seeking advice on my room- graphs attached

[Ian_S]

Hello,
What a fantastic thing...REW and room correction. My 5.1 tv room is a Meridian DSP in wall system, with a James inwall analogue sub.
I have taken measurements with a Umik-1 90 degrees cross spectrum labs calibrated mike.

I wanted to check with the experts on my methodology and results
The Meridian G68 processor only builds filters from 20-250 hz. So I built filters based on measurements taken for each speaker separately. Is this correct that I would not combine the sub to build filters, and make filters for the sub by itself too? What about surrounds? I would think surrounds do not need to have peq. Only SPL measurement would matter.

Another question is, do I need to set the target level for each speaker or do I keep it the same for all speakers?

I let REW build the filters. The bass is more integrated and overall the sound is not as localized as it was. The sound seems to be less directional and the dialogue is more intelligible.

Would RTA measurements with pink noise be helpful. I did sweep measurements with acoustic timing reference...
I have posted my measurements and just one channel with the REW filters. Comments would be most appreciated as I am only getting my feet wet with REW.



Thanks much...
Ian S.

 

[Matthew J Poes]

You actually should combine the sub with the main to build filters, they both contribute to the LF's and applying EQ separately doesn't combine to the predicted response.

You should also share seperate measurements of just the sub if possible, it looks like there could be an integration issue (related to the phase/time and crossovers along with placement), but I can't tell without seeing the sub seperately.

Also, take more measurements of the sub and mains at the same time from different positions around the room (in which you sit).

 

[Ian_S]

Thanks for the reply. Is there a way to combine 2 measurements to make eq filters? Or do I measure for example Left and Sub together and use that to make filters? I took measurements across all the seating positions. It just a couch and not rows of seats.
I guess I would average the measurements to make filters -yes?

What about setting target spl? Do I set one value for all the filters I am making.
Attached is my sub measurement. I would appreciate any suggestions.
Many thanks,
Ian

 

[Matthew J Poes]

Thanks for the reply. Is there a way to combine 2 measurements to make eq filters? Or do I measure for example Left and Sub together and use that to make filters? I took measurements across all the seating positions. It just a couch and not rows of seats.
I guess I would average the measurements tomato filters -yes?

What about setting target spl? Do I set one value for all the filters I am making.
Attached is my sub measurement. I would appreciate any suggestions.
Many thanks,
Ian

View attachment 10976

Hi Ian,

I prefer to measure the response by playing a main speaker (L or R) and the sub together, just as you would with music. You can use REW to combine, but I prefer this approach, it's simpler and more fool proof.

I apologize for the oversight, I was thinking those were different speakers. My bad, I wasn't looking carefully.

There are different approaches to creating filters for multiple seating positions. Personally, I don't love using averaging exclusively. While I do take it into account when generating filters, I prefer to look at each individual measurement position. The reason is that I want to personally evaluate the way peaks and dips move. I don't EQ things that don't remain relatively constant. My only exception is that if my main listening position has a HUGE peak and a near by seating position doesn't, I would probably apply EQ based more on that main position than the side positions. Who cares about the guests, I want the best sound! Because peaks and dips can be so position dependent, I think you need to consider taking more measurements around that seat if that ever happens. If you want to average a set of positions and simply weight those measurements closest to where you sit, average all equally, then re-average the seats closest to you. It weights them higher, so to speak.

For example, let's say you have 5 measurements, and 3 (1,2,3) are closest to you, then here is the process I might use:

• Average 1, 2, 3, 4, 5 = AllAve (All averaged)
• Average AllAve, 1, 2, 3 = WtdAve (Weighted Average)

As for the SPL target, if you use a subwoofer for your bass, then I would create filters for 20hz to 100hz to be applied equally across all positions. I always apply the same filters to the left and right speaker, unless there is something really off about one speaker vs the other (in which case I try to match the response of one to the other, but I would be very careful doing this). For surround and center speaker, I tend to apply separate EQ to those channels based on how they measure in their position, but I try to avoid anything wildly different between channels.

In terms of the target curve SPL, you need to align the target with each speaker's relative level. If using the same sweep type and volume levels, the LCR and S should all be the same level. If they are not, that should tell you something (but you want to fix that in the levels, not the EQ). For the most part I choose a target curve SPL level that hits the center of the peaks and dips and try to set it such that I am not needing to apply much if any boost.

 

[Ian_S]

very helpful.
Thanks,
Ian