Audyssey Reference curve and controlled directivity speakers?

[yepimonfire]

The reference curve targets a sloping response due to the fact most of the room reflections mirror most speakers off axis roll off, meaning the combined reflected and direct sound would exhibit their target curve roll off, rather than flat. The issue is my speakers do not roll off off axis and far field measurements show this. The FR at 11’ is mostly flat with no slope out to 14khz, where a roll off begins due to the waveguides lack of directivity control above 14khz.

After running Audyssey, I noticed my speakers sounded veiled in comparison to before, and the response now mimicks target curve vs the before eq flat response at the mlp. Note that my speakers DO NOT display brightness that would need correction, they are +-2dB flat anechoically as can be seen from this

Is the target curve appropriate with CD speakers?


Sent from my iPhone using Tapatalk

 

[TLS Guy]

The reference curve targets a sloping response due to the fact most of the room reflections mirror most speakers off axis roll off, meaning the combined reflected and direct sound would exhibit their target curve roll off, rather than flat. The issue is my speakers do not roll off off axis and far field measurements show this. The FR at 11’ is mostly flat with no slope out to 14khz, where a roll off begins due to the waveguides lack of directivity control above 14khz.

After running Audyssey, I noticed my speakers sounded veiled in comparison to before, and the response now mimicks target curve vs the before eq flat response at the mlp. Note that my speakers DO NOT display brightness that would need correction, they are +-2dB flat anechoically as can be seen from this

Is the target curve appropriate with CD speakers?


Sent from my iPhone using Tapatalk

I Don't use an Audyssey curve at all and good speakers don't.

I note they did a real hack job on the crossover, with a big wide null right on it. That has to be highly audible.

 

[yepimonfire]

I Don't use an Audyssey curve at all and good speakers don't.

I note they did a real hack job on the crossover, with a big wide null right on it. That has to be highly audible.

That’s the “midrange compensation “. They default it to 2khz, it’s supposed to correct a frequency response problem due to the directivity mismatch between the tweeter and woofer, Microphones do not pick up sound the way our ears do, so a dip is needed in eq corrections unless you want very bright midrange, however, a CD design doesn’t have a directivity mismatch. My speakers are crossed over at 1.5khz, the directivity of the horn and the woofer are identical at this frequency, so there is no mismatch obviously. This can be confirmed by off axis measurements

Everything from about 1khz onward is controlled in a 90 degree pattern, dropping approximately 2dB for each 15 degree increase off 0 degrees. Dome tweeters in two way designs often drop like a rock above 4-8khz off axis.

It’s basically making assumptions based on “typical” home audio speakers.


Sent from my iPhone using Tapatalk

 

[TLS Guy]

That’s the “midrange compensation “. They default it to 2khz, it’s supposed to correct a frequency response problem due to the directivity mismatch between the tweeter and woofer, Microphones do not pick up sound the way our ears do, so a dip is needed in eq corrections unless you want very bright midrange, however, a CD design doesn’t have a directivity mismatch. My speakers are crossed over at 1.5khz, the directivity of the horn and the woofer are identical at this frequency, so there is no mismatch obviously. This can be confirmed by off axis measurements

Everything from about 1khz onward is controlled in a 90 degree pattern, dropping approximately 2dB for each 15 degree increase off 0 degrees. Dome tweeters in two way designs often drop like a rock above 4-8khz off axis.

It’s basically making assumptions based on “typical” home audio speakers.


Sent from my iPhone using Tapatalk

That is corporate BS. Don't be a sucker. That is a lousy crossover transition.

 

[gene]

I concur with TLS Guy. Turn that 2kHz dip off if you're gonna use Audyssey at all. Also, if you have a recent Denon/Marantz, it's best to use the Audyssey editor app and limit the EQ correction to under 5kHz, preferably around 500Hz, around the room transition frequency where the speaker dictates most of the response, not the room.

 

[yepimonfire]

That is corporate BS. Don't be a sucker. That is a lousy crossover transition.

What is BS?


Sent from my iPhone using Tapatalk

 

[yepimonfire]

@gene

My room is well treated with absorption effective from about 500hz up. The only area I really have issues is between 80-200hz. I’ll likely limit the curve to that. Even the “flat” setting mucks up the sound. I never understood why someone would try correcting more than bass, I guess if you had a non-flat speaker, Audyssey could correct it, but I would hope most of us are using reasonably accurate speakers to begin with.


Sent from my iPhone using Tapatalk

 

[TLS Guy]

What is BS?


Sent from my iPhone using Tapatalk

They say one picture is worth a thousand words!

 

[everettT]

@gene

My room is well treated with absorption effective from about 500hz up. The only area I really have issues is between 80-200hz. I’ll likely limit the curve to that. Even the “flat” setting mucks up the sound. I never understood why someone would try correcting more than bass, I guess if you had a non-flat speaker, Audyssey could correct it, but I would hope most of us are using reasonably accurate speakers to begin with.


Sent from my iPhone using Tapatalk

Most people don't have Speakers with a flat response or accurate as you call them. The majority of people are hoping for the "room correction" in those frequencies where the room interactions occur .

Why do you say you don't need Correction below 80hz? I know plenty of people that spend countless hours on setting up their subs, and mains in the beginning, none of them assume that the RC is going to work magically or out of the box on the first run. When I changed to outboard amps in my subs from their plate amps that had boost @25hz with high pass filter, it took me a few days, several measurements and lots of math to achieve the response I was looking for. How many positions did sweep?

Also post before and after curves when asking advice or opinions as it helps form them.

 

[yepimonfire]

@everettT

I don’t know what you’d consider “flat”, but I do know many studio highly accurate monitors achieve around +-2dB anechoically, for the Klipsch to achieve that on axis, and a +-45 degree off axis response (-6dB) at +-3dB 1000hz-14khz, I’d consider it highly accurate.

If a speaker with a sloping response off axis (ie most of your run of the mill 2 way direct radiating speakers with 1” dome tweeters) is equalized flat in the far field, you will end up eqing the direct sound with a rising treble response and it will be extremely bright sounding and certainly not accurate. Our ears first hear the direct sound and then combine the reflected sound with it, in addition, unchanging aberrations in response are “filtered” psychoacoustically, IOW, our brains adapt to the response variations introduced by the room, effectively ignoring it after a few minutes of listening. Ever worn a pair of crappy headphones for awhile and eventually notice they don’t sound as bad after a few hours of listening? Same idea. A microphone and frequency response measurement does not differentiate between the two, since a majority of the off axis sound bouncing off the walls and ceiling has a response that declines rapidly with frequency above 4-8khz, the mic will see a sloping response from the combined direct sound and “rolled off” reflected sound. A speaker with a waveguide that gives uniform directivity that doesn’t change significantly with frequency will not have a sloping off axis rolloff, rather, the off axis sound will simply be attenuated across the entire band that the waveguide controls directivity. The original graph I posted is a far field measurement that is a combination of the reflected and direct sound along the gold 0 degree line. As you can see, the combined in room response (ignoring the weird dip at 2.5khz) is -5dB, +3dB, and only begins to roll off at 14khz, where the waveguide fails to control the dispersion at the top end.

Eqing for a sloping response used in the original Audyssey target curve simply dulls the direct response of the speaker. I may try playing around with the multeq editor and see if it will allow me to create a curve that slopes above 14khz, since the flat setting would likely lead to an eardrum piecing boost from 16khz-24khz. Either way, what’s the point in EQing what is already an impressively flat far field response above 500hz?

I do need some minor correction below 80hz. I have a very small (3dB) bump at ~ 50 hz but the bass above 80hz up until about 200hz is pretty whack. This is an older measurement but still applies

Multiple subs wouldn’t be able to solve the erratic behavior in the mid bass because it would require a 200hz xover, even with the sub placed dead center between the mains, not be localized, even a 120hz xover I was forced to use with a pair of Polk satellites at one time sounded bad, like the mid bass was disconnected from the sound stage, even though the sub was perfectly level matched and time aligned. The “mother in law” seat along the side wall experiences an 8dB boost at 47hz. While this doesn’t sound good with music, it’s great for rattling the innards of guests during movies, or as a remedy for occasional constipation



Sent from my iPhone using Tapatalk

 

[highfigh]

The reference curve targets a sloping response due to the fact most of the room reflections mirror most speakers off axis roll off, meaning the combined reflected and direct sound would exhibit their target curve roll off, rather than flat. The issue is my speakers do not roll off off axis and far field measurements show this. The FR at 11’ is mostly flat with no slope out to 14khz, where a roll off begins due to the waveguides lack of directivity control above 14khz.

After running Audyssey, I noticed my speakers sounded veiled in comparison to before, and the response now mimicks target curve vs the before eq flat response at the mlp. Note that my speakers DO NOT display brightness that would need correction, they are +-2dB flat anechoically as can be seen from this

Is the target curve appropriate with CD speakers?


Sent from my iPhone using Tapatalk

What speakers are you using? The dip centered at 1500Hz looks like a phase cancellation or a crossover that has the -3dB/-6dB points too far apart, assuming it wasn't caused by Audyssey (and I have yet to see it do that). Is it switchable? You shouldn't need to use anything that creates anything like this. Your response through that region should look like the surrounding areas. That removes a good chunk of a whole octave, near the most sensitive range of 'typical' human hearing.

The response above shouldn't sound 'veiled', if anything, it should lack a bit of warmth because there's less mid-bass than mid-range and treble.

I would try tweaking the speaker placement and look into using something to break up the nulls in the bass (assuming you can't just put bass traps in).

 

[TLS Guy]

What are the dimensions of the horn mouth of those speakers?

You are in denial about the awful crossover problem with those speakers. I have always felt that the crossover is too low. The low frequency cut off of a horn is entirely determined by the dimensions of the mouth. Those Klipsch speakers have always looked to me to have too small a horn mouth for the required FR.

I suspect the woofers do not have the FR to be crossed over where the horn needs to be crossed over. There may also be severe phase issues with the crossover.

That dip makes them not a good reproducer at all. I would never leave a design like that.

 

[yepimonfire]

What are the dimensions of the horn mouth of those speakers?

You are in denial about the awful crossover problem with those speakers. I have always felt that the crossover is too low. The low frequency cut off of a horn is entirely determined by the dimensions of the mouth. Those Klipsch speakers have always looked to me to have too small a horn mouth for the required FR.

I suspect the woofers do not have the FR to be crossed over where the horn needs to be crossed over. There may also be severe phase issues with the crossover.

That dip makes them not a good reproducer at all. I would never leave a design like that.

As I’ve previously pointed out, that dip is introduced by the room, not the speaker. The right side of my wall has a large window that is mostly uncovered except for blinds and 36”x24” acoustic paneling at the first reflection point, the panel is placed to catch reflections at a distance of 11’, not 4’, and therefore, the window reflection at the measurement spot is responsible for that dip. The left speaker measured at 1m gives this response, which does not display a dip.

As for whether or not the mouth diameter is appropriate for the xover point:

The tractrix horn in Klipsch speakers isn’t a pure tractrix design, but either way...the math should still apply since the cutoff is determined by the mouth area.

The cut off frequency of a tractrix horn is determined by the area of the mouth where the mouth area (in m^2)=c^2/(4*pi*fc^2). Where c=the speed of sound in m/s and fc= cutoff frequency. A horn should be crossed over well above the cutoff to avoid problems. Plugging the numbers into the above equation and applying it to a 1500hz xover frequency being an octave above the cut off, we get 343^2/(4*3.14*750^2)= .01665 sq meters or 25.8” square. the mouth of the horn measures 5”x5” which equates to 25 inches square, so I’m not sure why you believe the horns mouth is too small to properly support a 1500hz xover.


Sent from my iPhone using Tapatalk

 

[TLS Guy]

As I’ve previously pointed out, that dip is introduced by the room, not the speaker. The right side of my wall has a large window that is mostly uncovered except for blinds and 36”x24” acoustic paneling at the first reflection point, the panel is placed to catch reflections at a distance of 11’, not 4’, and therefore, the window reflection at the measurement spot is responsible for that dip. The left speaker measured at 1m gives this response, which does not display a dip.

As for whether or not the mouth diameter is appropriate for the xover point:

The tractrix horn in Klipsch speakers isn’t a pure tractrix design, but either way...the math should still apply since the cutoff is determined by the mouth area.

The cut off frequency of a tractrix horn is determined by the area of the mouth where the mouth area (in m^2)=c^2/(4*pi*fc^2). Where c=the speed of sound in m/s and fc= cutoff frequency. A horn should be crossed over well above the cutoff to avoid problems. Plugging the numbers into the above equation and applying it to a 1500hz xover frequency being an octave above the cut off, we get 343^2/(4*3.14*750^2)= .01665 sq meters or 25.8” square. the mouth of the horn measures 5”x5” which equates to 25 inches square, so I’m not sure why you believe the horns mouth is too small to properly support a 1500hz xover.


Sent from my iPhone using Tapatalk

I had no idea what the size of the horn was. I have modeled it taking a guess at the Tractrix horn length and the size of the horn does seem appropriate for that crossover point.

 

[yepimonfire]

I had no idea what the size of the horn was. I have modeled it taking a guess at the Tractrix horn length and the size of the horn does seem appropriate for that crossover point.

The throat of the horn is conical, without actually being an insider and having access to trade secrets, it’s very difficult to pin down what the design consists off. If you read through some of the posts on the Klipsch forum, you’ll see that even though the design is tractrix based, and the mouth has a tractrix flare, the throat and walls have been heavily modified for uniform directivity control, since a true tractrix horn would beam quite terribly, and tractrix horns aren’t exactly good for high frequencies.

Klipsch does not use a ubiquitous xover point or horn size for their various speaker models. The same titanium and aluminum diaphragm is used in nearly all of the premiere series and cheap reference series. The woofer is usually crossed over at a point where the directivity matches the horn at a ka=2. For example, the R-14m has a 4” woofer and is crossed over at 2250hz vs 1800hz for the 15m. My guess is the 15m is crossed over at 1800hz vs the 150ms 1500hz due to the fact the actual cone diameter excluding the surround is smaller.

Ignoring room interaction, you can see that, like every other Klipsch design, directivity control for the small R-14m starts at around 1.5khz. One benefit of a smaller horn in this case (I think it’s somewhere around 3” square?) is that the directivity extends all the way out to 20khz, vs 14khz in the bigger horns.

For a $100 (on sale) a pr of speakers, the response is surprisingly flat.

Every two way Klipsch speaker from the reference II series on up that I’ve measure thus far displays very similar results, with very consistent control of directivity beginning somewhere around 1khz, regardless of xover point.

I am not simply repeating marketing BS, in fact, none of the consumer speakers say anything in their marketing info regarding most of this stuff outside of vaguely mentioning controlled directivity, it was determined by taking measurements, and then, out of curiosity, researching the design of the modified tractrix ubiquitous in both professional and consumer models. There are posts from Roy Delgado (along with a few other Klipsch employees), the original engineer behind the modified tractrix design, floating around that touch on the design being based on both constant directivity and constant coverage (ie the directivity doesn’t change with frequency) and the fact that nearly all Klipsch speakers, from the reference II series, icon k series, the modified “Best Buy” Reference series, old synergy series, Reference Premier series, palladium series, and most of the heritage III series horns are based on this same design.


Sent from my iPhone using Tapatalk