Measuring In-Room with REW

[wkhanna]

jim1961 has been kind enough to offer me some help learning to use REW.

so I have started this thread so others may also gain some information.

my system specs can be seen if you check out my profile link.
basically I am using a 65 liter, bottom-ported version of Jon's Natalie Portman's with two HSU subs crossed at ~ 90 dB, placed next to each speaker running from discrete L/R outputs from my highly modified Carver C-19 pre.

This first sweep was made with one of those nice pre-calibrated mic's (forget the name ops positioned were my nose is as I sit in my listening sweet-spot around 9 feet from the front baffles.

[dar47]

Bill,

This a relative db level or you close with one of those sp meters?

[jim1961]

Well, this is your thread. So, what do you want to know or do?

[wkhanna]

Well, this is your thread. So, what do you want to know or do?

well, i thought we could discuss - as in you can tell us - basic set-up parameters (aka smoothing, et.al.) for using the software to generate graphs that give basic, useful information on system performance.

then talk about the specifics of my room (which has minimal treatments) & how my system is interacting with it.

then maybe about effects on some modifications on speaker & sub placement along with some minimal treatments (with measurements to show effects)

hows that sound?

[5th element]

The biggest issue I see with the measurement is the declining response from about 600Hz and down. Did you build a version with no baffle step compensation or something?

[jim1961]

well, i thought we could discuss - as in you can tell us - basic set-up parameters (aka smoothing, et.al.) for using he software to generate graphs that give basic, useful information on system performance.

then talk about the specifics of my room (which has minimal treatments) & how my system is interacting with it.

then maybe about effects on some modifications on speaker & sub placement along with some minimal treatments (with measurements to show effects)

hows that sound?

Maybe the best way to go about this is to show what I look for and how I display the data when I look at my own room, except I will use your data.

Can you link or post your mdat's ?

BTW, you mentioned you have a calibrated mic. Did you direct REW to the cal file when taking the measurement(s) ? Do you know at what mic angle the cal is for? Do you have multiple cal files for different mic angles? Did you match the measurement mic angle to the corresponding cal file?

Do you have measurements for the left and right channel separately (recommended)? Do you remember what sweep length (128, 256, 512, or 1M file sizes) you used? (recommend 256k-1M in most cases for data analysis).

Just making sure we have good data to work with.

[dar47]

I'm going to have to show you the weird warning that pops up on me. I didn't send my mic in just the dayton cal file that is 20 to 20k. I should have got one done that goes lower. Mine says nothing of mic angle I just use straight forward. Any value in doing spl level or is relative fine for most?

[jim1961]

I'm going to have to show you the weird warning that pops up on me. I didn't send my mic in just the dayton cal file that is 20 to 20k. I should have got one done that goes lower. Mine says nothing of mic angle I just use straight forward. Any value in doing spl level or is relative fine for most?

If you got only one cal file, its most likely a 0 degree file meaning you have to aim the mic at the source (speaker). 90 degree cal files are handier, IMO, cause you can point the mic strait up and therefore get an even response from the room in all directions on a horizontal plane.

[wkhanna]

Maybe the best way to go about this is to show what I look for and how I display the data when I look at my own room, except I will use your data.

perfect

Can you link or post your mdat's ?

yes, how should/can i email/PM them to you?

BTW, you mentioned you have a calibrated mic. Did you direct REW to the cal file when taking the measurement(s) ? Do you know at what mic angle the cal is for? Do you have multiple cal files for different mic angles? Did you match the measurement mic angle to the corresponding cal file?

Do you have measurements for the left and right channel separately (recommended)? Do you remember what sweep length (128, 256, 512, or 1M file sizes) you used? (recommend 256k-1M in most cases for data analysis).

Just making sure we have good data to work with.

getting this info now from my friend who did the sweeps.

[jim1961]

You can either upload your mdat's to dropbox or some other offsite server and provide a link to them, or upload the files directly to this site as a file attachment in zip format.

[dar47]

Thanks, Jim thought so. I'm going to try and get that error for you. Any value in spl readings?

[jim1961]

Thanks, Jim thought so. I'm going to try and get that error for you. Any value in spl readings?

Not sure what you mean.

[dar47]

I know some guys use those spl meters to calibrated the spl level but I don't have one and just used relative level and calibrating with the mic file and doing the sound card cal.

[wkhanna]

i have used the the Radio Shack SPL meter (with the LF corrected values) & white noise fq sweep.

crude but better than nothing.

REW is pro-level tech, however.

[jim1961]

I know some guys use those spl meters to calibrated the spl level but I don't have one and just used relative level and calibrating with the mic file and doing the sound card cal.

For most things, the shape of the curve is whats important, not the absolute magnitude. That said, you might borrow one some time to get it right. Once you get it correct once, all your measurements thereafter would be correct. You shouldn't have to calibrate after every measure. REW locks in the setting as long as you use the same mic and input every time.

Where correct magnitude is vital is when your seeking what SPL your sub or mains begin to show stress in the form of distortion. Or perhaps if your trying to gauge speaker SPL efficiency.

[wkhanna]

BTW, you mentioned you have a calibrated mic. Did you direct REW to the cal file when taking the measurement(s) ? Do you know at what mic angle the cal is for? Do you have multiple cal files for different mic angles? Did you match the measurement mic angle to the corresponding cal file?

Do you have measurements for the left and right channel separately (recommended)? Do you remember what sweep length (128, 256, 512, or 1M file sizes) you used? (recommend 256k-1M in most cases for data analysis).

Just making sure we have good data to work with.

we do have all the cal files.
we used the 90 deg angle to match the cal file.

We did not do separate left and right measurements.
pretty sure we used the 256k.

if you PM me an email address i can dropbox the files to you.

[wkhanna]

The biggest issue I see with the measurement is the declining response from about 600Hz and down. Did you build a version with no baffle step compensation or something?

I used Jon's standard Natalie Portman xo.
not sure yet why we see the roll-off.

[jim1961]

we do have all the cal files.
we used the 90 deg angle to match the cal file.

We did not do separate left and right measurements.
pretty sure we used the 256k.

if you PM me an email address i can dropbox the files to you.

No need. Just post the public link here.

Also, if both channels were run simultaneously for the measurement, we cannot know if the data is accurate above 1K-2K given the cancellation effect that takes place (which is why measuring each channel separately is important).

The premature rolloff starting at about 8K in the graph you posted is probably due to this.

So, do you want to analyze what you got anyway, or remeasure?

[wkhanna]

probably makes more sense to re-measure each speaker individually, then.

that might not happen right away, as in next time Dan (PewterTA) is over here with the mic.

not sure about the dropbox link, may be other stuff here that is not audio related.
if so just ignore & excuse my crash-course errors on using social media apps (such things are not in my Luddite area of expertise as i loath all if it, as in- i have never been to farcebook let alone set up an account)

come to think of it, it is a near a papal-level miracle that i can even link a pic from photo-bucket.

Bill on fron of table.mdat

[jim1961]

There is something wrong with the data. Not sure how to go about getting it right. Perhaps I could talk to Dan directly?

The reason I say this is that the data says you got 71% distortion at 800hz, and over 2000% at 41hz. That couldnt be right, could it? Other details also point to this being bad data.

I cant be sure whether the actual measurement is bad, or the data got corrupted somehow. There wasnt lots of room noise when the measurement was taken was there?

In any case, new measurements are needed.

[wkhanna]

a re-do is obviously in order.

as an engineer by trade, i can attest to the fact that there is nothing worse than "Bad Data"!

let the learning curve continue (15 minutes of poke&hope don't cut it) ops:

Thanks Jim!:T

btw, there were five or six 'lubricated' philes along with my my best Dawg in the room during an audio 'event' at my place when the sweep was made.
all of us newbees at REW.

[jim1961]

a re-do is obviously in order.

as an engineer by trade, i can attest to the fact that there is nothing worse than "Bad Data"!

let the learning curve continue (15 minutes of poke&hope don't cut it) ops:

Thanks Jim!:T

btw, there were five or six 'lubricated' philes along with my my best Dawg in the room during an audio 'event' at my place when the sweep was made.
all of us newbees at REW.

That could do it :lol:

[wkhanna]

Hope to have some new sweeps next weekend when Dan brings the mic over.
He has been busy lately....like going to Philadelphia last weekend to see Melody Gardot...I am soooo jealous!:B

[cjd]

I just went through this myself for a first-pass at sub tuning and EQ - kept having to re-do things. The biggest grump is I only found one calibration file (am using a UMM-6) which I assume is not a 90deg cal, but even the current result after working through some sub EQ (via miniDSP) is worthwhile. Still things to figure out, like how to measure sub + (single) main using ASIO drivers...

I found it interesting that I was successful tuning delay on my subs by ear using the standard level-setting noise generated by the Emotiva pre-pro (one in the front wall, 4x Tempest-X 15 in 300cu/ft+, one just behind the front row of seats using the riser as a box for 2x Ultimax 15 in 50cu/ft.) Delay was later confirmed by measurements (man, is that slow work.)

This will be helpful when I get to working on room treatments.

[PewterTA]

Hey Everyone, Bill told me about this thread I just didn't have time to go and find it. Wasn't sure where he put it on here.

From what I know/remember of doing his (I didn't do it our friend did) I'll best try to fill in any answers I can.

...And yes, we'll have to re-do his sweep just to make sure that we have all the right settings.

The mic is a Calibrated MiniDSP UMIK-1 that we have that was calibrated by Cross-Spectrum and we have multiple calibration files. We have the typical ones that come from MiniDSP plus we have both Narrow Band and one third octave calibration files for 0, 45, 90 degree.

Bill answered most of the questions it looks like you guys had, I think we'll need to redo the sweeps with each channel separately to know what we are really seeing. I (so far) have run all mine, another mic than what was used at Bills, but same thing, and am getting very similar results. I have done all 512 sweep length. I've done each set at the recommended degrees with the calibration files.

I didn't do each channel separately as I've not had a chance to sit down and really go through everything, but I plan on it and hopefully we can sit down at Bill's and do the test as best as we can the proper way with all the correct settings.

Just an example of what I'm seeing in my room...

https://www.dropbox.com/s/u2rxlq9xxt8x72c/Sep%2030%2020_05_20.mdat?dl=0

-Dan

[jim1961]

Hey Dan,

To my eyes, there is a world of difference between your mdat and WK's. WK's was polluted with noise. This one at least looks believable.

Measuring each channel separately is the way to go. Cant trust the top end otherwise. Settings wise, it seems you had it right.

This is optional, but IMO, more accurate, and that is to put the mic tip pointing up and using the 90 degree cal file where the ear entrance would be for each ear, for each channel. After all, ones ears are not in the middle of our head

Middle of the head for each channel is OK though.

[jim1961]

Any movement on this topic? New Measurements yet? Noticed it was 6 weeks ago when anyone last checked in.

[wkhanna]

sorry.....not yet.
heading to Dan's today, maybe he has some new sweeps to post.
I will give him a poke in ribs to get some data posted.

[fbov]

As someone who also has a pair of 65L, bottom-ported, NatP towers, and a measurement rig, I would suspect your data on other grounds... it doesn't look right, compared with mine.





That's not to say I'm well versed in audio measurement, but the data are consistent with the requirements for high quality audio, and a trip to MWAF showed that was the case. I bet your ears would notice if this were an accurate measurement!

HAve fun,
Frank