Subwoofer shootout in Ultimate AV (1 Viewer)

[Keir H]

Bringing this thread back up a bit, any new comments on Part III of his series? I wish he had a HSU or SVS to test and compare since I think his methods were very educational, although most of these subs are over $2500. I thought the B15 would do better at 20hz but I was impressed of what the Velodyne measured esp for the size. Of course these numbers would be elevated inroom correct? But by how many db?

 

[MingL]

I just picked up this mag....... have not read it yet, but in the line up are
1) Codrive CB1
2)Snell ICS sub24
3) Triad inroom platinium
4) Velodyne DD18

Will read it tonight.

 

[Edward J M]

Tough call.

You've got additional boundaries indoors (walls and ceiling). But rooms can be "lossy" and you typically won't see the full theoretical boundary gain from each surface.

You've got room gain, but that will vary greatly (in both amplitude and where it starts to manifest), with room size.

Then you have the distance from the sub to the listener, which typically is considerably longer than 1 meter or 2 meters (the typical ground plane test distances). And SPL varies inversely with the square of the distance, so what you gain in boundary effects and room gain, you might then lose with a long listening distance.

So, it depends.....

 

[Mark Seaton]

Ed's learned well. "It depends..." is the only correct answer here. A better question would be what range of average, relative levels will we see at the listening position. While hardly scientific, a good guestimate would be that with better placement in smaller rooms (maybe less than 12' x 18') you can see as much as 1-7dB through most of the range with gain increasing below the mid 20s. In larger rooms (as large as 25' x 36') you may see anywhere from 10-20dB less output than seen at 1m in half space. In all but the most open and leaky of rooms you will usually see this relative output start rising in the low 20s reaching -9 to -3dB in the mid teens. Construction will certainly have a major impact on lower frequencies. Unfortunately there is no ideal solution in a home unless you can afford lots of THICK absorption followed with a very solidly constructed outer shell. The idea would be to contain the lowest frequencies where gain is realtively smooth throughout the room, and reduce the reflected energy at higher frequencies to reduce the seat to seat variations.

 

[Keir H]

Thanks Ed, you made things clear for me on your great explanation. Indeed an impressive lineup of subs included on this test, just wishe there was a part IV with SVS,HSU,M&K(can they compete though?)Krell,Linn... I love comparison test as it helps you gauge where you are in your own woofage at home or hope to be able to obtain.

 

[MikeDuke]

Even though my review of the sw900 may be less then perfect, it is still a good subwoofer. I think the JM labs sw900 deserves a test like this.
Mike D

 

[Mark Seaton]

Hi guys... Correction to make here:

The numbers I first posted were relative to free space, not 1m on the ground. As such, you can subtract about 6dB to those numbers. I'll edit the post.

 

[Zack_R]

The third installment is now up on the website.

here

I really liked his test especially the waterfall output to source comparisons. I'd have liked to seen this method used for bass music as well the movies. The Genelec HTS6 in the earlier installment, really did an outstanding job of mimicking the source.

However, I wish he had made a slight adjustment to this test. IIRC, Keith played these scenes over and over and louder and louder until his ears detected audible distortion. At this point he backed the volume down one dB and recorded the graph.

I'd rather him back down a full 3 dBs to better ensure that the sub is still not struggling against itself during these passages. In this way the sub is still almost full tilt and we can really see how well it articulates these dynamic passages. By backing down only 1 dB I feel some subs still may have been playing above their limits if his ears didn't catch the distortion soon enough.

I thought the power compression was interesting but some what misleading.

For example, compare the power compression of the Triad to the DD18. The DD18 looks like it really walks all over the Triad. But if you compare both to their frequency response from which Keith normalized to the 90 db line, you will see that (for lack of a better phrase) the DD18 has built in compression (the response seems trail off faster below ~30 hz).

If you look at the 100 db graph frequency sweep for both subs you will see that the Triad hits 20 hz on this sweep at 98 db while the DD18 manages around 95 db. In fact the DD18 never really breaks 100 db at 20 hz on any sweeps but he Triad pushes 102 db on the 110 and 115 sweep.

I guess the discussion point I'm attempting to make is that if a sub were to be designed so that the output is artificially rolled off sharply below 30 hz where bass becomes more demanding but output stayed high above 30 hz. I think we'd see that the power compression could be very good so long as the curve maintains its shape.

But in this instance we would clearly see that this sub would not be considered a good sub for any application that calls for high output below 30 hz even though it scores high marks for maintaining it's shape as spl increases as far as power compression goes.

In other words for compression testing to be useful, shouldn't a sub have to hit a minimum dB at say 20 hz and only be down no more than x db from 70 - 80 hz to make the results meaningful?

BTW, I think THD is important too. No doubt about that.

 

[Mark Seaton]

I have to take a look, but there is a significant error in two of the compression plots. The first is seen in the "100dB" plot for the CoDrive CB1, which clearly should show more compression below 25Hz. Similarly the plot of the DD-18's power compression at "110dB" is mis labled as the "115dB" curve is in fact correct for the "110dB" level. Comparing the curves it is clear in looking at the "105dB" response and the "110dB" response that there is 5dB of compression by 20Hz, starting around 32Hz. The "115dB" curve starts changing shape below 45Hz and reaches a maximum of about 10dB of compression by 20Hz. Clearly the red compression plot does not reflect this.

I did have the pleasure of pointing this out to Keith Yates at CEDIA, who quickly confirmed that the response is correct and the compression curves did not match. Keith was frustrated that it happened, as it looked like the program written to average the many measurements he gathered referenced the wrong plot for these 3 examples. At this point there isn't much to be done spare maybe a correction which few will read. Of course that's what forums like this are for.

So far as the compression not telling you much, I would point out that no single plot in the review could tell you the usefulness of a given subwoofer. You really need to see at least TWO of the graphs to get any sort of meaning and interpretation. Much like in geometry, two points define a line, where we start to see some form, and 3 points define a plane and allow a much more complex comparison of where they intersect.

The successive frequency responses include the compression, but many don't really grasp that immediately in looking the responses over. Likewise, most readers are not used to seeing high resolution frequency response measurements, where the "straight line is good" concept of the compression plots is a good translation. Realize that there will never be a "complete" set of measurements for subwoofers. New designs and different processing and DSP control will likely mask interesting details. Likewise, the use or exclusion of EQ has a HUGE impact on what this data shows. Even in the section on our own ContraBass, had the recommended high pass of 10-14Hz been included, it would have likely withstood higher playback levels with those soundtracks with lots of super deep energy.

In reality these measurements are easy for any designer to relate and equate to, making for a great sampling of data for various current offerings. I know I learned a lot in looking through all the data, and will incorporate that into future product offerings.

Cheers,

 

[Keir H]

Mark, or anyone, can you tell me what it takes to get that chest thumping bass, the kind you feel when are watching action movies with for example gun fire or shotgun blasts? Is it a matter of db level when playing back dvd's..at or very near ref? Or, is it a certain frequency that your sub or mains will have to produce to get that effect. For instance i get the rumble all the time but i am looking for that bass shaker effect w/o having to buy em..just yet anyway. Is it program material dependent on how it was recorded? thanks

 

[EdwinK]

Is there any clue how the bigger SVS subs would do in these line-ups?

 

[Keir H]

I'm wondering as well how they would match up. My first guess is that the SVS's top 3 would be somewhere near the Genelec and Contra, perhaps edgeing out the Velo at lower freqs...but again thats just my first thought of what I've read.

 

[Robert Cowan]

you might edge out the velodyne in pure SPL, but you wouldnt come close to its distortion...

SPL isnt everything...

 

[Kevin C Brown]

Yes, but Velodyne's limit output to keep the distortion low. 6 of one, 1/2 dozen of another.

Plus, if you look around (Secrets, and Ed J M's sub reviews), SVS's are no slouch when it comes to distortion either.

And, I have owned mutliple Velodyne subs, and I have an SVS now, SVS gives you much better bang for the buck.

Velodyne makes great subs. No question about it. (Although they did have some QC issues at least late 90's, early 2000's.) But man do you pay for it.

 

[Robert Cowan]

this is true...

but, nothing in this hobby is what you could consider "cheap" when you start getting into the higher end stuff... its all pricey.

 

[Jon Lane]

With one caveat, I'm totally with ya, Craig: "Quickly enough" is not associated with cone mass. Other than that, not enough attention is paid to why systems don't sound "quickly enough" and that's a real problem.

To review, speed in this context is associated with system transient response, ergo, the system Q. It's been said critically damped systems sound "fast"; maximally-flat systems and those increasingly less damped less so.

But...

Then there's the sub's integration to the satellites, where those essential higher-frequency overtones lie. Lots of systems do this mesh exceedingly poorly -- ~80Hz sub xovers are the ones we all hit-or-miss on -- and there goes the sound of "fast" bass, TRAD, PRAT, what have you.

Speed is a useful subjective handle to describe how bass sounds, and it encompasses system transient response (both the sub system and the combined system) distortion, and even level. But it's not solely related to the "speed" of the driver per its cone's mass.

 

[Jack Gilvey]

People are under the impression that Velodyne uses some sort of magic to get lots of output with "low distortion". They don't. Limit output in the deep bass, you get less distortion and great reviews. SVS goes another route toward low distortion while achieving high output. "SPL isn't everything", of course, but neither is "THD" if the wideband power response suffers.

 

[Robert Cowan]

???

i wouldnt call it magic by any means, but by using feedback system like they do that compensates for distortion, and the addition of EQ'ing... its not just limiting bottom end.

everyone loves to hate velodyne because its a pricey product. im sorry, but i still believe in the age old saying "you get what you pay for".

 

[Jack Gilvey]

So...the more expensive sub will be the better one?

 

[Aaron Smithski]

Robert,

I don't believe you'll find too many Velodyne "hatas" here. Velodyne has some very nice products -- no question about it. Many folks on this forum have a pretty good idea about what it takes to build a serious subwoofer, and they would simply like to get much more output at a much lower cost; while still maintaining excellent low-distortion figures.

You don't always "get what you pay for"...you usually "get what you get."