How much should speakers roll-off at and below the crossover point set at the AVR?

[welldun]

Hello folks, I'm trying to learn more about how to best integrate my Main speakers with the Subwoofers in my HT, and after reviewing the measurements taken using REW, I'm left wondering how much of a drop/roll-off there should be from my Mains at the crossover point set in the AVR? The measured results don't show much of a change (maybe 1-2 dB difference in output) between the mains being set to Large vs them being set to Small with various crossover points. This small difference seems to be a contributing factor in the bass managed region looking different when measured with just the subs and then again with the mains set to small, plus the subs combined. The odd thing in that some of the frequency cancellations are happening below the crossover point which leads me to believe that the Mains are putting out a good amount of energy below the crossover point. So again I'm left wondering how much should my Mains be rolling off, and how (other than getting speakers with a higher roll-off point) can I improve on the integration? Maybe using a higher crossover point in the AVR?

For reference, the Main speakers are Andrew Jones-designed Pioneer Elite SP-EFS73 towers. the subs are dual SVS SB12-NSD. The AVR is a Pioneer Elite SC-99.

 

[Tony V.]

Hi Welldun, first of all can we see the graphs of your readings? When you take the readings with REW are you turning off the sub its self and leaving the crossovers on the mains where they are at set to small? Something does not seem right if you say you are getting the same readings even when selecting a crossover point and running the mains as small. Ideally the crossover should be set about 40Hz higher than what the speakers are rated for because a crossover is not a brick wall and the slope is quite gradual.

 

[JStewart]

@welldun you want the speaker response flat through the crossover region and your Pioneer Elite SP-EFS7 are listed on Pioneers website with a response of 38Hz to 20kHz (It does not say how many dB they are down at 38Hz however) and they are ported. Best bet is to start at 80Hz for the crossover, one octave above the cut off point listed in the manual. (same advice as Tony V.)
I didn't check the manual for your SC-99. It likely has a low pass of 12dB per octave and a high pass of 120dB per octave which in a perfect world would give you a little more bass in the region between 38Hz and 80Hz. If it does and you like like it great. If it does and you don't like it then lower the crossover. If the response is flat then also great, goal accomplished because that's what you wanted to happen. If its flat and you want more bass then just turn the sub's amp up to suite to taste. Every room is different so the amount of bass needed for it to sound right in a given room will vary.

 

[welldun]

Hi Welldun, first of all can we see the graphs of your readings? When you take the readings with REW are you turning off the sub its self and leaving the crossovers on the mains where they are at set to small? Something does not seem right if you say you are getting the same readings even when selecting a crossover point and running the mains as small. Ideally, the crossover should be set about 40Hz higher than what the speakers are rated for because a crossover is not a brick wall and the slope is quite gradual.

Hey Tony, thanks for replying. I'll take some new measurements in a minute and post them here. As for how I'm testing them, I usually take various measurements with the subs on and off, and the AVR crossover set to 80Hz or 100Hz. Per the manual, the speakers are rated at 38hz-20,000Hz. The manual also mentions that they included a foam plug to tailor the bass response for the particular environment. I don't have the foam plugs so I'm gonna reach out to Pioneer to see how I can get them.

 

[Grayson Dere]

I'm wondering if this has anything to do with the channel balance configuration during AVR speaker setup.

Is it possible that the SPL of the mains and subwoofer were matched very close to each other during channel balance (I believe this is a good thing) so that even at the LFE crossover point, ie: 80hz, there isn't much change in SPL when the frequency is shifting from mains to LFE? Also, just as Tony mentioned the drop-off at crossover should be pretty gradual.

Or do you mean the frequencies aren't transitioning from the Mains to the LFE channel when they're supposed to according to the REW graphs?

 

[welldun]

Attached are some pics plus the .mdat file from the REW measurements. It looks like below the crossover point when I measure each speaker individually with the subs active, the worst offender is the Left Main Speaker. The same could be said for the Center Channel. However, when I measure both Left and Right Main combined with the subs, the null at 71Hz disappears.

 

[Grayson Dere]

Thanks for the graphs, Welldun. It seems a little odd why the Green (left channel+subs@small) graph in the first chart has that dip a little after 70hz. All the other charts seem okay, though..as in I'm not seeing anything amiss due to software issues or measurement anomalies. If anything I believe room acoustics are playing a part in the dips and peaks of the various speaker readings.
Anyone else here have some thoughts?

 

[welldun]

I'm going to try plugging the ports on the Mains to see if there is a difference both in roll-off and if the dip at 71Hz disappears. At the moment, one of the subs is located in the front left corner of the room and the lower ports (it has 3 ports) of the Left Main speaker are aimed at the sub. Not sure if the location of the sub and speaker is causing the cancellation at the 71Hz. I've included pics for illustration.

 

[Grayson Dere]

I'm going to try plugging the ports on the Mains to see if there is a difference both in roll-off and if the dip at 71Hz disappears. At the moment, one of the subs is located in the front left corner of the room and the lower ports (it has 3 ports) of the Left Main speaker are aimed at the sub. Not sure if the location of the sub and speaker is causing the cancellation at the 71Hz. I've included pics for illustration.

I'd also be curious to see if it helps any by moving the sub away from the corner and seeing what REW measures.

 

[Tony V.]

Qurious what measurement mic you are using and are you using a calabration file?

 

[welldun]

I'd also be curious to see if it helps any by moving the sub away from the corner and seeing what REW measures.

The subs are currently positioned at opposite diagonal corners and have been time aligned and EQ'ed via MiniDSP prior to integrating with the rest of the system. So if you are referring to a sub only measurement, that one is already included in the post above. If instead you are referring to moving the sub to see how the Left main plus sub would look, then that will be another test I can run later after I measure with the speaker ports plugged.

 

[welldun]

Qurious what measurement mic you are using and are you using a calabration file?

UMIK-1 from Cross Sprectrum.

 

[Tony V.]

The in room response is really playing with your levels. Not that its a bad thing because in the end your final graphs look good with everything running. I am assuming that the mic is at the listening position at ear level and aimed at 90 degrees? what is the mic sitting on?

 

[welldun]

The mic is on a tripod boom sitting 90 degrees as you described. The room itself is not a perfect rectangle. It is an upstairs room in a Cape Cod-style home so the ceiling is somewhat barn-shaped, and the left wall has a dormer window close to where that left main speaker sits. I going to buy a curtain to close off that dormer space. That will give me a more symmetrical space. You can see the dormer window space in the attached pic. It's an older pic taken before I moved the subs, but the main speakers are pretty much in that same location now. So you can see how the Left Main is close to the dormer opening.

 

[Tony V.]

Well, as i said above your end results look good so I would not get to concerned about what the readings are individually.