Question for those familiar with Audyssey.

[Chris A]

Just now, Rich_Guy said:

 

Use ALL available positions, different versions of Audyssey have different amounts of positions available

 

Yes...but not the results...

[Rich_Guy]

14 minutes ago, Chris A said:

Yes...but not the results...

You can argue that with the Audyssey founder.

 

Personally, I run Audyssey and after that I prefer to make my own changes which include setting my speakers to SMALL (if needed) re-setting my speaker distances to the actual real distances and re-setting my speaker levels with my own SPL meter at my main listening position.

 

[Zen Traveler]

3 minutes ago, Rich_Guy said:

 

Use ALL available positions, different versions of Audyssey have different amounts of positions available

 

Fwiw, I was an early Audyssey user and posted quite a bit on the early pages of that AVS thread. Chris K used to post as well and then started the "AskAudyssey" FB page and then abruptly stopped at AVS. I remember him mentioning using all positions in the early days but haven't seen that referenced in quite some time so I understand Chris A's (not to be confused with "Chris K" the Audyssey CTO ) concern that there are no contemporary quotes or guidance from him since that thread's inception.

 

That said, I still have an older Denon AVR-4311ci that uses Multi XT and actually didn't even notice if the FB thread was still operational--If it is, then that would be a good source to find the answer, at least from the company itself.

[Chris A]

The point that I've been attempting to illustrate is not that Audyssey doesn't work at all (it does work under narrow conditions).  It's that its results depend quite heavily on the exact room and typical loudspeaker types that it is applied to.  Try to apply it to K-402s and Jubilee bass bins (or the K-402-MEH, for that matter), which control their polars down to the 100-200 Hz band and which sound quite spectacular without Audyssey, and you're in for an unpleasant surprise--no matter how many times that you run the algorithms and how many places that you place its little plastic-encased calibration microphone around your listening position(s).

 

In my case, Audyssey completely falls apart for some of the reasons stated above (and for others that I haven't discussed yet).  I'm not dissatisfied with my room or my loudspeaker setup--but only the performance of the Audyssey algorithms--which make assumptions that are clearly not valid for my setup and room. My setup requires no (i.e., zero) room curve or house curve compensation above 100 Hz. 

 

Audyssey is not a panacea and I believe that many of the AVR/AVP manufacturers are finding that little bit of information out the hard way...and are choosing to use other algorithms from other sources than this company's offering.  I applaud their decision to "spread the wealth" of engineering solutions, since I found the usage of Audyssey to be extremely brittle in actual usage. It costs too much for what it's doing to be that brittle in its application latitude.

 

You may find it useful, but note that it doesn't work for all setups.  It seems to be very limited in its applicability across differing audio applications, such as the one that I'm using.

 

Chris

[garyrc]

6 hours ago, Chris A said:

The results of Audyssey basically turns off the highs and the extreme lows below 20 Hz in some sort of "preferred Audyssey curve" (an oxymoron)

Once Audyssey calibration has been run, we have a choice of "Audyssey FLAT," which does not turn off the highs (or lows), or "Audyssey Reference," sometimes called just plain "Audyssey" by some manufactures, and "Audyssey Movie" by others, which starts rolling off the highs at 8K Hz, down 2 dB at 10K Hz, and down 6 dB at 20K Hz.  I much prefer "Audyssey FLAT," which produces significantly flatter response than with "No Audyssey."  With my sound system, IIRC, it makes many small corrections, plus boosting the response above 10 K, cutting down a peak centered at 8 K, curing a dip centered at 350 Hz, and generally boosting the bass a bit below about 38 Hz. 

 

In switching back and forth between "Audyssey FLAT" and "Audyssey OFF,"  "Audyssey FLAT" always sounds better, noticeably clearer, and, if you will, more musical.  Like many users, I boost the mid bass a bit, and the low bass (sub below 80 Hz) a bit more.

 

I have a moderately treated room, almost fully carpeted, 16.75 feet wide, 25 feet long, with a ceiling that slopes from 8.5 feet in the front to 11 feet 10 inches in the back. 

 

I believe that all of Chris K's early posts are still on AVS "Official Audyssey Thread (part I), or "Ask Audyssey" online.  The later ones are on Audyssey's Facebook.

[Chris A]

Gary,

 

We're not talking about the same thing.  The Audyssey MultEQ® XT that's distributed with the preamplifier does not have any choices except Audyssey EQ "on" or "off".  To do what you're talking about requires the insertion of an extra $500 to Audyssey to buy their after-market "MultEQ" firmware (not hardware)...which to me is outrageous.  I really don't recommend that path.

 

Instead I recommend REW (which is free) and a USB calibration microphone ($100) with a digital  crossover of your choice (a physical box that includes hardware and firmware of real value), or even a source-based parametric EQ (free).  Then you have the correct visibility and control over what is happening--not some bait-and-switch proposition.

 

For any company to put out any product that strips off the needed information for visualization and control of what is happening is for me not one that I recommend. 

 

Chris

[derrickdj1]

Subwoofer distance is the acoustical distance based on when the sound get's there.  It can be longer than the measured distance.  it should not be shorter than the physical distance.

[Chris A]

Tapped horn subs all have a low-amplitude "precursor pulse" from the side of the woofers near the mouth of the horn.  In the case of the SPUD design that I use (one in each corner of the room behind the Jubs), that's about 20 feet short of the actual maximum-amplitude pulse. 

 

                 

 

If Audyssey was designed such that it was keying on the high-amplitude low-frequency portion of the subwoofer's output via allowing the internal cross-correlation filtering left to run its full course of 30+ feet (or more) vs. the input electrical chirp, and using the maximum amplitude of the cross-correlation, there would be no problem. Instead they frequency filter the sub cross-correlation using only higher frequencies (which Audyssey thinks is the main pulse because it's not hearing the true low frequency pulse below 40 Hz that the subs are low-passed to--because of their little plastic microphone issues). 

 

So Audyssey goes along thinking that the subs are aligned, assuming that all subs are direct-radiating subs, and that they can be "better aligned" at frequencies well above 40 Hz--including completely ignoring those frequencies for which the sub is actually there to cover. 

 

This is the tip of the iceberg of the assumptions that Audyssey is making when it "calibrates" in-room response. 

 

Other assumptions that they are clearly making include: everyone uses direct-radiating midranges in their fronts, center, and surround loudspeakers that have issues with frequency-dependent polar control as you go lower and lower in frequency toward the subwoofer frequencies, that these all the loudspeakers in your 5.1, 7.1, etc. array all have issues around 1.5 kHz ± 500 Hz that require a "dip in response" around 1.8 kHz to mask the polar coverage mismatch that typically occurs due to the direct-radiating woofer handing over to the direct radiating midrange/tweeter (exactly where the perceived stereo imaging will dip anyway due to head-related transfer function...HRTF).  The list goes on and on. 

 

The problem is with these kind of assumptions is that oftentimes they're not valid.  But Audyssey (the corporation) is completely silent on these subjects because they fear their own customers finding out what they're actually doing-and that someone will figure out how to do it better.  And they have built in their assumptions into their algorithms and you, the user, cannot turn them off.

 

REW and a calibration microphone works a whole lot better--and you can actually see what you're getting at the same time.  It actually takes significantly less time to do it manually for all loudspeakers in the array using REW and an active digital crossover to balance the system...than to run Audyssey 16 times (like George Jetson pushing buttons for a living--and not thinking about what he's actually doing).

 

The results of doing it manually using REW and an active digital crossover are better than any surround loudspeaker array that I've ever heard.  Audyssey MultEQ® XT completely butchers the output at the end of its (very long) completion cycle.

 

Chris

[Coytee]

Chris, I'm presuming the results of a REW analysis would be entered into the active crossover?

 

To anyone:  Regarding Audyssey (or any other product that does similar), is the data saved inside the unit once it makes its mind up?

 

What I'm really wondering is....  (backstory)  I have some outlets that are switched so when I walk into/out of room, I can kill certain items via some wall switches.

 

If a processor with Audyssey/other, is plugged into a switched outlet and loses all its power, does it save the prior info or would you have to run it again?

[Chris A]

22 minutes ago, Coytee said:

Chris, I'm presuming the results of a REW analysis would be entered into the active crossover?

Yes.  In the case of Xilica, you can easily copy the results into XConsole application running on the same computer to update the Xilica XP crossover.  You have a visual representation of the aggregate EQ curve that you're using for each amplifier channel, and many saveable memory configurations to hold various combinations and permutations that you can name and describe for future reference/use via its nearly instantaneous recall function.

 

My version of Audyssey does not do that: if you lose power while it's running its "calibration", you get to start all over again.  If you want to change something in the setup, you get to start all over again "recalibrating", losing what you had beforehand.

 

However, once you run Audyssey, it saves its settings into non-volatile memory (just like Xilica does, and REW does when you choose to close a measurement or the application).  "Audyssey EQ" internal settings are not visible to the user, however.

 

Chris

[RoboKlipsch]

The unifying theory here is simple, at least in my opinion.

 

Audyssey is correcting an "area", defined by the measurements.  If it's one measurement, or it's a group of closely spaced measurements, you will find that the result is pleasing in that one location or in that small space.  This is because the "groupings" of problems that Audyssey is looking for are close together -- there isn't a huge variation.

This is why in some small rooms and certain circumstances, it is suggested even by Kris himself to closely space the measurements.

 

But ideally, Erskine's theory of "no bad seats" should be in play, at least for a HT application.

 

In doing so, Kris suggests the layout for a typical couch, attempting to illustrate a "bubble" space that he is correcting.  However large the bubble is, defines the space corrected.  As such, the room has huge influence on this, as a small room with a lot of widely spaced measurements will have a wildly varying response.

 

Audyssey will correct what it is given as best it can by grouping the problems and attempting to correct them.

Therefore -- REGARDLESS of distance, the more closely your measurements match each other BEFORE running Audyssey will determine how well Audyssey corrects the space for "no bad seats."

 

You simply have to get the room layout and sub positions/delays such that the response is as similar as possible within the measured area.

If you have wild variations going in (garbage in), you will have wild variations, somewhat corrected, going out (garbage out).

If you have some significant variations but not all over the spectrum at each position, you will find it does a better job.

And if you can get a set of measurements that have little variation going in, you will have a beautiful response coming back out.

 

There is no hard and fast set of rules here.  It is simply a matter of measuring a space within the room and attempting to correct it as best as possible.

Assuming one has measuring equipment, the easy way to deal with this is to work on sub positions in order to get the least variation possible in the space.

Correcting every seat is not always the solution -- it may be a smaller area within the listening bubble that is actually corrected.

 

One must decide if one great seat (2channel), or no bad seats (multichannel), is the goal.
if the goal is no bad seats, then spend all your time finding the least variation in your listening area by sub positions and delays.

 

Gary and Chris, you both know way more than you ever could possibly need to use this software (or not use it) to it's optimal response.

There is no doubt it does not work well in every room, for every application, especially to a highly trained ear that spends a lot of time listening to music.

 

 

[garyrc]

On 12/17/2016 at 2:59 AM, Chris A said:

Gary,

 

We're not talking about the same thing.  The Audyssey MultEQ® XT that's distributed with the preamplifier does not have any choices except Audyssey EQ "on" or "off".  To do what you're talking about requires the insertion of an extra $500 to Audyssey to buy their after-market "MultEQ" firmware (not hardware)...which to me is outrageous.  I really don't recommend that path.

Chris,

 

You know so much more than I do about audio and physical science in general that I'm hesitant to disagree with you ....but ... my Marantz AV7005 preamp/processsor of c..2011 contains a version of Audyssey MultiEQ XT that gives me a choice between "Audyssey" and "Audyssey FLAT, with no extra charge and no special firmware.  "Audyssey" imposes the roll-off and the midrange dip you speak of.  "Audyssey FLAT" provides flat response from the F3 of the speaker on up to its upper limit, providing that the correction needed is no more than +9 dB, or - 12 dB (IIRC), over "hundreds" of points of corretion.  It also gives me a choice of some other options, such as "Audyssey Byp L/R" which applies Audyssey correction to the center and the surrounds, but leaves the L/R alone, as well as "Manual," and "Off," which don't use Audyssey at all.  Since Audyssey v.s. Audyssey FLAT is discussed so often over in the AVS "Official Audyssey Thread" part I and part II, I can't help but think these options are included in many pre/pros and AVRs.

 

Could you have the "pro" add-on in mind?  It does cost extra, and provides a few other target curves.

 

Audyssey FLAT with about 6 dB bass boost superimposed sounds better in my set-up than any other configuration I've had, including EQ fever with a McIntosh C28, and, especially, a Luxman L580 with a wide choice of turnovers, separate bass boost, etc.

 

When you use REW, do you use more than one mic position?

 

Thanks

 

 

[Chris A]

9 hours ago, garyrc said:

When you use REW, do you use more than one mic position?

This is an interesting subject.  What I've found is that for frequencies above about one octave above the Schroeder frequency in-room (i.e., starting about 200 Hz in my room), one on-axis measurement position at one metre from in front of each loudspeaker works much better.  One measurement at the listening position to check overall responses serves as a check and opportunity to recognize an issue that you've missed using at least six one-metre-on-axis loudspeaker measurements (assuming a 5.1 surround sound configuration).

 

Below the Schroeder frequency in-room, there are apparently a lot of misperceptions about what can actually be done with EQ.  I've found the I have to pay close attention to excess group delay vs. group delay curves.  That will tell me where room modes are present (i.e. with microphone at one metre in front of loudspeaker).  You really cannot do any boost at those frequencies using EQ, but you can do these three things:

 

1) You can attenuate response peaks using EQ (but not their decay response),

2) You can tilt up or down the average bass frequency response between the peaks and nulls using EQ filters with low Q/high-bandwidth PEQ or shelf filters, and

3) You can see the effects on smoothing bass room mode response by having distributed low frequency drivers located at the room's anti-nodes (mid-wall, etc.) which fill up the response modes that cannot be filled using only corner-located subwoofers. 

 

The lower that you go in frequencies, the farther apart the room modes are located.  The Schroeder frequency is merely a rule of thumb to divide the diffuse mode region from the sparse mode region of the room.  It doesn't really demark a hard line between these two regions.  What this means is that you you can usually get away with more boosting PEQs at the lowest frequencies (below 30-40 Hz) than at frequencies from about 50 Hz (i.e., about half the Schroeder frequency) up to the Schroeder frequency, but these must be in non-room-mode frequency bands.

 

The sum total of the above is to say that the idea of moving the microphone around a lot doesn't help much if you know something about the room...more than Audyssey knows.  Audyssey is trying to find the measured boundaries of a dark room with internal acoustic reflectors in it in the form of furniture, equipment, and loudspeakers using lots of microphone positions to guess where they are.  You have eyes and probably a tape measure, and can do simple conversions from time delays in milliseconds: speed of sound converts into about 1 millisecond/foot and back again. 

 

The notion that you can "optimize" the low frequency settings for an area of the room is actually a misnomer.  Audyssey (the corporation) is trying to cover its tracks by using that concept like commercial psyops.  Using the notion of multiple microphone positions as a cover story, what it is really doing is trying to find out information that you--a human being--already have unambiguously, and that you can use in your calculations to determine which measured responses are room modes, which are loudspeaker direct arrivals, which are precursor pulses from tapped-horn subwoofers, and which are reflections off of internal room reflectors (i.e., not room boundaries) like coffee table tops, leather chairs, and nearby bookcases.  Audyssey is actually pretty stupid in that respect: it doesn't really know about those things, but rather it is guessing which responses are which from a multiplicity of measurements, and the more the better.  It's trying to sort that stuff out using lots of measurements, stuff that you can see by inspection...almost at one glance.

 

You really need only one measurement or perhaps two, and your eyes+a tape measure using simple calculations. I find that humans do a much better job a lot more efficiently.  However, if you can't figure out which responses are reflections from room modes and which are diffuse responses, you're going to be chasing your tail a lot trying to boost or attenuate those frequencies that can't be changed using EQ.  Audyssey wants you to believe that it can...but it can't cheat physics.  The only thing that it can really do is to try to figure out if it is EQing based on a "knife-edge" measurement conditions and avoiding those knife edges by using lots of measurements. 

 

That's the price you pay for "pushing a button and standing back" vs. using innate human intellect, senses, and a tape measure.

 

Chris

[DizRotus]

 Chris-

Everything you say above is fascinating, some of which, I think I understand. But, it's way more effort than most of the audio zealots here, or anywhere, will ever undertake.  

 

You're right, of course, about Audyssey trying to simplify the process to make it accessible to the lazy, but even Audyssey intimidates and confuses audio hobbyists.  I'd wager most people who have Audyssey enabled equipment don't use Audyssey at all, or they make an attempt, but fail to fully implement the "simplified" Audyssey program.

 

Please don't stop sharing your insights regarding room equalization and recording remastering, someday I might have the time and energy to follow your lead.  But, for now, I need to relax and listen to non-remastered music from my minimally Audyssey'd 5.1 system, including four tapped horn subs, that has been finally adjusted to taste by ear.

[Chris A]

2 hours ago, DizRotus said:

You're right, of course, about Audyssey trying to simplify the process to make it accessible to the lazy, but even Audyssey intimidates and confuses audio hobbyists.  I'd wager most people who have Audyssey enabled equipment don't use Audyssey at all, or they make an attempt, but fail to fully implement the "simplified" Audyssey program.

Neil,

 

I assume that if someone has enough knowledge and experience to prefer Klipsch to other non-horn-loaded brands, modify or even make their own passive crossovers, play with different amplifier types and taps (especially tube amplifiers) or even make them, play with different preamps, DACs, turntables, loudspeaker positioning in-room, acoustic treatments (diffusion and absorption), and buy different versions of music albums (some of which they already have owned for decades), and read "talking head" audiophile reviewers writing the most obscure and obfuscating reviews imaginable, and more-or-less blindly defend a company for which they own no stock and do not know any of the people or their real intentions (such as Audyssey), they can probably handle easily what I've said above.  Especially if they read and reflect on a book or two on the subject of audio (Toole's book would be one of them).

 

Not all audio enthusiasts are at the same place in their journey of knowledge and experience, and I do understand that there are some folks that are challenged by setting up a little microphone and punching a button a few times, but by and large, these aren't  typical "audiophiles" that spend time and really enormous monetary resources learning their hobby over the years.  I merely present information which I have found first-hand that is closer to physical reality than most of the audiophile equipment magazines that people pay real money for. 

 

I also do it for free.   (Caveat lector.)

 

Chris

[ClaudeJ1]

On 12/14/2016 at 4:23 PM, Khornukopia said:

My Audyssey microphone is attached to a hinged bracket on the end of a boom that clips onto the quick release of a camera tripod. This allows me to "float" the microphone to the exact position desired. If I plan to make any room or equipment changes in the near future that will require re-calibration, I will just test three positions and skip to calculate.

 

After having used all different flavors of Ausyssey up the XT/32, I concur with this simple advice. First one at your nose position, then swing left and right no more than 4-8 inches. The tighter the pattern the better the imaging. I have done this hundreds of times over the years and it's the best way. No more than 3!! No need for a tripod, you can just tape the mic. to a stick with a counterweight, if needed, and just balance it over your couch or chair. Easy.

[ClaudeJ1]

50 minutes ago, Chris A said:

Not all audio enthusiasts are at the same place in their journey of knowledge and experience, and I do understand that there are some folks that are challenged by setting up a little microphone and punching a button a few times, but by and large, these aren't  typical "audiophiles" that spend time and really enormous monetary resources learning their hobby over the years.  I merely present information which I have found first-hand that is closer to physical reality than most of the audiophile equipment magazines that people pay real money for. 

There are no experts in any field, only varying levels of ignorance................one of my fave expressions, but I digress...............................

Many of us have done all the things you described in your first paragraph, including audio prose, which I no longer read. I only read technical papers and many of the things your write, Chris.

 

But I must confess, after doing all of the possible things in the speakers, room, and choice of test methods, I admit to being a big fan of Audyssey and other similar room correction "Technology for Dummies" approaches. Being that they work in both the time and frequency domains, the fewer problems we give them to solve, the better job they will do. The simpler, the better.

 

As an example, room correction cannot make up for bad subwoofer placement just because the wife/girlfriend says it won't fit the decor or "look good there." So if you start with fixing that first, then have good transducers at the correct speaker distance to the sweet spot, toe-in angle,  to give good direct/reverberant ratio, etc. then Audyssey can do an amazing job.

 

Even the biggest and baddest heavy duty machinery in the mechanical world can choke on a piece of work that it's not designed to handle.

 

The same is true in Audio. Fix the acoustic problems acoustically as best as you can, then add the Audyssey, or other, "global feedback" methods as "icing on the cake"..........but you gotta bake a good cake FIRST.

 

 

[Chris A]

Note that all my issues have been with "Audyssey EQ", since my paid-for version of Audyssey that came bundled with my pre-amp/processor has been hobbled to do nothing else--"on" or "off" only.  This was, in my opinion, a terrible decision by Audyssey that they knew what they were doing at the time (i.e., it would have been far easier for them to include the functionality to run "flat" with perhaps a few lines of code, costing them nothing in recurring expenses).  This is the problem, not something that they have later done. 

 

I learned to judge a company by its least performance in the marketplace (...but not necessarily people by the same standard).  I've learned to do this over time since there are a lot of companies that are not really committed in business...as Deming would put it..."to fully serve their customers".  The tendency to do the wrong thing comes from the top of companies and their investors, and I've found that it doesn't ever really disappear once its been encased in the corporate culture.  I don't ever trust them again for doing something that they knew was deliberately a "bait and switch" or otherwise knowingly doing the wrong thing.  This occurred in my specific buying experience. 

 

Since the time of my purchase Onkyo has cleared its name in the case of infancy failures of HDMI boards (a third party piece of hardware), fixing them for free, although my hardware and non-Audyssey firmware has never had any issues.  I've not seen such a generous offer of free upgrades from Audyssey to fix their "bait and switch" scheme from earlier in their corporate life.  The price is $500 to get "Audyssey flat", something that should have been the default.

 

YMMV.

 

Chris

[derrickdj1]

I use MCACC but, some of the same principles apply.  I run a set of measurements at the MLP.  This may be done several times to get the sound at the MLP dialed in.  Once this is done, I changed the speakers to small and run a more detailed 3 position measurement around the MLP.  All measurements are from a rotating tripod so, one measurement in the middle and the other two measurements 1 to 1 1/2 ft to the left and right of the middle.

 

Since I don't have a dedicated HT or chairs in rows, no attempt is made to make every seat closely matched.  There are some other things that I do before and after this, which is in the Chicago GTG in Indiana under the general info. section.  The biggest problem in the small listening room is getting the bass correct.  Most of the calibration programs do fairly well with frequencies above the 200 - 300 Hz area.

 

Needless to say, other things such as, room treatments, speaker positioning and the use of a distributed bass system with multiple subwoofers is also needed.  There is also the point that each room can be made only so good from an acoustic standpoint.  This is related to the room, furnishings and the owner ability or willingness to modify the room.   I am not in the camp of throwing up room Tx's all over the place and favor using the room furnishings.

 

The guidelines for the various calibration programs are suggested as a starting place.  It's been a long time and  hundreds of measurements since I have calibrated the room with the mic at the MLP.  I usually measure 2 1/2 ft. in front of the MLP.  The key is to become really familiar with your calibration system and to understand why you are take measurements at certain places in the room, what can be influenced at that position and what things can be done post calibration to finish dialing in the system.  It takes a long time to know the room and the limits of a particular calibration programs.  This learning occurs by listening and measurement.  As alway, know your system.  Easier said than done in many cases.

[RoboKlipsch]

#Chris A 

 

I'm surprised by some of what you wrote.  Here's the chart from Audyssey:

 

 

Obviously the big difference comes in with the satellites, where once you go past EQ you get more filters.

They all share the same sub EQ except at the XT32 level where there are 4x the filters.

 

My concern with what you wrote, is how would the average person improve the response of their room without Audyssey?

On the sub, they can move it to the right position, they can adjust the delays, but how then would the average person be able to EQ it?