EQ Limit for Cutting - Subwoofer FR

[CK.]

The crossover will never be a brick wall, so you will get some response from the speakers as you seem to have noticed already. What you can do is choose a higher crossover, for instance 100 or even 120hz, to get less energy from the speakers in the lower frequencies.

I could choose a higher cross over but then I think the issue will just move to a higher frequency. Couldn't I just apply a broad EQ filter (or shelf filter) in the low frequencies of the front channels to further diminish the effect?

 

[CK.]

Then just run a standard Audyssey calibration with the Editor app and use the A1 Evo script. Your sub and mains (and all other speakers) will be optimally aligned and crossed over.

I did but it resulted in dips in the bass region (e.g. 70Hz) and overall it does not seem like a good response to me (see attached). Also, the script determined that I should reverse my sub's polarity and use LFE+Main instead of LFE.
I had the same issue with the response (dips in XO region or below) with earlier versions of the script.


That's why I thought of going the manual way cause I also have MEQ-X.

Forgot to mention, the XO it came up with was 90Hz which seems reasonable but still...

 

[sigbergaudio]

I could choose a higher cross over but then I think the issue will just move to a higher frequency. Couldn't I just apply a broad EQ filter (or shelf filter) in the low frequencies of the front channels to further diminish the effect?

So, your Denon is setting a crossover for both the subwoofer and speaker that will typically be correct to get a good blend. If you create a shelf, you are likely to make a gap in the response.

 

[CK.]

So, your Denon is setting a crossover for both the subwoofer and speaker that will typically be correct to get a good blend. If you create a shelf, you are likely to make a gap in the response.

Its all a bit confusing to me because from what I understand the choice of slope for the high pass filter has to do with how much the phase changes, and a 12db/octave that the avrs are using correspond to 180 degrees phase shift. The idea is that this will be combined with the natural roll off of the speaker assumed to be also 12db/octave and thus a total of 360 degrees of phase shit or essentially 0. But this is so ideal and theoritical that does not make sense to me.

I could be totally wrong though!

 

[sigbergaudio]

Its all a bit confusing to me because from what I understand the choice of slope for the high pass filter has to do with how much the phase changes, and a 12db/octave that the avrs are using correspond to 180 degrees phase shift. The idea is that this will be combined with the natural roll off of the speaker assumed to be also 12db/octave and thus a total of 360 degrees of phase shit or essentially 0. But this is so ideal and theoritical that does not make sense to me.

I could be totally wrong though!

May I ask how the result is if you just let the Denon do its thing? Why are you doing this manually?

 

[CK.]

Sure basically audyssey gets many things wrong, such as the cross over and the speaker type (small vs large). Audyssey has no way of confirming the actual parameters it sets, it is basically a shot in the dark.

But more importantly, it destroys my low end responce. During initial spl matching, because of a strong room mode at 48hz, it thinks this is the responce of sub and instructs me to lower the sub gain to 10%.

It then lowers the 48hz peak and determines the rolloff of my sub is at 40hz! The resulting responce is criminal.

 

[sigbergaudio]

Sure basically audyssey gets many things wrong, such as the cross over and the speaker type (small vs large). Audyssey has no way of confirming the actual parameters it sets, it is basically a shot in the dark.

But more importantly, it destroys my low end responce. During initial spl matching, because of a strong room mode at 48hz, it thinks this is the responce of sub and instructs me to lower the sub gain to 10%.

It then lowers the 48hz peak and determines the rolloff of my sub is at 40hz! The resulting responce is criminal.

But you can adjust most of this post correction. You can sett the speakers to small again, and you can also adjust the crossover up (not down). The 48hz-thing sounded a bit problematic though. There's no EQ directly on the sub so you can dampen that before running Audyssey?

 

[CK.]

But you can adjust most of this post correction. You can sett the speakers to small again, and you can also adjust the crossover up (not down). The 48hz-thing sounded a bit problematic though. There's no EQ directly on the sub so you can dampen that before running Audyssey?

Yes you can adjust a few things but lets not forget that by default audyssey corrects full range which in most cases is not desired, it can really mess up a good speaker.

No my rel sub does not have eq options unfortunstely and I think a lot of sub brands are the same

 

[sigbergaudio]

Yes you can adjust a few things but lets not forget that by default audyssey corrects full range which in most cases is not desired, it can really mess up a good speaker.

No my rel sub does not have eq options unfortunstely and I think a lot of sub brands are the same

Do you have the audyssey app? With that you can set the cutoff frequency so it for instance doesn't correct over 500hz or 200hz or whatever.

 

[CK.]

Do you have the audyssey app? With that you can set the cutoff frequency so it for instance doesn't correct over 500hz or 200hz or whatever.

Yes actually I have the meq-x pc app. There are also spinorama data that you can use to correct the respince above the schroeder frewuency.
So you see why I chose the manual way?

Every autocorrection system has limitations, some more serious than others.

 

[OCA]

I did but it resulted in dips in the bass region (e.g. 70Hz) and overall it does not seem like a good response to me (see attached). Also, the script determined that I should reverse my sub's polarity and use LFE+Main instead of LFE.
I had the same issue with the response (dips in XO region or below) with earlier versions of the script.


That's why I thought of going the manual way cause I also have MEQ-X.

Forgot to mention, the XO it came up with was 90Hz which seems reasonable but still...

All the script's recommendations rely on brute force calculations, and it's rare to find people who don’t notice a significant improvement over standard Audyssey calibration with Evo results. A flat frequency response graph isn’t a reliable indicator of good sound quality. If you focus too much on achieving a perfectly flat graph, you may end up chasing narrow local dips that aren't even audible. I recommend listening to the default Evo results first. Experimenting with different target curves and Evo options is the next step if there's an area you're dissatisfied with.

PS If you really have to, crossover frequency search range is customizable per speaker from within the script.

 

[CK.]

All the script's recommendations rely on brute force calculations, and it's rare to find people who don’t notice a significant improvement over standard Audyssey calibration with Evo results. A flat frequency response graph isn’t a reliable indicator of good sound quality. If you focus too much on achieving a perfectly flat graph, you may end up chasing narrow local dips that aren't even audible. I recommend listening to the default Evo results first. Experimenting with different target curves and Evo options is the next step if there's an area you're dissatisfied with.

PS If you really have to, crossover frequency search range is customizable per speaker from within the script.

@OCA, I am only applying a handful of filters per channel <300Hz and applying psychoacoustic smoothing to verify, but the with the EVo the FR @ psychoacoustic smoothing had a significant dip.

In any case, with all things being equal, and my speakers having a good directivity and a good off-axis response, and a smooth in-room FR, I am assuming you mean that I should also look at the time domain issues and namely the impulse response and RT60.

With the EVO script how can I determine the actual benefit I will get over the manual PEQ I am doing via REW-->Multeq-X ?
Do I look at clarity metrics? and how much impact does this really have on the perceived sound quality?

I would appreciate your insight on this as I am still on a learning path.

thank you

PS: One last question, when I use the time aligning tool in REW for sub to mains integration, it determines a large offset (>10 meters) which is not acceptable by audyssey/denon. I know you are handling it somehow, can you explain?