Home Theater Forum and Systems banner

1 - 20 of 25 Posts

·
Registered
Joined
·
1,018 Posts
Discussion Starter · #1 ·
Hi

You have your REW response curve up for your sub and it shows a fairly decent curve but with a couple of deep troughs due to room problems.

So you boost your sub's overall gain to lift the bottoms of the troughs to your target curve level. Then you cut back the response around the troughs with the BFD.

How does this differ from trying to fill in the the troughs themselves with BFD boost filters?

Thanks
 

·
Registered
Joined
·
7,514 Posts
How does this differ from trying to fill in the the troughs themselves with BFD boost filters?
It differs quite a bit. By introducing any gain internal to the BFD lowers its headroom and increases noise. Theoretically at maximum input level, basically any gain is not allowed if you don't want to clip at that frequency.

You want to take full advantage of every bit available in the BFD. The bit resolution defines the dynamic range of any digital system and so the signal to noise ratio is maximized by using as much of the dynamic range as possible. If the range from your weakest signal to the strongest signal was only half the required maximum input level, you would be robbing yourself of dynamic range and subsequently worstening your signal to noise level.

The BFD1124 is a 24bit device. This means the device theoretically can resolve 16777216 discrete voltage steps (translates to ~ = 6.125x24=147db). Well, that's the theory, but another 1124 spec is that the noise unweighted is >94dB. That's a little better than 15 bits of resolution. So about 9 bits (of the 24 bits) are lost in the noise - you have no contol of this and it's fixed. That leaves ~15 bits to define the input voltage levels or a possibility of 32768 voltage steps as long as you supply the input with a full scale signal.

You can't supply that full input signal if you've introduced gain in the BFD or you'll clip the signal at the frequency where you added the gain, so you must subsequently turn down the master input level (of all frequencies) to compensate and so lose dynamic range - and so on with more increased gain. Remember that the voltage signal feeding the BFD isn't subject to all those peaks and troughs you see on the REW screen. The level is perfect into the BFD.

The subwoofer amplifier on the other hand likely has quite a bit of headroom and since it's an analog device offers far more elegant clipping than a digital system does. When an analog signal just reaches clipping, a small amount of distortion will be generated as the peak flattens. The more signal we push, the distortion increases - not so awful really. This isn't the case with digital. It's fairly unforgiving, proceeding to maximum distortion at the initial clip level.

Anyway, there are quite a few reasons not to use too much gain in the BFD with the filters, specifically in those frequencies before the crossover area. After that you can get away with it a bit more.

brucek
 

·
Registered
Joined
·
1,018 Posts
Discussion Starter · #3 ·
Thanks for that very thorough response, brucek. Even if it was all bad news. :)

I now have to rethink my maximum (+16dB) BFD boost at 20Hz which helped to give me my housecurve on my IB.

I was thinking about your suggestion of a -15dB cut at 366Hz. Doesn't this also equate to a boost low down?

I have noticed very little lightbar activity on my BFD. A third bar is very rare even on the heaviest bass.
Even the second bar is quite unusual. Which suggests I am not pushing anything too hard.

Here's my raw REW response curve. Any thoughts please? I feel the line should continue straight to 96dB from 30Hz.

 

·
Registered
Joined
·
7,514 Posts
I now have to rethink my maximum (+16dB) BFD boost at 20Hz
Understatement...... :blink:

I was thinking about your suggestion of a -15dB cut at 366Hz. Doesn't this also equate to a boost low down?
Nope, not at all. And that technique is outdated since we now have the ability to use "what you see is what you get" with REW and can almost design any shape we want - just keep cool with the gain.
That 366 filter trick was needed when we were flying blind with the Excel sheets with tones only.

I have noticed very little lightbar activity on my BFD. A third bar is very rare even on the heaviest bass.
Even the second bar is quite unusual. Which suggests I am not pushing anything too hard.
Remember when in the user mode with the BFD, the LED's are monitoring the output and not the input.
Make sure your input is oK (using the in/out switch) first and set up the yellow LED to be coming on when you are playing a bass heavy DVD at its loudest you'd likely have it.

Then monitor the output (using the in/out switch) with the filters turned on and play the same DVD. Does the 16dB gain at 20Hz cause clipping? If so, you now have to reduce the input level to accomodate the gain filter. See the problem?

Also note that if you have a 16dB boost on a specific frequency and you've adjusted your input level to the BFD so that it just clips at that frequency, I have no doubt that the rest of the frequencies that you didn't boost are quite reduced in level and would provide the low reporting on the LED's that you're experiencing...............

feel the line should continue straight to 96dB from 30Hz
Well, convention says it should be fairly flat from 30Hz down to 15Hz (in a perfect world).....

brucek
 

·
Registered
Joined
·
1,018 Posts
Discussion Starter · #5 ·
Thankyou brucek

As usual your responses offer considerable food for thought and much to absorb.

I have ordered a new-style analogue RS SPL meter to compare with my 15-year-old model.

I want to confirm that my REW curves are repeatable between the two meters before attacking the BFD again. I have never fully trusted the old meter in the deep bass.

Once that hurdle has been cleared I shall been in a much better position to play with housecurves using more acceptable methods.

Once again I am most grateful for your feedback.

Regards
Chris
 

·
Super Moderator
Joined
·
9,280 Posts

The BFD1124 is a 24bit device. This means the device theoretically can resolve 16777216 discrete voltage steps (translates to ~ = 6.125x24=147db). Well, that's the theory, but another 1124 spec is that the noise unweighted is >94dB. That's a little better than 15 bits of resolution. So about 9 bits (of the 24 bits) are lost in the noise - you have no contol of this and it's fixed. That leaves ~15 bits to define the input voltage levels or a possibility of 32768 voltage steps as long as you supply the input with a full scale signal.
This is simply not true. The theoretical dynamic range of the 24-bit converters has nothing to do with a component’s ultimate unweighted noise spec. Its noise spec is largely determined by other circuitry in the unit, such as its power supply. This means that background noise levels are fixed no matter what the input signal.

Chris - to begin, I really can’t disagree that ideally we need to keep the BFD’s input signal as high as possible, to maximize dynamic range and S/N ratio. Equalization is not applied until after the signal has been converted to digital, so the real issue is not so much clipping the input as it is the output. If the BFD had an independent output level adjustment (a glaring omission), cutting vs. boosting wouldn’t be an issue: You could keep the incoming signal high to maximize dynamic range and S/N ratio, and reduce the output level as needed to avoid clipping if the sub required lots of boost.

But since the BFD has no independent level adjustment of any kind, our input level is at the mercy of the receiver, while the output level is fully dependent on any equalizing we’ve done. As such, if there is too much boost the only way to avoid back-end clipping is to reduce the incoming signal. Naturally, this will increase noise levels and reduce dynamic range, as the amount of functioning bits is reduced. But what are the real world ramifications?

Fortunately, noise issues are not terribly relevant when it comes to subwoofers, compared to the mains, since the only extraneous noise they can generate is – quite naturally - low frequency hum or rumble.

Approaching the boost vs. cut issue from another perspective, it’s impossible to dispute the scenario you presented, that cutting all around a couple of troughs gets you the same thing in the end as just boosting the troughs. Let’s take a closer look at that, and for the sake of presentation let’s assume that the rest of sub’s response is reasonably flat. To help make the discussion easier to follow, let’s say the two troughs are centered at 70 Hz and 45 Hz.

Being concerned about the potential downfalls of equalizer boosting, we cut response all over the place down to the level of our 70 and 45 Hz troughs. What is the end result? Well, you have to keep in mind that it’s the room that’s causing those troughs. The electrical signal is (or at least was) linear. So, when we cut everywhere except at 70 and 45 Hz, there’s no avoiding the fact that those two points are now at a higher level than everything else. That’s right – by default we’ve succeeded in electrically boosting those troughs without applying a single boosting filter! The result: Acoustically flat response, since we’ve eliminated the acoustically absent troughs at 70 and 45 Hz, but electrically those two points are now boosted.

You can see what I’m talking about in this graph, which shows electrical response of the signal where three cut filters were employed. As you can plainly see, between the cuts you have peaks!




As an aside, dealing with deep troughs using the “cut only” mantra can result in a problem that is unique to the BFD. I’ve seen at least one case where everything was cut so much it wouldn’t even pass a signal out the other end! Keep in mind that the BFD’s maximum cut value is a severe 48 dB. Go crazy with the gain reduction and you can literally obliterate the signal.

Theoretically, at this point the BFD’s meter should be reading the same on the input side (“In/Out” green light blinking) as it is on the output side (green light steady-on). This is because its output reading should be reflecting our lingering boosts-by-default at 70 and 45 Hz – which we should note, are unchanged from their original levels.

What would we be seeing if we had just boosted 70 and 45 Hz instead – a significant amount, since you said the troughs were “deep?” The output meter should show clipping. Since the BFD has no independent output level control like virtually every other ’ equalizer on the planet, we’d have to reduce the input level from the receiver to compensate.

So it seems we’re relying quite a bit on what the BFD’s meters are showing us. The million-dollar question is “Are they reliable?”

I’ve only recently acquired a BFD and begun experimenting with it (thanks again Sonnie, for dragging me out of the dark ages!), but it didn’t take me long to figure out that the accuracy of the meters is highly approximate.

I first observed it when I played a disc I had burned with sine waves from Sonnie’s test tone index, and noticed that the BFD’s meter was all over the map. By that I don’t mean the meter was fluctuating with the sine waves – it was staying steady, as it should. The problem was that different frequencies were delivering different meter readings – higher or lower. At first I thought it might a problem with Sonnie’s test tones, or that it might be the result of auto-leveling applied by the computer CD-burning program I had used to convert the MP3 files to WAV.

However, the CD player I was using is also a recorder, so it has signal level meters. And they were holding rock-steady no matter the frequency. To further verify that the tones were stable, I hooked up a cassette deck that has a much finer meter calibration: -3, 0, +1, +3. I adjusted the deck’s input level to +3, and it never moved up or down no matter what frequency tone I selected. That was quite a different picture from what the BFD’s meter was showing!

Since it was obvious a closer investigation was needed, I adjusted my receiver’s volume control so that the frequencies that appeared to be the hottest, in the 25 Hz range, were just hitting the BFD’s clip light. Then I ran the complete disc, which has 1/12-octave tones from 18 Hz to 125 Hz. Here’s what I found. (NOTE: Some frequencies are not at true 1/12-octave intervals. Since Sonnie’s index does not have tones at frequencies like 22.5 Hz, my “way around” has been to average 22 and 23 Hz readings.)

  • At 18 and 19 Hz, the meter read –10.
  • Between 20-22 Hz, it showed –6.
  • At 23 and 24 Hz, we got a reading of –3.
  • Between 25-27 Hz, the meter hit clipping.
  • At 28 and 30 Hz, the level dropped back to –3.
  • At 32 and 34 Hz, we got –6.
  • Between 36-47 Hz, –10.
  • Between 48-53 Hz, –6.
  • Between 56-75 Hz, –10.
  • Between79-89 Hz, –15.
  • Between 100-118 Hz, –20.
  • At 125 Hz we registered –30.

Needless to say, this was a real eye-opener, and I was intrigued enough to experiment further. This time I thought I’d see what kind of readings the meter would give with broadband pink noise, with boost and cut filters applied.

For this test I selected 25, 32, 40, 80 and 125 Hz as my frequency centers, because they were easy to get to (no “Fine” adjustment needed), and because they were in-range representatives of the various meter discrepancies: 25 Hz in the clip range, 32 Hz in the –6 dB range, 40 Hz in the –10 dB range, and so forth. I set the bandwidth at a fairly broad 1/3-octave, to help insure any EQ changes would show up at the meter. (Remember, filters boost and cut their bandwidth values in both directions from the center frequency, so a 1/3-octave filter affects on the order of a wide 2/3-octaves.) I fired up the pink noise and adjusted the input signal via the receiver’s volume control so that the BFD’s meter was reading –10. The results were even more shocking than they had been with the sine waves.

  • With the 25 and 32 Hz filters, it took a boost of +13 dB before the meter indication rose from –10 to –3 dB. That’s right – it took a 13-dB boost for the meter to register a 7-dB increase in gain.
  • With the 40 Hz and 80 Hz filters, the BFD took the full +16 dB in gain to get from –10 to –3.
  • With the 125 Hz filter, +16 dB didn’t even move the meter above than the starting –10 dB reading. Furthermore, switching the BFD’s meter from “EQ” and “Baseline Input” (i.e., the “In/Out” LED blinking) did not register any change at all.

Things got even more discouraging when I switched from boosting to cutting the same filters. To insure better meter resolution once the cuts were applied, I raised the incoming pink noise signal until the meter was reading –3 dB. Here’s what I found.

  • With the 25 Hz and 32 Hz filters, it took a cut of -20 dB to make the meter level change from –3 to –10. IOW, it took a -20 dB cut to register a 7-dB change.
  • With the 40 Hz filter, it took a gain reduction of –26 dB to move the meter from –3 down to –10.
  • With the 80 Hz filter it took –42 dB before the meter dropped to –10 dB.
  • The 125 Hz filter took the BFD’s full -48 dB and never hit –10. It only dropped from –3 to –6 dB.

Out of curiosity, I thought I’d see what would happen when combining boosting and cutting filters. I adjusted the meter level back down to – 10 and activated the 25 Hz filter to a full +16 dB. The meter was occasionally hitting the clip light. Then I applied a second filter at 80 Hz and cut –48 dB. The meter returned back to its original –10 dB reading. Furthermore, switching the meter from “EQ” and “Baseline” registered no change.

Next I thought I’d see what applying two filters at full boost would get. Anticipating the meter showing a lot of signal change, I started with it a little lower, at –15. Full boost at 25 Hz moved the meter up from there to –3. Applying a second +16 dB filter at 80 Hz had only a nominal effect. The meter would hit clipping, but only occasionally. Thus, two +16 dB filters registered about half that on the meter.

Until someone puts it to a precision distortion-measuring instrument to see what really happens when we boost or cut, we’re flying by the seat of our collective pants – and not very well at that. We don’t know if the meter is simply bogus, if it’s showing averages instead of peaks, or what.

So – getting back to the original question, is it okay to boost the BFD, or should we only cut? Will boosting result in reduced dynamic range and increased noise? Theoretically, yes it will. But practically speaking, the BFD can endure a full 50+ dB loss in the S/N department (the difference between the 147 dB the 24-bit converters get us and the BFD’s actual 94 dB noise spec). So I think we can reasonably assume it’s completely safe to either boost or cut the Feedback Destroyer. Just keep your eye on the meter readings. Ideally, there shouldn’t be a tremendous difference in the input and output readings.

Questionable as they are, we probably should pay attention to the clip indicators, especially on the output side, and especially if a lot of boosting or cutting filters are used at frequencies where the meter seems especially reluctant to respond in kind. If the BFD can take a pair of filters fully boosted and barely make a meter showing, it’s possible you could be clipping badly. But cutting filters could also mask a clipping problem, since they drag the meter back down despite the presence of boosting filters - there’s just no way of knowing anything for sure.

Until someone shows us some real scientific measurements, perhaps it’s best to simply let your ears be your guide. If you want to dial in some severe boost, I suggest paying attention to the way the sub sounds. Maybe play a sine wave at or near the boosting frequency, and pay close attention to what you hear when the meter, either input or output, reaches clipping. Even better, after you finish equalizing, run a full 1/12-octave sine wave test disc, paying special attention to areas where you used boosting filters. Bass distortion is audible, if it gets severe enough.

Regards,
Wayne
 

·
Registered
Joined
·
1,018 Posts
Discussion Starter · #8 ·
Indeed! Thankyou for that remarkable effort at educating us Wayne.

Since my IB shows quite a reasonable response without the BFD I may decide to give it (the BFD) a drop test out of the attic window. :D

More seriously: The troughs I was referring to were not of my own sub's making in my own listening room. My oddly-shaped, multilevel attic behaves impeccably.

It seems that typical living rooms (in the UK at least) often produce a couple of troughs like these. I was struggling between the devil and the deep blue sea when it came to recommending boosting with a BFD or not. Sometimes a trough behaves like a sonic black hole which will eat up a typical active sub's amp power.

Sub position is (unfortunately) not always a variable where doors, windows, room, system layout and partners often combine to do their worst.

I seriously doubt that many could make good use of brucek's or Wayne's suggestions. Few want to get that involved and only want a reasonable response so that they can relax and watch films.

Cutting to the core issue: What is the general recommendation for those suffering troughs? To boost the overall suboofer gain then drag the surrounding response down to the target curve?

Or is it really safer not to do anything about troughs and thus avoid clipping, distortion, driver over-extension and gently rising columns of acrid, blue smoke? ;)
 

·
Registered
Joined
·
162 Posts
acording to rew i have a couple of nasty trougths (probably due to the lack of phase control on my passive system)

but if i put it on manual mode and follow the cursor, and point to the dip, the spl reading on the left still reads low, but it doesnt "sound" much quiter than the other frequencies.

where there are peaks you realy hear the extra volume, even small 3db peaks i seem to notice, the 9db humdingers are as plain as day.

any way, my point was i got a very satisfactory sound by removing the peaks, not trying to fix the dips.

since your sub chriss has bonkers abounts of headroom, id give it a go just cutting your responce to the curve you want, if you notice the loss of head room, or you have issues with the bfd, try something else, thats the advantage of rew.

edd
 

·
Registered
Joined
·
7,514 Posts
Hey Wayne,

But practically speaking, the BFD can endure a full 50+ dB loss in the S/N department (the difference between the 147 dB the 24-bit converters get us and the BFD’s actual 94 dB noise spec).
We'll have to continue to agree to disagree on this one. :)

I adjusted my receiver’s volume control so that the frequencies that appeared to be the hottest, in the 25 Hz range, were just hitting the BFD’s clip light. Then I ran the complete disc, which has 1/12-octave tones from 18 Hz to 125 Hz.
But would the crossover in your receiver not produce exactly what you are describing? As the frequency advances from low to high, the crossover drops the subwoofer output level......

So it seems we’re relying quite a bit on what the BFD’s meters are showing us. The million-dollar question is “Are they reliable?”
I have to say the meter is very accurate. I had tested it originally with proper test equipment and it was accurate in both relative terms from LED to LED divisions and also in measured terms where the actual measured input level to the BFD from a proper signal generator and scope corresponded with the level indication on the BFD. (i.e. 0dB in registers the 0dB LED etc).

To verify, I just went and tested my BFD using REW as a signal source. The meter tests quite accurate from LED to LED. As I increase REW sine wave output by +/-10db, the LEDs follow accordingly through the entire frequency range. (I have the ability on my fancy shmancy processor to feed full range analog out my subwoofer port). The LED's indicates at around 6Hz a drop in level, but that's expected, given the BFD's response.

I admit with Pink noise it becomes a bit more difficult. I guess we can assume that the meter responds to the average AC signal it receives. If I feed pink noise from REW and the -10dB LED is just coming on and I vary the level of the REW output up and down, the meter is quite accurate in relative terms, although it jumps around as expected, given the nature of Pink noise. If I introduce a filter of 10dB that is very narrow, it indicates very little on the LED's. If I widen the filter the meter responds accordingly. Pink noise may not be the best testing tool here. To test the meter itself, it would be best to use a signal generator such as REW. I feel to set up levels for actual use, it's best to use DVD's as we suggest in the guide. There can easily be concentration of signal at specific frequencies and so using DVD signal is somewhat more appropriate.

Cutting:
So, when we cut everywhere except at 70 and 45 Hz, there’s no avoiding the fact that those two points are now at a higher level than everything else. That’s right – by default we’ve succeeded in electrically boosting those troughs without applying a single boosting filter! The result: Acoustically flat response, since we’ve eliminated the acoustically absent troughs at 70 and 45 Hz, but electrically those two points are now boosted.
Boosting:
What would we be seeing if we had just boosted 70 and 45 Hz instead – a significant amount, since you said the troughs were “deep?” The output meter should show clipping. Since the BFD has no independent output level control like virtually every other ’ equalizer on the planet, we’d have to reduce the input level from the receiver to compensate. That would get the output signal out of clipping, but as brucek has shown us, the penalty would be a reduction of digital bits, and therefore dynamic range and S/N ratio.
So we certainly agree on this major point. Both the cutting and boosting scenarios above result in the same output voltage waveform from the BFD, but the cutting example enjoys the best dynamic range and S/N. When using only cutting, the input is at maximum and the ADC has advantage of all the bits to feed the DSP. The DSP now has the most possible information to work with to create the filtered signal to send to the output DAC.
Even though the boost example results in the same output voltage waveform, it doesn't enjoy the advantage of a full input signal to the BFD.
I choose the cutting scenario....... :blink:

brucek
 

·
Super Moderator
Joined
·
9,280 Posts
Hey brucek,

But would the crossover in your receiver not produce exactly what you are describing? As the frequency advances from low to high, the crossover drops the subwoofer output level......
Well, slap me silly. Somehow that managed to get past me. :duh: :duh: :duh:

Crossover bypassed, all is well with the test tones now. :T

However, we still have the issue of how the meter responds with filters applied. Using pink noise I set three filters, 1/6-octave at 25 Hz, 1/3-octave at 50 Hz, 1/6-octave at 100 Hz, all boosted +16 dB. The meter registered minimal movement, from –10 to –6 dB. With all filters set to maximum cut, there was virtually no meter change.

I’m still trying to get the hang of the REW program, but using Sonnie’s sine waves got better results from the meter. Using a 1/6-octave 40 Hz filter @ +16 dB, the 40 Hz sine wave jumped the meter from –10 to –3 dB. It should have pushed it to clipping. Oh well, at least it registered something...

Both the cutting and boosting scenarios above result in the same output voltage waveform from the BFD, but the cutting example enjoys the best dynamic range and S/N. When using only cutting, the input is at maximum and the ADC has advantage of all the bits to feed the DSP. The DSP now has the most possible information to work with to create the filtered signal to send to the output DAC.
Even though the boost example results in the same output voltage waveform, it doesn't enjoy the advantage of a full input signal to the BFD.
In theory, this is correct. But you have to keep in mind that once you’re over 16 bits, input levels simply don’t matter much. This is because a 16-bit waveform, which has 65,536 amplitude or quantization “steps,” is considered the threshold of what is acceptable for hi-fi sound. While there may be some debate about that in audiophile circles, a 24-bit waveform has over 250 times more amplitude “steps” - 16,777,216. It should be obvious that you can feed a 24-bit A/D converter a low-level signal and it will still deliver dynamic range figures and low-signal quantization well in excess of 16-bit. In other words, input signals for 24-bit converters is virtually a non issue.

Also it occurs to me: The BFD equalizes in the digital domain, and knowing what little bit I do about how digital circuitry works, once the signal is digitized you can to anything to it void of analog-world constraints. As such, the signal can be manipulated without concerns of things like headroom. That only becomes an issue once the signal is converted back to analog. Right? Wrong? Enlighten me, o digital guru!

I feel to set up levels for actual use, it's best to use DVD's as we suggest in the guide. There can easily be concentration of signal at specific frequencies and so using DVD signal is somewhat more appropriate.
Well, after seeing all this, that kinda worries me. We typically use an action flick with a loud explosive scene to set the output. The problem is, that low freq rumble is a lot like pink noise – i.e., broadband. No, not 20 Hz – 20 kHz broadband, but certainly encompassing pretty much the entire bass region. Broadband low-passed, if you will. Since that’s what our most demanding movie program material is like, and since the BFD’s meter has problems showing us what’s going on inside its digital brain with such material, I doubt we’re getting an accurate picture when setting the levels with such material, or how the equalizing affected things.

I guess all that remains to be seen is if you get anything audible from the sub when you drive the BFD into clipping. I may give that a shot tomorrow.

In the meantime, I’m not worried enough about potentially loosing a few dB in S/N ratio to tell folks they should avoid boosting when it’s needed. There are just too many BFD users out there who should have reported back that they had problems doing it, if it were a problem.

Regards,
Wayne
 

·
Super Moderator
Joined
·
9,280 Posts
Hey Chris,
I seriously doubt that many could make good use of brucek's or Wayne's suggestions. Few want to get that involved and only want a reasonable response so that they can relax and watch films.
Yeah, I know - we’re a sad, sad bunch, aren’t we? :rolleyesno: I’ll have you to know we spend lots of time and money on our systems making sure the deliver the absolutely best-sounding sine waves! :laugh:

Just kidding, or course... :)

Cutting to the core issue: What is the general recommendation for those suffering troughs? To boost the overall suboofer gain then drag the surrounding response down to the target curve?
Well, I guess it all boils down to which “general” you want to go with! :laugh: My esteemed college likes to keep S/N ratio as high as possible. I’m not worried about it, since you can’t give up more than 16 dB, which still leaves the S/N at least 78 dB - plenty for a subwoofer application. Back in the days of turntables and cassette decks we would have thought we had died and gone to heaven if we could get the S/N ratio that high!

More specifically to your problem, I usually recommend a combination of boosting and cutting filters. Just look out for a null when boosting – i.e., a problem that doesn’t respond. Try to keep the input and output meters at about the same levels when you’re finished, and you’ll be fine. The output meter shouldn’t be drastically higher or lower than the input meter, which is what you might get using all boost or all cut filters. I also think I’d try to keep the levels well below clipping, since I’m finding the meter doesn’t seem to give us accurate readings with those booms and explosions in movies. I think I’d keep things no higher than the –6 reading until we can get this sorted out.

By the way, this that you wrote earlier...
I have noticed very little lightbar activity on my BFD. A third bar is very rare even on the heaviest bass.
Even the second bar is quite unusual. Which suggests I am not pushing anything too hard.
...tells me your levels are way too low. As noted, try to get them up to at least –6.

Regards,
Wayne
 

·
Registered
Joined
·
7,514 Posts
The BFD equalizes in the digital domain, and knowing what little bit I do about how digital circuitry works, once the signal is digitized you can to anything to it void of analog-world constraints. As such, the signal can be manipulated without concerns of things like headroom. That only becomes an issue once the signal is converted back to analog. Right? Wrong?
Well, you certainly have no access to the DSP data once its been passed. It handles the samples as they're passed from the ADC. If you feed the DSP a signal that is clipped on the input and is basically garbage, that's what you'll get out - just like any computer. The lowest signal to error ratio comes when the signal is using the most bits available, but even a small amount over that level and you are entering maximum distortion.
Alternately, if the softest signals are now down in the noise because your input signal is too low, what can you expect the DSP to do. The sample has no intelligible information - how can it process it to apply a filter.
The BFD doesn't appear to apply any AGC, in that if I apply a gain to a signal that results in a level over 0dB, it simply complies and the output will clip.

brucek
 

·
Premium Member
Joined
·
571 Posts
Hey, Chrisbee... You have an EP 2500 right? I'm having a brain fart. What level do you have your amp at? I'm trying to follow the recommendations in this thread but my amp clips and I'm assuming it's because I have a strong signal coming from my BFD. I'm at about 2:00 on my EP2500 and things are ok but if I increase my signal to my BFD to get the lights near -6 I tend to clip the amp...

Another question, at what frequencies should it be hitting -6? Just on the more dynamic scenes or what?
 

·
Registered
Joined
·
1,018 Posts
Discussion Starter · #17 ·
It is really rather pointless in giving level settings when our systems and rooms are are so very different. My amp sits out in the IB enclosure so it's a bit of nuisance going in an out just to change settings.

But if it makes you happy: :) I tend to leave my EP2500 controls at 3 o'clock.

I use the controls on my active crossover (sitting within easy reach in the rack) to raise and lower the bass levels relative to the speakers. The settings depend on whether I'm watching films or listening to music. I raise and lower the bass by anything up to an indicated 8dB according to the media and my mood. I am no longer a perfectionist when it comes to calibration. Though I will quickly change the settings if something sounds subjectively bass heavy or light. This very rarely happens. I hold the opinion that the better the sub the less sensitive one is to excessive bass. It may simply be that my room is kind to all bass levels. All bass just sounds incredibly clean and completely boom-free. I put this down to the lack of parallel walls, sloping ceilings, multiple levels and low dwarf/knee walls.

BTW: I have never seen the fourth bar on the BFD nor have I ever seen the EP2500 clipping lights come on.

I can't really answer your last question. Obviously large dynamics will cause clipping. I'm not sure of the frequency relationship. Presumably it would be BFD filter dependant.
 

·
Premium Member
Joined
·
571 Posts
Yeah, I realize a direct comparison is useless but I wanted to get a general idea of how your amp is set. The difference in the room and drivers makes the comparison useless but I get an idea of how your amp is being utilized when you don't get any clipping at that setting :)

I can easily clip my amp so I'm trying to fine tune things.
 
1 - 20 of 25 Posts
Top