Getting to the bottom of it all

Mark Seaton

WBF Technical Expert (Speaker & Acoustics)
May 21, 2010
355
74
435
44
Chicago, IL
www.seatonsound.net
#1
I thought I'd start a thread where we can collect some questions, discussions and thoughts on the deep end of the spectrum many know me for.

First and foremost, while you always need to listen and check you haven't missed something audible obvious, in small rooms, if you aren't measuring, you're guessing. I'll add to a list of various measurement systems and links in this thread over time.

Dayton OmniMic - Parts Express
Room EQ Wizard
Studio Six Digital's iAudiointerface
XTZ Room Analyzer
(list to be updated as time and info permits)

I don't intend to promote one over the other, but see all of the options suiting the preferences of different users. What ever system gets you measuring your system consistently and reliably is the way to go. Without it, you're at best making educated guesses.

For those who have seen me hunched over a computer at some audio gathering or installation, you would have seen the TEF25 system I use. Frankly I don't recommend it as an easy option for anyone who isn't doing loudspeaker development or commercial sound system setup. I myself have managed to figure out the system over the years and attending multi-day workshops. The TDS measurements do offer some useful qualities, and I use multiple platforms for different tasks.
 

Mark Seaton

WBF Technical Expert (Speaker & Acoustics)
May 21, 2010
355
74
435
44
Chicago, IL
www.seatonsound.net
#2
Hello Mark,

I saw some studies somewhere that we can not discern up to like 15% or more harmonic distortion at say 20hz. I realize that the second is then at 40Hz, but it seemed to say that we are not too good at picking distortion out at super low frequencies.

Tom
Hi Tom,

There has been quite a bit of research on distortion audibility, and the one concensus is that distortion types are not all equal in perception. I'm sure there would be some debate on exact thresholds, but the general concept is correct in that the closer the distortion is to the original signal the less easy it is to identify. In subwoofers, 2nd and 3rd harmonic distortion have much higher thresholds than do say 5-10th harmonic before becoming audible or offensive.

I would not suggest that lower order distortion is ever desired, but "good enough" is a much higher level than higher orders. Sometimes seemingly moderate 2nd order distortion will be acceptable when you can greatly reduce the more audible distortion products. To get a basic concept of high order vs. low order distortion, it is related to the cause of the distortion. The sharper the non-linearity, the higher the order. The softer the non-linear behavior, the lower order the distortion. In extreme, but tangible terms, imagine a woofer smacking the back plate of a woofer as an extreme example of a very sharp non-linearity (dead stop!). A low order non-linearity would be a sealed woofer in a very small box starting to compress the air in the box past the point of linear spring action. This produces a very gradual increase in stiffness against the woofer. This will round over the signal and have a very different, and subjectively softer sounding distortion than the extreme of a voice coil slamming into the back plate which by contrast produces a startling, knocking noise.
 
Aug 3, 2010
646
3
170
New Milford, CT
www.basspig.com
#3
My take on this is that while low order distortion in the bass is more tolerable (in fact, it's often used as a musical instrument enhancement in guitar amps), the ear is also more prone to detecting small amounts of it, where it's sensitivity to the harmonic frequencies is much greater, due to Fletcher-Munsen, than at the fundamental frequency. If you have a lousy woofer that, at 10Hz, has 5% THD at some useful SPL, you're going to hear the harmonics which will mask the fundamental. Down in the very low frequencies, ANY small harmonic distortion, no matter how small it may look on paper, is going to be audible, due to the ear being already starved of sub-20Hz sensitivity. The ear is at a 30+dB disadvantage down there, so woofers have to have a correspondingly lower harmonic distortion level. Where -30dB harmonics might be acceptable on a midrange driver, a subwoofer needs to have those harmonics down better than 55dB below the fundamental, or they will be easily spotted.
It's difficult to achieve clean bass anywhere inside of a structure, as nearly all structures produce distortion in the form of sympathetic vibrations. The rattles, creaks and grown of lamp fixtures, convectors and wood joists will always present a challenge.
That said, controlled harmonic distortion, as injected in a single track on a master recording of a bass, CAN be interesting. In fact, nearly all string bass is rich in harmonic content by nature of the instrument. Even a Barker Bass has rich harmonics. It's what gives a bass that 'fat' and 'round' sound. I like it myself, as long as it's part of the music. Some music goes too far in that respect, which is why I have (shudder) --the dbx 500 to reconstruct missing fundamentals in some 70s recordings of pop music.
 

Mark Seaton

WBF Technical Expert (Speaker & Acoustics)
May 21, 2010
355
74
435
44
Chicago, IL
www.seatonsound.net
#4
Hi Mark,

My take on this is that while low order distortion in the bass is more tolerable (in fact, it's often used as a musical instrument enhancement in guitar amps), the ear is also more prone to detecting small amounts of it, where it's sensitivity to the harmonic frequencies is much greater, due to Fletcher-Munsen, than at the fundamental frequency.
The question Tom asked was related to harmonic order, and I wanted to answer that question directly first. The complication of our reducing hearing sensitivity to very low frequencies is an additional factor that super imposes on top of the harmonic spacing factors.

Fortunately we have two factors working in our favor with very low frequency reproduction in small rooms (not concert hall). ANY space will provide some rising transfer function below ~20Hz, where some rooms will have an increase in level at the seats of almost 10dB from 20Hz to 10Hz given the same output from a subwoofer. The other fact is that the hearing sensitivity curves compress and are closer at low frequencies than they are at high frequencies. Once you do break the audibility thresholds, a 4-8dB increase down low can subjectively sound equal to a 10dB change at 1kHz. In other words, once it's loud enough to matter, we are very sensitive to changes in VLF levels.

If you have a lousy woofer that, at 10Hz, has 5% THD at some useful SPL, you're going to hear the harmonics which will mask the fundamental.
Thank you Mark, I laughed out loud at the notion of a lousy woofer with only 5%THD at 10Hz. In room gain can make 5% THD at 10Hz possible for moderate levels, but that's hardly an easy task for even the best of woofers.

My own experience both in hearing the TRW-17 a few times along with my own prototype experiments having delivered 105-115dB in the 5-15% THD at the listening position with direct efforts to reduce high order distortion have been evidence enough that this makes for a great reduction in perceived distortion. Of course the further complication is that without a high output, low distortion reference, most would never identify the higher distortion options as distorted.
 
Aug 3, 2010
646
3
170
New Milford, CT
www.basspig.com
#5
Something is very wrong with the configuration of the example woofers you've link to there. It's too damned small and it's sealed. Naturally the cone excursion will exceed Xmax below 20Hz. An excursion of more than 1/4" @ 10Hz is a red flag that the tuning isn't correct for that level of bass extention. A properly tuned Hemholtz Resonator will achieve high fundamental output (≥120dB) with almost no visible cone movement. In fact, there is a military patent that uses a plain old Electro-Voice EVX180B to produce 130dB @ 10Hz with well under the driver's rated RMS power input. That's a woofer with 6.4mm of Xmax. Now if you swap that out for a true subwoofer that can do that excursion or better with .1% THD, you'll have the sort of clean infrasonics I've described. 105-115dB is barely on the threshold of hearing, especially at 10Hz. What I call useful output would be ≥125dB at minimum. It would be a farce to have a system that could produce 140dB @ 50Hz, but not at the important frequencies below 20Hz. Unfortunately, the size of the boxes required make commercial production impractical, thus the very low frequencies will be restricted to the domain of the DiY fanatics and the super-wealthy. Small box = high THD at the lowest frequencies because the driver becomes unloaded down there. A huge box and appropriate tuning can add 30 or more dB of real fundamental output in the 14-8Hz range. I recommend 38 cubic feet for 8Hz tuning and about 400 sq inches of vent area at the mouth of the vent. Yes, 5% THD is WAY too much distortion, because the ear won't be able to pick out the fundamental for being masked by all that distortion. Our ears naturally do what a nulling network does to extend the noise floor of distortion analyzers when measuring amplifiers. ☺
 
Aug 3, 2010
646
3
170
New Milford, CT
www.basspig.com
#7
Is that an iPhone application? Looks pretty slick for a take-it-with-you portable solution. Over here, I use the freeware and very excellent Room EQ Wizard that Ethan Winer turned me on to a few years ago. I can't say it leaves anything to be desired. It seems to cover everything I need, and then some.
 

MACCA350

New Member
Feb 12, 2011
3
0
0
#8
Is that an iPhone application?
AudioTools is the main app, they also have optional in-app modules, to be used primarily with their iAudioInterface which provides an included calibrated mic input and stereo analog output to/from the iPhone/iPod/iPad.

You can also use some USB audio devices connected to the iPad as the interface. Just found I can use my Sencore SP495 with it, might have a play around.

Cheers
 

Mark Seaton

WBF Technical Expert (Speaker & Acoustics)
May 21, 2010
355
74
435
44
Chicago, IL
www.seatonsound.net
#9
Hey Mark, have you tried this one out, they now have the Smaart Tools app

Cheers
AudioTools is the main app, they also have optional in-app modules, to be used primarily with their iAudioInterface which provides an included calibrated mic input and stereo analog output to/from the iPhone/iPod/iPad.

You can also use some USB audio devices connected to the iPad as the interface. Just found I can use my Sencore SP495 with it, might have a play around.

Cheers
One measurement savvy friend has played with the App on an iPad already and was waiting for them to ready the iPad hardware interface (iTouch/Phone currently available). Seems like a slick option, especially for those who are less PC friendly. As I said above, it matters much more that anyone can take repeatable measurements more so than what tools they prefer to do so.

Of course once you start taking measurements, you have to then learn what they can, and more importantly, the things they won't always tell you.
 

Mark Seaton

WBF Technical Expert (Speaker & Acoustics)
May 21, 2010
355
74
435
44
Chicago, IL
www.seatonsound.net
#10
4 subwoofer locations. One customer built dedicated theater. I took these measurements in working on placing 3 SubMersives in a customer's home about a year and a half ago. I hadn't had time to post the measurements before but thought this might be a good place to just throw out some examples of what we encounter in real rooms. I did take measurements at all seating locations, and the results were relatively similar laterally for each of the two rows. Below is a sampling of measurements with a SubMersive at different locations, but driven at the same level. In other words, they are driving the room with the same energy from different locations. The differences you see in the magnitude curves represents how efficiently that subwoofer location transfers bass energy to the listening position:



You can see the marked difference between two of the curves where the green curve is 15dB more efficient at 25Hz as one example. Obviously the complication is that no one curve will be most efficient at all frequencies. I'll post up the curves from the integration of the subs later.
 

Mark Seaton

WBF Technical Expert (Speaker & Acoustics)
May 21, 2010
355
74
435
44
Chicago, IL
www.seatonsound.net
#12
Been a busy week here, but finally took a few minutes to parse out some other measurements stashed away on the computer...

I should remind those reading along that this is not a complete picture of the bass response in the room, as we'd need an array of measurements to see further what's going on. The way we can empirically sort things is by measuring across the locations of interest and noting the differences between them. The measurements above and below were after we had already hunted through the subwoofer locations and their interactions across the seating area.

Given the need for plenty of output and efficiency, and provided we had plenty of EQ available, we chose from the 4 locations shown above, where 2 of the SubMersives had to stay up front, and we had flexibility in the location of the third. Below the magenta curve shows the response of the front *pair* of SubMersives. The red curve shows the response from a single SubMersive at the back left corner of the room where it is getting the same drive level of each of the front two. All 3 are contributing equally.

If we clustered/stacked the subwoofers outdoors, going from 2 to 3 subwoofers we would expect to generally see a 3.5dB gain (20*LOG10(3/2) = 3.5dB) at low frequencies. Remember that 2 sources producing equal level and maximally combining (in phase) will produce a 6dB increase in level. The green curve is the simple combination (at the listening position) of driving all 3 subwoofers equally. We can see that the combination varies from actually being lower than either curve around 22Hz to being 4.5dB greater around 45Hz. This suggests there is potential for a more efficient combination. The trick is to find a balance of efficiency at a specific frequency vs. the smoothness and attenuation of big peaks (ie 20Hz).

With the rear SubMersive closer to the listener, I started with adding delay to this sub as this is the correct direction to insure we reduce any localization tendencies, especially when driven hard. Basic precedence effect tells us that the first sound to arrive dominates the perceived origin of the sound. Unfortunately low frequency sound does not follow a simple line-of-sight rules, so the exact setting requires some testing, and even a little listening to check that you haven't created any unintended issues (ie localization). The yellow curve came from adding just ~3ms of delay to the rear sub. At 45Hz this brought the combination to 5.5dB, and overall raised the level by 1-2dB over the majority of the range. Note the subwoofers are all producing the same level as in the green curve. The added delay just helped them combine more constructively while also better insuring there isn't any localization. As you can see with the cursors at 31Hz, adding just a single SubMersive in the right location yielded a full 9.5dB gain, which is the same as having the power of 3x the starting subwoofers (magenta curve). Since the magenta curve is a pair of SubMersives, that's akin to 6x SubMersives operating from the front of the room.

The lesson to be gained here is that as with many things, location matters, and can be dramatically more efficient than a brute force electronic correction:

 

Mark Seaton

WBF Technical Expert (Speaker & Acoustics)
May 21, 2010
355
74
435
44
Chicago, IL
www.seatonsound.net
#14
In your experience, how much granularity is needed for the distance setting in order to get the best results? Is 1/2 ft (approx .5 ms) fine enough?
Hi mojave,

0.5ms is about 1/2 period at 100Hz, and 1/4 wavelength at 50Hz, making for a reasonable starting increment. You can tinker with finer changes, but you will see most of the changes at the upper frequency range. The lower the crossover or the lower the frequency the cancellation is, the larger delay increments required to make noticable changes.
 

bwraudio

New Member
Jan 24, 2011
54
1
0
#15
Hearing is believing

Hi Mark,



The question Tom asked was related to harmonic order, and I wanted to answer that question directly first. The complication of our reducing hearing sensitivity to very low frequencies is an additional factor that super imposes on top of the harmonic spacing factors.

Fortunately we have two factors working in our favor with very low frequency reproduction in small rooms (not concert hall). ANY space will provide some rising transfer function below ~20Hz, where some rooms will have an increase in level at the seats of almost 10dB from 20Hz to 10Hz given the same output from a subwoofer. The other fact is that the hearing sensitivity curves compress and are closer at low frequencies than they are at high frequencies. Once you do break the audibility thresholds, a 4-8dB increase down low can subjectively sound equal to a 10dB change at 1kHz. In other words, once it's loud enough to matter, we are very sensitive to changes in VLF levels.



Thank you Mark, I laughed out loud at the notion of a lousy woofer with only 5%THD at 10Hz. In room gain can make 5% THD at 10Hz possible for moderate levels, but that's hardly an easy task for even the best of woofers.

My own experience both in hearing the TRW-17 a few times along with my own prototype experiments having delivered 105-115dB in the 5-15% THD at the listening position with direct efforts to reduce high order distortion have been evidence enough that this makes for a great reduction in perceived distortion. Of course the further complication is that without a high output, low distortion reference, most would never identify the higher distortion options as distorted.
I have owned the TRW-17 rotary subwoofer for almost 3 years in
combination with Magneplanar Tympani IV bass panels (modified)
and the low bass is so clean, powerful and realistic. Once heard
there is no going back to conventional subwoofers. The lower
perceived distortion is very easy to discern and one of the reasons
it produces this low bass is it
intrinsically operates in a low pressure, high volume mode, exactly like the original live bass sounds, and the opposite of conventional subwoofers. Sonically, it is instantly obvious that the bass quality from this low pressure TRW subwoofer sounds natural and has the correct bass quality, from all bass events, be they a bass viol, the large radiating sounding board of a grand piano, the huge volume moved by a bass organ pipe, a bass drum, or even a cannon shot from a film soundtrack. And, in direct comparison, it is instantly obvious that the bass quality from conventional subwoofers sounds all wrong in quality, like a high pressure, low volume pump flailing away in a vain attempt to produce some semblance of bass sound.
 

Attachments

May 9, 2011
97
0
0
#16
Fletcher-Munsen, than at the fundamental frequency. If you have a lousy woofer that, at 10Hz, has 5% THD at some useful SPL, you're going to hear the harmonics which will mask the fundamental. Down in the very low frequencies, ANY small harmonic distortion, no matter how small it may look on paper, is going to be audible, due to the ear being already starved of sub-20Hz sensitivity. The ear is at a 30+dB disadvantage down there, so woofers have to have a correspondingly lower harmonic distortion level. Where -30dB harmonics might be acceptable on a midrange driver, a subwoofer needs to have those harmonics down better than 55dB below the fundamental, or they will be easily spotted
Quite so! Although -30 dB harmonics should not be acceptable on a decent mid, it is true that bass harmonics have a huge audibility advantage over their fundamentals. Many people extol the imagined virtues of their systems' bass, when in actuality what they are perceiving is doubling and tripling. Furthermore, the temporal relationships of existing bass harmonics is often hopelessly mangled up, because of the cascaded effects of multiple low-pass filters in the recording and reproduction chains. Add the smearing effect of bass reflex ports and satisfactory bass reproduction at all seems like a small miracle.
 
Aug 3, 2010
646
3
170
New Milford, CT
www.basspig.com
#17
Definately true that -30dB would NOT be a quality midrange, but I pulled that number based on average listener tests of harmonic distortion on a musical waveform and most people no longer detect the distorted version as different from the undistorted version once distortion is reduced below -30dB. At mid frequencies, our ears have a somewhat even playing field. But at 10 hz, the field looks like a steep slope or even a cliff with the ease of hearing harmonics at the top and the fundamental at the bottom. Even 0.5% THD is going to produce audible sound that's very different than the 10Hz fundamental.
Woofer performance can be verified with a calibrated mic feeding a dual channel oscilloscope with the original sine wave signal on channel B and the microphone signal on channel A with inputs summed and channel B inverted. When the levels are matched between channels, nothing but a flat line should appear on the scope. Any junk that appears is either ambient noise or distortion.
 

Mark Seaton

WBF Technical Expert (Speaker & Acoustics)
May 21, 2010
355
74
435
44
Chicago, IL
www.seatonsound.net
#19
Indeed. Do you think a measurement like the one you describe is robust even if done non-anechoically?
It will make for an interesting display, but waveforms are rather hard for most to conceptualize into something we might hear. I would also point out that there are many more potential distortion mechanisms in a midrange driver vs. a woofer. In most cases the woofers are covering a range where the wavelength is much larger in dimension than the cone and the enclosure. This is not so in a midrange, and other issues creep into the picture. At very low frequencies we also have limitations of hearing thresholds which work both for and against us in terms of distortion audibility.

Of course it's easy to wave our arms about the evils of distortion, but combined testing and observation of real devices are what really matters. The TRW has its own set of limitations and major strengths. Not every situation or system is a good fit for this solution. I am not of the opinion that you can hear how bass is being reproduced, but rather we hear symptoms common to a given executions and some factors related to a given design. There are many ways to accurately reproduce bass, and you can find applications where one of the various approaches can be well justified.
 
May 9, 2011
97
0
0
#20
My point was directed not at hobbyists, but rather at speaker manufacturers who cannot afford an anechoic chamber (ie most of them). In essence they are trying to build a car without a wind tunnel. Any claim about stability at high speed will be on word of mouth ("trust me, I've driven it really, really fast, and it's really, really stable") coming from a financially interested party, themselves. The truth content of such a statement will then depend on luck, not data, and lots of corpses strewn along highways may ensue.

If the motor car analogy seems far-fetched to some, let me point out that price-wise it is often rather valid.
 

About us

  • What’s Best Forum is THE forum for high-end audio, product reviews, advice and sharing experiences on the best of everything else. A place where audiophiles and audio companies discuss existing and new audio products, music servers, music streamers and computer audio, digital to audio converters, turntables, phono stages, cartridges, reel to reel, speakers, headphones, tube amplifiers and solid state amplification. Founded in 2010 What's Best Forum invites intelligent and courteous people of all interests and backgrounds to describe and discuss the best of everything. From beginners to life-long hobbyists to industry professionals we enjoy learning about new things and meeting new people and participating in spirited debates.

Quick Navigation

User Menu

Steve Williams
Site Founder | Site Owner | Administrator
Ron Resnick
Site Co-Owner | Administrator
Julian (The Fixer)
Website Build | Marketing Managersing