Some help needed understanding MultiEQ measurements

[ABall]

Hi folks, I have some measurements that I would like some help with. I have been reading so much my head is spinning so I am reaching out.
I've just installed a Denon x3700 and these are the results of some measurements. The green scale is Pure Direct and the red one is the Stereo mode on the amp. I can see where Audyssey has made corrections but there's also parts where it looks worse, 6.5k for example with loads more spikes lower than PD, is this normal? Also, I limited the range to 500Hz so why does it look like its still been EQ,d above that Freq?
One more thing, my default curves are different for each speaker set, am I right in thinking that Audyssey has judged my surrounds can be corrected to 0dB at 20Hz? I know its not true but I am thinking its telling me my surrounds are way more capable than my fronts? The last picture is the curve I uploaded to the amp and should be what the REW results are based on?

Sorry if these questions are obvious, I just haven't grasped this measuring stuff out, its only taken 30 odd years to actually start measuring my room and I'm weeping at the money I have wasted trying to make audio sound good.

I have no Sub bass.

Many thanks

 

[LTig]

You need to apply smoothing, like 1/24th or 1/48th per octave to get rid of the dips in the higher frequency range.

 

[ABall]

You need to apply smoothing, like 1/24th or 1/48th per octave to get rid of the dips in the higher frequency range.

Thanks Ive added the smoothed response.

 

[LTig]

Thanks Ive added the smoothed response.

If you shift the red curve around 1.5 dB up the mids and highs would be a very close match.

 

[-Matt-]

I have no Sub bass.

Do you mean no subwoofer?

One more thing, my default curves are different for each speaker set, am I right in thinking that Audyssey has judged my surrounds can be corrected to 0dB at 20Hz? I know its not true but I am thinking its telling me my surrounds are way more capable than my fronts?

If no subwoofer, perhaps Audyssey defaults to a full range target for your surrounds. Are they auto detected (or manually set) to Large on the "Speaker Detection Results" page? If small, what are the crossover values?

Rather than posting the target curve screenshots, it may be more informative to share the before & after "Room Correction Results" pages.

Also, I limited the range to 500Hz so why does it look like its still been EQ,d above that Freq?

If you shift the red curve around 1.5 dB up the mids and highs would be a very close match.

Yes, with the smoothing it looks as though Audyssey is doing as its told and sticking to below 500Hz.

The main thing Audyssey is doing is filling in some gaps between 75Hz and 275Hz. I wouldn't say it looks worse anywhere with the eq. (It knocks a peak down at 45Hz, but that probably helps overall smoothness of bass).

 

[ABall]

Do you mean no subwoofer?


If no subwoofer, perhaps Audyssey defaults to a full range target for your surrounds. Are they auto detected (or manually set) to Large on the "Speaker Detection Results" page? If small, what are the crossover values?

Rather than posting the target curve screenshots, it may be more informative to share the before & after "Room Correction Results" pages.



Yes, with the smoothing it looks as though Audyssey is doing as its told and sticking to below 500Hz.

The main thing Audyssey is doing is filling in some gaps between 75Hz and 275Hz. I wouldn't say it looks worse anywhere with the eq. (It knocks a peak down at 45Hz, but that probably helps overall smoothness of bass).

Sorry yes no Subwoofer.
You are right the speaker results show that everything is been cut off at 500Hz, the default curves for each set must be a red herring. My speaker settings are all as set by Audyssey so no manual changes. Looking a bit closer at the results it does seem my right channels have been set at a slightly higher level, I am a bit closer to them but I thought EQ would level match them for the measuring position? Also looking at the results my rear speakers look to be much better than my fronts, would this be right? I have Club 27 Kurts up front, Tannoy XT6F's surrounds and IPL S2TLK CD3 as surround backs, centre is a Monolith THX-365C. Here are the results.

 

[-Matt-]

Let's look at the FL and FR first...

The dip downwards at 100Hz (R) - 120Hz (L) is likely caused by SBIR. Does it make sense, are the fronts of these speakers roughly something like 86cm (R) and 72cm (L) from either the front or side walls? If so acoustic absorption panels may help there. Do you have any room treatment currently?

Audyssey tries to fill in this dip but it doesn't want to/can't boost by more than 10ish dB. This is why the dip isn't completely removed (especially on the right, where it is deeper).

This asymmetry could be because the wall behind the speakers (or the side wall) offers a stronger reflection on the right side?

Audyssey may also be setting the level higher on the right to give an extra bit of boost in this dip (it should then attenuate everywhere else so that the measured levels at the mic position still match).

The peak at 40Hz is due to the custom curve that you made for the fronts, combined with your speakers natural cutoff below this frequency.

It might be worth experimenting with running the correction up to higher frequencies (1kHz, 2kHz or even full range). This will probably tidy up some of the peaks at about 700Hz.

Note: I'm just learning this stuff myself, so hopefully more expert members can verify this analysis.

 

[-Matt-]

Are SR and SL close to the corners of the room?

They seem to be getting quite a boost in the low frequencies which is possibly why Audyssey thinks they can go lower than your mains.


One thing you can try is adjusting the upper frequency limit for Audyssey correction. (You don't need to re-measure for this; just make a copy of the setup and edit that). On the after plots, some of your surrounds have quite a large offset from 0 in the flat region (eq'd region). This can happen when there is a peak in the frequency response that coincides with the Audyssey high frequency limit. Increasing the limit a bit (few hundred Hz) might allow the flat region to fall nearer 0. (Pretty sure the level adjustment is separate to this, and the graphs are normalised).

 

[ABall]

Let's look at the FL and FR first...

The dip downwards at 100Hz (R) - 120Hz (L) is likely caused by SBIR. Does it make sense, are the fronts of these speakers roughly something like 86cm (R) and 72cm (L) from either the front or side walls? If so acoustic absorption panels may help there. Do you have any room treatment currently?

Audyssey tries to fill in this dip but it doesn't want to/can't boost by more than 10ish dB. This is why the dip isn't completely removed (especially on the right, where it is deeper).

This asymmetry could be because the wall behind the speakers (or the side wall) offers a stronger reflection on the right side?

Audyssey may also be setting the level higher on the right to give an extra bit of boost in this dip (it should then attenuate everywhere else so that the measured levels at the mic position still match).

The peak at 40Hz is due to the custom curve that you made for the fronts, combined with your speakers natural cutoff below this frequency.

It might be worth experimenting with running the correction up to higher frequencies (1kHz, 2kHz or even full range). This will probably tidy up some of the peaks at about 700Hz.

Note: I'm just learning this stuff myself, so hopefully more expert members can verify this analysis.

Are SR and SL close to the corners of the room?

They seem to be getting quite a boost in the low frequencies which is possibly why Audyssey thinks they can go lower than your mains.


One thing you can try is adjusting the upper frequency limit for Audyssey correction. (You don't need to re-measure for this; just make a copy of the setup and edit that). On the after plots, some of your surrounds have quite a large offset from 0 in the flat region (eq'd region). This can happen when there is a peak in the frequency response that coincides with the Audyssey high frequency limit. Increasing the limit a bit (few hundred Hz) might allow the flat region to fall nearer 0. (Pretty sure the level adjustment is separate to this, and the graphs are normalised).

Thanks very much Matt, If your just learning this stuff I have a VERY long way to go! I just had to look up SBIR, I see what you mean and this is going to be a bit of a journey. My fronts are 63cm from the front wall to baffle, they are fairly deep at 36cm so I dont have a lot of wiggle room left in my lounge. The FR has 47cm to the side wall and a door opens onto my FL leaving less than a 20cm gap (a wide media unit fills the gap in between) although a fireplace juts out about half a meter in front of the FR which would put that boundary in about the same place as the door on the left. To be honest its the bass I would like to improve, its what has caused me to buy speaker after speaker, I stupidly thought that was the right thing to do. I am surprised and disappointed the Kurts roll off so high in the bass but I can't pull them out any further, I was told they could go a lot lower but at what minus dB wasn't mentioned,
The SR and SL are farther away from the side walls than the fronts but they are closer to the wall behind, the SB's are a bit enclosed in a bay window and only about 6" from the sofa which is fairly high, they were gathering dust for a long time due to my speaker search so I figured I would put them to use.

I have no room treatment, I think I have a very understanding wife but I need to draw the line somewhere. Do you know if the panels would need to be large, deep and wider than the speakers for instance?
I will run the correction up as suggested in your first response, is this what you also mean about increasing the HF limit for the surrounds in your second post?

Thanks again.

 

[ABall]

Having done a little reading 4 inch absorbers behind the speakers is a good idea but a member here stated in another thread that this would only work on bass above 200hz, Does anyone have any experience with this in regards to dropout around 120hz?

 

[-Matt-]

I will run the correction up as suggested in your first response, is this what you also mean about increasing the HF limit for the surrounds in your second post?

Yes, that's what I meant.

It sounds as though our rooms have a pretty similar set of problems!

I have no room treatment, I think I have a very understanding wife but I need to draw the line somewhere. Do you know if the panels would need to be large, deep and wider than the speakers for instance?

Having done a little reading 4 inch absorbers behind the speakers is a good idea but a member here stated in another thread that this would only work on bass above 200hz, Does anyone have any experience with this in regards to dropout around 120hz?

If my wife hadn't vetoed it I'd be looking at getting the monster bass traps from Gik. (They seem to have useful absorption down to 80Hz). Thicker is definitely better for the low frequencies. (The corner tri traps might suit your room even better).

The other thing that you could try is putting the speakers right back against the front wall (as close as you can, either with or without the panels behind). In theory this should push the SBIR dips up to higher frequencies, where the absorpion panel, if present, can be more effective. Hopefully it would also reduce the magnitude of the dips to something that is more manageable for Audyssey. (Positioning FL and FR nearer the front wall may also have the advantage of boosting the bass from the mains a bit too, especially without the acoustic panels).

If you go down this route please share measurements so that I can enjoy the room correction vicariously!

 

[ABall]

My plan is to push the speakers up against the wall first then listen to music via a stereo amp to see how it affects imaging, if that doesnt work I am going to put my IPL's back up front, they definitely have more bass low down and I think Audyssey will be able to sort the boom/muddiness out easier. I expect the issue is just that I have purchased another pair of speakers in the Kurts that dont work in my room very well. I will indeed post measured results along the way, I have pushed up the surrounds to 5k and the fronts to 10k but I dont think I will get any measuring in until Sunday.

 

[ABall]

Just a quick update, following on from the results I had with Amused To Death i got the tape out and measured my distances, as I thought, FL/FR are exactly the same distance from me. The app had the left at 2.78 and the right at 2.83, i made them both 2.86 which is what they are and upped one side half a dB so they were both the same level and bingo!, everything is where it should be. So dont trust the software to get it right, I know it's a tiny offset but what a difference it makes to Q Sound, if it's not doing what its intended to do then the speakers aren't set up right which Is why i have always used a tape measure and find that it helps with all music. Much happier now, I thought it was going to take me ages instead of 5 mins.

 

[Jon AA]

Audyssey uses the first mic position to calculate distances and levels so it's really important to have the mic positioned exactly where your head will be for that measurement. It usually does quite well, but sometimes a reflection or something of the sort can mess things up. One of the many nice things about MultEQ-X is you can re-do single measurements, even for just single speakers if you think it got it wrong the first time without having to re-do the entire calibration. And you can see the results instantly so if you see something that doesn't look right you can re-do them.

Also, I limited the range to 500Hz so why does it look like its still been EQ,d above that Freq?

The limit is not a brick wall. If you set it to limit correction to 500 Hz, it will gradually reduce the filter magnitude over the next octave. So, the amount of correction is gradually reduced until it becomes zero at 1000 Hz.

 

[ABall]

Audyssey uses the first mic position to calculate distances and levels so it's really important to have the mic positioned exactly where your head will be for that measurement. It usually does quite well, but sometimes a reflection or something of the sort can mess things up. One of the many nice things about MultEQ-X is you can re-do single measurements, even for just single speakers if you think it got it wrong the first time without having to re-do the entire calibration. And you can see the results instantly so if you see something that doesn't look right you can re-do them.


The limit is not a brick wall. If you set it to limit correction to 500 Hz, it will gradually reduce the filter magnitude over the next octave. So, the amount of correction is gradually reduced until it becomes zero at 1000 Hz.

Thanks! I didnt know you could do a single measurement again, I thought you had to do the whole lot. Could you please tell me how that is done?
I did pay very particular attention to the position of the MIC, I have always measured the sweet spot with a tape measure, I guess MultiEQ just got it wrong this time, I always check my speaker placement with Roger Waters so no big deal.

 

[Jon AA]

Yes, in MultEQ-X you simply go to the measurements section, right click on the particular measurement you want to re-take and select "measure." You can only do this with the new PC based MultEQ-X--you can't do it with the App. One of the perks of paying the $200 fee--but in my opinion it's a pretty big one. You can replace your sub or one of your speakers or move one of your speakers or move your main listening position slightly...and quickly integrate those changes without having to re-measure all 12 speakers all over again.... Just in recent days, I've added a sub, changed my center channel, removed MiniDSP HD's with "pre-EQ" from the LCR...all without having to remeasure the surrounds and height speakers. It saves a lot of time. And you can see the results of each measurement in real time, so if there's something phunky going on with a particular mic placement you can identify it immediately.

Something else it'll do to get you more accurate results on this subject--I don't remember the details but there has been some sort of calculation error in the chain between Audyssey and the AVR (always been there) which causes the distances to be slightly incorrect. MultEQ-X uses a correction factor to eliminate this error (you'll see the distances in your file don't match the distances on the AVR).

IMHO a tape measure is about the least accurate way to measure distances. Of course I use horn speakers (how far do you stick the tape measure into the horn?). When you think Audyssey got it wrong, I'd verify that with the impulse/time alignment measurements from REW and use that as a basis for fine adjustments.

One final thing MultEQ-X will do in this regard which is somewhat icing on the cake: Some AVRs only allows somewhat course trim level adjustments. There's an option in MultEQ-X that will lift/lower the entire target curves for each speaker a fraction of a dB to make up for the rounding error in the AVR.

 

[ABall]

Yes, in MultEQ-X you simply go to the measurements section, right click on the particular measurement you want to re-take and select "measure." You can only do this with the new PC based MultEQ-X--you can't do it with the App. One of the perks of paying the $200 fee--but in my opinion it's a pretty big one. You can replace your sub or one of your speakers or move one of your speakers or move your main listening position slightly...and quickly integrate those changes without having to re-measure all 12 speakers all over again.... Just in recent days, I've added a sub, changed my center channel, removed MiniDSP HD's with "pre-EQ" from the LCR...all without having to remeasure the surrounds and height speakers. It saves a lot of time. And you can see the results of each measurement in real time, so if there's something phunky going on with a particular mic placement you can identify it immediately.

Something else it'll do to get you more accurate results on this subject--I don't remember the details but there has been some sort of calculation error in the chain between Audyssey and the AVR (always been there) which causes the distances to be slightly incorrect. MultEQ-X uses a correction factor to eliminate this error (you'll see the distances in your file don't match the distances on the AVR).

IMHO a tape measure is about the least accurate way to measure distances. Of course I use horn speakers (how far do you stick the tape measure into the horn?). When you think Audyssey got it wrong, I'd verify that with the impulse/time alignment measurements from REW and use that as a basis for fine adjustments.

One final thing MultEQ-X will do in this regard which is somewhat icing on the cake: Some AVRs only allows somewhat course trim level adjustments. There's an option in MultEQ-X that will lift/lower the entire target curves for each speaker a fraction of a dB to make up for the rounding error in the AVR.

Thank you for all that info! I am not quite ready to punt up the cost of a PC based suit though, I may do when I understand everything a bit more, the PC software does allow for the calibrated version of Denons MIC to be used though which is nice.
I have to totally disagree with your opinion on the tape measure though, tape measure, string, whatever, its more accurate than the software, I already proved this using Amused To Death. You dont need to stick the tape down your horn, my fronts are horn loaded, all you have to do is take a ref point on one speaker, I use the centre of the top, it doesnt even need to be at the driver, then measure from the listening position to the same point on the other speaker, as long as the ref point is the same on both you will be at an equal distance from any part of the speaker providing they didnt build your speakers differently!
I am a bit peeved that Denon ask for more money to get better results but thats their business model and I was aware of this before I made my purchase, its still way cheaper than the alternative DIRAC AVR'S.

 

[ABall]

some more measurements, all using no EQ.
RED: original at 80dB
Blue: speakers pushed back, transformed the null at 120Hz but a little worse around 140 and 240Hz 80dB
Green: I added 2 cushions, single depth but full height almost. Definitely absorbing some peaks. 80dB
Black: I decided to do a run at 75dB leaving the cushions in, have to say I wasn't expecting the lower level to produce a worse result but maybe that's just the way it goes?

 

[-Matt-]

Thanks for sharing the measurements.

I'm glad that the suggestions seemed to have had a positive effect on that dip. Nice to see the theory actually working in practice!

I'm surprised that the cushions make that much difference. I'd have thought proper panels with larger area will be able to help a lot more with those dips at 140 and 240Hz.

 

[ABall]

Yes me too! I dont know what's it's done to my soundstage yet though.
I am hoping to grab a couple of traps tomorrow, we are doing a home cinema and GK specced a few panels for it, totally forgot to see if they were still in the office.