered on 300 to 400Hz. But that's for next week.

In the mean time, SWMBO was off researching her family tree. Back to 1732 (that's before we even founded you )and they've moved as far as the next village up the road - they're very local Therefore, I had some time to play.

The main problem I'm having with this very manual approach, is time. Adjusting a filter, running a sweep of test tones, typing the numbers into a spread sheet as you go, is slow. At least you see the graph change as it goes, so you know if you're in luck or you've cocked it up. But I'm limited to a 1/6th octave resolution and the troublesome frequencies aren't. AND each filter can have effects further up the frequency range, so you can't just sweep the bit you're trying to alter. No it's all 22, 10 second tones every time.

For instance, on the graph the peak at 36Hz. It's not, it's at 33.75Hz (according to the room mode calculator HERE). So I have to cross reference back and forth between the latest graph, room mode calculator, and ISO frequency charts.

Anyhoo, The yellow line is the Monolith with no BFD, the blue is last weeks effort and the red this weeks.

Peak @ 33.75Hz (Axial mode along the length of the room) is long gone,Peak @ 48.5Hz (Axial mode across the room) and the dip between '33.75Hz and 48.5Hz have been smoothed quite nicely. Dip @ 71.6Hz just disappeared while I was playing with the first two filters. I had being trying to target it previously with its own filter, but somethings working so I'll leave that well alone. The real sod is the peak around 100Hz. I think its because that it's a combined result of second and third harmonic resonances of the two main room modes all being near a lower order resonance of the axial modes. That and the fact the main speakers are dominant in output at this frequency. If I apply a big enough cut, no matter what the band width, to affect this peak it knackers things way down the scale. I can't EQ the mains except with Audessey MultEQ and I have no control over how it really works nor actually like the results of it being applied.

The short of it is, that I was mighty impressed with the Monolith after my MJ Ref 200. The thunderous nature of it's ablity to move the fabric of my house without effort was quite a revelation. But it's only now that I'm appreciating how deep it really goes. The major in room peaks I was suffering were all but masking lower volume, deeper sounds. I was feeling their energy, but now I'm hearing them as well. As I've tamed the peaks I've been turning the Monolith up. I've gone from 9 o'clock on the low level gain to about 10:30 but it doesn't sound louder, just much deeper and tighter. If you have a Monolith, you'll know that turning it up 1 1/2 notches equals quite a lot louder. I placed my hand on the door to the room at 22Hz to stop a rattle and was astonished to find I could feel it flexing under my hand - not buzzing but actually flexing with the pressure in the room.

If anybody has any suggestions based on this crappy graph, I'm more than happy to listen.

Now, where did I put U-571?

Russell

Edit: better graph attached. 02/06/07
Edit: Unfiltered response re-run -10dB. 03/06/07

[Sonnie]

Looks almost like you have somewhat of a built in house curve there already with A if you just cut 36hz by a few db.

I don't guess you can use Excel huh?

What about corrections... have you implemented correction values?

[JohnM]

On the issue of Midi interfacing with REW on a Mac, you will need the code from the Plumstone project, can get it here: http://www.cems.uwe.ac.uk/~lrlang/plumstone/

[russ.will]

Sonnie wrote:

Looks almost like you have somewhat of a built in house curve there already with A if you just cut 36hz by a few db.

I don't guess you can use Excel huh?

What about corrections... have you implemented correction values?

I have Excel at work, but at home I've only got Appleworks. All the values are corrected using the values from your Spreadsheet, which of course only works at work. I normally email myself in the values and then plug them into the Excel spreadsheet, but it's just a funkier, smoother looking version of the Apple one. I've managed all this with about 4 filters, So I'll use another to smooth the 36Hz value into the curve.

Any suggestions about the 100Hz spike? I've tried plugging the front and rear ports on the main speakers, but as expected this had made no difference.

Russell

[Chrisbee]

Hi Russell

I'll sneak in here while our American friends are sleeping!

You can't cut at 16Hz with a BFD. It only goes down to 20Hz. You can only affect VLFs indirectly with broad filters.

You don't need Midi to set BFD filters. It's probably just as easy to manually set them on the advice of REW [IMO].

I found that many iterations were ncessary to get the desired response curve.

REW tended to suggest far too narrow a filter to kill specific peaks. This was probably my fault for doing something wrong.

Setting one filter doesn't change anything in isolation. It affects frequencies other than those you are actually working on. The broader the filters you set the wider the effect.

The slow motion approach to sub testing needs the patience of a saint and you still haven't a clue what is really happening on live programme material. Been there, done that. (for decades!)

Have you considered getting a modern basic PC and SB Live! soundcard for home use? It would probably cost you less than your subwoofer these days and offer bells and whistles for other uses apart from AV testing. Do you have any family that needs a useful prezzy quite soon that you can "borrow" now and then?

Afterthought: If your firm is anything ike mine they are constantly replacing PCs. Get yourself on the top of the list for the next clear out.

[brucek]

I'll sneak in here while our American friends are sleeping

No fair using the fact that the world turns to get in ahead of us.....

apply a massive narrow cut centered on 16Hz,

As Chris says, the lowest filter allowed is at 20Hz, although it extends as any filter would above and below the centre frequency. On this same issue, I guess your concern is that you don't have a subsonic filter and you want to limit, so as to protect the sub? It looks like it's naturally rolling off, so not likely a concern.

Adjusting a filter, running a sweep of test tones, typing the numbers into a spread sheet as you go, is slow.........................

Again as Chris says - get a PC for pennies anywhere. My main computer is the latest and greatest, but my REW computer is a crappy old 350Mhz PC that I added some extra memory to. It runs REW and Excel fine. They're literally giving them away....... No one wants them and they are ideal for REW.

The real sod is 'D' around 100Hz.

I guess this is a sub + mains response issue you're referring to? Sometimes you can make large cuts in the sub at these higher frequencies, although by the time the sub is at 100Hz, the crossover has lowered the level so its influence on the overall sub + mains output at that frequency is much less.
Also you may attempt changing the phase of the sub. This has its affect around the crossover (which 100Hz is certainly a candidate). Of course if you had REW you could just flick the phase switch and watch its affect.

If anybody has any suggestions based on this crappy graph

Modify it to be ~45Hz to 105Hz.....

brucek

[russ.will]

Thanks for the advice guys. I was expecting a complete mauling, but I appear to be heading in the right direction and I've got loads of filters to spare.

Definitely going to play with the phase to see if I can iron out the 100Hz peak. I set the phase initially by playing an 80Hz test tone and wound the knob back and forth until I got the maximum reading from the SPL meter. This turned out to be 0deg.

Duh! Of course I knew the filter frequency filter is 20Hz, so lord knows what planet I was on suggesting 16Hz. But the info that filters still have effect below 20Hz, was something I had wondered about.

brucek - The graph is now much better, now I've been to work.

Many Thanks

Russell

PS. If you think it's annoying that the worlds spin favours the British, it's worse than that. The Australians always get there first.

[Sonnie]

I'm with these other guys on the computer Russell. I have a few myself ranging from a 350mhz up to an 800mhz that we no longer use. They are just sittin' around, monitors and all. Some nice computers but nobody seems to want them. Drive on over and pick one up.... lol. What I'm sayin' is I bet if you checked around other businesses you'll find a computer that will be perfect for REW and they will probably thank you for getting it out of their way.


Now is that blue line your original response and the yellow your filtered response? If so it appears to me you've cut way too much. I'd suggest cutting that area from 25hz to 36hz down to 95db and then work on 80hz and above to get it down to about 80-85db and you should look pretty good. You'll have to decide if you like it or not.

[russ.will]

Oops! I forgot to rejig the post to reflect the new graph. All the lines have had the correction values applied. The blue line is the original unfiltered response. The red line is where I stand right now after applying 5 filters (not 4, I checked). The Yellow line that has snuck in there is the filtered response AFTER rerunning Audissey for the heck of it. If you can make out what the dickens it's doing, good luck to you. Funny thing is, I did a before and after profile of Audissey before I had the BFD. It produced a curve that was almost exactly 3dB down on the unEQd line. Now the BFD's about it's all over the shop.

Russell

[brucek]

If so it appears to me you've cut way too much.

As Sonnie says, did you cut too much?

Or, perhaps you adjusted the volume control on your processor during the response test (a big no-no).

Either way, it seems as though you have a wholesale drop of about 10dB that just isn't necessary.

Perhaps a picture will explain what we are talking about.

See the blue and magenta lines at the highlighted area.
There appears to be no filtering there, but it's 10dB down. How come?

-----------------------------------------



-----------------------------------------

brucek

[russ.will]

Dagnabit. You've spotted the deliberate error. The first sweep was run yonks ago after I'd set up reference level to give 85dB from each channel using DVE. It was so very loud for running test tones I reduced the master volume by 10dB which is actually about the level I listen at. All BFD filtering has been done at this level ever since. Hence the -10dB drop at 16Hz where I haven't applied any filters either. I figuered as I went along that the shape of the lines were exactly the same where I hadn't applied any filters, so it was the shape of the red line I should work on rather than the absolute levels.

I'm just off to try playing with the phase adjustment, as you suggested earlier, to see if it affects the 100Hz peak. I'll run an unfiltered sweep while I am about it as this is obviously causing you some discomfort.

I will look into the cheapo PC option, although it's bad enough dealing with the devils handywork 9 to 5. I'm not sure I'm ready to turn to the dark side at home.

Russell

[brucek]

You've spotted the deliberate error

Busted.........hehehehe.....

[russ.will]

And I'd have got away with it too, if it hadn't been for you pesky kids.

Just re-run the unfiltered sweep. Surprise, surprise it's virtually the same except for, 1- it's 10dB quieter and 2- the 71Hz (approx) dip has lifted reatively by 0.5dB.

Tried playing with the phase and got all manner of weird things happening up and down the range, but the 100Hz peak refused to budge by more than a couple of dB. Until I've done as I've been told and blagged a PC, or bought an external sound card, I'm quite happy with the results for now. Or until the weekend.

Russell

[Sonnie]

Something else you might tinker with is the crossover. I'm not sure how well you Kef's will handle, but you might try measuring with a 70hz crossover point and a 60hz crossover point.

[russ.will]

Good one. They measure pretty flat to about 40Hz in room. They're currently set to large with a 60Hz crossover which I'd completely forgotten. I'll try them with a higher crossover at the weekend.

Russell