First REW Sweep - How does this stack up?

[zinger084]

Hello,

Using a laptop, an external USB Soundblaster 3 soundcard, and a IMM-6 microphone, I performed the first sweep of my 10"/1.25" two-way system.

I believe I calibrated the soundcard correctly after a lot of trial and error. I also loaded the mic cal file into REW.

I'm wondering what others think of these sweeps. It's my first time doing it, so wanted to get some quick opinions.

Thank you.

[wogg]

It's a little hard to tell with 0 smoothing in place. Play with smoothing a bit, 1/24 octave is a good way to see plenty of detail without the fuzzy line. Aggressive smoothing like 1/6 octave is closer to how our ears will perceive the balance, but won't show you small issues like a phase cancellation and what not. Have you applied gating, and what's your measurement position relative to walls, floor, ceiling etc?

[billfitzmaurice]

1/6 octave is probably the best, that way you're not trying to fix every little zit, which will all change if you move the mic/listening position anyway. http://ethanwiner.com/believe.html

[fpitas]

Yes, I like to play with 1/24 oct and 1/12 oct a bit, reducing obvious peaks, but if the 1/6 oct doesn't look good it won't sound good.

[devnull]

And don't forget to uncheck the mic and soundcard calibration

[DeZZar]

Are you measuring for a final in room response here or are you looking to validate axial response of the speaker itself?

If it's the later you'll need to have measured on axis at a given distance (at least 3x baffle width). This distance in conjunction with height from the floor, ceiling and walls will give you a time window to work with. For me it's generally around 4ms for far field measurements.

To apply the gating (time window), click on IR Window at the top, enter the left/right window and apply it. You generally won't have to worry about any smoothing when gating.

Then click the All SPL tab to get a clear view of the quasi anechoic axial response which will be showing you somewhere around 200hz on up.

[zinger084]

Here is the latest captures smoothed at 1/6. I did 4 sweeps of each.

Frankly, i'm not sure what i'm validating... i'm doing this in the room at the actual location they will remain (at least for now). I figured the response should closely match my VituixCAD sim, which i'm not sure how well it does from what i see...

I'm a little concerned about the right speaker (red) having a big dip at the 120hz level... not sure what's up - any thoughts? I know the left speaker has a few boxes by it (one being my 18" Ultimax ;) so maybe that's helping increase (falsely) the left speaker at 120hz... i don't know...

I'm measuring these 1m in front of tweeter on the same horizontal.

I'll follow what DeZZar says here and try that out... real work just keeps getting in the way ;) I have no idea what the IR window is, but i'm willing to research and learn more...

[zinger084]

also, not sure why my phase differs so much between them either...

[DeZZar]

Your mic needs to be the EXACT same distance, angle and height from the listening axis (assume tweeter) for each speaker and preferably measure with a timing reference loopback.

[zinger084]

got it. i did a reasonable measurement, but not sure the plus/minus on the accuracy. sounds like something I'm not currently doing well enough. i'll see how i can adjust this in the future...

[DeZZar]

...i'm doing this in the room at the actual location they will remain (at least for now). I figured the response should closely match my VituixCAD sim, which i'm not sure how well it does from what i see...

Not at all - two completely different considerations.

To validate the design of the crossover for your speakers and how they are performing - you need anechoic axial response data. Here is one of the best guides I've found on how to go about doing this:
https://www.audiosciencereview.com/f...tuixcad.21860/

The other is how those speakers AND YOUR ROOM are performing. The latter introduces room reflections, resonances, cancellations etc. For instance, the 120hz dip is likely a cancellation node.

I personally don't believe in building the room into the crossover (and in a lot of instances you can't - boundary effects (Allison effect) cant be EQ'd out) - but there are others that have a different view - likely driven by the fact they are in their forever home and forever listening room by now so why not build the environment into a bespoke speaker. I take the view that a speaker should be built as an accurate reproduction tool that can be implemented equally anywhere. Dedicated listening environments should be treated first and EQ exists for the rest that can be EQ'd.

It looks like you are essentially +-5db IN ROOM response (at least on the white measurement) - I don't think that's too bad and will improve if you consider some room treatments (or perhaps you have already?).

[zinger084]

room treatments... interested in learning more. i'll continue the duckduckgo searches...