Jump to content

what makes a sub "sound good"


jcr159

Recommended Posts

hopefully this hasn't been asked here before, I couldn't find anything....

 

What makes one sub "sound" better than another?  vague question, and i'm sure going to cause a flame war...  at least it would on AVS, hence asking here...

 

I know all the data and research about how much less sensitive we are to distortion at lower frequencies, etc.  So I don't exactly need pointed in that direction.

 

 

Here's where I am...   Was planning to go DIY for my theater build out that is in the design stage (the whole basement is actually...)  The original plan was to DIY 2 or 4 subs depending on size and layout.  I'll be using opposing corner or opposing wall placements as well which line up roughly with the harman and toole research.  not looking for boosted output, but more even room response in that sense.

 

based on the room volume, i'll be in the SVS PB 13 ultra or 16 size range, but know i can build some sealed 18's or heck even the SI 24's sealed for smaller and less money than that.  What i'm looking for is whether there would be any sound quality difference to be had by picking different drivers or from an SVS (or HSU/outlaw/PSA/etc.) route.

 

I'm a data guy to be honest...  so don't hit me with subjective hooey gooey...   This is data-bass after all where we look at charts and numbers, etc...

 

To end cap things, I created this post because on a whim i was comparing the distortion measurements here (thinking it was an indicator of sound quality in some respect at least but not sure how much).  I was looking at the Dayton UM-18-22 i think sealed vs. the SVS PB SB13 Ultra which is quite expensive, and was shocked how much less distortion was measured on the Dayton.  I didn't think that particular Dayton was considered fantastic either.  It shocked me...  I then went looking at the Stereo Integrity 18 models (HT and HST) and was shocked how much more distortion they had as well...   So I then wondered if any of the tested drivers could do less than 5-10% distortion down to 20Hz and below, and clicked on the SI 24.  it hit that mark easily.  Made me think of the adage "there's no replacement for displacement".

 

So now here I am asking for advice, lol.  I know the distortion numbers don't matter as much in the bass/sub-bass as they do in the midrange as the ear is much less sensitive, but if i can build and get significantly better quality bass out of some dayton parts than a pb13, i just saved a lot of cash to go buy a front stage...  

 

Thanks for any thoughts!

-j

 

 

Edit: changed to reflect I was in fact comparing the SVS SEALED not PORTED...   I was comparing Sealed to Sealed...

Link to comment
Share on other sites

Sound quality is so subjective it's really difficult to say anything definitive even with all of the data for 2 systems being compared gathered here. The reason is that what sounds good or "right" to one person may be completely different from another. Some people prefer less bass overall, less extension, a peaky response in the 40-70Hz range even or might prefer a bit of compression or 2nd harmonic distortion, as that is what they are used to hearing and sounds best to them. Really there's nothing wrong with that. I prefer the sound of a bass system that goes very deep, is run a bit hotter than flat, has a relatively smooth response shape and has monumental amounts of headroom. I've had systems like that for a while and it's what I'm used to.

 

You'll often hear people say that there's more to it than sheer SPL capability and that's true but having insane amounts of headroom is a major component of "sound quality" for a number of reasons. 1 of which is it is an indicator of being able to reproduce dynamic peaks easily and keep distortion and compression effects minimized. Almost no subwoofer has what would be considered low distortion when it is being run near it's limits.

Link to comment
Share on other sites

Thanks for the reply. I gravitate towards "accurate" I think.... I like my paradigm studio towers.... I like my Etymotic headphones, so I'm likely to prefer the most accurate and least distorted bass... that said, I'd prefer enough headroom so whatever I end up with doesn't need to struggle to hit peaks... that "should"? Lower distortion and bad sounds overall I think...

 

I think the kicker for me in all of this was the realization that some of the popular DIY drivers are probably popular for cost and power/output reasons more than sounds quality reasons....

 

All that said, I realize the slight ridiculousness of talking about bass/sub-bass sound quality since the human ear is much less sensitive to it... on top of that the room sizes(small room acoustics) are probably so dominant that driver and sub design may only account for a fraction of what you are hearing anyway, but figure if I'm going to build or buy, why not get something good... lol

Link to comment
Share on other sites

 

 

To end cap things, I created this post because on a whim i was comparing the distortion measurements here (thinking it was an indicator of sound quality in some respect at least but not sure how much).  I was looking at the Dayton UM-18-22 i think sealed vs. the SVS PB 13 Ultra which is quite expensive, and was shocked how much less distortion was measured on the Dayton.  I didn't think that particular Dayton was considered fantastic either.  It shocked me...  I then went looking at the Stereo Integrity 18 models (HT and HST) and was shocked how much more distortion they had as well...   So I then wondered if any of the tested drivers could do less than 5-10% distortion down to 20Hz and below, and clicked on the SI 24.  it hit that mark easily.  Made me think of the adage "there's no replacement for displacement".

 

So now here I am asking for advice, lol.  I know the distortion numbers don't matter as much in the bass/sub-bass as they do in the midrange as the ear is much less sensitive, but if i can build and get significantly better quality bass out of some dayton parts than a pb13, i just saved a lot of cash to go buy a front stage...  

 

 

You are comparing the distortion of sealed subwoofers to a ported subwoofer, and that is an apples-to-oranges comparison. Sealed subwooofers will always have much more distortion at the lower frequencies. Another thing is you are using SVS subs to compare them against, and SVS is very strict with their limiters. Most manufacturers are more permissive of distortion quantities than SVS. If you placed a Ultimax 18 or SI18 driver in a ported box, you would very likely end up with less distortion for the same output level than SVS, so long as the ported enclosure is semi-competently designed and built. About the SI24 driver, in order to achieve that low distortion level, Josh used a very large cabinet: 43"x36"x23". If you can not handle a cabinet that large, you will have more low-end distortion. That said, the distortion measurements for the SI24 are really terrific, and, in that enclosure, do resemble that of a ported subwoofer. 

Link to comment
Share on other sites

Thanks for the response, but I was reading the DVD pb13 ultra sealed results against the ultimax 18-22 which I thought was also sealed... I may have read the wrong thing and will double check, but thought I was comparing all sealed including the pb13... maybe the confusion was that the "on" should stand for "ported"?

Link to comment
Share on other sites

Are you saying that the ported designs generally have lower distortion?(presumption on my part is that would be above tuning)...

Yes. Near port tuning, the moving parts do not travel far, although there is great pressure on them. Since the moving parts are in the 'comfort zone' and well within linear travel, they do not generate much distortion. The further the moving parts (cone, voice coil, former) have to move, the more tension they put on the soft suspension parts, and the further away the voice coil gets from the magnetic field that it is designed to operate nicely in. That all adds up to distortion. In a sealed subwoofer, cone travel has to increase four times to achieve the same SPL for dropping one octave. That isn't the case at all with ported subwoofers. 

Link to comment
Share on other sites

Yes. Near port tuning, the moving parts do not travel far, although there is great pressure on them. Since the moving parts are in the 'comfort zone' and well within linear travel, they do not generate much distortion. The further the moving parts (cone, voice coil, former) have to move, the more tension they put on the soft suspension parts, and the further away the voice coil gets from the magnetic field that it is designed to operate nicely in. That all adds up to distortion. In a sealed subwoofer, cone travel has to increase four times to achieve the same SPL for dropping one octave. That isn't the case at all with ported subwoofers. 

 

But are there potential issues with phase shift??

 

 

 

Either way, I will have to put my hand up now and shout "Ooh, ooh!  PPSL!  PPSL!" :P lol

 

I posted a few links to some distortion data sources on page 1 of my build thread :)

 

http://data-bass.ipbhost.com/index.php?/topic/179-push-pull-slot-loaded-quad-si-ht18s/

Link to comment
Share on other sites

From what I understand, PPSL designs can reduce even-order harmonic distortions. However, for all the sealed subs I have dealt with, the even-order harmonics weren't the worst offenders. Nonetheless, less distortion is better, so if one were to mount two drivers on a sealed enclosure, why wouldn't anyone want a PPSL design (aside for maybe aesthetic reasons)? Is there some other performance trade-off for PPSL designs? What is the catch?

Link to comment
Share on other sites

PPSL's use larger cabs because of the reverse mounted driver and mounting configuration. Also that driver puts the motor and moving parts outside of the cabinet and mechanical noises can become an issue if large excursions are used. These are much more clearly audible to me than harmonic distortion, even though I like to think I have a pretty good ear for HD as well.

 

Harmonic distortion is only one form of distortion to worry about. There are many types. Driver mechanical noises, cabinet rattles, buzzing, port wind noise, dynamic compression, ringing, response shape, extension, etc...

Link to comment
Share on other sites

I was reviewing subwoofers for AH at the time that was published. It mainly focuses on harmonic distortion though and misses the point I was making above. Focusing on HD as the primary indicator of SQ or accurate reproduction is a mistake IMO.

Link to comment
Share on other sites

I would say the difference between harmonic distortion (and also IMD) is that it is inevitable, unlike the types of distortion you listed (excepting compression). It will crop up no matter what, except if the driver is very underpowered. The other types of distortion, like cabinet vibrations, flutter noise, etc., may or may not be there in any significant amount. Harmonic distortion also gives one a good idea of when the cone is at the edge of linear excursion. For example, the CEA-2010 distortion thresholds seems to be more tied in with maximum linear excursion than it is with the audibility of distortion. The distortion of CEA-2010 thresholds are easily audible, but the subs don't usually have much left to give after surpassing those thresholds. 

Link to comment
Share on other sites

To answer the question in the title of the thread:  What makes a sub sound good is to be part of an optimized system.  The optimized system includes:

  •  One or more subs placed at room locations that provide good coverage across the listening area and minimal excitation of modal resonances.
  •  Possible electronic DSP/EQ to help shape the in-room response and further control resonances.
  •  Possible Room treatments, to the extent that they are useful.  (Most aren't in the sub bass range.)
  •  Enough excursion and amplifier headroom to ensure reasonable linearity at all times during operation.

I say reasonable linearity because all subs exhibit slight non-linearity, even at very small excursions and small amounts of power.  Linearity is important for accurate sound but is especially important when using EQ to ensure the filters behave as expected.  All subs also reach a point beyond which non-linear distortion rises rapidly and the excursion is at its limit.  This ultimate limit is approximately revealed by CEA testing and might be described as "usable Xmax" to distinguish it from the Xmax that the manufacturer might rate it.

 

Where distortion measurements may be of more interest is when considering what happens as excursion approaches the usable Xmax.  In some systems, distortion stays quite low right up to the excursion limit, but in others, it rises quite a bit well before the excursion limit.  Likewise, it's also worthwhile to review the power compression data, which reveals a more subtle form of non-linear distortion in which output is diminished relative to what would be expected at high levels.  Both of these depend on the cabinet alignment (e.g. sealed vs. ported and box size) as well as the driver.

 

Another issue is that the THD measurements only reveal distortion that arises from sine wave content.  Complex content may produce inter-modular distortion (IMD), which is regarded to be more audible and offensive, and it's not straightforward to predict how much IMD will be present with a particular complex signal given a certain measured THD.  As a particular issue, the sine sweep THD measurements do a poor job of revealing non-linear inductance with excursion because inductance primarily affects high frequency signals but high excursion usually requires strong low frequency signals.  Therefore, you may not observed IMD from inductance non-linearity unless testing with content that contains both low and high frequencies at the same time.

 

Non-linear distortion may arise due to the presence of high currents in the coil or fluctuations in the magnetic field caused by the coil (flux modulation distortion), in addition to high excursion.  These other sources of non-linearity may or may not appear in the THD sine sweeps and may be difficult to distinguish from other sources of non-linearity in those measurements. 

 

Many drivers feature sleeves, shorting rings, and other mechanisms to try to control flux modulation and/or inductance non-linearity.  Using a driver in a cabinet alignment with higher efficiency within the intended bandwidth can reduce the amount of current in the coil, as well as heat production, which cause distortion and/or power compression.

 

Distortion is a complex subject, as you can see here, but if you go back to my bullet points, you'll see that distortion is actually a pretty small part of a good, accurate sub sound.  While a bad sub will probably sound bad no matter how it's used, a good sub can still sound quite bad if it's not set up to perform well in the room.  If it's an accurate sound you are after, we can discuss strategies for achieving that in the space that you have.  Plan on buying or building multiple subs and investing in some kind of EQ solution for best results.  While some people are reasonably satisfied using Audyssey MultEQ and related room EQ systems, I recommend going for manual EQ if you are willing to take the time to learn how to do measurements and make adjustments.

Link to comment
Share on other sites

Another issue is that the THD measurements only reveal distortion that arises from sine wave content.  Complex content may produce inter-modular distortion (IMD), which is regarded to be more audible and offensive, and it's not straightforward to predict how much IMD will be present with a particular complex signal given a certain measured THD.  As a particular issue, the sine sweep THD measurements do a poor job of revealing non-linear inductance with excursion because inductance primarily affects high frequency signals but high excursion usually requires strong low frequency signals.  Therefore, you may not observed IMD from inductance non-linearity unless testing with content that contains both low and high frequencies at the same time.

 

 

IMD in subwoofers has not been the subject of extensive research, but Ed Mullen and Ilkka Rissanen did a study on it and found that the amount of IMD correlates to the amount of THD: the more THD there is, the more IMD. The obvious takeaway from this is that if one is concerned with IMD in subwoofer bands, then just get subs that are low in THD.

Link to comment
Share on other sites

IMD in subwoofers has not been the subject of extensive research, but Ed Mullen and Ilkka Rissanen did a study on it and found that the amount of IMD correlates to the amount of THD: the more THD there is, the more IMD. The obvious takeaway from this is that if one is concerned with IMD in subwoofer bands, then just get subs that are low in THD.

 

 

I agree that this makes sense. All subs produce high amounts of distortion once the excursion gets high enough in the deep bass. Pay attention to the distortion levels above 40Hz at moderate to high output. This area usually has comparatively low excursion levels and is a good way to gauge which drivers are lower distortion / lower inductance and likely to have lower IMD. This tracks with inductance effects quite well. High inductance or non-linear inductance drivers exhibit much higher distortion in the upper frequency bands even at modest drive levels and I would expect them to exhibit higher IMD as well.  

Link to comment
Share on other sites

thanks for all the well crafted discussion.  still getting my head around everything and learning here.  

 

I think one of the takeaways i'm getting is that if quality and accuracy of reproduction is critical, it is probably best to design for more "sub" than i need to hit the volume levels i'm looking for...

 

for me it will be a given that i'm building 2 subs, placing them mostly optimally to get the best coverage across the room.  but if i'm reading things correctly, and need say an average 15" dayton driver in each box (sealed will be my choice for size reasons most likely) to hit reference, I may be better off building a pair of 18" subs that don't need to be pushed as hard.  that seems to make sense to me, but i don't know...

 

If that is correct, and one "over designs" the system to only use a small part of the output capability, will it keep things more accurate as a result?  i.e. - only using 30%-50% of xmax to acheive the output desired.  If doing that, does it ensure better quality output by eliminating the non-linear distortions, inductance and power compression, etc?

 

Sorry for so many questions, but trying to learn the best i can.  It seems like the only advice i can get on AVS is to buy a 10,000 gigawatt amp and buy the driver with the biggest xmax to get output...   Seemed like it is a bit more nuanced then that to get a good result, and appreciate the education i'm getting here.

 

thanks again,

-j

Link to comment
Share on other sites

 

If that is correct, and one "over designs" the system to only use a small part of the output capability, will it keep things more accurate as a result?  i.e. - only using 30%-50% of xmax to acheive the output desired.  If doing that, does it ensure better quality output by eliminating the non-linear distortions, inductance and power compression, etc?

I think that your understanding here is largely correct. But what will cause the most distortion is the room itself, ie linear distortion of the frequency response. The audibility of THD is not much compared to what the room acoustics do to the response. I would say yes, its good to do overkill with the subs, but take measures to achieve a flat response as well. In fact, you might want three subs instead of two. I would rather have three 15"s than two 18"s, since the former will go further towards getting a flat response (so long as the subwoofers are intelligently placed).

Link to comment
Share on other sites

I think that your understanding here is largely correct. But what will cause the most distortion is the room itself, ie linear distortion of the frequency response. The audibility of THD is not much compared to what the room acoustics do to the response. I would say yes, its good to do overkill with the subs, but take measures to achieve a flat response as well. In fact, you might want three subs instead of two. I would rather have three 15"s than two 18"s, since the former will go further towards getting a flat response (so long as the subwoofers are intelligently placed).

 

Ok, thanks ShadyJ.  I was thinking along the same lines...  Room will dominate what i hear a lot, but if the reproduction is crap from the sub itself, no overcoming that...  to that end though, the room will need treated properly to get the best sound quality and i haven't worked out that plan yet.  as for 3 subs, i actually was planning 2 or 4, as my reading so far has shown that i'm more likely to get a good result when planning out ahead of time.  in other words, if i need to design locations into the room layout without have a built room to test first, i'm less likely to screw it up by sticking to 2 or 4 subs...  since 4 might take a bit too much real estate for spousal acceptance, i was leaning toward 2 knowing that i'm leaving something on the table...

 

one will be located behind an acoustically transparent screen.  Maybe i'll get creative and build the second into some furniture in the back of the room...   i'd love to double up on 18's and put 2 in each enclosure dual opposed, but then maybe i should just go to a 21 or 24...   lol

 

-j

Link to comment
Share on other sites

You are on the right track. Yes a speaker operating at 25-50% headroom is almost always going to have lower distortion, compression, self noise, etc, than a speaker operated at 75-100%. The room is a BIG factor. It makes major impacts on the frequency response, relative distortion, decay characteristics and overall headroom. In other words it modifies almost everything. Some for the better generally, such as lowering deep bass distortion and others badly, such as decay rates. Some of this can be improved by using multiple sub placements, smart seat positioning, and equalization and other dsp. Room treatment can help too, but is often limited to the upper bass range / mid bass, due to the physics involved with wavelengths any longer than that. Equalization is used as the final step quite often and it can cause the sub/s to be driven harder over certain frequency bands. This again goes back to having as much headroom as possible so that sub/s and amps will have plenty of headroom for any EQ applied.  

Link to comment
Share on other sites

Room treatment is most effective at and above 100 Hz, and then only if the absorbers are thick enough.  They should ideally be 4"-6" thick and installed with some air gap if possible.  With 6" of the right material with an air gap, you may be able to get a little bit of absorption down to almost 60 Hz or so, but they will still work much better at higher frequencies.  Some bass traps designed for corners are up to 12" thick, which is enough to actually put a dent in a problem at 60 Hz.  You can also find specialty products like tuned membrane absorbers that absorb in a very narrow range of frequencies, hopefully at a problematic room mode, but you still need to cover a fair amount of surface with them for good results.

 

So as a practical matter, you'll get a lot farther with strategic sub placement and EQ to tame subwoofer bass.

 

At the same time, a lot of the bass you hear and can feel is above 100 Hz, so room treatments *do* make a big impact on the accuracy and enjoyment of bass.  I believe it's also important to have capable mains with a lot of capability for the 100-200 Hz octave, and then these need to be adequately balanced with the subs so that one doesn't drown out the other.  Ironically, too much sub can totally kill the bass, just as too little sub can.

 

On that note, what kind of mains do you have?

Link to comment
Share on other sites

Good stuff SME... right now I'm running Paradigm studio 60's and the matching LCR center, but when I build out the theater was thinking about changing up to the sierra RAAL's... I'm torn between bookshelves and towers since I'll only be about 8-9 feet, but the space will be open to about 5200 cu ft total (no real doors to other areas).

 

Makes me think I'll need some driver area than the bookshelf models have.... but the towers are so much pricier

Link to comment
Share on other sites

Archived

This topic is now archived and is closed to further replies.

×
×
  • Create New...