Does an (S)LLT require a subsonic filter?

[SteveCallas]

bosso wrote:

I wish the measurements could have been shown from 2Hz to 100Hz. The graphs that are cut off at 10Hz are a bit frustrating. Of note, it seems by this thread that suddenly, 10Hz is very important whereas a year ago, 10Hz was not viewed as important at all.

The graph partially shows that by an octave below 10Hz, the advantage has flipped back to the sealed sub, which clearly displays the flatter response when viewed further down. Maybe by this time next year we'll lower our threshold another octave.

He did measure down to 2hz, and the ported is stronger and flatter to about 7hz. The IB has a less steep rolloff below 7hz. Rodny's setup is perhaps the most perect comparison possibly available - two drivers on one side of the room in a LLT, two of the same drivers mirrord on the other side of the room in an IB, both seeing the same amount of power. All previous comparisons I have read about are illegitimate, like 4 drivers in an IB with kilowatts of power vs one driver in a LLT placed somewhere else in the room. Even when I eventually get a house, I couldn't have hoped to have set up a more perfect experiment.

His results and impressions go to show that staying flatter with more clean headroom in the infrasonic range is preferred to rolling off sooner but less steeply. This is what you and I have been debating for ages, and here in the perfect comparison lies the proof.

bosso wrote:

As I've said many times, having looked at the differences very carefully at great length, to retain the advantages of the sealed sub and match the output (in this case from 7Hz to 14Hz) just double the displacement and power, keeping in mind that excursion can be controlled by the box.

Ermm, yeah, you still aren't grasping it. Get 4 more drivers and a ported implementation would still be better, just use a lower tune. He didn't care for the performance of the 4 driver in the IB EQ'd flat, tht's what led him to start this project.

[bossobass]

Ermm, yeah, you still aren't grasping it. Get 4 more drivers and a ported implementation would still be better, just use a lower tune. He didn't care for the performance of the 4 driver in the IB EQ'd flat, tht's what led him to start this project.

Is that it? I'm not grasping it?

As I've said, I'm not much of an IB fan for several reasons. That's not to disparage those who enjoy them for one of their positive attributes, it's just to say that I disagree with several of the alignments downside elements.

To be 'logical' about this comparo, there's really nothing new here that we haven't hashed out before. A ported sub has output advantage over a one octave band using same driver/amp compliment. The price is extreme size and moving to a resonant system. Below that one octave band, the advantage flips to the IB to DC.

We can clearly see this in the graph, which, you have to admit, and by your own standards, is all we have to look at. I'll just say that 'perfect comparison' is your term.

I'm reading the graph from 7Hz to 80Hz as the IB being +/- 8dB and the ported being +/- 9dB. Expanding the BW to from 5Hz to 80Hz, that disparity widens. What does that mean? not much, IMO, but it certainly doesn't mean that the ported sub is flatter.

It's much appreciated information, and the effort is certainly appreciated, we just need a more objective interpreter of the data, IMO.

Bosso

[SteveCallas]

bosso wrote:

A ported sub has output advantage over a one octave band using same driver/amp compliment.

Assuming one wants to maintain a relatively flat FR even at spirited levels and wants low end extension (that's what we all want isn't it?), a ported implementation will actually have more output from just below tuning all the way on up, not just in one octave. Use a small enclosure to keep the sealed safe from overexcursion down low and the distortion advantage of ported gets that much larger.

[brucek]

Rodny says:

, I know that my ECM is not calibrated to 5Hz, maybe when Sonnie gets his mic back I can get a real measurement!!

Yeah, the graph you posted doesn't mean a lot below 10hz until you apply the new calibration file now avaiable for the ECM8000 on the download site.

You don't need to redo any measurements because of the new mic calibration file though, since the soundcard and meter calibration data isn't included in the impulse response. The mic/meter and soundcard calibrations are only applied when calculating the frequency response.

So you are able to use your old mdat file and simply apply the new ECM8000 calibration file that now extends to 5Hz.

To apply or remove a mic/meter or soundcard calibration for an REW measurement after it has been taken, simply load or clear the cal data as required and press the Apply Windows button (invoked when you click the IR Windows ICON) to recalculate the frequency response with a new calibration file.

Rodny says:

I'm using my ECM8000 and the UB 1002 and the REQW spl meter

One of the variables that use to be insignificant when measuring down to 10Hz now begins to rear its head since the wacky world of <10Hz measurements has begun. Since the ECM8000 calibration file extends to 5Hz, I think it appropriate to include the response of the ECM8000 preamp into the soundcard calibration file so as to remove its low frequency response effects.

The UB802 that I use begins to drop off at 10Hz and so I have created a soundcard calibration file that included the preamp in the loopback measurement when creating the Soundcard.cal file. This inverse offset then removes the premps influence from the measurements.

It's no big deal, simply create a soundcard.cal file with the preamplifier in the loopback cable. The line-out of the soundcard connects to the Line in-1 1/4" TS plug below where the microphone normally plugs and the 1/4" TS jack connects to the soundcards line-in. The levels can be set with the Main Mix as usual to satisfy REW for the measurement........

Here's a graph showing my soundcard.cal file with and without the UB802 in the loop. It's important to nullify as many test equipment variables as possible when taking measurments this low.



brucek

[Ilkka]

brucek wrote:

Rodny says:

Yeah, the graph you posted doesn't mean a lot below 10hz until you apply the new calibration file now avaiable for the ECM8000 on the download site.

I don't want to sound negative, but even that doesn't help since it's Sonnie's ECM8000 which got calibrated, not Rodny's. I've seen the calibration data for five professionally calibrated ECM8000's. The difference between the best and worst mic was already around 7.5 dB at 10 Hz. Below that it gets even worse.

You can't use someone else's calibration data unless the mics have identical frequency responses.

[brucek]

You can't use someone else's calibration data unless the mics have identical frequency responses.

Well exactly, but why not remove as many variables that you do have control over....... I have no illusions that Sonnies cal file can be applied accurately to anyone elses microphone. I think the adventure into sub 10Hz measurements is a bit silly myself, but we supply the file for what it's worth to at least take the edge off those low measurements without any cal file.

Actually, I was amazed though at the consistency between all the Galaxy meters that Sonnie painstakingly tested. They were all identical. Why wouldn't the ECM be so..........

brucek

[Ilkka]

brucek wrote:

Well exactly, but why not remove as many variables that you do have control over....... I have no illusions that Sonnies cal file can be applied accurately to anyone elses microphone. I think the adventure into sub 10Hz measurements is a bit silly myself, but we supply the file for what it's worth to at least take the edge off those low measurements without any cal file.

brucek

Which variable exactly are you removing by using an unknown calibration data? You simply don't know whether it makes Rodny's measurements more or less accurate.

[Ilkka]

brucek wrote:

Actually, I was amazed though at the consistency between all the Galaxy meters that Sonnie painstakingly tested. They were all identical. Why wouldn't the ECM be so..........

brucek

The Galaxys are amazingly consistent. The data I have on the ECM8000 shows that they are not alike.

Here are five different ECM8000 mics. Sonnie's is ~9.2 dB down at 10 Hz so it's pretty similar to the red one.

[brucek]

Yeah, but I have begun to question some of the previous "professional" calibrations since Sonnie sent his to West Caldwell labs and he got a certificate and data on the calibration with some very detailed work. These guys appear to be the real deal. I'm sure they use standards modules to determine the calibration rather than comparison methods.

This Calibration is traceable through NIST test numbers 822/274345-07

The expanded uncertainty of calibration: 0.5db at 95% confidence level with a coverage factor of k=2.

-39.61 dB re.1V/Pascal

10.46 mV/Pascal

(1Pascal = 94dB SPL)

Ambient Temperature: 21.9 C

Ambient Humidity: 43.4 % RH

Ambient Pressure: 99.57 kPa

Calibration was performed by West Caldwell Laboratories Inc. under Operating Procedures intended to implement the requirements of ISO10012-1, IEC Guide 25, ANSI/NCSL Z540-1, (MIL-STD-45662A) and ISO 9001:2000, ISO 17025.

Specifically, it's interesting how the new calibration is quite different from the one we used before (and the ones I've seen like you show in your graphs).
I trust the latest one is quite accurate. It is the one we used for the creation of the latest Galaxy CM140 using the comparison method (for what its worth).

Here's a comparison of the new ECM calibration file over the full bandwidth compared to the old one.
The red is the latest calibration file supplied by West Caldwell labs.



brucek

[brucek]

Which variable exactly are you removing by using an unknown calibration data? You simply don't know whether it makes Rodny's measurements more or less accurate.

Of course I do. The response of the preamp is a variable until I measure it with REW and apply the inverse response in the soundcard file. If the preamp was down -2dB at 5Hz then that influence is removed. That makes it more accurate by 2dB.... Not accurate, just 2dB more accurate than it was without..

brucek

[Ilkka]

brucek wrote:

Of course I do. The response of the preamp is a variable until I measure it with REW and apply the inverse response in the soundcard file. If the preamp was down -2dB at 5Hz then that influence is removed. That makes it more accurate by 2dB.... Not accurate, just 2dB more accurate than it was without..

brucek

I meant the mic calibration file. Naturally one has to calibrate the whole measuring chain, including the preamp.

[Rodny Alvarez]

This is the soundcard cal file with the UB 1002 and the ECM8000 with Sonnie's cal file, I guess this look right ..................

[brucek]

I guess this look right

Looks exactly right.

brucek

[Rodny Alvarez]

Thanks Brucek!!

I was thinking of making some measurements .......
Computer, receiver, BFD and amp.
Computer, receiver and amp...................no BFD.
Computer and amp...........no BFD and no receiver.
Do you think this will help some or not??

[brucek]

Do you think this will help some or not??

No, given what you are trying to accomplish.

The goal in taking measurements is to remove any influences from the test equipment. Once the test equipment is compensated for, you are left with the unit (or units) under test that will determine your overall response.

The units under test are your receiver, plus the BFD, plus the sub amplifier and its speaker(s), plus the rooms influence at the listening position. The combination of these items is what determines your systems response. These are the items you'll be using to listen to your system, so you don't want them inversely compensated for in a system response measurement.

Other than having the knowledge that one of these items is holding you back, there's little to be gained by 'breaking out' the individual devices that make up the system response. You've compensated for the soundcard, preamp, and microphone deficiences in your test equipment. Job done.

brucek

[Chrisbee]

I mentioned the plug of air in the huge port in the other thread as a potential oscillator. From Rodny's SQ comparison this worry would seem to be unfounded.

But could the air in such a large port offer enough inertia to make the enclosure behave like a sealed box below the tuning point?

It might modify the behaviour of the enclosure in a completely different way to classical (ie. small) reflex enclosures where port air mass is relatively tiny.

Rather than allowing the drivers to completely unload below the tuning point it might be protecting the drivers with a degree of progressive enclosure compression to resist extreme cone movements. A pneumatic high pass filter, if you like.

We are in relatively unknown territory with the LLT. The means to measure such low frequencies in a domestic situation, suitable loudspeaker drivers and the programme material to push such extreme subwoofer designs to high levels are all relatively new. That Rodny modified his existing IBs certainly adds to the interest.

Should the conventional ported enclosure design software still hold up in such extreme examples of the art?