advertisement


Upgrading my acoustical room treatments

I finally got my subs out of my room, so I now have more space to experiment with speaker positioning!

First stage of testing: measuring effect of front wall.

Speakers placed 103cm from side walls (25% of room width) and in line with front of hifi rack (speaker baffle 55cm from front wall). Speakers then moved back towards wall in 5cm increments until they can't get any closer to wall due to skirting boards. Mic is at my usual LP, 120cm from the back wall and 103cm above the floor.

As for room treatment, as well as the usual GIK corner traps running floor to ceiling in all four vertical corners and GIK 244 panels at side wall midpoints, I also have two GIK Monster traps on the back wall, centred behind the listening seat and stood about 15cm off the wall.

This isn't exactly representative of how I'd be listening to the speakers, I normally toe them into aim at the LP and I normally tilt them backwards to optimise phase at the MF/HF crossover frequency, but it would've been too time consuming to maintain these exact angles across multiple tests. For quickness these measurements are with the speakers parallel to the front wall and with no tilt.

The measurements in the graph below are for both speakers playing together, not individually, and with no averaging. The results suggest that the null at 135Hz gets better as the speaker is moved closer to the front wall.

I'll continue these tests tomorrow by moving the speaker forward to see what happens when the distance to the front wall is greater than 55cm.

EDIT - For quickness these tests were conducted using pink periodic noise and REW's RTA with 1/48 smoothing. I then applied additional 1/12 smoothing in post-processing. This makes the overlays easier to read at the cost of evening out the size of the differences between the different speaker positions.

attachment.php
 
Last edited:
The results suggest that the null at 135Hz gets better as the speaker is moved closer to the front wall.

Not only that but 210Hz comes down a bit smoothing that area out quite a bit. Interesting that nothing at all happens at <75Hz. I wasn’t expecting that. To my eyes the green line looks the flattest over the majority of the range, which is no surprise as I’m pretty sure that is where they are designed to work. I’d be inclined to run with that for a while and next try altering the distance from the corners, and also seeing where the listening seat now belongs.
 
First stage of testing: measuring effect of front wall (continued).

Picking up where I left off, I continued to measure the speakers from 35cm to 160cm from the front wall in 5cm increments. Graphs below show stereo speaker measurements in 10cm increments, the overlays are too messy if I include every 5cm increment.

Summary of results:
- The 75Hz null is smallest the closer the speakers are to the front wall.
- At distances greater than 100cm, the 75Hz becomes too deep to accurately measure (the noise floor of my room is louder!), hence the measurements from 100cm to 140cm all look similar at 75Hz.
- The 75Hz null starts to improve again at 160cm, but this distance is too close to the middle of my room and catches the axial length mode null at 43Hz.
- The upper bass response improves as the speakers get further from the front wall. Between 85cm and 95cm produces the flattest response, 100Hz-400Hz +/- 7.5dB with 1/48 smoothing.

So at the moment it seems there’s a trade-off between smooth bass from 100Hz-400Hz and a big null at 75Hz.

Next step will be measuring the effect of varying the speakers’ distance to the side walls, and lastly varying the distance of the listening position.....and then I'll watch the FR go to shit when I toe-in and tilt the speakers! :D

attachment.php


982645d1632942570-room-treatment-advice-help-interpreting-giks-published-lab-results-varying-stereo-speakers-distance-front-wall-pt1-speakers-103cm-side-walls-mic-120cm-fr.jpg


982646d1632942570-room-treatment-advice-help-interpreting-giks-published-lab-results-varying-stereo-speakers-distance-front-wall-pt2-speakers-103cm-side-walls-mic-120cm-fr.jpg


982647d1632942571-room-treatment-advice-help-interpreting-giks-published-lab-results-varying-stereo-speakers-distance-front-wall-pt3-speakers-103cm-side-walls-mic-120cm-fr.jpg


982648d1632942571-room-treatment-advice-help-interpreting-giks-published-lab-results-varying-stereo-speakers-distance-front-wall-pt4-speakers-103cm-side-walls-mic-120cm-fr.jpg
 
Last edited by a moderator:
Second stage of testing: measuring effect of side walls

Keeping the speaker as close to the front wall as possible (35cm), I varied the distance to the side walls in 5cm increments from 60cm to 150cm. 150cm means the speakers are almost touching my hifi rack!

Graphs below show stereo speaker measurements in 10cm increments because again there are too many overlays for every 5cm increment.

Summary of results:
- The 75Hz null is also affected by side wall, not just front wall.
- The 75Hz null is deepest when speakers are 120cm from side walls.
- Trade-off between smooth bass response below 100Hz and smooth bass response above 100Hz.
- Smoothest upper bass response is obtained when speakers are around 100cm from side walls. 100Hz-400Hz +/- 9dB with 1/48 smoothing.

Next step is measuring effect of changing the distance of the LP to the back wall.

attachment.php


attachment.php


attachment.php


attachment.php


attachment.php
 
...absorption between 100Hz-250Hz, so I’d expect it to help here. Unless the only way to improve the 150Hz-250Hz dip is to absorb the fundamental standing wave frequencies at 42Hz, 45Hz and 53Hz? This goes back to my original question in post #1...
sorry I did not read all too thoroughly and might get this wrong.. This one thing is banging in my head.. what if the corner tritraps are effectively (too) effective in the 100-250Hz range and are causing the dip? Don't be mad at me.. I reckon in your case I would like to see the situation with no treatments whatsoever (maybe you showed this ready?) in order not to mix the problem and the cure and then work from there..
Me Also thinks that the peaks at about 42/45 and 52Hz (I'd consider maybe 43,5Hz at first tentative) could be addressed with a Helmholz resonator in the closet.. I'd bet my 5c that you will get smoother response in the multliplies of the Helmholz frequency too..
 
what if the corner tritraps are effectively (too) effective in the 100-250Hz range and are causing the dip?
This thought crossed my mind also, a couple of years ago. I thought the corner traps could be over-absorbing the tangential modes and not absorbing enough of the axial modes. But I did not test my suspicion by removing all the corner traps (I am disabled so need physical assistance with all this stuff).

However I've since done lots more research and, almost everyone who uses acoustic treatment, almost always uses treatments in the corners, since it is a hotspot where all room modes gather.

If I can convince my dad to help me, I might still try removing the corner traps, but I'll have to ask him very nicely! :D

Me Also thinks that the peaks at about 42/45 and 52Hz (I'd consider maybe 43,5Hz at first tentative) could be addressed with a Helmholz resonator in the closet.. I'd bet my 5c that you will get smoother response in the multliplies of the Helmholz frequency too..
I asked on Gearspace if absorbing the fundamentals at 42Hz/45Hz and 53Hz would also help the multiples. The replies I got said I have wishful thinking, and that's not how room acoustics work. I still wonder if it would work though?
 
Which one, 75Hz, 95Hz or the broader one at 150Hz-200Hz?

The big 150-200Hz one. That’s the deal breaker to me, the rest looks pretty benign on paper unless you get obvious honks or booms. I don’t mind a bit of tonal unevenness, my main system certainly has a bit of a ‘loudness contour’ thanks to the rear wall, but it translates as low-end heft and scale at the lowish volumes I listen at, not boom or honk. I actually like it!
 
@ToTo Man .. I am no expert here I am afraid, however I think that until we are pros a lot of empirical try and see is necessary to (try to) get a handle of it.
 
I've been looking back over your 66 refurb thread, the first few far-field Hf/mf axis measurements are interesting when compared to where you are now.
 
I've been looking back over your 66 refurb thread, the first few far-field Hf/mf axis measurements are interesting when compared to where you are now.
Can you link me to said graphs, Frank? I took a lot of those measurements with the speakers raised on 25cm plinths, which significantly changed the response around the LF/MF crossover area. I also currently don't have the speakers aiming directly at the mic or tilted backwards, the combination of this plus the fact that the grilles are on result in substantially attenuated output at the MF/HF crossover area (plus the fact that the measurements I'm taking at the moment are with both speakers playing together which is causing comb-filtering havoc in the HF!).
 
@ToTo Man could you post the "left speaker" and "right speaker" plots for the 35/100/120 plot below together with the "both speakers", pehaps applying 1/12 octave or Variable smoothing?

attachment.php
 
@ToTo Man could you post the "left speaker" and "right speaker" plots for the 35/100/120 plot below together with the "both speakers", pehaps applying 1/12 octave or Variable smoothing?

attachment.php

51542895841_9281f0250b_o.jpg


And this is the measurement with the speakers the same 100cm distance from the side walls but toed-in towards the listening seat and tilted backwards a few degrees for improved treble response (this means the inside edge of the speaker is closer to the front wall and the outside edge of the speaker is further away from the front wall):

51543659199_2cdc3c0098_o.jpg
 
An interesting but somewhat obvious side note musing:

In order to make experimenting with different positions and layouts easier I removed clutter from the room, including side tables with headphone amps on, a fabric reclining chair that lived in a corner, and also my two big (unused) subwoofers. The low end gets worse each time something is removed from the room, becoming less damped and taking longer to decay. I've said it before but I continue to believe that, in lieu of extensive room treatments and dense soft furnishings, having unused speakers in the room makes a positive contribution to low-end control. I noticed a huge difference the instant I removed my huge Tannoy Edinburghs from my room last year. Decluttering has been good from a practical POV as it's made the room tidier and more spacious and has reduced my risk of tripping over stuff, but the low frequencies have definitely suffered. It'll be interesting to see how many additional bass traps I'll need to restore the balance, and how much my wallet will take a hit as a result!...
 
51542895841_9281f0250b_o.jpg


And this is the measurement with the speakers the same 100cm distance from the side walls but toed-in towards the listening seat and tilted backwards a few degrees for improved treble response (this means the inside edge of the speaker is closer to the front wall and the outside edge of the speaker is further away from the front wall):

51543659199_2cdc3c0098_o.jpg

Thanks.
Could you confirm that you measured the two speakers together instead of using REW's "average responses" feature?

The circa-15kHz peak in the first both-speakers graph is a bit strange, as is the dip in the left speaker at 150Hz.
There's the shelved down region between 2kHz and 10kHz but the mid- and upper-midrange looks (much?) better.
Which one do you prefer? I would probably try to bring down the 500Hz to 2kHz region perhaps by reducing the toe-in and/or getting rid of the tilt. I don't think that the changes above 10kHz are really worth worrying about, it's "air" territory.
 
An interesting but somewhat obvious side note musing:

In order to make experimenting with different positions and layouts easier I removed clutter from the room, including side tables with headphone amps on, a fabric reclining chair that lived in a corner, and also my two big (unused) subwoofers. The low end gets worse each time something is removed from the room, becoming less damped and taking longer to decay. I've said it before but I continue to believe that, in lieu of extensive room treatments and dense soft furnishings, having unused speakers in the room makes a positive contribution to low-end control. I noticed a huge difference the instant I removed my huge Tannoy Edinburghs from my room last year. Decluttering has been good from a practical POV as it's made the room tidier and more spacious and has reduced my risk of tripping over stuff, but the low frequencies have definitely suffered. It'll be interesting to see how many additional bass traps I'll need to restore the balance, and how much my wallet will take a hit as a result!...

Why not try a touch of digital (or even analogue) room EQ to tame the peaks in the low- and sub-bass?
 
Thanks.
Could you confirm that you measured the two speakers together instead of using REW's "average responses" feature?
The circa-15kHz peak in the first both-speakers graph is a bit strange, as is the dip in the left speaker at 150Hz.
There's the shelved down region between 2kHz and 10kHz but the mid- and upper-midrange looks (much?) better.
Which one do you prefer? I would probably try to bring down the 500Hz to 2kHz region perhaps by reducing the toe-in and/or getting rid of the tilt. I don't think that the changes above 10kHz are really worth worrying about, it's "air" territory.
Yes, both speakers together, and this is why I dislike measuring both speakers together when interpreting HF information because millimetre changes in speaker location cause huge changes in HF comb-filtering. I prefer looking at individual left and right speaker measurements for HF analysis.

The above measurements show the speakers tilted backward 15mm. I have now tilted them the full 30mm like I had done before I started doing all these new measurements and the output around 5kHz is now significantly better.

The differences in the <200Hz response of the two speakers is probably due to room geometry (remember I have a bay window opposite the left speaker and a door to the right side of the right speaker.

The following graph compares the EQd response of the speakers in their old position (64cm from side walls and toed-in to cross 2ft in front of listening position) vs their new position (103cm from side walls and toed-in to cross 4ft BEHIND listening position):

51546333123_ea8e556397_o.jpg


The centre imaging in the new position is much more consistent and convincing. However, the mid frequencies are more coloured sounding. I think adding absorption to the front wall and side wall reflection points, and possibly also the sides of my hifi rack will help smooth out the mids. I'm also finding the speakers' tonality much warmer in this new position, probably because the 170Hz-300Hz area is no longer scooped out. It's taking my brain some time to adjust to this new balance.

Why not try a touch of digital (or even analogue) room EQ to tame the peaks in the low- and sub-bass?
I am already using EQ. Additional bass trapping will (hopefully!) bring the benefit of reducing decay times in the low-bass to make it tighter.
 
Last edited:
Yes, both speakers together, and this is why I dislike measuring both speakers together when interpreting HF information because millimetre changes in speaker location cause huge changes in HF comb-filtering. I prefer looking at individual left and right speaker measurements for HF analysis.

I’d not understood that significance until now, and it makes total sense. I’ll start doing the two speakers then average thing from now on whenever I measure anything.
 


advertisement


Back
Top