Neptune EQ vs SVS EQ1

[chipperman]

I have compared the Neptune EQ, correcting for subwoofer only, directly against the SVS EQ1 (which I own), using an audio precision signal generator to perform a sweep and then measured and set up a response curve on an inexpensive small KLH subwoofer, using each units included microphone and set up procedure. Another signal sweep was then performed and the actual corrected response of the speaker was measured and plotted. This was performed in a workshop with absolutely no thought given to proper acoustics, nor any attempt made to correct them. However, a couple of intresting differences arose. I have discussed them with Tech Support at SVS, and as I undertand them to the best of my ability, here is what I observed.

It is interesting to note that the SVS unit has the capability to correct 2 subwoofers independently and then "blend" them by correcting both together to get the best overall response. Set up is fairly quick and easy, but it helps to read the manual a few times to figure out exactly how this all works. The measured response curve that displays on the laptop is a highly smoothed version of what actually goes on. The "corrected" curve that displays at the end is not the actual measurement, but a projection of what it would ideally be. This is misleading. Actual measurements deviate quite a bit from this.

The unit detects where the natural rolloff of frequenies occurs in the driver and does not supply any eq past these points. While it may protect the driver, especially in a vented enclosure, it also mandates the the subwoofer have adequate range as built. In the case of our woofer test this occurred at about 52 hz.The top end began its descent at 130 hz. It did not show any correction ability above 150 hz, although it is claimed to go to to 1khz. It also requires the use of a laptop computer.


The Neptune unit is so very user friendly. No laptop is required, and everything is built in to the unit.
Its measured response to an input signal sweep is about the same from 40 hz to about 130hz but then the difference is really apparent, with the Neptune boosting to 200 hz on the high end, and adding almost 6db of boost at 35 hz, while staying with a 4 db boost down to 10 hz. The user also has the option of selectively boosting frequencies.

How about the real world results? With our little KLH sub, the Neptune gave an average flattened curve from 35 hz to 170 hz, while the SVS went from 52 hz to 130 hz. The big difference is in the low end, where the SVS unit simply would not provide boost beyond the natural rolloff, which is how it is designed. For those with sealed box units, the ability to push this curve to the left is much welcome.

So how did it sound? Well, to be honest, all I had to play was pink noise through the unit, but it sounded much more even with frequency changes using the Neptune. While the SVS unit claims to have many more points of correction than the Neptune, out measurements did not indicate that this made any measurable difference, at least with the equipment used.

SVS pros: Does 2 subs at once, works well within its range.
cons: Requires laptop, doesn't boost beyond natural rolloff. If you want more, add additional eq.

Neptune pros: Easy setup, no laptop, can manually add boost, extends range. Don't need another eq.
cons: Only does one sub at present, but they are working on it.


Winner: Nepune

ps: The SVS unit indicated it was almost twice as far from the driver as the Neptune unit. Both were tested in the same place. This would indicate that the delay is being read differently between the two units. I am not sure what to make of this.

[Kal Rubinson]

1. I agree with your analysis of the basic differences in how these two units approach sub EQ.

2. I do believe that these units should be tested with a more appropriate sub. I cannot imagine anyone with a KLH sub considering spending $800-$4000 on an EQ. A more potent sub would moderate the differences you found and, perhaps, raise some others.

3. The differences in delay are interesting because the SVS must compensate for its DSP latency only of the sub channel and this adjustment is made in the prepro before the SVS. The Neptune setup presumes that all the channels are subjected to the same internal DSP latency and so these cancel. One result is that they are each correct in normal use but if, in your tests, only the sub channel went through the Neptune, the indicated setting for it would be incorrect when used with the main channels.

4. Extending the FR of a sub upwards is a chancy prospect and should only be done where absolutely necessary. I cannot imaging such a situation with quality components.

5. Finally, I have to ask if you did any tests of modal decays in addition to your FR measurements. I find these are generally more audibly significant below 100Hz.

BTW, my (favorable) reviews of the Neptune and SVS EQs appear in Stereophile's July and September issues, respectively.

[Mike P.]

The comparison was done on a small KLH subwoofer with probably no output from 25 hz down. It would be interesting to see how both units would perform on a sub that has 16 hz capabilities.

[Sonnie]

Let's keep in mind we are comparing a $799 retail unit with a $3999 retail unit. I would probably expect to see some differences in favor of the Neptune. Now... if Neptune were to develop a sub eq at the same pricing of the SVS sub eq... then it would be more of an apples to apples comparison.

What would be interesting is to compare the two units with a quality subwoofer that someone would actually use, which is not going to roll off as low as the KLH, which I do not see anyone using. A proper comparison would include an REW measurement of the sub before eq, with SVS sub eq and with Neptune sub eq. However, as stated previously, price has to be considered.

[chipperman]

I need to clarify a few things here. The units were run standalone without the use of a prepro, so any signals output or measured were direct from the eq's. It is easy to do this with the SVS as you just ignore the mains matching function.

Sure, the sub used was a beater in the shop. Measuring the response of a better unit would not have made a difference because I was only trying to see the limits of either eq. A better sub would or course have more range, but nonetheless, the SVS unit senses the natural resonance and doesn't boost beyond where the curve begins to get steep, and the Neptune does. As regards the cost ratio to cheap subs, well, come on, this isn't how it would be most likely used, but it was a sub nonetheless. And keep in mind that the Neptune unit equalizes all channels in a surround system, so it does cost more, but I only tested the subwoofer portion of it. The distance that each unit calulated to the speaker was significantly different. Apparently the algorithms for delay are not the same in each unit, and the microphone used with the Neptune was probably of higher quality. This wasn't a real scientific test obviously, but is WAS a chance to use the two units side by side. I didn't measure any decays or such, that was not the purpose of the test. I have used the SVS unit in my house, without measuring anything, and it does make a huge difference. It just appears that one would need an LT or something to get lower end boost/equalization. You cannot fault the unit for its clever packaging and customer service.

[JohnM]

chipperman wrote:

The distance that each unit calulated to the speaker was significantly different. Apparently the algorithms for delay are not the same in each unit

You may have misunderstood Kal's response on that. The SVS AS-EQ1 "distance" figure includes 8.5ms (approx 9.5 feet) to compensate for the unit's processing delay.

[chipperman]

Well, now you're way ahead of me technically. What I was referring to (probably the same thing in reality) is that the units display the distance in feet that you are to enter into the AVR or prepro. I guess the SVS unit has more internal delay itself?

[Kal Rubinson]

chipperman wrote:

Well, now you're way ahead of me technically. What I was referring to (probably the same thing in reality) is that the units display the distance in feet that you are to enter into the AVR or prepro. I guess the SVS unit has more internal delay itself?

No. You didn't read the manuals carefully.

With the SVS, you enter the delay into the prepro and, therefore, it must incorporate the delay introduced by the SVS (which follows the prepro) as well as the real distance to the sub.

With the Neptune, the delay is set in/by the Neptune (you are instructed to set all delays/levels to "0" in the prepro) and, since the Neptune's delay is the same for all channels, it cancels out, leaving only the sub's real distance.

In other words, they are not seeing different distances but, rather, compensating in the different ways necessitated by their different designs.

I really wonder what you heard if you misunderstand this difference. Did you actually transfer the Neptune settings to the prepro?

Kal

[chipperman]

Kal, I didn't use a prepro. Both units were connected directly to the sub. Your explanation makes perfect sense.

[Kal Rubinson]

chipperman wrote:

Kal, I didn't use a prepro. Both units were connected directly to the sub. Your explanation makes perfect sense.

OK. But did you run your test and auditioning signals directly into each? If so, there was no distance/level compensation applied by the SVS.

Kal

[chipperman]

Each unit was plugged directly into the sub, the appropriate microphone was set up in the same location, and the signal generated by each eq was used. There was no external signal fed into the units. If I understand you correctly, are you saying I needed an external signal? Each eq generates its own test tones. I didn't write it down, but I think the SVS read something like 16 feet and the Neptune around 9 feet. The actual physical distance was about 6 1/2 feet. Even if the units read differently, wouldn't setting the delays in your prepro to 0 with the Neptune, give you the same reading as the summed settings of the SVS? I still don't get the distance discrepancy when both units are set up as standalone. Based off what you discussed above, it would look like there is no delay in the Neptune? I would think each unit would correct for this and display the same apparent (not actual) distance. What this testing DID point out to me was that you aren't going to even come close to doing it by entering manual distances and using a sound level meter. I suppose, if you had the perfect room...but both of these units measure the delay and compensate for it. How could you ever do that with a tape measure and meter? In the end all that counts is what it sounds like. And it is amazing what these units do.

[Kal Rubinson]

chipperman wrote:

Each unit was plugged directly into the sub, the appropriate microphone was set up in the same location, and the signal generated by each eq was used. There was no external signal fed into the units. If I understand you correctly, are you saying I needed an external signal? Each eq generates its own test tones. I didn't write it down, but I think the SVS read something like 16 feet and the Neptune around 9 feet. The actual physical distance was about 6 1/2 feet. Even if the units read differently, wouldn't setting the delays in your prepro to 0 with the Neptune, give you the same reading as the summed settings of the SVS?

Without a prepro, the Neptune will still implement all the settings (delay/level) but the SVS would not. It only measures and recommends. Thus, you would be comparing the two with different level and delay settings.

I still don't get the distance discrepancy when both units are set up as standalone. Based off what you discussed above, it would look like there is no delay in the Neptune?

What I said was that the Neptune software assumes that all channels are delayed the same amount (by the Neptune DSP) and does not need to add any differential to any channel, including the sub. The SVS software assumes that the SVS delay is unique to the sub channel and it recommends the shorter delay (longer distance setting) to bring it into time alignment with the other channels.

I would think each unit would correct for this and display the same apparent (not actual) distance.

Sorta but they need to be implemented differently.

What this testing DID point out to me was that you aren't going to even come close to doing it by entering manual distances and using a sound level meter. I suppose, if you had the perfect room...but both of these units measure the delay and compensate for it. How could you ever do that with a tape measure and meter?

Agreed. SLM is inadequate for this task.

[chipperman]

I hope this works. Here is the KLH sub's native response curve, equalized with the SVS/Audyssey, and finally with the Neptune. You can see how the Neptune shifts the curve to the left. And no, the voice coil didn't come through the dust cap.

[Kal Rubinson]

chipperman wrote:

I hope this works. Here is the KLH sub's native response curve, equalized with the SVS/Audyssey, and finally with the Neptune. You can see how the Neptune shifts the curve to the left. And no, the voice coil didn't come through the dust cap.

It is as you described and does not surprise me.

[neptuneEQ]

Hi Mike, nice to see you here. It was a pleasure meeting you last month!

These curves were created at a meeting we had, when Mike was at our place at Thilo of TC sounds' suggestion. We're both good friends of his, whom I respect above most people in this industry. Mike came to us because he wants to achieve excellent bass; extending down to the nether regions. That's why I suggested he come here to get input from the knowledgeable folk at this forum.

All I can do is explain our philosophy regarding the curves. To me, one thing is imperative regarding bass: Extend the frequency response much lower than audible! That doesn't mean boost that low, but by all means, don't hinder either! If I ever detected any audio equipment that wasn't flat below 20Hz, I pulled it from my system. Like the SVS unit, we detect the corner frequency of the subwoofer. That is something that is mandatory. The philosophy comes into play when determining what to do at that point.

Some people would say let it roll off without change. Others might say what happens below that point doesn't matter. I say give some help without going too far, then go back to flat below that. Here's my reasoning:

As we all know, woofers have limitations. We would also all like to think that subwoofer designers are "on top of it" and eek the utmost out of their designs; but, we do have some experience in that area... Before we became Neptune audio, our company manufactured subwoofer amplifiers OEM for 20+ years. We interacted with many of the top name manufacturers and know quite well how they think, which is pretty unique depending on the individual company. There is a consistency, though, regarding frequency response in the mainstream companies: they like to maximize output while sacrificing extension. Bass sells, and the louder and boomier, the better. Keep in mind that most people can't tell the difference. Mike here is amongst the few who can.

Another point to keep in mind is that any well designed subwoofer can't be damaged by any signal supplied to it! It may bottom out earlier, it may distort, but it can't be damaged. If it can be, that is simply a design flaw.

With those points in mind, we decided that boosting at the -3db point (by 3db) is acceptable. If it causes a rare problem, it can be removed manually. Below that, we do nothing. The response falls to a 0db flat line. That small boost, and the subsequent maintaining of output below that corner frequency has the ability to extend the response nicely as can be seen from the graphs.

It has been pointed out that the problem is that the cheap KLH subwoofer's response is too poor to be worthy of this test. I feel (as an engineer who has had many years experience making measurements) that extreme test subjects yield more obvious results. In other words, the reactions of the Eq will be the same with a better woofer, it will simply happen at a lower frequency and thus will be less obvious. I feel this test was quite valid, and I feel our product performed admirably, as would be expected from a $3995 piece of gear. To compare the two units is not as unfair as it seems though, because only the subwoofer portions were compared. If you use straight math, the subwoofer portion of our unit is 20 out of the 230 bands, so comprises approx. $350 of the cost.

The bottom line: Both units work very well. They both have their place. Where they fit depends upon what the individual is looking for.

Ken

[epereira]

Well stated Ken. Each unit performs very well, just as designed - I don't think we are comparing apples to oranges; both are "apples" but of different varieties.
Erle

[tpaxadpom]

Does SVS apply its correction curves only up to 150Hz? I wonder how the upper frequency band correction of the sub is being determined with Neptune or is that a preset value. I understand that one might want to have a flexibility though this is unlikely the case with $4000 and really small satellite speakers. I think it would be more beneficial for Neptune EQ to utilize 20 bands of correction dedicated for sub in the lower frequency spectrum rather than dealing with frequencies higher then 120Hz (imho).

[neptuneEQ]

Your point is well taken Andrey.

We detect all corner frequencies of all loudspeakers, both upper and lower. With this data we can determine the optimum crossover point, taking more than just the corners into account, but by all means giving those corners their proper weight. If a woofer starts to roll off above 100Hz for instance, and the full range loudspeakers reproduce down to 40, we'll select a lower than average crossover point, though also the fact that subwoofers are designed to reproduce much higher SPLs at low frequencies than full range speakers is taken into account.

Mike and my observations of the SVS unit was that above a certain frequency there was no more EQ applied, though the output was still higher than the bulk of the remaining response. Obviously I'm not privvy to their algorithms, but I can tell you that the neptuneEQ will go ahead and boost the higher frequencies of subwoofers all the way up to 200Hz if that's what it determines the sub needs. It may seem frivolous to "waste" bands in the stopband of the sub, but it does help assure a clean splice, and help maintain proper phase throughout the crossover range. Since we use 20 bands between 20Hz and 200Hz, we have a lot of resolution anyway.

Ken

[tpaxadpom]

Ken, thank you for your time explaining how Neptune EQ operates. It does make sense to maintain proper phase as the sub will have some audible output above the cut frequency. So if sub is crossed at 80Hz we are looking at -24dB at 160Hz in case of the fourth order crossover. I think most sub drivers are capable of reproducing frequencies up to 200Hz unless there is a built-in low pass filter in subs amp/crossover section. I'm patiently waiting for Kal's review in Stereophile.
The reason for my original comment was the difference I saw in plotted graphs. KLH sub with SVS help achieved more uniform frequency response in 52Hz-130Hz range despite Mike's comment:

chipperman wrote:

While the SVS unit claims to have many more points of correction than the Neptune, out measurements did not indicate that this made any measurable difference, at least with the equipment used.

[chipperman]

Andrey,

You are correct as to that narrow range of frequencies. However, we wind up with a nice low/mid bass driver, not really a subwoofer. I agree that the curve looks a little more ragged upon reinspection, but the Neptune unit can also be corrected manually, and we didn't do this for this test. My main concern was to see if the extension of the curve could be pushed lower. In that respect, we achieved success. With time and tweaking, I'll bet we should be able to get the Neptune response very close to the SVS, and still have the benefit of the extension. Isn't this hobby fun?

[neptuneEQ]

tpaxadpom wrote:

Ken, thank you for your time explaining how Neptune EQ operates. It does make sense to maintain proper phase as the sub will have some audible output above the cut frequency. So if sub is crossed at 80Hz we are looking at -24dB at 160Hz in case of the fourth order crossover. I think most sub drivers are capable of reproducing frequencies up to 200Hz unless there is a built-in low pass filter in subs amp/crossover section.

You're very welcome :-). I agree with your assessment. My experience has been that most subs roll off substantially above about 120Hz (or lower) despite the "bypassed crossover" status, and that that will affect the splice (however subtly)...so should be corrected.

tpaxadpom wrote:

I'm patiently waiting for Kal's review in Stereophile.

It hasn't come out yet? I thought it had long ago. The editorial world moves at a measured pace. ;-)

[Kal Rubinson]

tpaxadpom wrote:

I'm patiently waiting for Kal's review in Stereophile.

neptuneEQ wrote:

It hasn't come out yet? I thought it had long ago. The editorial world moves at a measured pace. ;-)

It appeared in the July issue which hit the streets and mailboxes in June.

[tpaxadpom]

That's great! I should have it in my zinio library then. It would be a good read in Jamaica (leaving tonight).

[Kal Rubinson]

tpaxadpom wrote:

That's great! I should have it in my zinio library then. It would be a good read in Jamaica (leaving tonight).

Or here: http://stereophile.com/musicintherou...37/index1.html

[tibuan]

Nice Information.