Pyschoacoustics and the LAS vs. Khorn debate

[DrWho]

I know this is going off topic...(perhaps we could continue this in a

PM?), but what did you do to model/anticipate the room gain? Right now

I'm using a linkwitz-transform using a formula I picked up somewhere,

but I don't think it's too accurate (though it's not too far off from

the predictor in the demo version of lspcad).

[sfogg]

"but what did you do to model/anticipate the room gain?"

Nothing very fancy. I used Linkwitz's Modes spreadsheet and his

statement that room gain starts below the lowest mode in the room. If

the room is well sealed/solid it is a second order boost which

corresponds with the second order rolloff of a sealed woofer. If the

room has a lot of losses in it then the gain won't be that strong. My

room is pretty small so the gain kicks in at a reasonable frequency and

all the walls are double drywall glue/screwed together, same with the

ceiling and the floor is poured concrete.

I had a bit of luck in this as my main goal for the subs was the Q

value as I have a ContraBass (f3 14hz) for the LFE channel. While I was

modeling it it also worked out that the low frequency rolloff looked

like it should be close to where the room was going to start kicking in

so that was what made me decide to go with that layout.

Shawn

[edwinr]

Shawn, that second graph you posted, measured at your listening position, seems to me to be a better representation of what we might be expected to hear in our home. Again this would still vary depending on room size and other variables such as materials used in construction, etc. The little peaks and trouphs would be room anomolies, I would suggest.

Measuring the La Scala at, or near the horn throat, would probably have more validity than trying to measure the Klipschorn in a similar fashion. I guess using known acoustic models, you would be able to factor in room gain to extrapolate the La Scalas true in-room performance. The Klipschorn must, by the very nature of it's design, be more difficult to measure.

Dr Who. I am not suggesting that all measurements are bunk. But merely posting some graphs with wiggley lines on it [] does nothing for me. I agree that measurements may give us a general idea about what we may be hearing, but it's got to be applicable to the actual environment we are listening in. We need to identify what we are measuring first. Then we need to know how these measurements reflect what we are hearing. Not such an easy task. []

[sfogg]

"measured at your listening position, seems to me to be a better

representation of what we might be expected to hear in our home."

It is certainly a better representation of what I hear at my listening

position then a free space measurement or a near field measurement.

The problem of course is what the system does in my room might be very

different in someone elses room. For example in a much larger room, or

an open room, I wouldn't have the room gain extending my bass response.

Or if the other room was very live (mine is very dead) they might sound

far to bright.

"The little peaks and trouphs would be room anomolies, I would suggest. "

The big dip at 500hz looks to be a room problem. I ran the mid-range

down to 400hz and still saw the dip at 500hz so it isn't something

specific to the LaScala. Some of the other peaks/dips are likely room

too, but I'm sure others are also the response of the drivers/horns

themselves.

"you would be able to factor in room gain to extrapolate the La Scalas true in-room performance. "

Taking advantage of room gain to extend response really works best with

sealed enclosures. At best room gain adds 12dB per octave. A sealed

enclosure (depending upon the Q) rolls off at 12dB per octave. If

implemented well the room gain can basically apply the inverse of the

sealed enclosure to keep the response flat to the point that room gain

stops being effective.

For a vented enclosure or a horn then roll off 24dB/octave. Even if

sized so that they are rolling off at the point room gain is kicking in

you will still end up rolling off 12dB/octave so you can't use room

gain to extended the response flatly.

By far the best way to see what the in-room frequency response of a

speaker is is simply to measure it using a RTA, MLS, FFT..etc..etc.

Measuring response with a SPL meter can be very misleading as the SPL

has no way to differentiate frequency.

Shawn

[maxg]

Now I dont want to start the great amp debate again but I cant help wondering if those of you that are using a sub to fill in the bass on a Khorn are merely exploring an alternative to using a bigger amp.

Now I understand that the KHorn bass response is not exactly flat down to 35 Hz - but it is most definitely there and the peaks that we see in the graph are probably nothing in comparison to the sort of variations you would expect to see in a typcial listening room.

Many moons ago when Tony had a pair of KHorns he ran them - after much switching and testing - with an Accuphase E406 amp (170 wpc into 8 ohms). The last thing that ever crossed my mind - or more importantly his, was that there was insufficient bass!!

This is entirely my opinion but it seems to me that as the KHorn has a 15 inch woofer one would expect to need an amp with some power reserves in order to control it properly. That was certainly our experience as Tony searched for an amp to match. He must have tried dozens of amps - including many low power tube options before settling on the Accuphase.

The LaScala, in comparison, is not a bass heavy speaker, and the mere fact that is sits lower to the floor means that it has a rather different presentation from the KHorn - despite having the same tweeter/squarker combo.

It would be very interesting to compare the LaScala with bass bin mod to the Khorn as that would both raise the LaScala height comparible to that of the Khorn and fill in much of the missing bass.

All of this is a round about way of saying that IMHO the big difference you are all experiencing when adding a sub is not really the additional driver - so much - as it is the additional amplification that goes with it. In effect you are bi-amping and replacing the bass driver in the process. Its not wrong persay - just a means of bypassing the horn loaded woofer which, as the defining characteristic of the Khorn, is an odd thing to do IMHO.

OTOH - if you tell me that this is the best way to get the sweetness and midrange bloom of low power tube amps without sacrificing the low end oomph who am I to argue?

[NOSValves]

Btw, rejecting ANY kind of measurement based on pure emotion is a complete waste of time. If you disagree with anything, you must certainly have a reason for it and then be able to back it up. I might as well just claim the khorn doesn't dig below 80Hz based on a listening test. I know I'm wrong, but how do I know? Because I can easily measure the speaker and say hey, look how it digs deeper. And then I can confirm this by listening to some music and noting the correlation. Anyone claiming that measurements don't correspond to what we hear is essentially claiming that the models used to design speakers is completely false. And this means you are up against the entire audio industry where these models are being used on a daily basis.

First off who died and put you in charge of what people say and how much backing up of what they say has to be provided? I have a reason for dismissing those graphs and it has nothing to do with emotion. Those graphs are completely tilted in the Khorn's favor. I suggest you go to the Klipsch head quarters and see for your self how they test the Khorn compared to the rest of the speaker line. The Khorn is mounted in its optimum low frequency position while all other speakers are in there worst possible position for low end response. I completely understand why they do this but the graph's are still very much misleading in the real world.

Craig

[DrWho]

You have quite a bit to say about the measurements and how they're

invalid, but have yet to objectively support your own claims. Allow me

to provide the true interpretation of the post that you quoted: "you

are full of crap"

(until of course you can provide real proof other than your faulty ears...hey, my ears are faulty too).

Room gain can be very accurately modelled and that is why anechoic

charts are useful. (there is no transfer function to be undone with the

measurement, so all you have to do is apply the transfer function of

the room the speaker is going in). You would have to have one fricken

small room to start noticing room gain at 50Hz...in fact, you would

have to be in a room about the size of a car. And even then, you aren't

going to get that much more low end out of the speaker. Go model it for

yourself. Oh wait, we gonna start claiming that T/S is full of crap now

too?

[NOSValves]

Dr Who,

I suggest you watch your mouth my friend there is no need to be nasty. The point is in the Klipsch chamber and also when placed outside the Khorn is tested with full benefit of corner placement while all other speakers are not. The Klipsch engineers have come into the forum and backed up my statement. The graphed are misleading in this case. I understand completely why they do this and that there is no fair way to all speakers involved in THIS unique case. But none the less the graphs are very much misleading. Place the Lascala in the same corner in the chamber or out doors and it will test very differently. All speakers gain low end response with corner placement.

I never said the charts are useless I said or meant the comparison is useless in this case.

YOU ARE THE ONE FULL OF CRAP. It's reached levels above your eyebrows at this point.

Craig

[DrWho]

And I'm saying you can model corner and room gain and it doesn't make a

big difference. How do you account for Shawn's measurement? That's a

measurement of a lascala in a corner. I know the mic was very close, but still...

Gosh, all it'd take would be one measly graph (measured correctly of course).

[NOSValves]

I'm saying why bother.

[sfogg]

"How do you account for Shawn's measurement? That's a

measurement of a lascala in a corner."

For the record the first measurement of the LaScala was not in a corner.

Shawn

[NOSValves]

Records are always better when there straight. I hate them darn warped ones.

[edwinr]

Dr Who, you're letting this measurement thing get to you. Without doubt certain acoustic models can give an idea of how a particular loudspeaker may work in a particular room. But the problem with any model, is that it is only as accurate as the parameters included for the sake of measurement. In addition, many acoustic models are not perfect. They are constantly being refined and improved. You go on about the importance of measurements, and I'm not going to dispute this necessity. I feel the need to point out however, that many high end loudspeaker designers tune by ear. The reason for this is that the materials used in the construction of the cabinets, transducers and crossovers have tolerances which may or may not be affected by their interaction with each other, and/or the environment they have to work in. And there may very well be other factors not fully understood and explained by current acoustic models. Loudspeaker design has moved on from the days when it was seen as a 'black art' - but not that far beyond. []

[PrestonTom]

Dr Who, you're letting this measurement thing get to you. Without doubt certain acoustic models can give an idea of how a particular loudspeaker may work in a particular room. But the problem with any model, is that it is only as accurate as the parameters included for the sake of measurement. In addition, many acoustic models are not perfect. They are constantly being refined and improved. You go on about the importance of measurements, and I'm not going to dispute this necessity. I feel the need to point out however, that many high end loudspeaker designers tune by ear. The reason for this is that the materials used in the construction of the cabinets, transducers and crossovers have tolerances which may or may not be affected by their interaction with each other, and/or the environment they have to work in. And there may very well be other factors not fully understood and explained by current acoustic models. Loudspeaker design has moved on from the days when it was seen as a 'black art' - but not that far beyond. []

Certainly the topic has drifted. But this issue of how much should we rely on measurement, especially as the first step, is of a personal interest to me.

It is probably best considered as a continuum. Clearly there are some folks who are toward one extreme and discount many of the advantages afforded by physical & performance measurement. Dr Who tends toward to the other end of the continuum, where measurement is the key.

Similar to Dr Who, I also favor the measurement approach. Quite a bit can be learned from this approach. Quite a bit of nonsense can also be avoided (the typical audiophile voodoo etc).

When the argument sinks to the level of name calling, then the Dr Who's of the world seem even more reasonable in their reliance on measurement rather than notions from the "black art".

Good Luck,

-Tom

[3dzapper]

How can a $45K speaker measure this bad yet sound so good? PWK had it right: low distortion. Frequency response charts mean nothing if a flat response speaker sounds like a muffler has been added to get that line.

"Fig.3 Wilson Audio Specialties MAXX2, anechoic response on axis 36" from the floor, corrected for microphone response, with the complex sum of the nearfield midrange, woofer, and port responses, taking into account acoustic phase and distance from the nominal farfield point, plotted below 300Hz."
(stereophile.com)

Here is what JA had to say about the measurements:

"Some of the benefits of David Wilson's design philosophy, such as the MAXX2's undoubtedly low distortion and its high maximum loudness capability, go unremarked in my suite of measurements. However, the speaker's impressively extended and powerful low frequencies were evident, as was its relatively smooth in-room responsewith the exception of the elevated upper midrange, which concerned me. But I must say that, measured performance aside, I was very impressed by what I heard in Mikey's listening room.John Atkinson "

Craig, As shipped the Khorn is incomplete, the false corner merely "completes" the Klipschorn for measurement purposes. Although it is better than nothing, a false corner does not live up to the full potential of the corner horn on good, solid corners.

Shawn, When are you coming down for a Khorn audition? You can bring your mic and stuff and we can measure Khorns in good corners. Some may be surprized, I might be chagrined.

Rick

[NOSValves]

Craig, As shipped the Khorn is incomplete, the false corner merely "completes" the Klipschorn for measurement purposes. Although it is better than nothing, a false corner does not live up to the full potential of the corner horn on good, solid corners.

Rick

Rick,

I believe I said "I completely understand why they test as they do" [] Still very much misleading when you compare the two side by side. I mean these graph's and tests are only so usefull no matter what IMHO. If your going to go by them I think we all just better dump the Heritage speakers all together since there has to be many speakers with better graphs out there. I bet Klipsch has current lines of speakers that have better graphs then the Heritage line

Craig

[sfogg]

Rick,

"When are you coming down for a Khorn audition? "

Probably couldn't make it till after the holidays. Things are a little nutty right now.

"You can bring your mic and stuff and we can measure Khorns in good corners. Some may be surprized, I might be chagrined."

Sure, how much in the way of measurements are you interested in?

We can do simple FR very easily. THD and THD+Noise at different frequency and SPL points is pretty straight forward too but takes more time the more points you want to look at.

The other one that could be interesting is impulse response and time domain behavior. I have the software to do that too but it has been acting fairly flakey on me lately.

Shawn

[DrWho]

"Dr Who, you're letting this measurement thing get to you. Without

doubt certain acoustic models can give an idea of how a particular

loudspeaker may work in a particular room. But the problem with any

model, is that it is only as accurate as the parameters included for

the sake of measurement. In addition, many acoustic models are not

perfect. They are constantly being refined and improved. You go on

about the importance of measurements, and I'm not going to dispute this

necessity. I feel the need to point out however, that many high end

loudspeaker designers tune by ear...."

Perhaps...I always do my best to listen to everything without bias and

decide upon what I hear. I've grown up thinking about sound in terms of

numbers so perhaps that gets a bit misleading at times (for example,

instead of saying "man, that kick could use a lot more oompf to it" i

say "man, that kick could use more 80Hz and a touch of 40" - I guess

you could say I describe things in the manner I would go about

addressing them"). Whenever what I hear actually corresponds to what I

think I'm hearing, then I note the correlation and move on. It's funny

how close you can get as long as you don't take it too far. (Like with

the khorn...I've never seen any frequency response plot and yet I

pictured a very similar to curve to the one klipsch

provided...coindidence? perhaps, but there is certainly a correlation

between frequency response and what one hears. There is also a

correlation between the average output and all the other factors that

affect the sound of a speaker as well). Anyways, the point is that it

always comes back to comparing what you hear with the measurements. If

the measurements don't line up, well then go take more measurements and

find a reason for it. But if the measurements do, then I see no reason

not to draw a correlation between the two. For example, the lascala

drops fast below 50Hz. When listening to music with a bass line

centered between 30 and 50Hz (as can be shown on a waterfall plot) the

entire bass line just drops out and you can't hear it. Also, kick drums

lose their impact (again, a 40Hz-ish frequency). The conclusion? The

lascala just don't dig below 50Hz. Now let's pretend that

corner-loading results in a flat response down to 40Hz...would you

prefer a speaker that can just barely satisfy the requirements or would

you rather have a speaker that digs a little lower and catches all the

frequencies without a problem.

"How can a $45K speaker measure this bad yet sound so good? PWK had it

right: low distortion. Frequency response charts mean nothing if a flat

response speaker sounds like a muffler has been added to get that line."

I would bet good money that the Wilson speaker would sound very bloated

at louder volumes, but I have no doubt that it sounds great at lower

volumes. In fact, it's curve reminds me very much of the cornwalls

response. The first thing that comes to mind are the F-M curves. Also

at loud volumes the natural low end will be too loud which will be

compounded by the voice coil getting hotter. I have a feeling though

that the speaker starts to break up before this starts to happen. How

loud is loud? well that will be dependant on the specific listener in

question [] Now if it doesn't behave like I predict (which is a

pretty loose and stretched prediction that I would never bias myself

with), then I'm sure we can find some other behavior that explains it -

like perhaps the interaction of the off-axis response and the room

coupled to the direct sound at the listening-position. Heck, this is

why science is so fun....you get to predict things, test them, and then

use the same predictions over and over until proven false (or are

better modified to fit all corner cases). This is where the "black

magic" comes in because we aren't able to fully model acoustics

(despite how simple of a subject it really is).

crap, time for finals so i gotta end here....

[3dzapper]

Quite the contrary, Who. The reviewer's guest asked him to turn down the volume and was impressed that the Maxx's sounded as good at low volumes as they did at high.

David Wilson as a speaker designer belongs to an exclusive group up there with Allison and Kloss. As did PWK, Wilson voices his speakers by ear, then measures them.I have no doubt that properly set up and with sufficient power the Maxx 2 would put my Khorns to bed wet.

Shawn, Just an RTA or two that we can we can post. Good or bad I'm sure it will interest the rest of the forum. I'll upgrade my RTA subscribtion for the event. After the hoildays is ideal. I will be stuck home for a while convelescing. You can lug my different amps to switch them out.-)

Craig, PM me your phone number I need to talk with you.

Rick

[DrWho]

Well now I wanna go hear them to see for myself. I don't tend to trust

reviews unless I know the person (and their musical tastes). I very

much believe a speaker needs to be built around the source material

being listened to (even on this forum there seems to be somewhat common

trends in speaker preference based on the musical tastes of the

listener).