• WANTED: Happy members who like to discuss audio and other topics related to our interest. Desire to learn and share knowledge of science required. There are many reviews of audio hardware and expert members to help answer your questions. Click here to have your audio equipment measured for free!

Trying to simulate presence of body at listening position for measurements

sejarzo

Addicted to Fun and Learning
Forum Donor
Joined
Feb 27, 2018
Messages
961
Likes
1,066
My apologies if this has been thrashed out elsewhere.

I have seen various opinions on mic measurement angle, and most tend to lean toward 90 degrees. I thought that might also help with the fact that there is obviously a body at the main listening position when the system is actually playing music, so I ran a quick test. First, some background...

The listening room is not the best when it comes to dimensions, 21 feet long, 17 wide, and 7 feet high, in a basement where there's a post almost directly in the center. The only way to get a reasonable symmetric arrangement is to place the sofa up against the post. The speakers are 76" center-to-center and 91" from the center of the sofa. Due to enormous slap echo, I put in a lot of 2" thick rigid fiberglass absorbers (homemade...LOL no way could I afford to have bought all of it otherwise!) All things considered, I get far less difference left vs right than I see in other room measurements.

IMG_20230208_142211190.jpg
IMG_20230208_142222422.jpg


Here's the mic placed without any pillows...

No pillows.JPG


and here it is with some pillows placed to roughly simulate the presence of a body. I didn't touch the mic at all, and while it looks as if it's not at 90 degrees, that's not the case.

Pillows.JPG



I ran multiple sweeps in REW (matching was dead on up to 15 kHz) and applied psychoacoustic filtering. Here are the resulting traces:

res crop.JPG


Ok, yeah, I know this is a single point...but look at the massive differences in SPL from 600 Hz and 1200 Hz between the two.

I measured this prior to implementing the sub crossover on the DDRC-24 and running Dirac Live. My plan was to run Dirac for a tightly focused central listening position. We're empty nesters, and I am the only one who uses the system. My understanding is that regardless of which correction algorithm is used, they tend to give more weight to the central measurement.

If the response in that critical frequency region can be shifted by as much as 6-7 dB by the presence of those pillows, does measuring response over a reflective leather sofa without any bodies present, even using the mic at 90 degrees, create anomalies that present significant problems in developing a proper correction?
 
OP
sejarzo

sejarzo

Addicted to Fun and Learning
Forum Donor
Joined
Feb 27, 2018
Messages
961
Likes
1,066
My first attempt at implementing Dirac Live on the mains only, no sub, following the 9 point process recommended for my desired result, resulted in a huge smearing of the central image with vocals pushed back way in the mix. Will try again, but my experience in room correction to date has been hugely frustrating.

Back story, for what it's worth...

This is how the room was configured originally:

HT and Music Room C.jpg


I couldn't get it to sound "stable" with the right speaker only a couple feet from the side wall and the left around 13-14 feet from the wall on that side. The absorber on that side was carefully placed at the first reflection point, but really didn't make a noticeable difference.

When the room and gear were in that configuration, I bought a used TacT Audio RCS 2.0 in an attempt to correct just the left and right without the sub, just to see how much it might fix things. It was a disaster. Odd out-of-phase issues at certain frequencies made everything sound worse, and resetting the thing and redoing the measurement and correction process three times didn't change the outcome. I even had the guy at their HQ in the US on the phone with me on my final attempt, and he could hear how bad the result was even over a landline in mono. I sent the unit back, and he installed it in a system there and claimed it was just fine, running perfectly, but he'd run it for a few days to be sure there wasn't some interim issue. Two days later, he called and said that when he turned it on that morning, it had the same weird phase problem I had complained about...and to make matters worse, he said they didn't have a clue why, or even any ideas how to attempt a fix. Eventually the seller took it back from me and refunded my money. No idea what he managed to do with it after TacT sent it back to him.

I ended up selling that MF CD-Pre, Outlaw ICBM, and the monoblock amps shown there that I used for music listening and bought an Onkyo TX-SR875 on closeout in early 2010 to replace the older Onkyo AVR that was only used for movies. I obviously hoped that the Audyssey version implemented in that unit would end my woes. Nope. End result was a smeared image and overly bright, which I soon found was a common complaint on the AV forums. Ergo, the system fell into disuse as I turned to headphone listening.

The HDMI board in the Onkyo succumbed to heat just as it has in virtually every TX-SR875, followed by the TV giving up the ghost. I finally decided to get a couple of Emotiva PA-1 amps a few years ago on sale, and picked up a used DDRC-24 just before the home theater became a storage room for an elderly parent's junk. My current effort is the last attempt to make everything sound good enough to enjoy it rather than constantly hearing the problems before we downsize out of this place in a couple of years or so.

I've followed the Obsessive Compulsive Audiophile channel on YouTube, and he has recently presented a correction protocol that uses only free tools to create a combined convolution .wav that implements a virtual bass array, frequency correction, crossover correction, and excess phase correction. His latest recommendation for correction at a single point listening position call for measurements of left channel and right channel at only three points: dead center, left ear position, and right ear position. The center measurement gets double the weight of the left ear and right ear when they are averaged to develop the correction. That's why I thought I would see how much the center measurements are affected by a simulated body in the listening position.
 
Last edited:

alex-z

Addicted to Fun and Learning
Joined
Feb 19, 2021
Messages
910
Likes
1,684
Location
Canada

First, I assume you loaded the corresponding 90 degree calibration file when the mic is orientated as such?

2" absorbers are not bad for general purpose use, but you should go thicker when possible, particularly for bass trapping the corners. Low frequencies are the most problematic in residential rooms, and fixing them via EQ doesn't solve the time domain issues that accompany them.

Something which can be particularly effective with drop ceilings is removing some of the panels, stuffing the cavity with mineral wool, and then using fabric covers. That turns your ceiling into a major bass trap while being cosmetically benign.

Pillows are a poor form of emulating a body, they absorb substantially more energy than a human, as seen by the smoothing effect in the 500-10000Hz region. I would honestly just recommend placing the mic atop your head, pointed at the speakers. You will see a difference in the response compared to the "bare mic" response, but that is expected. Our brains process the audio received by our ears, to compensate for the time and frequency shift caused by our skulls and ear structure. That is the magic behind stereo imaging. "Psycho-acoustic" smoothing does not actually mimic our ears, it mimics the bandwidth limitations we have.

I have a personal policy of not correcting any issues above 500Hz unless they cover at least 1/3rd octave. Often, the results are negative, or insubstantial at best.

OCA on Youtube has good information about the more advanced uses of REW. However, these correction techniques are not a replacement for the fundamentals, such as low frequency treatment, multi-sub integration, and speakers with good directivity.

I keep mentioning low frequencies because they have a surprising impact on your stereo imaging. Excess energy at 60-300Hz causes perceptual masking that extends well into the mid-range.

Something else, it is worth checking the speakers individual response with the grills removed. Virtually all speakers have grills which detract from the sound quality.
 
OP
sejarzo

sejarzo

Addicted to Fun and Learning
Forum Donor
Joined
Feb 27, 2018
Messages
961
Likes
1,066
Thanks, Alex. Yes, the proper cal file was used for 90 degree measurement. The main speakers are Paradigm Studio 40 v3, which were designed to be used with the grilles on, and you can see why. The grilles "fill in" that area between the protruding driver frames and the edge of the front baffle with a smooth curve. Many folks commented over the years that they look a lot cooler with the grilles off. These are ca. 2003 speakers, and I recall Paradigm commenting on various forums at the time that the grilles should be used--and I think I recall they posted graphs to prove the point.


40v3 no grille.JPG


The bass trap on the left corner is 4" thick, and mounted at a 45 degree angle with an air gap behind it. Unfortunately, I couldn't make the one near the door on the right side that thick due to the door.

I should probably emphasize I'm just trying to get to a reasonable solution for the limited time we will be in this home before I have to sell off pretty much everything I have. Given my previous post, I'd at least like to glean a bit of enjoyment from all the time and money I spent before I will no longer have a dedicated listening space due to liviing in a condo.

Meet "Buddy"...yeah, he's just kind of a modified pillow, he's certainly no ISO certified apparatus.

Buddy.JPG



Results showed a little less variation, but there were still swings of +/- 5 dB from 600 to 2000 Hz with Buddy in place versus an empty sofa. Buddy was not entirely absorptive--he created a couple of 4 db peaks where the "empty sofa" was flatter.

Here are the filters that REW calculated for the right channel only when I averaged left ear, center, and right ear positions as suggested by OCA plus his new suggested target curve and other settings, with Buddy in place versus an empty sofa:

Different filters.JPG


The filter around 500 Hz is similar, but REW calculated significantly different filters in the 700-2000 Hz range.

Perhaps this is just proof that trying to implement precise correction above a certain point is not a simple matter?
 
Last edited:

thewas

Master Contributor
Forum Donor
Joined
Jan 15, 2020
Messages
6,741
Likes
16,174
If the response in that critical frequency region can be shifted by as much as 6-7 dB by the presence of those pillows, does measuring response over a reflective leather sofa without any bodies present, even using the mic at 90 degrees, create anomalies that present significant problems in developing a proper correction?
I have a personal policy of not correcting any issues above 500Hz unless they cover at least 1/3rd octave. Often, the results are negative, or insubstantial at best.
Currently I correct above 500 Hz only based on direct sound (=listening window from anechoic or gated measurements) so its not an issue anymore.
 
Top Bottom