Below is a frequency analysis of a large sample of music from "Exodus". It's quite evident that this particular track has peak bass output around 55Hz or so. If your listening position happens to have a null in its room response at that frequency, then boosting the 55Hz area won't produce as much effect as would otherwise be expected.
A frequency analysis of the Herbie Hancock track "Dis is da Drum" is presented below. It has a major peak in the bass response at around 34Hz, with smaller but still significant peaks at 38Hz, 49Hz and 52Hz. That signal at 34Hz would give most subwoofers a reasonable workout. 🙂
Last two pieces are gonna go on today for Subwoofer No. 1.
Then I could do some measurements. I have:
My workshop is about 40 m², I could do the measurements in there. I could also go outside, but, despite this being a small village, it's pretty noisy here, including a damn small airplane flying in circles all day (spraying pesticides or whatever).
If someone could give me some hints on how to setup for measuring, and how to setup REW, I would greatly appreciate it. I don't have lots of experience doing proper measurements.
Then I could do some measurements. I have:
- Laptop with REW installed
- MiniDSP USB Mic with calibration data
- Tripod for the Mic
My workshop is about 40 m², I could do the measurements in there. I could also go outside, but, despite this being a small village, it's pretty noisy here, including a damn small airplane flying in circles all day (spraying pesticides or whatever).
If someone could give me some hints on how to setup for measuring, and how to setup REW, I would greatly appreciate it. I don't have lots of experience doing proper measurements.
Alright then... here some measurements.
What I did: I put the subwoofer upright onto my table saw (an old cast-iron beast, so a sturdy stand).
Put the MiniDSP USB mic onto a tripod, set it up so the mic pointed directly into the center of the cone.
Started my Laptop with REW, loaded the mic calibration data, and made sure the levels were high enough.
Then I ran two sweeps, with all default settings, except the range limited to 0 - 200 Hz
The first measurement was taken with the mic about 10 cm from the cone:
The second with the mic 1 meter from the cone:
As one might expect, the room heavily influenced the second result.
Anyway, I do not know much about measuring.
Please let me know how to interpret this... or what I can do better.
What I did: I put the subwoofer upright onto my table saw (an old cast-iron beast, so a sturdy stand).
Put the MiniDSP USB mic onto a tripod, set it up so the mic pointed directly into the center of the cone.
Started my Laptop with REW, loaded the mic calibration data, and made sure the levels were high enough.
Then I ran two sweeps, with all default settings, except the range limited to 0 - 200 Hz
The first measurement was taken with the mic about 10 cm from the cone:
The second with the mic 1 meter from the cone:
As one might expect, the room heavily influenced the second result.
Anyway, I do not know much about measuring.
Please let me know how to interpret this... or what I can do better.
Attachments
The measurement with the mic about 10 cm from the cone is in the near field, not representative of it's response in the far field.
The far field distance is reached where the path length difference for wave arrivals from points on the device on the surface plane perpendicular to the point of observation (the speaker baffle) are within one-quarter wavelength at the highest frequency of interest.
A subwoofer with a one square meter face measured at one meter with a measurement mic on the ground plane would be in the far field up to 210Hz.
An outdoor measurement well away from any reflective objects (buildings, cars, etc.) with both the subwoofer and mic on the ground plane, mic 2 meters from the baffle will be as accurate as the test mic response is.
The far field distance is reached where the path length difference for wave arrivals from points on the device on the surface plane perpendicular to the point of observation (the speaker baffle) are within one-quarter wavelength at the highest frequency of interest.
A subwoofer with a one square meter face measured at one meter with a measurement mic on the ground plane would be in the far field up to 210Hz.
An outdoor measurement well away from any reflective objects (buildings, cars, etc.) with both the subwoofer and mic on the ground plane, mic 2 meters from the baffle will be as accurate as the test mic response is.
I was afraid that only an outside measurement would be meaningful.
I heard about gated measurements... those will not help either?
It will be some time before I can try to setup an outside measurement. Noise is also an issue.
For now I will spend time building the second one...
I heard about gated measurements... those will not help either?
It will be some time before I can try to setup an outside measurement. Noise is also an issue.
For now I will spend time building the second one...
The reflected low frequency wavelengths are too long to "gate out" in rooms with dimensions smaller than those wavelengths.
Outdoor measurements can be done at fairly low levels, just need to be above the noise floor by +10dB or more.
Wind noise can be reduced using wind screens.
Yeah, do that. OSB is not as internally strong...if it gets wet it swells up a lot more than plywood. Kiss of death in my opinion. Plus sometimes you can get some nicely veneered plywoods. What did you finally use for the build? I used to do business with McCauley, way out there in Puyallup, nice folks.
As for " It still puzzles me why this driver, which is supposed to have 95 dB sensitivity, needs so much boost" I must note those sensitivity (whether measured on whatever the heck the manufacturer is using for a setup, or calculated from Thiele-Small) are MIDRANGE sensitivity. The actual sensitivity at very low frequencies is quite a different matter for most subwoofers, whose boxes are barely big enough to fit the woofer. Yours is getting to a proper box size so per the sims should keep the sensitivity down fairly low, plus room gain. BUT
- Measure the actual in-room response, with REW+UMIK or such. Maybe some rooms modes are killing out some of the bass.
- Yes you really should not "notice" a subwoofer until you turn it off. Most music just does not go that low. U-571 depth charge scene, that should be a different story. I think Enya's "The Long Ships" has a very low tone. Somewhere there is a database about such...
Last edited:
@head_unit @drtebi Below is a frequency analysis of Enya's "The Long Ships", taken from the "Watermark (Remastered bonus track version)". It's apparent that there are some quite significant low-frequency peaks at 27.5Hz, 32.1Hz, 36.9Hz, and 43.8Hz. This supports the contention that this track has a very low tone. Interestingly, at least for me, the track doesn't seem to have a particularly large amount of low-frequency bass energy. In any case, the 32.1Hz low-frequency content can give most subwoofers a good workout.
That's absolutely correct, and something that is possibly forgotten amongst the plethora of data. With drivers that tend to be well-suited to standard alignments, it's generally the case that the vented-box alignment's F3 is going to be pushed up quite a way higher than the driver's free-air resonance frequency, Fs. That's just the inherent nature of vented boxes, which give the benefits of a "smaller" enclosure for a given bass response with any particular driver. In the case of the McCauley 6174 driver, its Qts=0.34 lends itself to a natural QB3 vented-box low-frequency alignment. To some extent, this helps to keep its box volume in check, as its Vas=733litres is quite large.
By way of example, a traditional QB3 alignment with Vb=357litres and Fb=23Hz will produce an F3=28.3Hz. Although a quite reasonable result, it's still hardly real "subwoofing" territory. Below is the resulting frequency response plot. Of course, the passband sensitivity is nice and high, being about 93dB.
In this instance, and also many, many others, there is no escaping the fact that a non-standard tuning will offer more low-frequency extension. Those that are less familiar with vented-box alignments will find that somewhat counterintuitive. To get more low-frequency response out of this enclosure, we can tune the system to a lower Fb.
Below is the same vented-box enclosure (Vb=357litres) tuned to Fb=12Hz, with a parametric EQ of +6dB with a Q=1.20 added at 23Hz. We have now achieved a substantially lower F3=21.5Hz, which is more in keeping with the expectations of this driver. Note that, with this driver, we have been able to maintain its natural passband sensitivity.
If we want to use the full potential of the driver in a 357litre enclosure, we can make use of an equalized closed-box alignment. We can achieve an F3=15.6Hz if we apply a shelving filter coupled with two lots of parametric EQ, one of which provides a boost of 9.5dB at 17Hz with a Q=1.9, while the second is applying a cut of 0.4dB at 47Hz with Q=1.0. Of course, with this heavily-equalized design, we have obtained excellent low-frequency extension by reducing the maximum SPL that can be achieved within the passband. Nonetheless, for some applications, this might be quite acceptable.
Last edited:
Thanks for the explanations. I guess one thing learned is that one should really measure a driver before designing... I would have realized that the 95 dB comes from the midrange sensitivity.
I will do some in-room measurements once the second subwoofer is done. I suppose with two subwoofers running, I should have a bit less room modes also.
What worries me about this kind of setup is, that I may get some distortion if I do the EQ and shelving filters digitally. Of course I can avoid it by taking the entire signal down before boosting, but that means less overall output. It's something I have often experienced when experimenting with EQ for open-baffle bass.
An analog shelving filter however did wonders. I happen to have a Sony SE-P900, and every time I add it into the "mix", it does wonders to the low frequencies due to the low shelving filter. Only that this takes a lot more setup... the EQ runs on 100V, requires a step-down transformer, more cabling etc.