Audyssey's Next Generation of Room Correction (MultEQ-X)

[Chromatischism]

Your subs will sound a bit round with that hump there, best to see if you can reduce that a little with placement and EQ so that you see a gradual rise all the way down to 20 Hz instead of a hump. I also see a problem in the bass region from 120-180 Hz in your right channel. Above that, your left channel is up to 4 dB quieter than your right channel, so I doubt you have a centered stereo image at this point.

 

[Acerun]

Your subs will sound a bit round with that hump there, best to see if you can reduce that a little with placement and EQ so that you see a gradual rise all the way down to 20 Hz instead of a hump. I also see a problem in the bass region from 120-180 Hz in your right channel. Above that, your left channel is up to 4 dB quieter than your right channel, so I doubt you have a centered stereo image at this point.

Thank you for the feedback. I'm battling the audio demons and doing the best that I can.

1. Yes, I have a sofa over on the right side near the speaker that is an issue...but nothing I can do there.
2. I will play with getting that hump down. It "may" have to do with the fact that I'm crossing at 60 and Multieq-x will only cut at 80 so there is overlap? Maybe?
3. What is the target in db that I want, 75? No one has ever told me what the SPL target is...but I guess I'm using 75db for white noise calibration...
4. When you mention the 4db difference, you are referring to the frequencies starting around 380?

 

[Chromatischism]

Thank you for the feedback. I'm battling the audio demons and doing the best that I can.

1. Yes, I have a sofa over on the right side near the speaker than is an issue...but nothing I can do there.
2. I will play with getting that hump down. It "may" have to do with the fact that I'm crossing at 60 and Multieq-x will only cut at 80 so there is overlap? Maybe?

It's usually sub placement and its interaction with the walls of the room. At least you don't have any nasty dips (at least, none that we can see from this resolution). Can you run a 15-300 Hz and scale the graph accordingly so we can see your bass region?

As long as you've addressed the big cancellations, Audyssey will bring down the bumps.

3. What is the target in db that I want, 75? No one has ever told me what the SPL target is...but I guess I'm using 75db for white noise calibration...

Well, you're in the Audyssey thread, so target dB doesn't matter. Audyssey will level match all of your speakers and line them up such that Reference Level will output if you set your volume to MV=0. However, it can't do that across the entire spectrum. There is a reference point that it uses, so you can still end up with imbalanced highs causing a stereo image shift, for example.

With that said, I would try to address whatever is causing the imbalance and get them as close together as possible before running room correction, for best results.

4. When you mention the 4db difference, you are referring to the frequencies starting around 380?

Yes, and especially above 3 kHz.

 

[Acerun]

Thank you for the ideation. What I'm finding is that what I have to do to hit the crossover (adjusting the speaker level via test tone) to hit X marks the spot for L and R at 60, it creates the 4db speaker imbalance. I think I need to start from scratch with new measurements and then go from there!

 

[Acerun]

New measurements, tried to simplify with just a -.04 tilt for target curve.

 

[Steve Dallas]

That looks pretty good to me. The bump centered at 55Hz will not be problematic. It is common to lose energy below 40Hz even with good subs and a good calibration. To tame that bump, you probably need steeper crossover slopes, which are likely not available in your AVR.

Hot tip: You can click the camera/capture button just to the left of the SPL and phase button to create clean graph captures without the rest of the UI showing.

 

[jbattman1016]

Apparently there is some video with steps on how to use REW with the Audyssey systems? I may try this.
Here is it....

 

[Acerun]

That looks pretty good to me. The bump centered at 55Hz will not be problematic. It is common to lose energy below 40Hz even with good subs and a good calibration. To tame that bump, you probably need steeper crossover slopes, which are likely not available in your AVR.

Hot tip: You can click the camera/capture button just to the left of the SPL and phase button to create clean graph captures without the rest of the UI showing.

Hi Steve. Thanks for the note. I've got it set for a 2nd order crossover but I can choose 3rd, 4th, 5th, 6th order crossovers which look to be steeper. Is that what you were thinking? EDIT: Probably not, that seems to destroy the response in Multieq-x filter view.

 

[Steve Dallas]

Hi Steve. Thanks for the note. I've got it set for a 2nd order crossover but I can choose 3rd, 4th, 5th, 6th order crossovers which look to be steeper. Is that what you were thinking? EDIT: Probably not, that seems to destroy the response in Multieq-x filter view.

Yes. I was thinking of experimenting with 3rd or 4th order (not necessarily the same for each curve if possible). What the editor shows and what you measure in room do not necessarily match, so it might be worth trying. Still, I do not think what you see there will be objectionable.

 

[Acerun]

I'd also be curious on opinions as to where I should cut the EQ. I've tried 250, 1000, and 2000. Do I want to try to knock down those humps up to 1000? I've heard people say to only EQ to 250 or 500 max too.

 

[Palmspar]

I'd also be curious on opinions as to where I should cut the EQ. I've tried 250, 1000, and 2000. Do I want to try to knock down those humps up to 1000? I've heard people say to only EQ to 250 or 500 max too. View attachment 341066

Looks good to me, try 300 and let the speakers do their thing.
For music dont eq to high indeed.

 

[Jpert]

I'm sure this is old news to this crowd (I haven't scrubbed the forums), but one of the major benefits I've found in MultiEQ-X is the ability to measure channels individually at each position. Since the microphone has a 90-degree calibration file, I've found improvements setting up the mic aimed 90 degrees from each speaker tweeter using a laser and doing individual measurements. The difference is not trivial like I thought it would be, and is very impactful for height channels if you plan to correct them to 20 kHz . I am even enjoying listening to my mains running full-band correction. (Edited with correct graphic: Top Left Front speaker w/ mic in vertical config (left) and 90 degree config (right))

 

[Theriverlethe]

Any chance a version of Audyssey could integrate moving mic method? I'll see myself out.

 

[kejar31]

So wanted to share what I did.. Sorry if this has been discussed but thought it was pretty awesome and gave good results..

Created a 90 degree calibration file for my Audyssey mic

• I plugged my Audyssey mic into my Mic input on my PC that has REW.. with that I tool a measurement with both my UMIK-1 mic and another with the Audyssey mic (being sure to keep them far enough away from the speaker to get a good summation without any major dips etc. Double checked the measurements for proper placement by checking the impulse response to ensure they were basiccly exact. Both were on mic stands so this was a bit easer but still took multipe tries.. Was sure to also take a super long 4M measurement (0 to 24khz) of each mic..
• After this I followed the guide over on OSA's youtube channel (This Is How I Calibrated My Stock Audyssey Microphone)
  • Set the Full Display Window (FDW) in IR to 4 cycles (apply and keep refrence time)
  • SPL align the two mesurements (agerage at 1khz center with 5 octave alignment span)
  • Last but not least used a simple tracearithmic A / B (Audysee mic / UMIK-1 mic) to get the difference
  • Exported the A to B measurement to text in REW saving the file for 20 to 20000 hz
  • This export is my new calibration file. (see attachment for an image)
  • Sanity check: Tested the mic's one after another to ensure the calibration file created the same results (they did)

Now that I had my mic calibrated I used OCA's guide (Audyseey Art) on youtube to pull down my measuments from MultEQ-X and calibrate and EQ my setup the only difference is that I adjusted all the measuremnts using my calibtation file, another simple tracearithmic A / B (measurment / calibration file) before starting the spl or time aligments.

Then when I was done with REW I uploaded the finalized files to MultEQ-X using them as target curves on each speaker (also be sure to disable all measurements) see pics below

When all was said and done the results are pretty spot on with the predicted ones from REW

Anyways I thought this was both cool and powerful.. If anyone wants to know more or wants pointed in the right direction, I will do my best to do so.


Basically when all was said and done I was able to allow MultEQ-X to take all the measurements of my speakers including the hight ones, get those measurmemts and apply the correction curve myself, then EQ to my harts content in REW.. After that I was able to force MultEQ-X to implement those filters which returned results that align with what I expected by measuring with my UMIK-1 when all was said and done.

 

[kejar31]

I'd also be curious on opinions as to where I should cut the EQ. I've tried 250, 1000, and 2000. Do I want to try to knock down those humps up to 1000? I've heard people say to only EQ to 250 or 500 max too. View attachment 341066

I am pretty sure the point is.. Anything below 200 - 500hz or so you are mostly hearing the room and not the speaker. Anything above 500 and you have now reached an area where you hear both the speakers and reflections of the room. Your brain has the ability to filter this very well, the microphone not so much. It is possible to EQ for higher frequencies but you should probably use a gated measurement for that. (time window to capture the measurement before it has time to reflect off walls and return to the mic) This can be done in REW but not MultEQ-X

 

[kejar31]

Any chance a version of Audyssey could integrate moving mic method? I'll see myself out.

lol you can def do your own moving mic measurements in REW, EQ and upload either the PEQ or use the reverse house curve method.. But yes, I agree it would be great if MultEQ-X allowed for single point measurement at the LP to get distances etc.. Then enable a MM measurement to be used for EQ. But again while great it would be excessively difficult to walk people though the method lol..

 

[Datec]

So wanted to share what I did.. Sorry if this has been discussed but thought it was pretty awesome and gave good results..

Created a 90 degree calibration file for my Audyssey mic

• I plugged my Audyssey mic into my Mic input on my PC that has REW.. with that I tool a measurement with both my UMIK-1 mic and another with the Audyssey mic (being sure to keep them far enough away from the speaker to get a good summation without any major dips etc. Double checked the measurements for proper placement by checking the impulse response to ensure they were basiccly exact. Both were on mic stands so this was a bit easer but still took multipe tries.. Was sure to also take a super long 4M measurement (0 to 24khz) of each mic..
• After this I followed the guide over on OSA's youtube channel (This Is How I Calibrated My Stock Audyssey Microphone)
  • Set the Full Display Window (FDW) in IR to 4 cycles (apply and keep refrence time)
  • SPL align the two mesurements (agerage at 1khz center with 5 octave alignment span)
  • Last but not least used a simple tracearithmic A / B (Audysee mic / UMIK-1 mic) to get the difference
  • Exported the A to B measurement to text in REW saving the file for 20 to 20000 hz
  • This export is my new calibration file. (see attachment for an image)
  • Sanity check: Tested the mic's one after another to ensure the calibration file created the same results (they did)

Now that I had my mic calibrated I used OCA's guide (Audyseey Art) on youtube to pull down my measuments from MultEQ-X and calibrate and EQ my setup the only difference is that I adjusted all the measuremnts using my calibtation file, another simple tracearithmic A / B (measurment / calibration file) before starting the spl or time aligments.

Then when I was done with REW I uploaded the finalized files to MultEQ-X using them as target curves on each speaker (also be sure to disable all measurements) see pics below

When all was said and done the results are pretty spot on with the predicted ones from REW

Anyways I thought this was both cool and powerful.. If anyone wants to know more or wants pointed in the right direction, I will do my best to do so.


Basically when all was said and done I was able to allow MultEQ-X to take all the measurements of my speakers including the hight ones, get those measurmemts and apply the correction curve myself, then EQ to my harts content in REW.. After that I was able to force MultEQ-X to implement those filters which returned results that align with what I expected by measuring with my UMIK-1 when all was said and done.

Yea I wanted to do the same thing. Any info would be helpful.

 

[kejar31]

Yea I wanted to do the same thing. Any info would be helpful.

What would you need help with? MMM, REW, how to import PEQ into MultEQ-X?

 

[Dovydas]

Are there any rumours when and if they are going to add UMIK-1 support for MultEQ-X?

 

[Fahzz]

Does the license still cost $199?