BFD for full-range room equalization

[gsmollin]

Is the Beringer Feedback Destroyer, model DSP1124P, a suitable processor for full range room equalization. It appears to have capability to perform this function, yet I haven't seen it being used for this purpose.

If you have another suggestion, I would be interested in that too.

[Sonnie]

Welcome to the Shack, gs!

From what I understand... no... it would not be a suitable unit for most people on higher than sub frequencies. Although I have actually heard (read) of others using it to their satisfaction.

The DEQ2496 might be a consideration. Rane makes some good products but they are very expensive.

[gsmollin]

The DEQ2496 should work as a HT room equalizer, but it's a heavy hammer for that application. On the other hand, if I were sound man for a road show, it would come in handy with it's own 1/6 octave RTA, graphic and parametric equalizers, and feedback destroyer.

What about the FBQ2496?

[Sonnie]

Well it has better specs than the 1124p and should be better, but I'm still not so sure it would be clean enough.

[gsmollin]

The common specs are almost identical. THD, sampling rate, and word length are identical. Dynamic range is slightly less on the FBQ2496, but that corresponds to the lower signal level in the FBQ's I/O. If you derate the maximum signal level in the DEQ to match home theater levels (-10 dBV), the dynamic range is the same as the FBQ.

Of course the DEQ has a host of features lacking in the FBQ that justifies the higher cost. The digital I/O, for instance, would be just great except that I don't have a digital processing loop.

Thanks for your input, Sonnie. I wish somebody else had some experience and would add a post.

[Sonnie]

I'm not that familiar with the technical aspects of either unit. I know Ken (brucek) has a very solid technical review of the 1124 that is linked in the BFD Quick Links sticky thread. That might help you some. I'll also see if I can get him to respond here. He may be able to help.

These are from the manuals...

1124

• Bandwidth 20 Hz to 20 kHz, -3 dB
• Noise > 94 dB, unweighted, 20 Hz to 20 kHz
• THD 0.0075%
• Crosstalk <-76 dB
• Converters 24-bit Sigma-Delta, 64/128-times Oversampling
• Sampling Rate 46,875khz

2496

• Frequency Response <10hz to 44khz
• Dynamic Range 107db
• THD 0.007%
• Crosstalk <-100db
• Digital Processor Ultra-high resolution 24-bit/96 kHz A/D and D/A converter powered by a 32-bit DSP
• Sample Rate 96khz

EDIT: OR Wayne!

[Wayne A. Pflughaupt]

Is the Behringer Feedback Destroyer, model DSP1124P, a suitable processor for full range room equalization. It appears to have capability to perform this function, yet I haven't seen it being used for this purpose.

People have tried it, and the general consensus that it it’s not suitable for hi-fi reproduction, due to its poor-quality AD/DA converters.

I believe the DEQ2496 has better converters, so it might be an option. My only beef with it is that since it’s a digital product, it adds another AD/DA conversion to the signal chain, which I find unappealing. I’d be much more inclined to try it with a fully digital connection, between the DVD player and receiver.

Regards,
Wayne

[brucek]

I completely agree with Wayne. The 1124 is unsuitable. The 2496 is a better device, but only when used at full line level or as Wayne suggests, in digital input mode from a source such as a DVD or CD player.

brucek

[gsmollin]

Well, OK, nobody thinks its suitable. Since I already have one equalizing my subwoofer, I decided to remove it from the downstairs HT, run some tests on it, and then try it as a full-range equalizer in the fully analog upstairs stereo.

Bench tests that I was able to run, given time constraints and the nosey technician staff who make themselves scarce when I want them to wind a transformer, but were all over me as soon as they spotted the FEEDBACK DESTROYER PRO on the top of the black box, were transfer function tests, and distortion spectrum from a 1 kHz sine wave.

Transfer function control, as measured by the network analyzer, was pretty impressive. I set the filters for 1/3 octave, and alternated -12 dB and +12 dB, beginning at 63 Hz and ending at 16 kHz. This gave a very cute comb filter response. There were a couple of small errors I noted, such as only hitting +/- 8 dB instead of 12, however that may have been filter interaction. I didn't get a chance to try a single filter boost or cut because... my supervisor comes out into the lab (normally I might not see this guy for a week) and wants to know if I'm testing the new Solartron analyzer. So I have to explain what a BFD is... He wanted to know if he could use one to cut heterodyne howls from SSB receptions (radio amateur), So I programmed a preset with a 1/60 octave 48 dB cut at 1 kHz. It was quite an impressive notch, and had no overshoot in the skirts. He moved on.

Now I was able to get on with distortion. I put a -10 dBV signal at 1 kHz in. The 1 kHz filter was set for +12 dB, so the output was at -2 dBV, and the yellow LED was on. I measured a single 3rd harmonic line at -83.5 dB (.0084%). All other harmonics were in the noise floor. Unfortunately, I had a 20 dB attenuator in the input probe, and a 20 dB attenuator on the channel amplifier to keep analyzer hamonic distortion at a very low level. This meant my noise floor was largely analyzer noise, and I didn't get to sniff the noise floor of the BFD with the attenuators set to 0.

I also ran the same test with the filters set to flat, and got the same result.

I took it back home and plugged it between the preamp and power amp in the stereo. I discovered I had to set the input switches to +4 dBu, and even then I couldn't turn up the volume too much on the preamp. My preamp normally puts out in the +20 dBu range. I did a few A/B comparisons with the BFD plugged in, then plugged out. It was clear that the BFD added an audible hiss, which I could hear with the power amp's gain control turned all the way up and my ear cocked towards the tweeter about 3 feet away in a quiet room. With the gain control turned down to a normal listening level, I could not hear the hiss anymore. With the BFD plugged out, there was no audible hiss, period. I left the BFD in the stereo for the weekend, to see if anybody complained about it, but all they did was ask what that thing was perched on the power amp. Wherever it goes, the BFD seems to generate interest.

I am hoping to get a chance to finish the noise floor testing on the BFD Monday, early, before the techs come in. I'll post my findings.

Impressions to date are that the BFD is slightly noisy, and that would limit its applications. It may be a special problem with some HT equipment that operates at a low level, and become objectionable. Distortion seems to be very low. As part of the noise testing, I hope to do a low-level distortion test with all the analyzer attenuators at 0. I do not anticipate a problem with the 24-bit delta-sigma A/D, since its quantization errors are about 140 dB below full scale, assuming it is working correctly in the BFD.

[Wayne A. Pflughaupt]

It may be a special problem with some HT equipment that operates at a low level...

All home gear operates at a low level compared to pro equipment, so I expect the high level of background noise is a problem that everyone will experience.

Regards,
Wayne

[gsmollin]

I forgot to add that I switched the I/O level to -10 dBV, and the noise reduced to the point where I could no longer hear it in my tweeter. That indicates that the noise issue should be no worse at -10 dBV than it is at +4 dBu. However, that does leave it audible. I will measure the exact noise floor levels for both positions of the switch.

[gsmollin]

I got a 90 dB dynamic range, unweighted, 20 Hz-20 kHz. The noise floor switches up and down 12 dB with the level switch on the back. There is no noise advantage to using the high level setting, except that it may better match your signal level. The noise is pretty flat broadband, with a 1/f region starting at 4 kHz. There were some line spectra at 60 Hz, 120 Hz, and 180 Hz, but I did not pursue that because I was running the test single ended, so I couldn't be sure it wasn't a ground loop. I did not see any switching feedthrough in the 46 khz range.

The spec claims 94 dB, close to what I got, but disappointing because it measured worse, not better than spec.

I also conducted a low level distortion test, looking for anything in the A/D or DAC that would have been sensitive to that. I used a 1 kHz, -54 dBV signal. There was no harmonic distortion visible above the noise floor at -114 dBV (10 Hz RBW). This is not surprising, since 24 bit systems should be accurate below -140 dBV, but it is nice there wasn't a problem.

Behringer's other offerings claim better dynamic ranges, but the improvements are not large.

[Josuah]

I've tried a few pro GEQs on my L/R speakers for full-range EQ. They always introduced way too much noise. Not a problem for subwoofers which can't reproduce the higher-frequency noise, but a huge problem for the mains.

I also don't like the idea of inserting another set of DACs into the signal path, after I paid a bunch to get good DACs in my processor to begin with. You could go completely analog on the EQ though.

Another problem is that there's way too many frequencies once you leave a subwoofer. And ideally you should use PEQs because 1/3 octave is too coarse. I might give a go at things with the FBQ2496, but not interested in doing that right now. Building acoustic panels instead at the moment, for my surrounds.

[Wayne A. Pflughaupt]

Just curious, Joshua, which pro equalizers did you use for full-range?

Regards,
Wayne

[Josuah]

I tried a dbx 231 and an Alesis DEQ830. The dbx 231 is analog, 2x31-band GEQ. The Alesis is digital 8x31-band GEQ. I wanted to try an Ashly because supposedly those are top quality, but it was too expensive for a trial type thing, I think. dbx and Alesis are lower-end, in the opinion of the pro audio guys I talked to. But much more affordable for me. Now I have a Behringer FBQ2496 that I could try, although Behringer is also considered lower-end to my knowledge.

The DEQ830 was not too bad, if you could accept a noise floor. Which is certainly acceptable in a concert venue for example. But not for my home theater. People will less sensitive speakers might not have a problem.

What I actually did with the DEQ830 was use my Yamaha RX-V1400's YPAO to figure out the amount of boost or cut at the 1/3 octave center frequencies. Then adjusted the DEQ830. Then re-ran the YPAO. Eventually I got YPAO to only have one or two filters, even though it could apply up to seven, with the remainder in the DEQ830. My cats actually seemed to pay attention to the speakers for a short time after I made this adjustment, when sitting in my lap.

The 231, however, introduced way way too much high-frequency noise, proportionally increasing with the amount of boost or cut.

I wonder how the software managed DSP-based equalizers might fare, since doing things that way should theoretically not introduce noise. But those are even more expensive. And you still end up with the problem of an extra set of DACs. On the other hand, I suppose that must be what's in the mixing studios, so I guess it shouldn't actually be a problem.

[Wayne A. Pflughaupt]

Interesting info, Joshua. I kinda expected they were lower-end EQs, but didn’t want to say anything and maybe shoot myself in the foot if you came back with “BSS!”

Your experience confirms what I’ve long told people, that you really shouldn’t cheap out on an equalizer that you’re going to use for full-range in a home stereo, especially if the rest of the gear is high quality. I’ve also been leery of digital EQs, because of the extra digital/ analog conversion factor.

I wonder how the software managed DSP-based equalizers might fare, since doing things that way should theoretically not introduce noise. But those are even more expensive.

I don’t see any reason to use them, really. The digital auto EQ in late-model receivers is the way to fly, IMO, except perhaps for the lower end receivers where it might not work as well. The signal that’s being equalized is digital to begin with, and once it’s in the digital domain, you can do just about anything you want – equalize, delay, etc. – with none of the penalties of analog processing.

By the way, if you really want analog EQs for a good price, you might keep an eye out on eBay for some of AudioControl’s discontinued C-131 third-octave equalizers. They have much better specs than the best pro models – we’re talking $1500 + equalizers here. They were a good deal at their list price of $500, and at eBay prices in the $200-250 ea. range, it’s practically criminal!

Regards,
Wayne

[Josuah]

I hadn't heard of AudioControl before. I'll keep my eye out. Thanks.

Unfortunately, for higher-end processors, I haven't heard anything really good about auto-EQ except for TaCT or Audyessy (Denon). I've heard rumors of Audyessy coming out with a standalone product which might be interesting. The research paper is certainly persuasive. But these systems are just too expensive for me, except maybe the Denon. Which actually might be a good pre-processor choice now that you've made me think about it. Of course TaCT does more than just EQ.

Right now I'm using an Emotiva DMC-1 and MPS-1, and this was a significant audio improvement over the Yamaha RX-V1400 and Outlaw 7100, despite no EQ. Although Audioholics benchmarked the Yamaha RX-V2600 as better than the DMC-1 in terms of frequency response and THD, I don't recall seeing measurements on DAC accuracy.

[Sonnie]

The Audyssey Sound EQ is beginning to ship I believe... but MSRP is $2500 ... not to mention they are pushing for custom install. There is a dealer install kit that cost $325 that is not supposed to be sold to customers. I don't expect you'll see a host of big discounts on it either considering the low dealer margin.

[Josuah]

Hm. Well, $2500 might be worth it if it does multiple channels with enough parametric filters, but a custom install requiring that changes to the room brings out another dealer visit would definitely kill the product for me. Do you know where any additional information on this can be found?