Any Ideas?

[joorge]

I just finished my first build and completed some initial adjustments. I've studied this forum and tried to apply all the stuff I've learned here.

I set the gain on all of the speakers with a meter. I also did some testing with the MCAAC EQ, but it didn't help any and was off for all the charts. Cross over is 80 Hz and all speakers are set as Small.

I'm using the Shack's test tones with an RS meter and manually entering the data in REW with a correction file for the meter. REW helped me get 2 filters that were entered in my 1124. I'm attempting a house curve with the results in the following chart.



Next I tried all speakers with only a change in the phase setting on the plate amp. Here's what I got. I included the sub only chart for reference.



Then I took the best phase setting and ran a test out to 1kHz. Here it is:



That big hump looks ugly. I haven't had time to see how it sounds, yet. I tried to knock it down by using a custom EQ setting in the receiver. I set the filter at -6 dB @ 400Hz for all speakers. It didn't do much. I wonder if this is a room effect that can't be cured?

Any comments or advice is welcomed.

[Anthony]

At 360 Hz, it's not entirely into the main speakers, but it's pretty close. You're two full octaves away from the crossover point, so at a minimum you're looking at being -27dB from the subwoofer at that point.

So that is most likely a main speaker/room interaction. Given where it is, it could be positional. Try moving the speakers out another 6" or 1' and remeasure. My first thought was baffle-step. Most speakers have a compensation circuit in them to accomodate low frequency rolloff of the speakers at wavelengths less than the baffle width. They are tuned for a particualr distance from the wall. Too close to the wall and you get more bass out of the speaker, too far out and you get too little. Sadly, unless the manuf. tells you exactly what they tuned it for, you are just sorta guessing. Most know how their speakers will be in-room, though, so 6" to 2' from the back wall is usually a good place to start.

It could also be a resonance, in which case moving it may help.

Next, I would try room treatments meant to absorb broadband around that frequency and that would definitely help.

Broadband EQ will not fix this if it is a resonance. You need narrow, targeted filters (like in the 1124) to spike those down.

Good luck.

[joorge]

Thanks Anthony. I guess you are correct. I went back and looked at old charts that I made when learning to use REW. The hump is in all of them regardless of xover, phase or receiver EQ settings. I'll move the speakers around and see what happens.

[brucek]

I'm using the Shack's test tones with an RS meter and manually entering the data in REW with a correction file for the meter.

Why aren't you using REW to take the measurements? They're much more accurate, and it takes only a second compared to the time it takes to do the manual method.

brucek

[joorge]

Because I don't have all the necessary equipment. I don't expect to be doing this again for awhile, so I'll do it the hard way.

I'm happy I can get this far with a $25 used meter. The resources and expertise provided at the shack are rather extraordinary when ya stop and think about it.

[joorge]

Moving the speakers a much as possible was worth less than -1 dB in the peak area. I plugged the rear ports and that had no effect.

I tried it without the center speaker and tamed the big peak, but got a smaller one around 153 Hz. Here's a chart with and without the center speaker:



I'm out of time for now. I wonder if changing the phase or distance on the center speaker will do anything?

[brucek]

I tried it without the center speaker

Why is there a center speaker being played at all?

Measurements should only be taken with mains and a sub, both when using REW and when using the manual tones method. They are both mono signals, so any processing with a multi-channel processor soundfield will not be valid.

brucek

[joorge]

Thanks Bruce. You really have me confused. I have read everything I could find and have never seen anything about not using the center. Not doubting you for a minute, I'm just amazed that I didn't see that somewhere.

More confusion. How is it mono to have 2 mains in the test? Seems like that is stereo. Am I supposed to set the receiver to mono?

Doesn't the chart represent the SPLs where the measurement was taken? Isn't it really almost 10 dB louder around 350 Hz when I listen to music with all the speakers working?

Why wouldn't you want to look at the response with the setup you actually listen to? I'm not disputing your statements, just listing the questions that immediately popped up.

Obviously I'm missing something here. My knowledge is so basic that it must be inadequate to even get started. I hope you will take the time to help me understand.

[Wayne A. Pflughaupt]

joorge wrote:

More confusion. How is it mono to have 2 mains in the test? Seems like that is stereo.

No, you only have stereo when you have two distinct signals. If you send a single discrete signal to both speakers, you merely have mono in two speakers.

Why wouldn't you want to look at the response with the setup you actually listen to? I'm not disputing your statements, just listing the questions that immediately popped up.

When you have the center channel running, that means you’ve engaged Dolby processing. With a mono non-digital signal like your test tones, the default processing is Dolby Pro-Logic. Pro-Logic is not a discrete process like Dolby Digital, which means some signal bleeds into the L/R speakers in addition to the center. This can affect your measurements, as the mic is picking up a signal from all three speakers.

If you want to check the center speaker together with the sub, your best bet would be to switch the receiver to stereo mode, move the balance control to one side, then connect the center channel speaker directly to that channel.

Regards,
Wayne

[joorge]

Thank you, wayne. The bulb has a dim glow, now.

I am remeasuring everything. Do I turn the unnecessary speakers off with the receiver or can I just unplug them?

[brucek]

I have read everything I could find and have never seen anything about not using the center. Not doubting you for a minute, I'm just amazed that I didn't see that somewhere.

I suppose it's just assumed. A mono signal cannot be directed to all speakers equally with a DSP surround soundfield that uses phasing and level differences to accomplish the steering of the signals.
You must select stereo mode and only use the left and right mains with REW or test tones.

How is it mono to have 2 mains in the test? Seems like that is stereo. Am I supposed to set the receiver to mono?

Stereo mode will direct an equal signal to left and right mains when it receives a mono signal.

Doesn't the chart represent the SPLs where the measurement was taken?

But you don't know what you're measuring since the signal is not equally being sent to all speakers. REW and test tones are not designed to measure a multi-channel system. You may check each speaker on its own in your system by either using the direct mode using one channel at a time, or you may connect each speaker up to a right or left main output.

Why wouldn't you want to look at the response with the setup you actually listen to?

Can't be done with a mono signal.

brucek

[joorge]

That helps. I'll see what happens when I do it correctly.

[joorge]

I think I did it right this time. Here's the sub and mains. Does the 100-200 area need attention?

[brucek]

Does the 100-200 area need attention?

Yes, but first you need to test each main and sub individually to see if it's a (main versus main) or (main versus room) problem.

brucek

[joorge]

This chart is the Mains only overlaid on the Sub + Mains chart. I can't find the sub only chart, but it is virtually identical to the one posted above.

Did you mean for me to measure each main separately?

[brucek]

Well, you can see that the main that you posted alone above enjoys the same dip and peak as the mains+sub. That says that the offending dip and peak isn't an interaction between that main and sub, nor is it likely an interaction between that main and the other main. This leaves the culprit as that main and the room. You want to see what the other main alone looks like. If it doesn't exhibit this dip and peak, then you can assume that the other main needs to be moved a bit.

Anyway, you get the idea of the process of elimination. It's a lot of measuring. You can see why I recommended REW.

brucek

[joorge]

Thanks Bruce. Those are both mains. I didn't measure them separately, but did ask if I should. Sorry for the confusion. I think I understand how to proceed.

I'll do some more measuring tomorrow and let ya know what happens.

I wouldn't argue about how nice it would be to have REW do all of the work. It's not too bad doing it manually for one sub. Each test takes about 10 minutes to complete and enter in REW. It will get old in a hurry if I keep building subs.

[brucek]

Those are both mains.

Yeah OK. I think you get the idea though. Measuring two mains doesn't reveal where the problem is. If you do each main by itself, you might find both mains alone measure fairly good, revealing that it's their interaction between themselves that is the problem. If only one measures bad, then it's that speaker and the room that is the problem.....etc, etc..

brucek

[joorge]

I tried to run the mains independently. I get no sound with just one. Apparently the receiver requires both of them.

I have tested 10 different speaker locations (both mains) in the problem frequency band. Most of them didn't do much at all. Three show promise, and the difference is quite noticeable.

I'll do some full range testing when I get a chance. It looks like that peak at 145 is a room effect. Nothing seems to budge it. Here's a chart of Both mains with the testing overlaid.

[brucek]

I tried to run the mains independently. I get no sound with just one. Apparently the receiver requires both of them.

Just feed the left or right channel separately with your tone to get a single main.

brucek

[joorge]

Duuh.