Guys,

I wrote a small review of my Behringer Feedback Destroyer DSP1100P "BFD" earlier in the year. I've dragged it out and edited and updated it a little. You might be interested. It has since been replaced by the newer model DSP1124P.

Personally I wouldn't use a sub if it wasn't equalized - it's a must.

Stryke Audio has a rather good test tone CD that can be purchased at the web address below. Click on test CD's.

www.stryke.com

I made my own CD from a tone generator, but this one at Stryke looks good.

Well, I've had this device for about 6 months and even though it's already been discussed quite a bit on most forums, I thought I would give my opinions and concerns on its effectiveness in my HT / audio system.

It was a result of many posts on DTF and the AVS board that convinced me to try out this device in my HT system. I certainly got lots of great advice from people who were already BFD users.

For those few who don't know what the heck this thing is, below is the Behringer web site, but I'll give a mini explanation because simply reading the web site may still leave some people wondering..

BFD Website

Generally, this piece of equipment would be used by semi-professional studios and more frequently by anyone who sets up a P.A. system at a concert. Ever wonder why that huge wall of speakers directly behind the stage microphones doesn’t cause massive amounts of feedback. Well, this device stops that feedback automatically. No need to discuss this any further because it's the other feature that this unit has that many people in home HT and audio are using it for.

It has two separate (stereo) channels consisting of 12 parametric filters each, that can independently be used to tame subwoofer(s) response. If you've ever done a frequency response using a simple Radio Shack SPL meter and a 1/6 octave tones CD on your system from 20Hz to 100Hz you'll know there is a big problem mostly caused by either - your sub(s), your sub(s) placement, or mostly your room itself.

Myself, I have two different subwoofers, stacked and corner loaded. This seems to be the best location after much movement. But still, after doing a full frequency system response from 20Hz to 20KHz, I find the area from 20Hz to 80Hz to be very poor, with lots of peaks and valleys, as much as 15db difference. What happens of course is that this 'hot' peak frequency dominates, and you tend to adjust your sub level to it, which means the other frequencies are reduced to a much lower level. The goal would be to smooth out the response, so that all the frequencies from 20Hz to 80Hz are about the same level (depending on the sound you like) and the low end response will become very smooth and even sounding.

Enter the BFD. Each of the two channels has 12 parametric filters that can be individually controlled via the front panel in frequency, bandwidth, and gain. You can adjust the gain from -48db to +16db, but realize that increasing the gain of a particular frequency has a corresponding decrease in your subwoofers dynamic range. So you hope you only have peaks and not too many valleys to tame. You can adjust the filters to just about any frequency and its bandwidth is adjustable from 1/60th of an octave to two octaves.

The unit has both balanced and unbalanced inputs and outputs. Myself, I have the luxury of using the balanced inputs, but I use unbalanced out to my two subs. The unbalanced outputs are realized by plugging in a mono 1/4" to RCA converter plug you purchase at Radio Shack for $3. The unit senses the unbalanced output and adjusts its level accordingly.

I especially like the feature of being able to bypass the units' filters with a single button. It allows you to quickly compare your work.

You can save your filter settings in any one of ten recallable profiles. Normally, it's best to use Program 4 or 5 in the BFD because these are already set up as "PA" (parametric) filters.

I have to admit it to be a rather daunting task (as I'd been warned) to equalize two different model subs, but once you get going it isn't too bad. I set up 4 Excel graphs on my computer and entered the appropriate data, first un-equalized, so I would have a reference baseline from where to start. The 4 graphs were all from 20Hz to 160Hz and were:

(1) Sub-1 alone

(2) Sub-2 alone

(3) Both subs together

(4) Both subs with main speakers.

This allows you to visually see each subs response on a graph and then how they interact together and then how the total system response is, including the pre/pros bass management. It quickly became obvious that the two subs response was very similar and was being affected mostly by the room and not their individual characteristics.

For my tones I used a CD-R I made of every frequency from 10Hz to 80Hz and then 1/6 octaves from there to 160Hz. It's useful to have every frequency in that low area because it allows you to narrow in on exactly the offending frequency between octaves so you can dial the BFD to that frequency. Sixth octave measurements are generally enough though. I used a Radio Shack SPL meter at my listening position for the readout and compensated for its inaccuracies.

I feel I'm getting quite close to where I'm satisfied, but I can see the tweaking going on forever. The difference in sound is quite remarkable. Like others told me, you first think you don't have the bass turned up enough because you're so darned used to hearing the 'hot' spots. I was very successful in lowering all my peaks without any trouble, including ones that were a fair distance past my pre/pros crossover. There is still lots of signal going to the subs at 60HZ to 100Hz even though I have them crossed over at 60Hz. My worst room mode was at 80 Hz. The mains are getting 80Hz but so are my subs, and lowering the sub response at that frequency will drop the room response very nicely. My 4 graphs really helped out. Each time I changed a filter I would completely re-enter the data for the 4 graphs to see the overall effect. I filtered my subs right out to 125 Hz even though they're crossed over at 60Hz. The sound is very smooth now.

As suggested I didn't equalize from 20Hz to 160Hz absolutely flat. I did a slow rise to the lower end. It ain't ruler straight yet, but it's darn close, especially compared to the roller-coaster graph of the un-equalized system.

I'm a bit of a fanatic (read nuts) when it comes to noise floor and minimum crap in my audio chain. It's the reason I have the slightly unpopular pre/pro that I own (not enough bells and whistles to satisfy the appetite of the 10.2 crowd I'm afraid).

When I first read about this BFD I scoffed and said, "In a million years I'm not putting a gizmo like that in my audio chain". And I still say that about my 5 main channels, especially my sacred two analog main channels. But, in a subwoofer chain, I decided it was worth the price of admission, which is incredibly cheap for what you get. If I didn't like it, no big loss, it also makes a swell door stop.

After receiving and using this device and then taking it apart and doing bench measurements on it, I have to say it is the bargain of the century when used as a sub EQ, although there are a few problems I'll discuss later. I took it to my lab at work and did some tests mainly in the low frequency range. There were several things I was interested in checking out. I was interested in how well the ADC/DAC conversion system responded to very low volume inputs. No doubt the unit handles optimum line level inputs fine, but, does it move into distortion with very low level inputs which will certainly be the case since this unit isn't being fed a full line level (as from a CD player), but instead it's a signal that is fed from a volume control where sometimes we might be listening to very low volumes with full line signal only being in the tens of millivolts. ADC's tend to suffer when presented with this situation. Also, I wanted to know the low frequency response range. No sense using subs that go less than 20Hz and limiting their low end input with a piece of electronics in the path. How well does this unit respond to 10HZ, etc... Also, I wanted to see if the filters introduced any noise or distortion, and generally how accurate was their unity gain. And I wanted to see what type of phase shifting the unit caused with and without filters. All of my testing was generally only at frequencies of interest (0Hz to 200Hz).

The specs appear to be as good as or better than advertised.

Output signals remain very linear even with a very low input level (a problem with ADC's) across the full audio band and doesn't distort until you're slightly above the red clipping led. Using the BFD -10dBV sensitivity setting, the lowest LED turns on at 17mvpk and the yellow LED turns on at 1.2vpk. The red LED comes on at 1.5vpk and clipping occurs at 1.6vpk (which generally corresponds to their claimed max input level of +2dBV). The 4th green LED is about 300mv. The -10dBV nominal switch setting is certainly the sensitivity setting that should be used. The output, (after ADC/DAC conversion) tracked the input quite nicely down to 10mvpk. The lowest level LED isn't even on at this point – so it’s quite good in that respect. The filters’ gain / frequency / and bandwidth is very accurate. 0dB gain is indeed unity on all filters that I tested for all the frequencies that I tested, although I was mostly playing with the lower spectrum.

Frequency response appears to be greater than spec'd. The lower end continues to be linear until at least 5Hz which is certainly as far as I care about. The output level drops off below that, but continues easily to 1Hz. I'm certainly not too concerned about digitizing my low frequency chain. The sampling rate is 46KHz. At these low frequencies, that's a heck of a lot of samples. The 20 bit samples are certainly enough.

My HP distortion analyzer only measures accurately to 0.01%. I definitely read this or lower and this is in line with their claim of 0.0075% THD. Again, who cares in the sub chain - this exceeds any requirement I have. The noise floor was as good as claimed, and although good, I wouldn't put their 94dB noise floor in my 110dB S/N Bryston main chain. But for a sub, it's perfect.

Now, for my problem with this equalizer and any other equalizer as well. Phase shift. My field is electronics - I know very little about sound, so others can discuss this issue and its effect. Here's the deal. As I expected, with no filters on, the output exhibited a progressively greater and linear phase shift from 20Hz up to 200Hz. This is the area I was interested in. I won't type out all the results, but at 20Hz there is a 5 degree phase shift increasing in a linear fashion to 200Hz where the shift has grown to about 72 degrees. Actually, it increases all the way up to 20KHz at a linear rate. At 500Hz there is a full 180 degree shift and at 1000Hz it's back to 0 degrees. This repeats on and on. The reason for this is the DSP (digital signal processor) in the BFD, like any DSP, has a fixed processing time. It happens to be about 1 msec. This causes a fixed delay of every signal that passes through it and will exhibit the same result as a phase shift, because as the frequency increases, the processing delay remains the same. So, since the speed of sound is roughly 331m/sec or around 1000ft/sec, then the 1 msec DSP processing delay in the BFD would account for about a foot in distance. If you add a foot to the distance you tell your receiver that your speakers are away from your ears, then it will advance the sound the 1 msec required for that speaker. What does all this mean? When you use a BFD, add a "foot" to the value you enter into your processor when you tell it how far the listening area is from the subwoofer.

Of course, when you add in my filters, as expected, the phase differences aren't linear any more. There is some major phase changing going on as I move from 20Hz to 200Hz. Here is an example of my 20Hz to 100Hz phase data: [20Hz=7deg],

[30Hz=56deg],

[40Hz=175deg],

[50Hz=56deg],

[60Hz=28deg],

[70Hz=35deg],

[80Hz=50deg],

[90Hz=12deg],

[100Hz=12deg].

Now, since I have succeeded in creating a system response that is within a few db all the way from 20Hz to 200Hz, I guess I've correctly compensated for the phase problems, but it's interesting stuff anyway.

There is a FAQ on the BFD that is quite informative and well written at:

Behringer FAQ

I do want to caution that in my opinion there are a few inaccuracies to be aware of. I'll repeat these below with my view on them....

********************************

FAQ: If I am feeding the BFD a mono signal, should the filters be coupled?

ANS: While some have debated the benefits of linking the left and right engines in series, I have always coupled the filters for subwoofer use. It gives you more potential gain and cut for each filter, although it limits the number of filters to 12. If you absolutely need 24 different filters for smaller tweaks at different frequencies, then putting the engines in series might be of some use, but the signal will go through the DAC's twice, which could degrade sound quality. Also, I have yet to have a situation where more than seven filters were required to balance out the sound.

MY TAKE: The couple mode does not give you any more or less gain and it does not link the engines in series and it doesn't go through the DAC's twice. Couple mode allows you to set the independent left and right channel filters at the same time with only one entry. In "not coupled" mode, you have to set both left and right channel filters independently. Of course, this lets the left and right channels have different settings, but either way, both channels are completely independent and ADC / DAC processing is done independently and only once. “Couple” is handy if you are driving two similar subs and feel that you'll be setting all filters the same for each sub. In couple mode, you only have to enter a filter once and it will be copied to the other channel. In single mode, you'd have to enter it twice, once for each channel.

******************************

FAQ: Will the BFD process signals under 20 Hz?

ANS: No, the BFD has a digital brick wall at 20 Hz for all the filters. It will pass any information below that through, but without any filtering, even if there is a substantial boost right at 20Hz.

MY TAKE: It does not have a brick wall filter at 20 HZ. It will indeed pass signals down to 5Hz quite nicely. The lowest value filter that can be set is indeed 20Hz, but affects the bandpass normally as set. I.e. if a 20Hz filter with

-5db @40/60BW is set, then 15Hz is suitably reduced.

******************************

FAQ: What position should the input switch be in?

ANS: Still researching. I've had it in both successfully, although one will input a much higher voltage and possibly clip the DSP's.

MY TAKE: Correct, but the -10dBV sensitivity setting would always be used for home - line level equipment. This will ensure the system is receiving the spec'd "nominal" level. The spec indicates max input and output levels of +2dBV at the -10dBV sensitivity setting. This is quite accurate from my bench tests. +2dBV translates to 1.26vRMS; the yellow LED comes on at 1.2vpk and the red LED comes on at 1.5vpk and clipping occurs at 1.6vpk. Use the -10dBV setting. The +4dBu setting will result in nominal voltage swings far too low for the ADC system to respond properly.

******************************

I do have a few beefs with BFD that I'll list below:

The front panel lights are quite bright. If you used all 24 filters you would have 24 'really' bright red LED's shining at you, not to mention the program number and an assortment of other bright green LED's. This may wow the bells and whistle crowd, but they need a dimmer or need to reduce the current through their LEDS. I give them full marks for panel layout though - nicely done.

The garishly large white lettering on the top panel advertising that you have a Feedback Destroyer Pro has got to go. My cabinet slot for it is a narrow height, so it's hidden, but if it's on an exposed shelf, yuk.

My unit had a small mount of mechanical transformer hum. Typical stuff, but since I've paid a lot for a really quiet system, it kind of bugs me to hear a transformer humming. I removed the top panel and lined the contact area with rubber to see if that would help, and it did, but not enough. I then removed the power supply transformer and rubber mounted it to the case and that made the sound completely go away. Most sane people will not need to do this mod.

In keeping with the cheap price of this unit they have provided no soft turn on. A relay delay circuit would likely add 15 bucks to the price. That's about 10%, so it didn't happen. The result is quite a horrible pop when you turn the BFD on in any down line speaker, so this unit needs to be turned on before it’s down line amplifier. You'll maybe need an AC sequencer or easiest yet, leave it turned on. It draws about as much current as a couple of night lights.

It will also produce a ground loop hum if you use the three prong plug. This is standard fare whenever you insert a grounded unit in the middle of a chain like this. The fix is to use a two prong 'cheater' plug and that problem is eliminated, although a 'cheater' is never a permanent fix and not ever recommended.

Anyway, I bought my unit at a local sound shop. They had a ton of them. Look in the Yellow Pages under Sound Systems & Equipment. These kinds of places rent, sell and install sound and lighting systems for bands.

Add a cheater plug and a few Radio Shack adapter jacks and you're ready for some interesting stuff.

I want to comment on subwoofer output level from your receiver/ preamp if that's ok. When setting up the input levels to the BFD you have to remember it isn't a normal fixed line level situation. Consider the output of a CD player for instance. It is a line level that has a varying output depending only on the music being played, but it still has a fixed maximum line level output that isn't passed through a volume control. This line level would be very easy to match to the input of the BFD because it's predictable. This unfortunately isn't the way we're using the BFD. We are feeding it a line level signal that is affected by our preamps volume control. See the problem? I can play a DVD at a very low level or a very high level and the BFD will be fed different "maximum" inputs. One of the problems with ADC's (analog to digital converters) is that they tend to become fairly non-linear and rather ineffective at low input levels. They operate best if they are fed a signal which has a maximum that is just below their maximum input level. This ensures that the maximum signal does not clip and the quietest signal gets properly evaluated and assigned the appropriate low order bits to correctly represent its actual level when converted back to analog by the DAC. This is where quantization error (noise) exists, at these low levels.

Your goal then, is to decide what the probable maximum listening level will be for your volume control of your receiver or preamp while playing a bass heavy disk like U-571 or TPM and then set the "subwoofer output" level of your receiver to where you see the red LED's lighting up on the BFD. Now back the level off slightly. The BFD does not clip when the red LED turns on, but it does clip with a signal slightly greater than that. I've done quite a few bench tests on mine and I found the red LED turns on at about 1.5volts peak and the output begins to clip at about 1.6volts peak (slightly under +2dBv). This is the case when you use the operating level sensitivity switch set at -10dBv (which is the only setting you should use).

The final setting, is one where the red LED doesn't come on, even when listening at maximum levels for your system. This may be reference for one person and below reference for another. This may be an iterative approach, coming back to this setting after you've lived with it a while.

Anyway, once you have this subwoofer output level set, then you know the maximum sent to the BFD will be a maximum signal that it can tolerate and you'll know normal levels will be in the linear region. Now you can set the output level of your subwoofer with its own volume control to whatever level you enjoy.

Note that the BFD IN/OUT switch will enable and disable your filters with its’ light "ON" and "OFF" respectively. In these two conditions the LED indicators are showing the BFD's output level. But, if you push and hold the IN/OUT switch and it begins flashing, you bypass the filter section completely and the LED indicators are showing input level. This is the condition to set it up in and that I am discussing above.

With regard to filters, it depends on how high and how wide a peak is that determines how the filters are entered and how many you use for a particular peak. Generally, if a peak isn't too wide then you would only use one filter and adjust a bandwidth to remove it. Remember the bandwidth adjust is from 1/60 (very small) up to 120/60 (2 octaves). That's a lot of adjustment. You don't want it too wide or you'll lower frequencies you don't want to lower. Test after each filter change you enter and then readjust. Always start with the lowest frequencies first and work your way up to higher frequencies when entering filters.

brucek