or Connect
AVS › AVS Forum › Audio › DIY Speakers and Subs › Hey Ricci...
New Posts  All Forums:Forum Nav:

Hey Ricci... - Page 4

post #91 of 194
Quote:
Originally Posted by Mark Seaton View Post

If Josh had simply not attempted levels any higher than the green curve below, it would have fit your ideal perfectly, although we wouldn't know the full capabilities and would probably be left wondering why it sounds more powerful in real use than some other subs which top out at a similar sweep level:
478

This is my point. Anyone who spends 4 figures for a sub should not be left to wonder anything.

What happened when Ilkka applied +10dB of boost EQ to the LMS? Like the Paradigm, it's way too much boost for the LMS, but there you have it. The engineer decided on +10dB and had to stop the sweeps accordingly, because that's all such a contraption will give. Of course, he also ran the tests without boost, showing the maximum broadband output >using a different curve, not limiters<. The proper next step is to provide various curves in between those 2 curves.

How many times have we said "SPL is accomplished by adding multiples"? We've never said SPL is achieved by limiters.

How many Paradigm sub buyers will read and correctly interpret Josh's results? Instead, they'll assume the published FR is intact at any level (or the published max output level) they choose to operate the sub at. The same is true with every commercial sub available.

"My 'X' sub handles WOTW with no sweat".

I'd love to see a review of a mains tower that compresses a huge chunk of its BW like these subs compress the low end. wink.gif Would the reviewer say the tower gives 'X' SPL, which sounds really loud and clean due to the clever limiters that compress the midrange -6dB?

The only reason P'digm (and Velo and everyone else at the party) gets away with that with subs is that people have no idea what a subwoofer is supposed to sound like after decades of no subwoofers, followed by decades of grossly distorted versions of subwoofers accompanied by grossly distorted "mixing" of the sub with the sats by the user.

10dB hot is 'mixing' the low end to 10 times the rest of the bandwidth. +15dB hot is 32 times. And the trend continues, in part because of published output numbers that have zero to do with the designed FR and are based on use of limiters that yield a completely different FR.

I would like to see a post that admits giving up half the subwoofers BW to be able to 'hit 'X' dB during 'Y' scene".

Having said all of that, a question:

Do you believe that if P'digm offered a curve that equals the last sweep trace in Josh's graph that it would result in the same performance from the sub as 10dB of limiter-induced compression?
post #92 of 194
Quote:
Originally Posted by noah katz View Post

LTD was talking about efficiency; you just changed the subject to output.

Sensitivity? Then the fact that 2 drivers are required to equal one driver should have been the main topic.

Output IS the subject. Always was, always will be. How one gets there is the question being raised in this thread.

The OP presents a pair of unnamed 18s against a single known and tested 18 as having equal sensitivity at 10 Hz, and presents it with "blow a hole in this theory'... 'HONK!'.

The subject was changed when box size was injected and made to sound as if it was the original point. The XXX needs a HUGE box to display at least a .7 Q picture of its 'sensitivity' at 10 Hz. It's shown in a high Q box vs a pair of low Q drivers in the same box, or rather, the XXX is stuffed into the ideal box for the pair of low Q drivers and nothing else is stated about box size.
post #93 of 194
If you model a sub in WinISD, one can see that all drivers can handle more power as you go up in frequency. Bossobass seems like he is saying that he prefers a 1 speed transmission that never provides more power than can be used by the lowest frequencies. If this doesn't provide enough output, his solution is to add more drivers so you can utilize more power at the lowest frequencies. He is always leaving headroom on the table at the higher frequencies.

Seaton is saying that he likes a multispeed transmission because it utilizes the power handling of driver more consistently. You provide the power that can be used by the highest frequencies. The sub system will then "shift gears" so that the power output is still matching the driver's capabilities when it reaches the "rev limit" in the lower frequencies. This "shifting of gears" is done by limiting. Shifting to the next gear limits the frequencies below the shift point. To me this makes much more sense and maximizes the driver across its frequency band.

478
post #94 of 194
Dave,

Using sealed subs you either accept the raw response shape as is and maintain that shape all of the way to the limits of the system or you EQ the shape by boosting the low end and know that at some point if the level gets high enough there will be either A. A change in response reverting back to the native shape as the low end compresses, B. Blown drivers or C. The output is stopped at the point that the low end reaches its maximum output at which point the top end maximum output is still unutilized / unknown. I see no way around this.

Having multiple preset curves will not change maximum output of the system at any point. What it will change is the relationship of output at the top of the range to the bottom and at what point the response shape starts to deviate from the low level shape and compress. How does multiple EQ presets help you avoid this? Do you switch between curves altering the bass FR depending on the depth and level of content? How do you know what types of bass the content contains ahead of time? Do you switch depending on how loud you will be playing the system? A good limiter system is virtually inaudible in operation if implemented right and will modify the FR on the fly to allow the system to make use of every bit of its headroom without endangering itself in the process. You don't need to alter the response to another shape manually with a modified EQ curve since it allows the top end to utilize extra headroom on the fly as the low end compresses. IMHO no powered sub for sale commercially or professionally should be without it.

If we are talking about systems which have enough firepower to not have to worry about compression at all then the limiter on commercial powered subs is never even a factor to begin with and I'm not sure why we are talking about it. As Mark said if you never push the Sub2 past the green curve on the graph posted then there is no compression to speak of. You could achieve enough deep bass headroom to never have compression with ULS-15's or Submersive's or SB13's or whatever with enough of them just the same. When looking at output measurements just ignore any of the higher level measurements past where the limiter engages and do the math for multiples. I don't see how allowing the subwoofer to accept ever increasing input levels past the point where the driver is out of excursion and protecting itself from damage while still providing as much headroom as possible over the entire bandwidth of use can be considered as a design flaw? Most users purchase one or maybe 2 of the commercial subs and that is all they are willing to spend or fit into their rooms. This is not enough units to provide full range uncompressed bass playback in most cases. The limiter allows the sub to produce maximum extension at moderate volumes but allows the sub to produce as much output as possible to help fill a large space with only one or two units when pushed hard as well. The alternative with sealed systems is to severely limit the potential headroom in the octaves made heavy use of with popular music, to allow the driver/s to be overdriven or potentially damaged with low bass, or to allow the amplifier to clip heavily if a sub bass transient of too high an amplitude comes through (This is what happens to my system if I push it to hard with the EQ engaged, most sealed DIY systems either do this or have the potential to damage the drivers through over excursion). Some day I plan to attempt to dial in a good limiting circuit for my subs using the DCX2496.

If I were designing a powered sealed sub for sale I would definitely make use of a limiter and probably shoot for an eq'd 6dB octave roll off starting near 40Hz anechoic. I think that would work well with room gain in most spaces while keeping the amount of boost modest. . Going for extreme flat extension in an anechoic environment with a sealed system is a wasted move and probably more about a pretty FR graph or claims of extension to some obscene frequency than anything else.
post #95 of 194
I think Bosso is saying it doesn't matter what second gear has in store because all that matters is first gear Like a drag car starting in second. Why have more output above where it's limited? Would you like a speaker that has a dramatic output shift above say 95db such that the tweeter suddenly starts outputting more but the woofer gets left behind???
post #96 of 194
Quote:
Originally Posted by tuxedocivic View Post

I think Bosso is saying it doesn't matter what second gear has in store because all that matters is first gear Like a drag car starting in second. Why have more output above where it's limited? Would you like a speaker that has a dramatic output shift above say 95db such that the tweeter suddenly starts outputting more but the woofer gets left behind???

Because we are talking about the limiting employed in commercial subs which are smaller, less powerful on average and usually only used in ones or twos. This is not a scenario where endless headroom is on supply like with some of the big DIY systems nor the amount of sheer displacement.

Hypothetically if you make a commercial 1000w 15" sealed sub in a 19" cube and you EQ it flat to 20Hz anechoic and you limit it to the point that it gets as loud as possible at 20Hz before damaging the driver ( about 100dB maybe.) and then allow none of the rest of the response to get any louder from that point so that the same response shape is maintained perfectly you end up with a sub that is going to start getting reviewed as being subjectively very weak in output and people will look somewhere else. Lets say you are watching a concert special from a group that you like and you want to crank it up to get some of that concert feel. It is not going to be very fulfilling when the sub clamps at 100dB peaks despite the fact that there is nothing below 40Hz to speak of in the content. If manufacturer B's similar sub employs the limiter scenario commonly used but is other wise similar you are probably going to be much more happy with it since it will allow peaks nearer to 110dB during the same concert playback. I'd take manufacturer B's sub myself.

You might ask well why limit the sub at all? If it's not limited the first sub will produce the same levels as the second sub during the concert. that would be correct because there is no deep content involved. However if you were attempting that playback level and there is no limiter and a big 20Hz event comes along, the result is either damaged driver from overexcursion, or severe amp clipping. Both of which result in either damage, bad noises at least, or limiting of output. I'd argue that the limiter allowing the system to do its best while protecting the system is a better approach.
Edited by Ricci - 6/29/12 at 9:32am
post #97 of 194
Quote:
Originally Posted by Ricci View Post

Because we are talking about the limiting employed in commercial subs which are smaller, less powerful on average and usually only used in ones or twos.

Well, now this is the real problem isn't it tongue.gif
Quote:
Originally Posted by Ricci View Post


Hypothetically if you make a commercial 1000w 15" sealed sub in a 19" cube and you EQ it flat to 20Hz anechoic and you limit it to the point that it gets as loud as possible at 20Hz before damaging the driver ( about 100dB maybe.) and then allow none of the rest of the response to get any louder from that point so that the same response shape is maintained perfectly you end up with a sub that is going to start getting reviewed as being subjectively very weak in output and people will look somewhere else.

I see where the manufacturer is coming from. But I agree with Bosso that it's unfair advertising (not a new thing, I know). And this system would dissapoint me for the money spent.

Good explanation though. Thanks.
post #98 of 194
Quote:
Originally Posted by tuxedocivic View Post

I see where the manufacturer is coming from. But I agree with Bosso that it's unfair advertising (not a new thing, I know).
Are you talking about the Paradigm Sub 2? I looked at their website and can't figure out what advertising is unfair. They show a flat response up to 90 dB, but Ricci measured it flat up to about 102 dB. After that the limiting kicks in. The limited 111.6 dB output curve is actually what some other commercial sub companies try for as their native response at lower output levels because it matches well with room gain in most rooms and gives a flat in-room frequency response. Put the Sub 2 in most rooms and it will be completely flat at 111.6 dB all the way down to 10 Hz and lower. At lower volumes it will be louder at the lower frequencies and will provide the "house curve" that many like.
Quote:
And this system would dissapoint me for the money spent.
For the money, I would expect the drivers to be utilized to their potential across the frequency band. That is what takes a lot of time and testing. wink.gif Some companies amp limit the driver which is less work, but yields less satisfactory results overall IMO.
post #99 of 194
Thread Starter 
"Sensitivity? Then the fact that 2 drivers are required to equal one driver should have been the main topic."

the point of the op was sensitivity/efficiency, which is why i said sensitivity/efficiency. maximum output is another topic.

in the 8 cubic foot enclosure in the op, you can put a whole bunch of drivers and it ends up being pretty much the same low end sensitivity, whether you use multiple re xxx's or multiple jbl 2226s (high or low q drivers, it doesn't matter). i found that interesting and is why i started the thread.

it was then pointed out that this is what the hoffman iron law states--for a given enclosure size, there is a limit to how sensitive it will be on the low end.
post #100 of 194
Quote:
Originally Posted by desertdome View Post

If you model a sub in WinISD, one can see that all drivers can handle more power as you go up in frequency. Bossobass seems like he is saying that he prefers a 1 speed transmission that never provides more power than can be used by the lowest frequencies. If this doesn't provide enough output, his solution is to add more drivers so you can utilize more power at the lowest frequencies. He is always leaving headroom on the table at the higher frequencies.
Seaton is saying that he likes a multispeed transmission because it utilizes the power handling of driver more consistently. You provide the power that can be used by the highest frequencies. The sub system will then "shift gears" so that the power output is still matching the driver's capabilities when it reaches the "rev limit" in the lower frequencies. This "shifting of gears" is done by limiting. Shifting to the next gear limits the frequencies below the shift point. To me this makes much more sense and maximizes the driver across its frequency band.
478

Sorry, you've got it bass-ackwards. smile.gif

Look at the graph; there are 4 speeds (curves) shown. That's the result of a single speed transmission with variable governor. It would be much better if the sub had those 4 speeds (curves) as presets, selectable by the driver, according to his driving preference for any given day.
post #101 of 194
Thread Starter 
"Do you believe that if P'digm offered a curve that equals the last sweep trace in Josh's graph that it would result in the same performance from the sub as 10dB of limiter-induced compression?"

the last sweep may very well be the natural response of the sub. who knows.

just thinking out loud here...but i suspect it's also not really limiter induced compression. it is more of just not amplifying the signal fully or turning down the electronic boost that is employed to make the frequency response flatter. the algorithm probably is something like at low levels of spl, apply lots of boost. as the spl rises, apply less boost. so its not really limiting, but that is just a guess.
post #102 of 194
Quote:
Originally Posted by Ricci View Post

Dave,
Using sealed subs you either accept the raw response shape as is and maintain that shape all of the way to the limits of the system or you EQ the shape by boosting the low end and know that at some point if the level gets high enough there will be either A. A change in response reverting back to the native shape as the low end compresses, B. Blown drivers or C. The output is stopped at the point that the low end reaches its maximum output at which point the top end maximum output is still unutilized / unknown. I see no way around this.
Having multiple preset curves will not change maximum output of the system at any point. What it will change is the relationship of output at the top of the range to the bottom and at what point the response shape starts to deviate from the low level shape and compress. How does multiple EQ presets help you avoid this? Do you switch between curves altering the bass FR depending on the depth and level of content? How do you know what types of bass the content contains ahead of time? Do you switch depending on how loud you will be playing the system? A good limiter system is virtually inaudible in operation if implemented right and will modify the FR on the fly to allow the system to make use of every bit of its headroom without endangering itself in the process. You don't need to alter the response to another shape manually with a modified EQ curve since it allows the top end to utilize extra headroom on the fly as the low end compresses. IMHO no powered sub for sale commercially or professionally should be without it.
If we are talking about systems which have enough firepower to not have to worry about compression at all then the limiter on commercial powered subs is never even a factor to begin with and I'm not sure why we are talking about it. As Mark said if you never push the Sub2 past the green curve on the graph posted then there is no compression to speak of. You could achieve enough deep bass headroom to never have compression with ULS-15's or Submersive's or SB13's or whatever with enough of them just the same. When looking at output measurements just ignore any of the higher level measurements past where the limiter engages and do the math for multiples. I don't see how allowing the subwoofer to accept ever increasing input levels past the point where the driver is out of excursion and protecting itself from damage while still providing as much headroom as possible over the entire bandwidth of use can be considered as a design flaw? Most users purchase one or maybe 2 of the commercial subs and that is all they are willing to spend or fit into their rooms. This is not enough units to provide full range uncompressed bass playback in most cases. The limiter allows the sub to produce maximum extension at moderate volumes but allows the sub to produce as much output as possible to help fill a large space with only one or two units when pushed hard as well. The alternative with sealed systems is to severely limit the potential headroom in the octaves made heavy use of with popular music, to allow the driver/s to be overdriven or potentially damaged with low bass, or to allow the amplifier to clip heavily if a sub bass transient of too high an amplitude comes through (This is what happens to my system if I push it to hard with the EQ engaged, most sealed DIY systems either do this or have the potential to damage the drivers through over excursion). Some day I plan to attempt to dial in a good limiting circuit for my subs using the DCX2496.
If I were designing a powered sealed sub for sale I would definitely make use of a limiter and probably shoot for an eq'd 6dB octave roll off starting near 40Hz anechoic. I think that would work well with room gain in most spaces while keeping the amount of boost modest. . Going for extreme flat extension in an anechoic environment with a sealed system is a wasted move and probably more about a pretty FR graph or claims of extension to some obscene frequency than anything else.

Yes to both questions; you select the curve based on playback levels and/or source. These days, you have a darned good idea what content is on any given disc before you spin it. If you don't, limiters aren't THAT idiot proof. biggrin.gif

The problem with such hard limiting is that it IS audible, unless you believe the distortion numbers for the P'digm are insignificant.

Why do you need a limiter for your system? Does it not monitor clipping at the input and output? Can you not tell when your system is amp-limited or excursion-limited at the levels you may be playing it with the source you may be playing?

Let's look at the evidence:

1) Look at the P'digms sweeps, compression and THD graphs. The distortions shoots up exactly where the limiter is centered.

2) The P'digms distortion rises for each of the last 3 highest sweep, but the output does not.

3) Next, look at Ilk's Tumult, SDX and LMS graphs. In each, the distortion rises gradually as frequency decreases and the sweeps show output continues to rise.

4) There's no such limit at any one slice of BW.

Then look at his test of the LMS with 10dB of boost. Same thing. He stopped the sweeps because the system told him to and he isn't an idiot. cool.gif
post #103 of 194
Quote:
Originally Posted by LTD02 View Post

"Do you believe that if P'digm offered a curve that equals the last sweep trace in Josh's graph that it would result in the same performance from the sub as 10dB of limiter-induced compression?"
the last sweep may very well be the natural response of the sub. who knows.
just thinking out loud here...but i suspect it's also not really limiter induced compression. it is more of just not amplifying the signal fully or turning down the electronic boost that is employed to make the frequency response flatter. the algorithm probably is something like at low levels of spl, apply lots of boost. as the spl rises, apply less boost. so its not really limiting, but that is just a guess.

It's definitely not the subs naked response at any of the sweeps.

The compression is most definitely limiter-induced. It's a simple case of denying the signal-demanded (EQ'd or not) amp power. That's what a limiter does.

We see the effect in the sine sweep regimen, but the more interesting question is what effect it has with program source. IOW, you can't (at least I can't) tell what sort of limiting is used with sine sweeps.
post #104 of 194
Quote:
Originally Posted by bossobass View Post

Quote:
Originally Posted by LTD02 View Post

"Do you believe that if P'digm offered a curve that equals the last sweep trace in Josh's graph that it would result in the same performance from the sub as 10dB of limiter-induced compression?"
the last sweep may very well be the natural response of the sub. who knows.
just thinking out loud here...but i suspect it's also not really limiter induced compression. it is more of just not amplifying the signal fully or turning down the electronic boost that is employed to make the frequency response flatter. the algorithm probably is something like at low levels of spl, apply lots of boost. as the spl rises, apply less boost. so its not really limiting, but that is just a guess.

It's definitely not the subs naked response at any of the sweeps.

The compression is most definitely limiter-induced. It's a simple case of denying the signal-demanded (EQ'd or not) amp power. That's what a limiter does.

We see the effect in the sine sweep regimen, but the more interesting question is what effect it has with program source. IOW, you can't (at least I can't) tell what sort of limiting is used with sine sweeps.

A clipping amplifier does the nearly the same thing. The sweep stopping in level makes no clear indication of why it stopped. It could be electronic limiting of power that varies per frequency, it could simply be amp clipping, or it could be a limiter softening the clipping of the amplifier.

Like you, I prefer to have the sealed enclosure result in a 12dB/octave roll off on the bottom such that constant Voltage input results in a plateau in the excursion below some frequency. Smaller boxes and higher Fb for a given driver = less excursion for the same power, or more power to reach that excursion plateau. In Josh measured your subs with the outdoor curve set for +0/-3dB at 20Hz wink.gif we could produce similar curves that resulted in a 12dB/octave slope at the bottom if we keep increasing the level to see what is in the tank until we maxed out above 40Hz. Now bypass the BossoBassis and again increase sweeps till we max out. The maximums will be similar, with clipping earlier on the sweeps. Your drivers are amp limited, so there won't be any driver damage.

So now my question is this... Consider a case where we had a really heavy and powerful 15" woofer such that in-box it matched the response you electronically corrected your sub to for the +0/-3dB @20Hz curve in the same box size with the same sensitivity below 15Hz. If the drivers have the same Xmax/Xmech capabilities and used the exact same amplifier would there be any real world, observed performance differences between your implementation with your signal shaper and the driver which did not need any correction?
post #105 of 194
Thread Starter 
"So now my question is this..."

as best as i can determine, there are actually two approaches. one uses motor power and the other uses low inductance.

the tcsounds pro 5100 and the jbl w1500h both are very good, but have poopy le/re while at the other end of the spectrum are drivers such as the jbl w15gti and dayton 390hf which suck on motor strength, but makes up for it on lr/re.

the jbl 2242h, ev 180b and ae td18 all seem to incorporate the best of both...at the expense of xmax of course. but within their limits, folks seem more happy with them than any of the low sensitivity drives.

flip a coin. heads you win. tails you win. :-)
post #106 of 194
Quote:
Originally Posted by Mark Seaton View Post

A clipping amplifier does the nearly the same thing. The sweep stopping in level makes no clear indication of why it stopped. It could be electronic limiting of power that varies per frequency, it could simply be amp clipping, or it could be a limiter softening the clipping of the amplifier.
Like you, I prefer to have the sealed enclosure result in a 12dB/octave roll off on the bottom such that constant Voltage input results in a plateau in the excursion below some frequency. Smaller boxes and higher Fb for a given driver = less excursion for the same power, or more power to reach that excursion plateau. In Josh measured your subs with the outdoor curve set for +0/-3dB at 20Hz wink.gif we could produce similar curves that resulted in a 12dB/octave slope at the bottom if we keep increasing the level to see what is in the tank until we maxed out above 40Hz. Now bypass the BossoBassis and again increase sweeps till we max out. The maximums will be similar, with clipping earlier on the sweeps. Your drivers are amp limited, so there won't be any driver damage.

So now my question is this... Consider a case where we had a really heavy and powerful 15" woofer such that in-box it matched the response you electronically corrected your sub to for the +0/-3dB @20Hz curve in the same box size with the same sensitivity below 15Hz. If the drivers have the same Xmax/Xmech capabilities and used the exact same amplifier would there be any real world, observed performance differences between your implementation with your signal shaper and the driver which did not need any correction?

We're getting off track here by creating an impossible situation for the sake of an argument. If there was a driver that had same sensitivity <15 Hz in the same box with the same X-max and had the same FR without signal shaping, I'd be using that driver instead of hoeing the hard row. wink.gif

That really isn't the question here, which is: If the signal shaped version used the sort of limiters the P'digm and most commercial subs use, would there be a real world difference? And the answer is yes. I don't see how you can deny that.

You posted the P'digm example, so let's continue to look at that before we create or look at other scenarios. We have the measurements. We know limiters are the cause of the "compression". We can't know what sort of limiters are employed from sine sweep testing, but we certainly can make an educated guess (and our guess will be right).

Looking at the last 3 sweeps, why does the distortion continue to increase despite the fact that driver excursion does not?

Would that be the case if, instead of limiters, the signal shaping was altered 3 times to mimic the last 3 curves?

I already know the answer to these questions. There is no room for debate here, except to argue one type of limiter vs another and the precise implementation for the P'digm box/drivers/amp. But, all that would be is dancing around the facts of the matter, which won't change.

Now, I'd like to offer another similar example. Look at the Velo DD-18 test results:

Untitled-1.jpg

So, we have 5 curves, dictated by the servo limiting. Keep in mind that these curves are not the same as the curves I would affect through signal shaping and that's part of the argument, but I'll let that go for now.

A whopping 16dB of compression. Output at 15 Hz is static over the last 4 sweeps (thee were only 5 sweeps performed), yet distortion increases with progressive sweeps.

So, the servo, like the P'digms limiters, is only making sure the sub "keeps its composure" and the price is a continuously changing frequency response and loads of distortion.
post #107 of 194
Quote:
Originally Posted by bossobass View Post

Quote:
Originally Posted by Mark Seaton View Post

A clipping amplifier does the nearly the same thing. The sweep stopping in level makes no clear indication of why it stopped. It could be electronic limiting of power that varies per frequency, it could simply be amp clipping, or it could be a limiter softening the clipping of the amplifier.
Like you, I prefer to have the sealed enclosure result in a 12dB/octave roll off on the bottom such that constant Voltage input results in a plateau in the excursion below some frequency. Smaller boxes and higher Fb for a given driver = less excursion for the same power, or more power to reach that excursion plateau. In Josh measured your subs with the outdoor curve set for +0/-3dB at 20Hz wink.gif we could produce similar curves that resulted in a 12dB/octave slope at the bottom if we keep increasing the level to see what is in the tank until we maxed out above 40Hz. Now bypass the BossoBassis and again increase sweeps till we max out. The maximums will be similar, with clipping earlier on the sweeps. Your drivers are amp limited, so there won't be any driver damage.

So now my question is this... Consider a case where we had a really heavy and powerful 15" woofer such that in-box it matched the response you electronically corrected your sub to for the +0/-3dB @20Hz curve in the same box size with the same sensitivity below 15Hz. If the drivers have the same Xmax/Xmech capabilities and used the exact same amplifier would there be any real world, observed performance differences between your implementation with your signal shaper and the driver which did not need any correction?

We're getting off track here by creating an impossible situation for the sake of an argument. If there was a driver that had same sensitivity <15 Hz in the same box with the same X-max and had the same FR without signal shaping, I'd be using that driver instead of hoeing the hard row. wink.gif

Yet you've argued in the past that choosing a smaller box to control excursion of a driver to allow a higher input power without over-excursion has benefits in total capability.
Quote:
That really isn't the question here, which is: If the signal shaped version used the sort of limiters the P'digm and most commercial subs use, would there be a real world difference? And the answer is yes. I don't see how you can deny that.

Agreed there will be a difference. Next points of interest...
1. Which one do you feel will have the advantage subjectively
2. Which will require more power from the wall playing back the same levels? (at levels not hitting any limiting)
post #108 of 194
Quote:
Originally Posted by Mark Seaton View Post

Yet you've argued in the past that choosing a smaller box to control excursion of a driver to allow a higher input power without over-excursion has benefits in total capability.
Agreed there will be a difference.

Well, that's because it does. I guess you're insinuating that limiters are the same as a smaller air spring?

Quote:
Originally Posted by Mark Seaton View Post

Next points of interest...
1. Which one do you feel will have the advantage subjectively

No question whatsoever, the non-limiter sub presents a higher degree of accuracy, and this goes to the crux of my argument. As can be plainly seen, the limiter controlled subs display an infinitely varying frequency response during playback, at almost any playback level. With my method, the frequency response (whichever one is chosen) remains intact and therefore predictable.

An occasional peak limiter clip light flashing doesn't change the presentation, verified by SpecLab and anyone's ears. A constant clip light tells the 'idiot' to back off or select a compromise (subjective) curve.

Subjectively, there are people who think the P'digm (probably set +10dB hot) is "the best subwoofer in the world, period", so I'm not gonna touch that one. Honestly, at this point in the game, for me, I have no subjective preferences. That's a path that only leads to more and more distortion. FR linearity and flat calibration lead to accurate playback and, quite frankly, that sounds subjectively pretty darned good to me vs the alternatives.

Quote:
Originally Posted by Mark Seaton View Post

2. Which will require more power from the wall playing back the same levels? (at levels not hitting any limiting)

My method requires more power from the wall, there's no doubt about it. The war is against dynamic peaks at the extreme low end. I bring every available weapon to that war. otherwise, I would have built a ported sub tuned to 20 Hz and stopped posting in this forum years ago.
post #109 of 194
Thread Starter 
bosso, is your "signal shaper" an active device that uses something like op-amps or a passive device with lcr (inductors, capacitors, and resistors)?
post #110 of 194
Quote:
Originally Posted by LTD02 View Post

bosso, is your "signal shaper" an active device that uses something like op-amps or a passive device with lcr (inductors, capacitors, and resistors)?

Active.

I'm not aware of a passive solution, are you?
post #111 of 194
Quote:
Originally Posted by bossobass View Post

Active.
I'm not aware of a passive solution, are you?

A passive line level contour filter would do the same thing, assuming your pre-out is up to the task.
post #112 of 194
Thread Starter 
"I'm not aware of a passive solution, are you?"

as ev mentions, if you have enough signal to begin with, linkwitz has an rc solution on his site for a low shelf filter.
http://www.linkwitzlab.com/filters.htm#5

i was just curious.
post #113 of 194
Getting back to the conversation on limiters in commercial subs. smile.gif
Quote:
Originally Posted by bossobass View Post

No question whatsoever, the non-limiter sub presents a higher degree of accuracy, and this goes to the crux of my argument. As can be plainly seen, the limiter controlled subs display an infinitely varying frequency response during playback, at almost any playback level.

This is not true.

Exactly how does a non limited sub present a higher degree of accuracy? If you utilize similar displacement capabilities and amplifier power for both systems and equalize to the same FR there won't be much of a difference at all except that the limiter will squelch the last 1dB or 2dB or so of maximum output.

What the limiter does when properly set-up is provide long term protection to the sub, prevent the woofer from being driven into over-excursion and prevent the amplifier from being clipped to death if the input levels are too high. In the meantime it allows the system to produce its maximum useful output over its full bandwidth of operation. The output being limited is output that is either damaging or very distorted so it is not something that most people would consider as useful to begin with. This type of thing is a requirement for a commercial sub in order to maximize performance and prevent returns of damaged systems when some end user dumps 6 volts of 15 Hz into the amplifier input with the gain at 85%, or runs the amp at heavy clipping for 4hrs with their favorite dubstep mix.

Small sealed subwoofer systems do not have unequalized flat response down into the 10Hz range. Any sealed system that is EQ boosted on the low end will run into its output limitations where the boost is first. The maximum output of the system over its full bandwidth will not be flat and will track the naked response closely. There is no way around this. Very simple modeling and measurements bear this out. Any sealed system driven at xmax exhibits a 12dB/octave decrease in output towards lower frequencies. Hence the old 4X displacement needed to maintain a given output for each octave lower rule. Amplifiers are also primarily constant voltage sources these days and while there is some variation on how much voltage (and power) they can swing before clipping depending on impedance, they will clip WAY earlier wherever the boost is as even a moderate 6dB of boost requires 4X the power from the amp. These are the two main limitations of output for any design and both result in the system reverting back to the native system response more or less at maximum drive levels.

The change in response shape is not the product of the limiter at all it is a product of the physics of the sealed alignment itself and the signal manipulation or EQ used. No sealed system that has boosting EQ done to it will maintain the same response shape at maximum output unless it is purposely set-up to throw away a ton of headroom using a limiter system to limit the top end output above some arbitrary corner set by signal manipulation. Almost no one does this for obvious reasons. If you want to maintain the same response shape at maximum output as you do at lower volumes you either use no EQ boost, severely limit the top end with limiting or develop special drivers with incredibly low top end sensitivity and low Fs.

Saying that my DIY system will not change its FR during my use because I have so much headroom that it is never pushed hard enough to run into output limitations in the boosted region is not the same thing as determining what maximum output limits and behaviors are. This is apples vs baseballs. Comparing a system purposely tested for max output against one that is stopped at coasting and then saying that one exhibits a behavior that the other does not is not a correct comparison. If you have a system with more headroom than needed all of the way down into the deep bass and you then add a good limiter into the system, the FR will not change, the sound will be indistinguishable and the dynamics will be just as good because the limiter will never engage in the first place. It is a safety measure for when systems are overdriven. You are saying that systems should not be operated outside of their limits, which I think we can all agree on but the fact is that they often are, so they are necessary in commercial systems. If you use enough subwoofers having limiters to get the headroom needed there will be no morphing of the FR shape the same as with DIY systems not having them.

You mentioned Illka's test of the LMS with the low end boosted. If he would have had a limiter set up on it he could have kept going with higher levels in this test and the FR shape would have morphed into the native shape but without any danger of damage to the driver or amplifier. If he would have kept going with the testing with no limiter, response would have also reverted back to its native shape eventually as the amplifier clips into the boosted region first while still having plenty in the tank at higher frequencies. But if the driver could not handle the full amplifier power without exceeding mechanical limits something bad would have happened. That is why limiters are employed by commercial sub manufacturers and pro audio companies both. This is a positive not a negative. If the limiter is set-up well you don't want any output past the point where it is set.
Edited by Ricci - 7/2/12 at 9:14am
post #114 of 194
Ricci, is Bosso's system equal to the apples or the baseballs? Bosso's system does not need to employ a limiter, so why should anyone else's system?


http://www.youtube.com/watch?v=7pKyh6gB1bI


Now if Bosso would talk about the use of only one of his subwoofers instead of 8 of them, I doubt that he would play the same tune regarding limiters and/or soft clipping cuicuits.


I guess Bosso is only interested in full bandwidth performance with no regard to power requirements and size of the full complement of subwoofers required to obtain his performance goals.
Quote:
Originally Posted by bossobass View Post

Quote:
Originally Posted by Mark Seaton View Post

2. Which will require more power from the wall playing back the same levels? (at levels not hitting any limiting)

My method requires more power from the wall, there's no doubt about it. The war is against dynamic peaks at the extreme low end. I bring every available weapon to that war. otherwise, I would have built a ported sub tuned to 20 Hz and stopped posting in this forum years ago.

Edited by J_Palmer_Cass - 7/2/12 at 12:48pm
post #115 of 194
Quote:
Originally Posted by Ricci View Post

Comparing a system purposely tested for max output against one that is stopped at coasting and then saying that one exhibits a behavior that the other does not is not a correct comparison.

+1000. Great post!

Quote:
Originally Posted by J_Palmer_Cass View Post

Ricci, is Bosso's system equal to the apples or the baseballs?

I vote that it's baseballs because his arguing of semantics has nothing to do with this thread (and for some reason baseballs seems like the one that doesn't belong lol)
post #116 of 194
Thread Starter 
there is also the room gain. bosso posted in room. ricci tested outdoors. the outdoor test was not benefitting from any room gain.

just going from 25hz down to 12.5hz with a q=0.8 sub requires about 20x power for the same spl and 4-5x the driver excursion.

so the limiter on the paradigm sub whether it is based on power or excursion will kick in much later if it is eq'd to take advantage or a room with some gain.

not all rooms have the same room gain, so that has to be backed out of the in room measurement in order to know what the sub itself is actually doing.

ricci, that was a pretty good write up. you might consider a place on your site where you post some of your longer/more development forum posts that aren't quite full blown articles.

one of my pet peaves about forum discussions is that you can get some really good posts going, but then they can get buried in a massive sea of other stuff. as a result, new people ask all the same questions all over again.
Edited by LTD02 - 7/2/12 at 7:00pm
post #117 of 194
I think bosso is saying that the output at 20hz is the same but the thd gets worse at 20hz as the spl increases in the higher frequency. So a limiter isn't a totally free lunch.
post #118 of 194
Thread Starter 
"I think bosso is saying that the output at 20hz is the same but the thd gets worse at 20hz as the spl increases in the higher frequency. So a limiter isn't a totally free lunch."

he is saying that if you spend $9k on a sub, the frequency response should be flat and have enough headroom, so that you will never need a limiter.

ricci is saying that if you spend $9k on a sub, having a limiter so that the sub doesn't destroy itself should you happen to hit a low note in the wrong way won't leave you with a broken system is reasonable.

i'm saying that the paradigm sub was tested outdoors, so didn't benefit from the 10+db headroom that you are likely to get in the lower bass from room gain, so the headroom in the paradigm was much understated as was the point where the limiter might kick in.

it was also said that even bosso's system has something like a limiter built in anyway in that the cabs are designed to clip the amp before the drivers overshoot.
post #119 of 194
Quote:
Originally Posted by Ricci View Post

Getting back to the conversation on limiters in commercial subs. smile.gif
This is not true.
Exactly how does a non limited sub present a higher degree of accuracy? If you utilize similar displacement capabilities and amplifier power for both systems and equalize to the same FR there won't be much of a difference at all except that the limiter will squelch the last 1dB or 2dB or so of maximum output.

Hmmmm 1 or 2 dB? Although it's always nice to have the mechanics of a small sealed EQ'd sub explained to me as though it's my first day, most of what you posted has little to do with what I posted.

I see 16dB of limiter-induced compression at 16 Hz from the DD-18 in order to get to its max output. As I explained, and is perfectly clear as a bell in your test results (and Slarti's and Ilkka's and Mullen's) of any non-limiter (DIY) sub vs any commercial sub, the DIY subs keep their FR to the top sweep and NONE of the commercial subs do.

If you boost a sealed sub with a +6dB L/T, you aren't gonna see 16dB of compression from the amps peak limiters or voltage peak limiters.

The commercial subs present an infinite number of FRs under use with program source (as opposed to 5 progressive sine sweeps with a dynamic range of 20dB), the DIY subs do not.

Try to pay attention here, if possible... 9 curves, selectable according to preferred output levels and/or source, governed by voltage peak limiting and peak limiting. Where did I say at max boost EQ any sub won't reach its EQ'd curve limit? I tried to say, apparently to a tree in the forest, that selectable curves beat frequency-specific limiter-induced curves. Your test results are obvious. With a 1 curve + limiter sub, you get far more distortion and compression vs selecting the curve that suits your systems capacity at preferred playback level. This is a no-brainer that I'm kinda shocked I even have to debate.
post #120 of 194
Quote:
Originally Posted by LTD02 View Post

"I think bosso is saying that the output at 20hz is the same but the thd gets worse at 20hz as the spl increases in the higher frequency. So a limiter isn't a totally free lunch."
he is saying that if you spend $9k on a sub, the frequency response should be flat and have enough headroom, so that you will never need a limiter.
ricci is saying that if you spend $9k on a sub, having a limiter so that the sub doesn't destroy itself should you happen to hit a low note in the wrong way won't leave you with a broken system is reasonable.
i'm saying that the paradigm sub was tested outdoors, so didn't benefit from the 10+db headroom that you are likely to get in the lower bass from room gain, so the headroom in the paradigm was much understated as was the point where the limiter might kick in.
it was also said that even bosso's system has something like a limiter built in anyway in that the cabs are designed to clip the amp before the drivers overshoot.

Feel free to speak for Josh, but I'm not talking about headroom at all.
New Posts  All Forums:Forum Nav:
  Return Home
  Back to Forum: DIY Speakers and Subs
AVS › AVS Forum › Audio › DIY Speakers and Subs › Hey Ricci...