The case againts subwoofers

[SME]

Welcome @RCAguy !   Frequency response specs for a sub, even if done "honestly", often don't tell you much about its real capabilities.  A DSP (including the one built-in to most powered subs) can re-shape the frequency response, and a lot of commercial subs rely on DSP bass boost to achieve their specs.  Of course EQ can't compensate for the physical limitations of the driver and available amp power, but in practice it is often very beneficial to design a sub whose innate response starts rolling off much higher than its intended bandwidth and then compensate for this as needed using DSP/EQ.

Since you're new here, let me recommend you check out the main site at https://data-bass.com/ where you'll find measurements of various subs taken to their limits at frequencies across the spectrum.  Not only is DSP bass boost very common in commercial integrated designs but it's also becoming very popular among DIY sub builders as often the better drivers have natural roll-off that can be easily compensated for.  What really matters for bass performance is high displacement and sufficient amp power.  And doing it "cleanly".

Distortion is an extremely complex subject, one that can't be understood by simply considering THD% at different frequencies.  The relative distribution and structure of harmonics is also important as is the particular content and its characteristics.  There's no simple answer to how much THD is acceptable, and I don't think there ever will be because the notion represents a misunderstanding of how perception works.  In my opinion, *linear* response of the installed system is the most important attribute for sound quality.  For low frequencies in small rooms, this is a major problem with no easy solution as yet.  Strategic acoustic treatments can help but are only a partial remedy.

Linear response directly impacts the structure of both naturally occurring harmonics and harmonics produced by distortion (often on purpose during production---to improve sound quality).  Objectionable linear response flaws likely introduce irregularities in the presentation of harmonic structure, which interferes with the brains means of isolating sounds from one another in the acoustic scene.  This causes listener confusion, hearing of spurious auditory content (i.e. "hearing treble" or "hearing bass" that's disconnected from the rest of the sound), and harsh or uncomfortable sensations felt in the ears or even other body parts.  If music at 75 dB "hurts your ears", it's probably your brain telling you it's hearing something unnatural that it doesn't know how to process.  The "harsh treble" you think you're hearing might actually be caused by problems in the bass.

Anyway, this this stuff interests me a lot, and I'm working on new methods to improve linear response for low frequencies (and high frequencies too).  I have a DSP technique that sounds *very promising*; it seems capable of almost entirely neutralizing the acoustic signature of the listening room in addition to the speakers themselves.  It is in this kind of situation that I think low THD for subs in particular may be important.  Linear DSP can be used to alter the linear response to whatever I like, but this processing cannot touch any distortion produced the amp or subs.  For a speaker, that's not too big of a deal, especially if the driver response is fairly clean without EQ.  Most of the distortion produced is likely to be quite regular, so even if it becomes audible it's likely to sound natural and pleasing.  However, for bass and especially the sub range, the room interactions make the distortion much more irregular and therefore much more likely to sound objectionable, and there's no way to fix this with linear processing.  I have no idea how important this is in practice.  My system uses subs with very low distortion as far as subs go and has enormous headroom.  It might be a bigger problem for small speakers pushing their bass limits all the time.  Or pro speakers under similar circumstances.  Servo control might be a key solution.

Please forgive my ramblings.

[maxmercy]

The article listed on that page:

http://www.filmaker.com/papers/RM-WhtPpr_Subwoofer Camp.pdf

Brings up something that few do, except MonteKay:

http://www.mfk-projects.com/Home_Theatre/theatre_woofer.html

It is a very important point.

Due to the way our ears 'hear', a near infrasonic sound must be played back more cleanly than a midrange sound in order not to have upper harmonics 'color' it due to equal loudness curves.  I did several experiments with my first decently low distortion subwoofer system and it is enlightening seeing total harmonic distortions of just above 1% significantly coloring a pure sinewave tone upon playback at louder levels.  Of course, at loud enough levels, things rattling in the room 'color' the sound far more, until you start fearing for the structure.

Clean reproduction of bass is very difficult.  If you can play a 20Hz tone and with your eyes closed you can point to where the subwoofer is in the room, it is anything but a clean reproduction.  

JSS

[SME]

13 hours ago, maxmercy said:

The article listed on that page:

http://www.filmaker.com/papers/RM-WhtPpr_Subwoofer Camp.pdf

Brings up something that few do, except MonteKay:

http://www.mfk-projects.com/Home_Theatre/theatre_woofer.html

It is a very important point.

Good to see you back on the forum!  See my above post.

I would have agreed with all of this like three years ago.  Now I'm a heretic.  I believe this kind of analysis is seriously flawed.  Equal Loudness Contour data is interesting, and I think it describes real trends in apparent loudness vs. frequency.  However, it is a big mistake to generalize about complex music signals from data obtained using pure tone sine waves.  I'd argue this is *especially important* when trying to reason about audibility thresholds.  Pure tone sine waves are very unusual in nature, and our brain's reaction to them may not be a reliable indicator of how it reacts to complex signals.

For example, I believe most people hear to 20 kHz and beyond, but not sine waves.  The same is likely true at the bottom of the range.  If auditory perception is an inferential process as I argue, then the simultaneous presentation of associated harmonics (or sub harmonics) increases the listener's ability to accurately hear at the extremes.  That's because those associated harmonics are almost always present to a degree or other, so it makes the information about frequency extremes more "trustworthy".  It's a kind of redundancy of information that increases confidence.

On the flip-side, we're highly dependent upon harmonic structure to make sense of anything that we hear, so it's very hard to pin down the perceptual consequences of different kinds of distortion.  I wish I could remember who it was, but I saw a lecture from (IIRC) an old Bell Labs perceptual coding research engineer who would play two audio samples at his seminars, one with like "10% THD" and one with like "0.1% THD".  The latter one sounded far worse than the former being that he cherry picked the examples to illustrate his point.  The one thing I *know* about distortion is that real life content  comes loaded with it and "how that distortion sounds" on a playback system is highly dependent on how neutral the playback system's linear response is.

Ironically, the harshness that listeners often associate with "distortion" (not an unreasonable association, given that harshness does audibly increase if the amount of added distortion increases) is likely not the fault of the distortion per se but of the linear (i.e. "non-distorting") aspects of the playback system which make a mess of the harmonic structure.  I hope to be able to demonstrate this for interested audiences some day soon.  A key seems to be in the low frequencies because the lows are the foundation of the harmonic structure, to which everything else is perceptually compared.  The lows are also the primary source for distortion harmonics because that's where most of the energy is.  So if the lows are out of whack, as they almost always are in small listening rooms, then the mids and highs will be out of whack too.  I'd bet audiophile listeners often blame the "mids" and "highs" for many sins committed that are actually committed in the bass.  I've naively done it.

[maxmercy]

I see what you mean, I was only talking about ELC and how it relates to my experiences with both sinewaves and real content, in a "high fidelity" system.  I agree, linearity is a huge deal.  Listening/feeling experiences with an old system of mine with large peaks at 30 and 60Hz were significantly different from a later system that had much better control of the room via acoustic treatment, mode cancellation via multiple subs, proper headroom and proper seat placement with some EQ thrown in to tame the largest leftover resonances...that system was very good.

What is strange is that in some movie scenes, the tactile effect of the old, peaky system was far more present.    

Complex subject indeed.

JSS

[SME]

1 hour ago, maxmercy said:

I see what you mean, I was only talking about ELC and how it relates to my experiences with both sinewaves and real content, in a "high fidelity" system.  I agree, linearity is a huge deal.  Listening/feeling experiences with an old system of mine with large peaks at 30 and 60Hz were significantly different from a later system that had much better control of the room via acoustic treatment, mode cancellation via multiple subs, proper headroom and proper seat placement with some EQ thrown in to tame the largest leftover resonances...that system was very good.

What is strange is that in some movie scenes, the tactile effect of the old, peaky system was far more present.

Peaks at 30 and 60 Hz are harmonically related, which means they likely potentiate each other perceptually.  And I wouldn't be surprised if those movie scenes happened to have peakiness at those same frequencies.  I suspect this is a trick used on some live sound stages---EQ boost at the kick drum fundamental and a few of its harmonics, tweaking the relative levels by ear as needed---to get a very powerful slam effect, albeit at the cost of lost clarity of content elsewhere in the spectrum.  This may be the "easy" way to get slam, but the result is very much conditioned upon the content, which is easy to control in a live setup.

The "hard" way, which delivers excellent (and might I add *detailed*) tactile sensation and slam for a variety of content, is to target total neutrality.  I don't believe this is possible to achieve in a "normal" listening room using conventional methods including near-field listening and treatments.  Those things can certainly help, but even tiny residual issues can profoundly alter perceived response, including tactile aspects.

This is the primary reason it's taking me so long to debut my tech.  My own listening room is not stable enough to give consistent sound from day-to-day, and much of my time has been spent figuring out precisely which variable acoustic features in my room are most responsible for changes.  This is much harder than it sounds because of how erratic perception behaves under these circumstances, and that's also something I've had to discover and learn about as I go.  Ultimately this whole experience has profoundly altered my understanding of how auditory perception works.  All these seemingly non-intuitive things make way more sense under an alternate paradigm, one which I hope will be able to offer a rational explanation for a lot of other confusing and mysterious things about audio.  We'll see.