Calibration conundrum

[agarwalro]

I have been using the REW + Umik1 setup and always used the 0 Degree Narrow Band mic calibration file. I kept the mic on axis and fixed for the duration of the testing to prevent variation across measurements.

Today out of sheer curiosity (and being sick of moving the mic for each speaker) I decided to try the 90 cal with the mic vertical. The difference in readout is significant,

The conundrum is: Which measurement correctly describes what I am hearing?

I only found a couple of threads on the Dirac forum what confirmed that 90 Degree cal needs to be used with vertical mic orientation. There is also some opinion that room response is best measured with the mic vertical. But, I was not able to find any concrete information.

I understand that gated tests should be done with the on axis orientation using 0 degree calibration file. I assumed this orientation would also serve as the best for room response measurements. Is this an incorrect assumption?

 

[crossedover]

I'd say the 90 degree is more accurate to what your hearing with more of the room interactions. At what distance is the 90 degree taken from?

 

[agarwalro]

The setup is pictured in this post, http://forums.audioholics.com/forums/general-av-discussions/86935-what-new-stuff-have-you-bought-if-you-care-share-thead-29.html#post1038858

The mic is about 5 ft. from the speakers. I have angled them so that the tweeter axis matches my ear height. In the vertical position, the mic's top is at ear height.

 

[crossedover]

If your running measurements again, try them not angled and see how the response looks.

 

[agarwalro]

I'll try it tonight.

 

[rojo]

Also, I recommend trying three different measurements at each angle, and compare the averages of each set. For each set, first measure in the primary spot, then about a foot to the left, and a foot to the right for the third. There could actually be room response variations like you graphed within a few inches span. How wide is your head? Also, make sure you turn off your hvac and reduce the noise floor as much as possible as you measure.

Seems like I read somewhere that 90° is most appropriate for measuring bass response, and 0° for full range. Seems also that I saw graphs comparing the two, and the graphs were pretty much identical below 1kKz. I'm pretty sure the difference you see is more likely explained by room response variations than with measurement angle. Also, I think many dips and peaks that look drastic on a graph are largely inaudible.

 

[agarwalro]

Also, make sure you turn off your hvac and reduce the noise floor as much as possible as you measure.

Always, I even wait for the fridge to turn off. Makes for a long measuring session, my last one started at 11am and at 4pm I was too burnt out (read, drunk) to continue. The noise floor is about -50 to -55 dB. My linear measurement levels make the noise floor irrelevant.

Also, I recommend trying three different measurements at each angle, and compare the averages of each set. For each set, first measure in the primary spot, then about a foot to the left, and a foot to the right for the third. There could actually be room response variations like you graphed within a few inches span. How wide is your head?

I'm not so worried about the exact measurement as much as which one to rely on. Also, the setup has a seating of exactly 1. I me and my music/movie. The recliner is exactly wide enough to accommodate up to 2 people squished together .

Seems like I read somewhere that 90° is most appropriate for measuring bass response, and 0° for full range. Seems also that I saw graphs comparing the two, and the graphs were pretty much identical below 1kKz.

Bass frequencies being omni directional in a small room setting, I'm confident mic orientation and calibration file use will make little difference to the measured room response. Makes for a good exercise in confirming the physics.

I'm pretty sure the difference you see is more likely explained by room response variations than with measurement angle.

I feel there is more than that going on. This evening's tests will help better understand what is going on.

Also, I think many dips and peaks that look drastic on a graph are largely inaudible.

I'm perfectly happy with the sound, it is hugely improved from the previous setup. The measurements that have got me a bit OCD . They tell me there is room for improvement. That said, I'm not chasing measured perfection. I'm just trying to use the measurements to ensure I have optimally set up the speakers.

 

[fuzz092888]

Honestly, I don't see enough variation there to get worried about. Differences in the mic calibration files, plus catching room modes and reflections a little differently, and even small variations in the mic itself can all be possible reasons for those differences.

If you really want to know what you're getting at your ears you'll need one of the below. Short of that, I'd be pretty happy with either measurement (not that they're hugely different) sans room treatments.

 

[mtrycrafts]

...

The mic is about 5 ft. from the speakers. ...

That is only 1.5 meter. I would think that this is too close to get room responses and may reflect the mics response in the two positions.
Is 5 ft your normal listening position? If so, I would think you are mostly getting direct waves.

 

[agarwalro]

Honestly, I don't see enough variation there to get worried about. Differences in the mic calibration files, plus catching room modes and reflections a little differently, and even small variations in the mic itself can all be possible reasons for those differences.

I read your post and was closely looking at the graph again. I think above 500 Hz they are close enough. In fact, I had an epiphany that the wavy/rolling hills aspect of the red trace is probably due to reflections from the couch back since it rises over my listening height. These are absent from the black trace since the mic is pointed at the tweeter. Below 500 Hz the two traces vary a fair bit and at this point it is an academic exercise to understand the measurement process's systemic flaws better. I suspect your assessment regarding the mic (actual condenser capsule) variation may partly be responsible.

I'm going to test the couch back's reflection hypothesis and see what it produces.

That is only 1.5 meter. I would think that this is too close to get room responses and may reflect the mics response in the two positions.
Is 5 ft your normal listening position? If so, I would think you are mostly getting direct waves.

Yes, 5 ft is my normal listening position. The speakers are about 3.5 ft from sides and back wall and my listening position is close to 38% away from front wall (the room is 17 ft long).

I'm pretty sure a majority of the sound (above 200 Hz) is direct since I have them heavily toed in and angled down to control room reflections without treatments. The black trace is a study in room response impact. As one goes from 20KHz to 10 Hz, the magnitude of the peaks and dips increases for every decade. By the time one is at 75 Hz, we get the biggest dip of -10 dB.