Ask an Acoustic Engineer (me)

[Jdunk54nl]

1) you are on 1db increments. That is too fine of a resolution. Scale out to 5db and repost picture.
2) turn off the bars on rta and record in 1/48 then post smooth to 1/3
3) what are your electrical crossovers set for on each mid and tweeter?

[jtrosky]

I only have a USB mic and in my reading of HOLM it sounds like you have to have a phantom powered microphone in insure you don't have a time lapse. Hoping to change topics if possible and ask another question:

I have summing of my midrange drivers from about 200hz to 1.25khz. After 1.25khz my summing seems to disappear meaning that the individual levels of the left and right side can be greater individually than when combined. I believe it is a phase issue but my crossover point is not until 5000 hz. Is there anyway to address a phasing/cancelation issue than only takes place half way through it's intended range. My tweeters seem to sum but the upper half of the mid range seem to cancel.

I have attached an image. I use a Bit Ten D so I have the look set at 1/3 octave.

Thanks for all of your help each one of you.

Attachment 13111

It looks like an Allpass filter may be able to address that issue. Does your DSP allow for Allpass filters?

[jsalinasii]

Sad Face Emoji - No it does not it is an old graphic EQ with 31 bands per channel

[jsalinasii]

Jacob,

1. Yes it is one 1 db resolution
2. Here is what the combined curve looks like in 1/48 and zoomed out to 5 db increment marks. I will have to remeasure each side to remove the bars and get you an accurate photo.
3. Both Left and Right are set at 24 db 5000 hz between TWS and MRS and 24 db 150 hz between MRS and MBS and LMB is at 24 db 50 hz and RMB is at 24 db at 70 hz.

[Jdunk54nl]

Now do that second graph with all 4 selections like the first (individual speakers and combined).

What I'm getting at here is, in 1db increments it looks like an issue, in 5db increments it's not a problem. You are seeing a problem only because you are zoomed in so much.

[jsalinasii]

Here is the request format.

Target Curve
All - Means Front Stage plus Sub
Front Stage - Front 3 Way both Left and Right
Left - Front Stage Left
Right - Front Stage Right

All ran on 1/48 and then smoothed to 1/3 on a 5 db scale.

[Justin Zazzi]

Thank you for posting more information.

I think you're right that it is phase-related. This could be any number of things. I like breaking complex things into smaller chunks and looking at them individually. You have at least 4 speakers playing at the same time (left and right woofers and also tweeters) so lets make it less complex. What happens if you measure the left woofer, right woofer, and then the two woofers combined? (no tweeters, no sub, just two speakers). Then what happens if you do the same thing with just the tweeters only? Can you find a combination that works correctly, and a combination that misbehaves?

Can you isolate the behavior to a certain pairing of speakers? If you can, then try adjusting simple things like time delay of one speaker only. Or move the mic to the other front seat. Or change one small thing and see what happens.

[jsalinasii]

Justin,

Thank you very much. I will try this and send the results tomorrow. Very much appreciate all that have taken the time to try and help me, this is not an easy game to play and it always seems like one step forward two steps back.

Does anyone know if there is a way to upload a REW file as part of a post? I know they can get fairly large in size but it might be easier than snipping images and hoping they are clear and descriptive enough for others to look at.

[jsalinasii]

Justin,

Here are the readings. To me it looks like the upper side of the Midrange is where the phasing issue lies. They sum up to a certain frequency and then stop summing or even cancel out. These were measured on 1/48 each with over 100 averages and then smoothed to 1/12.

LXX = Left
RXX = Right
TWS = Combined Tweeters
MRS = Combined Midranges
MBS = Combined Midbasses

[jsalinasii]

I have another question:

I just finished watching the new Youtube video of PSSound and he mentioned that when running paired speakers such as midranges to use Stereo Pink Noise (uncorrelated) is that what I should be doing? I use the PN generator on REW for consistency but I don't know how to change between correlated and uncorrelated PN in the generator.