Home Theater Forum and Systems banner

21 - 24 of 24 Posts

·
Registered
Joined
·
66 Posts
Discussion Starter #21
Thanks jtalden. I am amased how much people are able to read out of one measurement. I will certainly have to dig in deeper to get the broader picture.

The calibrated mic file for the Dayton UMM-6 is an on axis measurement only. When I am going to measure my room response the measurment will be off axis, and technically the calibration file could be off. Just something that I have read on this forum. Could I make an off axis calibration file by doing the same measurment with a calibrated on axis aim, and an off axis aim, and then extract one off the other to create an off axis calibration file?

When I am going to do the room response with my 5.1 channel setup, am I supposed to be doing a measurment for every single speaker at listen position with the mic in the same postion, which the recommended postion seems to be aiming at the ceiling or at the corner of the front wall and the ceiling.

My apologies for asking simple questions. I haven't read that far ahead yet.
 

·
Premium Member
Joined
·
1,562 Posts
Measurements for EQ done at the LP do not need to be at the same position. The mic can be pointed in the general direction of the speaker using a 0° cal file. Pointing it within 30° of the direction of the speaker has almost no influence in my room so I don't think small errors in this regard would normally be expected to be a problem.

It is even preferable to average several measurement at different positions around the listening position. If we are interested in optimizing for one seat we can stay within 0.5m of that LP. If we spread the measurement area to accommodate 2 or more seats, it will be at some detriment to the main seat. The amount of detriment is related the particular room setup.

The advantage of averaging is in smoothing out the SPL measurement in the midrange where reflections occur. This makes it slightly less likely we would be tempted to over EQ for a particular SPL aberration. The advantage can be small and unnecessary depending on the room setup however. It also can help slightly above 1k particularly in case where there are strong early reflections/diffraction from the speaker box.

There is also the "Moving Mic Measurement" (MMM) method that does the same thing by using "forever" averaging within the RTA feature. A browser search will detail this. Here is one link.

The point is that small mic position changes can influence the measurement and averaging can improve the repeatability of the measurement while finding the central tendency. Averaging is not required for good results but it does help explain why the mic can be in slightly different positions for different speaker measurements.

See here regarding the other question about creating a 90° cal file. It can be done very satisfactorily if we use good techniques. I have compared my results to a supplied 90° cal for my mic with close agreement. My method was basically similar to this reference although I used other methods for smoothing the results.
 

·
Registered
Joined
·
66 Posts
Discussion Starter #23
Thank you all. I will read through all the information and do a complete system measurement once I have everything complete again. At some stage I will probably get a different mic when I understand all of this better.

One last question, does smoothing the curve have any influence on frequency limit like with gating? I remember reading somewhere where they gated and smoothed the curve, which pushed the frequency limit way up into the midrange.
 

·
Premium Member
Joined
·
1,562 Posts
One last question...
We've heard that before. :D

...does smoothing the curve have any influence on frequency limit like with gating? I remember reading somewhere where they gated and smoothed the curve, which pushed the frequency limit way up into the midrange.
That's correct, gating the IR will limit the lower frequencies. Assuming left window is very close to the start of the IR REW will properly truncate the SPL and Phase chart at the lowest freq that is usable. The first octave or 2 displayed is overly smoothed so its best to keep that in mind.

If the left window is set large then chart shows more low freq than there is data for. REW can't tell if you are including meaningless (zeroed) data before the IR starts. In your Post 3 measurement with windows set to 1 and 3 REW will show the low frequency limit as 250 Hz. If 0.1 and 3 are used then the low limit is calculated as 323Hz. With either setting I would suggest the low limit is 323Hz and the data below 2x323 (646Hz) is overly smoothed. As natehansen66 suggested it is convenient to use 1ms for left window for several reasons. I use that setting most often.
 
21 - 24 of 24 Posts
Top