Does Audyssey cause weird dips in the response

[jwc]

Ok.  So this is another thread on Audyssey.  I must have something set wrong as I'm finding Audyssey to make a "suck-out" just above 100Hz.  I've noticed with another set of speakers too.  Not going to go into that.

 

So.  I have a curve here with my speaker in 2 channel with subs.  I have a new Marantz HT SR6011.  Sub XO at 80 Hz and mains as "small".  The fronts is a homemade concoction with a 3 way crossover with my minipunches and fastlane horns.  I would like to minimize discussion of these as much as possible please.  XO points 400/5000.

 

The curve is in "Direct Mode" which is with Audyssey off.  I have the amp for the subs turned off.  The speaker for fronts is on top of one of the subs laying down.  In the picture, I usually have another set of speakers next to my MAC amps......that is why there is a gap there.  That is my analog setup.  Waitin' to put something else there in future.

 

So the "Direct" or "Pure Mode" have the same curve as you see here and sounds decent.   There is no EQ in this system.  Just the marantz...amps...crossovers...and front speaker......but when I turn Audyssey on......there is a huge drop in the FR just beyond 100Hz.  Music sounds thin and tinny.....no surprise.  I haven't saved that curve....but is happens over and over.  Same mic placement.  I'm tired of measuring the room with the mic.  

 

With a movie or TV programming, I really could care less as I'm not picky with this stuff.  But when playing music (which is my kids stuff mostly on PC server)....I'm forced to leave Audyssey off.

 

jc

[RoboKlipsch]

Post a graph with and without audyssey.  That may be an sbir suckout where distance from the front wall will change that dip in frequency.

 

If its not there without audyssey then its not sbir but audyssey correcting what is probably a peak at some measuring positions.

 

If run with only one measuring position as a test it should fix that.  If u use several measuring positions there may be a peak there at some locations being fixed at the mlp.

[jwc]

Will it let me get by with one position....run one time or run several times at one position...

 

never thought of that.

 

With audyssey off, there is never a dip like that wherever i put the mic.

 

 

[Chris A]

13 minutes ago, jwc said:

With audyssey off, there is never a dip like that wherever i put the mic.

Isn't that interesting?  You don't need no stinkin' bass traps in the corners--Audyssey will take care of that for you (without asking...). 

 

I wonder what else Audyssey thinks it's doing to "help" your room acoustics? 

 

Chris

[jwc]

Well, the room is "treated"  It might be over treated.....but my "analog" stuff sounds great.

 

I'm trying to keep an open mind here...this will be a slow thread to develop cause it may be next week for me to show the anomaly I'm speaking of on graph.

 

I had friends over a few weeks back.....3 systems to play with and "party" basically.  Lots of Stereo time.  The system with HT for 2 channel listening...to be honest...was a little embarrassing.  

 

But I'm still willing to make it best it can be.  It isn't my mission to encourage others to thing Audyssey sucks.  It hasn't worked for me real good for music.

[Chris A]

Which version of Audyssey are you using?

 

Where are you putting the microphone when running TrueRTA? 

 

How high off the floor is your Audyssey microphone?

 

Chris

[jwc]

Which version of Audyssey are you using?

 

http://us.marantz.com/us/Products/Pages/ProductDetails.aspx?CatId=AVReceivers&SubCatId=0&ProductId=SR6011

 

Advanced Audyssey MultEQ XT32

 

Where are you putting the microphone when running TrueRTA? 

 

Listening position..about 15-17 ft away from front wall.

 

How high off the floor is your Audyssey microphone?

 

ear level when sitting on seat.

[Chris A]

Here is a table that I found on the different flavors of Audyssey:

 

 

You might have to order Audyssey Pro kit where they clip you for another $550 for the privilege of actually seeing and controlling what it's doing.  I hear good things from Gary Camp (Garyrc) and Rudy (Rudy81) using this, I believe.  You're looking for "Audyssey Flat" settings and turning off any other corrections that Audyssey tends to throw in for grins--especially Dynamic EQ (something which attempts to correct for Fletcher-Munson loudness effects and loudspeaker compression issues - without actually measuring at different loudness levels (which is very weird--to say the least).

 

The 190 Hz dip that you show also corresponds to a microphone 1/4 wavelength dip at ear sitting height, too.  Curious?  (Likely.)

 

The TrueRTA microphone at the listening position will usually screw up the phase data measured.  I recommend a microphone placement one metre in front of each loudspeaker--measuring one loudspeaker at a time (I know that you know that, but I mention that for anyone else that may be reading this) in order to see intelligible phase data.  When you're at the listening position, you're going to get room early reflection issues at lower frequencies.

 

I've gone to REW, a calibration microphone and a good active crossover (Xilica XP-8080).  I have to tell you, I've never heard anything come out of my MultiEQ XT Audyssey version that's anywhere close to doing it by hand -- not by a country mile. 

 

YMMV.

 

Chris

[jwc]

The TrueRTA microphone at the listening position will usually screw up the phase data measured.  I recommend a microphone placement one metre in front of each loudspeaker--measuring one loudspeaker at a time (I know that you know that, but I mention that for anyone else that may be reading this) in order to see intelligible phase data.  When your at the listening position, you're going to get room early reflection issues at lower frequencies.

 

I am aware of this.   It is intentional to measure at the listening position or close to it.  If I were to actually measure the speaker....I agree.  That isn't what I'm doing.

 

No way I'm spending more money on HT setup.  I mean...if a new receiver with the latest Audyseey can't get it right...then screw it.  I shouldn't have to buy additional software or more hardware.  I'll leave it off.  I thought maybe there might be some setting or ideas about the microphone audyssey setup.  I may try different mic positions.............. different than what they suggest through the software.

 

13 minutes ago, Chris A said:

I've never heard anything come out of my MultiEQ XT Audyssey version that's anywhere close to doing it by hand -- not by a country mile. 

 

Well, I would agree that it wouldn't be my audiophile dream to use it and rely on it....but to completely screw up the sound and to measure weird.....what a waste.  I used to manually put in distances and use a cheap SPL meter with an Old Onkyo HT receiver.  Worked pretty good.

 

I must be doing something stupid in the settings.

 

[RoboKlipsch]

First I would say that response is not really that bad!  The dip has been pushed up by Audyssey because it looks at the average of the measurements and groups them and tries to correct for major issues.  Somewhere around your MLP you have this dip, and it was measured multiple times.  When correction was done it seems to have moved up that area of the mid-bass in order to create an overall average that fits where it should be...so there is a dip, but it's not dipping 12db below the average line...which is what is almost for sure looks like without Audyssey.

 

So try 2 things.  First, with Audyssey off, measure the same places you use with Audyssey with your microphone and find out which measurements have this dip to begin with. That is a speaker placement/broadband absorber issue.  That will show you where the problem is.  You can even run each speaker individually to find out which one causes the problem, or if it's both.    The other thing you can do to see what is causing that dip is change your crossover up and up and up as high as you can and watch the response, and see if the sub(s) at those positions overcome the issues of the main speakers.  That again tells you it's placement.  As Chris says, you can test the speaker up close to be sure the speaker design you made does not have a native response dip.

 

But you mentioned also earlier that another set of speakers had this issue.  If placed in the same locations and you listen at the same place -- you have an interference issue without a doubt.  Whether it's the front wall (my guess) or the ear height (Chris's guess) you can find out with some testing.

 

That dip exists in your space, and so when you measure and don't see it, it leads me to believe that the other measuring positions have this dip...and the MLP does not.  So when it corrects all the positions based upon grouped averages, it decided to move this section up a bit in order to remove that dip.   The response really -- again -- is not terrible by any stretch.  But if you want it gone, realize it's not Audyssey, it's the room and setup and speakers.  Audyssey will happily make a smooth graph when it can.  

 

To measure just one spot, measure that one spot.  If Audyssey requires say 3 measurements minimum, then leave the mic in the exact same spot and simply repeat the measurements.  Then you can see if that response is flat.  If it is, realize that Audyssey is doing it's job....making a space more flat, not one position.'

 

[Chris A]

JC - my thoughts exactly.  But, the channel delays seem to be pretty good out of Audyssey--within a few inches of being physically correct in my experience (like 4-6 inches).  I'd use their settings for delays by copying them down on a piece of paper after it completes, then turn off everything and reset the delays to the Audyssey MultiEQ XT32 settings by hand.

 

Channel gains are a little variable from Audyssey in my experience.  I use a hand-held SPL meter and the pink noise generator that's in the AVP and the remote to quickly set the gains.  I can set the gains to about 10x better resolution using REW, but then I have to drag out the microphone and turn on REW to do that.  The hand-held thing works like a charm.

 

Chris

[Chris A]

Hey Robo, I'm not discounting anything that you're saying...it's just that I've been around this tree probably 100s of times before using every trick that I can think of.  Perhaps some of the things that you mention will work for JC. 

 

It's just that I've settled on another approach to correct for loudspeaker EQ and room mode EQ, and the results have been outstanding.  It also takes about the same amount of time as running Audyssey ad infinitum to achieve significantly poorer results by Audyssey in my room.  When you factor in that the cost of Audyssey Pro is about 70% of the cost of the Xilica XP-8080 loudspeaker processor that I bought on ebay (unused in the box brand spanking new) plus a Behringer ECM8000 calibration microphone, then you might see my point of view.

 

Chris

[jwc]

41 minutes ago, RoboKlipsch said:

First I would say that response is not really that bad!  The dip has been pushed up by Audyssey because it looks at the average of the measurements and groups them and tries to correct for major issues.  Somewhere around your MLP you have this dip, and it was measured multiple times.  When correction was done it seems to have moved up that area of the mid-bass in order to create an overall average that fits where it should be...so there is a dip, but it's not dipping 12db below the average line...which is what is almost for sure looks like without Audyssey.

 

So try 2 things.  First, with Audyssey off, measure the same places you use with Audyssey with your microphone and find out which measurements have this dip to begin with. That is a speaker placement/broadband absorber issue.  That will show you where the problem is.  You can even run each speaker individually to find out which one causes the problem, or if it's both.    The other thing you can do to see what is causing that dip is change your crossover up and up and up as high as you can and watch the response, and see if the sub(s) at those positions overcome the issues of the main speakers.  That again tells you it's placement.  As Chris says, you can test the speaker up close to be sure the speaker design you made does not have a native response dip.

 

But you mentioned also earlier that another set of speakers had this issue.  If placed in the same locations and you listen at the same place -- you have an interference issue without a doubt.  Whether it's the front wall (my guess) or the ear height (Chris's guess) you can find out with some testing.

 

That dip exists in your space, and so when you measure and don't see it, it leads me to believe that the other measuring positions have this dip...and the MLP does not.  So when it corrects all the positions based upon grouped averages, it decided to move this section up a bit in order to remove that dip.   The response really -- again -- is not terrible by any stretch.  But if you want it gone, realize it's not Audyssey, it's the room and setup and speakers.  Audyssey will happily make a smooth graph when it can.  

 

To measure just one spot, measure that one spot.  If Audyssey requires say 3 measurements minimum, then leave the mic in the exact same spot and simply repeat the measurements.  Then you can see if that response is flat.  If it is, realize that Audyssey is doing it's job....making a space more flat, not one position.'

 

I will try some of that.

 

Just to clarify, the curve I'm showing is the listening position with Audyssey Off.  I haven't showed the curve with Audyssey on.  A Belle used to sit in those speaker positions....and Audyssey being turned on would cause the massive suck out too.  Something inside the Marantz is doing this.  

 

So, I need to give a curve with Audyssey on.  Then....I guess I'll redo with the mic in one spot.  To show my situation further.

[Khornukopia]

Testing with RTA, I would run sweeps with the mic in various spots near the MLP, looking for a frequency response peak that Audyssey is compensating for by trimming down. If you find a peak at the MLP, try to find a spot along the centerline of the room with a flatter response, then start your Audyssey measurement there.

 

If using Audyssey only, I would try a different microphone position for the first measurement, maybe a couple feet behind or in front and above your MLP. Leave the mic in the same spot as you click thru the remaining sequence during this search to save time.

[jwc]

14 minutes ago, Khornukopia said:

 

Testing with RTA, I would run sweeps with the mic in various spots near the MLP, looking for a frequency response peak that Audyssey is compensating for by trimming down.

 

 

Ok.  I'm starting to follow yalls line of thinking.  You are right, there may be a "bad spot" that Audyssey is trying to fix.  It isn't the listening position for sure.

 

I have a little hope on this.....I'll give it a try.

 

[RoboKlipsch]

22 hours ago, Chris A said:

Hey Robo, I'm not discounting anything that you're saying...it's just that I've been around this tree probably 100s of times before using every trick that I can think of.  Perhaps some of the things that you mention will work for JC. 

 

It's just that I've settled on another approach to correct for loudspeaker EQ and room mode EQ, and the results have been outstanding.  It also takes about the same amount of time as running Audyssey ad infinitum to achieve significantly poorer results by Audyssey in my room.  When you factor in that the cost of Audyssey Pro is about 70% of the cost of the Xilica XP-8080 loudspeaker processor that I bought on ebay (unused in the box brand spanking new) plus a Behringer ECM8000 calibration microphone, then you might see my point of view.

 

Chris

I have no problem at all with this.  I think a better calibration could be done manually as you are doing.  At least they are giving some flexibility now without Pro with their new app.

 

I would like to know how you manually set your gains when you do use REW.  Do you use the EQ to tell you what the channel level is or something else?  What range do you use, etc.

[RoboKlipsch]

21 hours ago, jwc said:

 

Ok.  I'm starting to follow yalls line of thinking.  You are right, there may be a "bad spot" that Audyssey is trying to fix.  It isn't the listening position for sure.

 

I have a little hope on this.....I'll give it a try.

 

If you're saying that measurement with the big dip in it is before Audyssey, then you already know what the issue is I think.

Look at that dip and ask yourself how Audyssey could or should correct it.  it's tough, especially considering that whatever it does to that dip it will change the EQ everywhere.

 

So the real question becomes (imo) can you either remove the dip through placement, treatment, or can you make that dip similar across all or some of the measuring positions, that way Audyssey can correct them all and get rid of it.  See where I'm coming from?  

 

It's so easy to run Audyssey and look at the results but the real trick is to do a lot of up front placement testing...closer to wall, away from wall, toe them a bit more or less and see how it affects each position.  The TV itself can also cause a lot of interference in the 100-500 range with sound bouncing around behind it.  There's a lot of things that can be addressed.  The more you research and find these issues, the better your sound will be.  Even if you never use Audyssey the same research can go a long way.

 

In looking at my main theater for movies, I realized that even though I had 20+ bass traps/abosrbers, the TV itself (flatscreen, not AT) causes interference.  A couple panels between the fronts and the TV itself cured a lot of it.  That discovery of where and why there's a problem can be really satisfying.  I find problems by running one speaker at a time and moving a sound panel around the speaker to show me which side the problem is on, then I look further.  Some advanced users use REW an readouts at specific frequencies...I'm not there yet.  

RK

[jwc]

11 minutes ago, RoboKlipsch said:

If you're saying that measurement with the big dip in it is before Audyssey

No.  The response is "flat" without Audyssey.  The curve I posted is without Audyssey.  Turning on Audyssey is a big suck out.  I haven't posted that yet.

 

I'm gonna work on this some more next week and update the thread.  Someone above mentioned this...as a thought.  One of the mic positions for Audyssey setup may have a huge peak....and Audyssey is trying to compensate.  That makes some sense....or it could be....I have something wrong with the Marantz settings.

[Chris A]

2 hours ago, RoboKlipsch said:

I would like to know how you manually set your gains when you do use REW.  Do you use the EQ to tell you what the channel level is or something else?  What range do you use, etc.

 

First, I run a sweep with the microphone one metre in front of each loudspeaker under test (one at a time), with the microphone approximately centered top to bottom on the loudspeaker, pointed slightly up but directed at the loudspeaker (i.e., NOT 90 incidence angle).

 

Then I iteratively run REW upsweeps and dial in PEQs into the crossover for each channel based on the REW equalizer predictions (it usually takes only two sweeps per loudspeaker to get a 90% solution using REW's EQ prediction facility).  The objective is the get the frequency response as flat as possible between the high pass and low pass points of the loudspeaker, while using "psychoacoustic smoothing".  This is repeated for each of the loudspeakers in the array, including the two TH subs.  I use a 1.5 dB/octave rise in loudness below 100 Hz as my "nearfield LF correction curve".

 

The upsweeps are usually done at 100 dBC at one metre (or higher) when I'm getting everything dialed in.  After that initial set up, I find that upsweeps of 80-85 dB/1 m are okay and give consistent results.

 

Next, I run pink noise from the AVP using the remote and adjust each channel gain up or down to get the handheld SPL meter loudness to within 1 dB.  For the subs, I run an upsweep either on the left or right channel in parallel with the subs to see the resulting loudness curve below 40-50 Hz (I cross at 40 Hz), and adjust the sub amplifier gains accordingly.

 

If I changed the placement of any of the loudspeakers in-room, the listening position, or switched out the active crossover to something with a different insertion time delay, I run Audyssey with its microphone at the main two listening positions to determine the channel delays, write those down on a piece of paper, turn off Audyssey, then hand hack those delays back into the AVP.

 

Voila! Done.

 

Chris

[Chris A]

Note that the AVP is not needed for delays and channel gains--all those could be dialed into the active crossover input channels or added to the existing output channel delays.  It's merely convenience that I use the AVP to do those things.

 

The AVP does everything else like decoding input digital streams, perform channel DACs, volume control, bass management, and switching among input devices (players and computers).  The AVP is controlled using a remote (Logitech Harmony One).