advertisement


Measuring the effect of acoustic treatments in the room

Having more or less concluded measuring the effect of treatment on each of the four walls of my room I returned to experimenting with loudspeaker placement.

I changed from the “rule of thirds”, which had speakers and listening position 138cm from front and rear walls respectively, to the “Real Traps 38% Option 2”, which brings the speakers and listening position even closer to the middle of the room. The speakers are now 147cm from the front wall and the listening position 157cm from the rear wall. The speakers are also closer to each other in this setup.

The arrangement looks very claustrophobic from the listening position, and this is largely how it sounds. Even after dialling in the EQ to look good on paper, the presentation is very forward and is verging on being shouty in the mids. Treble output has also increased relative to bass, but the presentation doesn’t actually sound any brighter up top to my ears. In fact, I heard more top end clarity and sparkle when I was sitting further away, so the perceived forwardness of the mids must be masking the highs. Centre imaging is on a par with the “rule of thirds” configuration but the soundstage is narrower and closer to the listening seat. It’s possible that toeing them in more might alleviate the ‘shoutiness’, but I don’t really have the enthusiasm to persevere with this configuration given that the null at 70Hz has become a sink hole that's swallowed the drummer's kick drum!

I’ll continue experimenting with different positions, including Tony’s diagonal suggestion if it’s feasible without moving too many treatments out of the way, but I think we all know where the speakers are going to end up… (HINT: two words, first word begins with ‘F’, second word begins with ‘W’). ;)

Real-Traps-38pc-Option2.jpg



01-FR-L-R-Avg-Real-Traps-38-Option-2-Spkrs-147cm-from-front-wall-mic-157cm-from-rear-wall.jpg
 
Last edited:
I was reluctant to try the ‘Cardas' placement as I thought I’d be wasting my time. It seemed counterintuitive to place the speakers so far away from the front wall and the listening position closer to the rear wall. I’m glad I did try it because I was rewarded with a presentation that’s fluid, wide open and spacious. This is by far the most ‘invisible’ the speakers have been in the soundstage. I’m not sure how much this should be attributed to their placement or the fact that I’m sitting just two and a half feet way from the QRD diffusers on the rear wall (which is much closer than the minimum recommendation), but it’s very addictive! The phantom centre image is at least as good as in the other placements I’ve tried, if not slightly better. The only obvious attribute this placement is deficient in is bass, it’s the leanest sounding configuration I’ve tried. On the positive side, there are no nasty low end peaks that need to be flattened with EQ apart from my 53Hz axial height mode. As you can see in the graph below, there’s quite a difference in the low end response between the ‘Cardas' placement and ‘Real Traps 38% Option 2’ placement:

Cardas.jpg


01-L-R-Avg-Real-Traps-38-Op2-vs-Cardas.jpg
 
The journey to the front wall has begun with a leisurely pit stop at the "Rule of Thirds - 29% Version". This arrangement looks more conventional and is definitely more pleasing on the eye as the speakers don't dominate the room as much. In terms of presentation I'd say it's on a par with the standard rule of thirds. Centre image is strong and the soundstage is wide, but not as 'wrap-around' immersive as the Cardas placement. In terms of frequency response it's an improvement on the rule of thirds as it's less lumpy between 100Hz-300Hz (I expect to see the FR continuing to improve as I move the speakers closer to the front wall). The graph below shows the frequency response attained in each of the four placements tried thus far, I've separated each measurement by 10dB as it's quite difficult to decipher four responses piled on top of each other:

29pc-placement.jpg


01-L-R-Avg-33-vs-38-Op-2-vs-Cardas-vs-29.jpg
 
Could you produce a plot with "33% rule" and "Cardas", L & R single speaker, psychoacoustic smoothing?
 
Could you produce a plot with "33% rule" and "Cardas", L & R single speaker, psychoacoustic smoothing?
Lspk measurement for "33% rule" overlaid onto Lspk measurement for "Cardas" and Rspk measurement for "33% rule" overlaid onto Rspk measurement for "Cardas"? Or do you want L+R Avg for "33% Rule" overlaid onto L+R Avg for "Cardas"?
 
This is what happens when I push the Rspk towards the front wall while keeping the mic fixed at 155cm from the rear wall (i.e. the listening position for the "29% rule"). The 80Hz-160Hz area becomes progressively weaker. I suspect this is partly due to reflections/diffractions from the stack of two Scopus boxes that become increasingly closer to the back of the speaker as I move it back. 60cm is as close as I can get the speaker to the wall without moving the Scopus boxes. It'll be interesting to see how much 80Hz-160Hz changes when I move the Scopus out of the way.

Moving-right-spkr-closer-to-front-wall-Scopus-stack.jpg


Below shows how the speakers (L+R Avg) measure 70cm from the front wall after they've been toed-in a little and tilted backwards to time-align the mid and tweeter, and moving the mic forward to 205cm from the rear wall to form an equilateral triangle. 205cm is the centre point of the room, hence the 42Hz axial mode cancellation. I'm not sure why but the phantom centre imaging isn't as convincing or consistent between tracks in this position. I can also hear a chesty colouration/resonance in the lower mids. Apart from the "RealTraps 38% Option 2" this is my least favourite of the placements I've tried.

01-L-R-Avg-spkrs-70cm-from-front-wall-directly-in-front-of-Scopus-stacks.jpg
 
Lspk measurement for "33% rule" overlaid onto Lspk measurement for "Cardas" and Rspk measurement for "33% rule" overlaid onto Rspk measurement for "Cardas"? Or do you want L+R Avg for "33% Rule" overlaid onto L+R Avg for "Cardas"?

L-spk "33% rule" overlaid onto "Cardas" and R-spk "33% rule" overlaid onto "Cardas" if possible. Perhaps a single speaker if that is possible.
 
Green curve R 90FW 155RW looks like a good compromise; you could bring 54Hz, 150 Hz and 205Hz down a bit with EQ (and maybe add a slight up-shelf filter to the whole bass region if needed).
How does it sound?

I have found that in some rooms sticking to the perfect isosceles triangle has not necessarily been the ideal option.
 
33% looks more balanced to me, both individual speaker and average.
Does it sound good with EQ'ed bass peaks?

Maybe you can tame down the BBC dip a little with a judicious mix of toe-in and side panels? (if you don't like what it does perceptually)
 
33% looks more balanced to me, both individual speaker and average.
Does it sound good with EQ'ed bass peaks?

Maybe you can tame down the BBC dip a little with a judicious mix of toe-in and side panels? (if you don't like what it does perceptually)

The online calculator will give you an idea of the width of the reflection zone (how wide an area you need to treat).

When you say "BBC dip" do you mean the 'saddle' between 2kHz-4kHz? That's part of the Dittons tuning, treating the side walls reduces midrange and treble output including the 2kHz-4kHz area so I'm not sure how that would help? I usually just apply a +1.5dB EQ boost there but TBH once I've cut 400Hz-2kHz and 4kHz-6kHz with EQ the saddle between 2kHz-4kHz is barely noticeable. ;)

Looking at the Psychoacoustic plots "29%" looks more balanced to me than 33% in the bass and mids? Of course the Psychoacoustic smoothing glosses over the fact that both positions have a -14dB dip at 70Hz!
 
When you say "BBC dip" do you mean the 'saddle' between 2kHz-4kHz? That's part of the Dittons tuning, treating the side walls reduces midrange and treble output including the 2kHz-4kHz area so I'm not sure how that would help?

The "saddle" happens mostly because of the tweeter's wide dispersion at the bottom of it's passband, or just above the "saddle" which is just below where the mid crosses to the tweeter (5kHz).
If you absorb that off-axis energy you will get a flatter response at the listening spot (as much as it is possible).

This plot is from my own speaker, and you can clearly see the excess energy of the lower tweeter off-axis where the mid-woofer transitions to it, at around 4.5kHz.
This is quite typical of speakers with non-waveguided tweeters, particularly when a large mid or mid-woofer is used.

t6ITild.png
 
The "saddle" happens mostly because of the tweeter's wide dispersion at the bottom of it's passband, or just above the "saddle" which is just below where the mid crosses to the tweeter (5kHz).
If you absorb that off-axis energy you will get a flatter response at the listening spot (as much as it is possible).

This plot is from my own speaker, and you can clearly see the excess energy of the lower tweeter off-axis where the mid-woofer transitions to it, at around 4.5kHz.
This is quite typical of speakers with non-waveguided tweeters, particularly when a large mid or mid-woofer is used.

t6ITild.png
That's interesting, I'll need to read more into this as I'm still not sure I fully understand its implications.

I think in the case of the Ditton the FR of the MF500 mid driver is also contributing to the dip. I measured lots of different MF500 drivers during my Ditton refurb and found significant variation in how early they roll-off up top. Here's two of the most extreme examples, measured nearfield from 1 metre in the with the tweeter disconnected from the crossover. I, of course, chose to install the MF500 drivers with the strongest output above 2kHz! ;)

47971688908_eb30ea30dd_o.jpg


47971679197_91bd3f32b1_o.jpg



This is how a typical MF500 measures in a 12" x 12" cardboard baffle from 1 metre (raw response without crossover):

MF500-raw-response-mounted-in-12-x12-cardboard-baffle.jpg
 
Well, this just goes to show how unpredictable in-room acoustics are... I expected the recessed 80Hz-160Hz bass to improve when I removed the 'obstacle' stack of 42Hz tuned Scopus traps from behind the speakers on the front wall, but the opposite happened:

01-L-R-Avg-Scopus-stack-vs-no-Scopus-stack-behind-speakers.jpg
 
As expected, however, moving the speakers back against the wall does smooth the low end:

01-L-R-Avg-speakers-against-front-wall-mic-205cm-from-rear-wall.jpg


It looks a bit different to the 210RW measurement in post #18, but I think that's due to the different mic and side wall distances used for that measurement.

We'll see things continuing to improve tomorrow as I move the mic closer to the front wall and tweak the speakers' distance to the side walls and toe-in.
 
Last edited:
After more than a few minutes of head-scratching over why I was unable to reproduce the measurements in post #18 the penny dropped… those measurements were taken WITHOUT treatments in the room! You can clearly see the effect of the corner traps still ‘eating’ the upper bass / lower mids despite me having turned them around so that the absorptive side faces into the corners. I think I'm just going to have to live with this until I’m able to DIY a bass trap that rolls-off above 100Hz.

01-L-R-Avg-spkrs-against-front-wall-toed-in-and-tilted-no-treatment-vs-GIK-Tri-Traps-reflective-sid.jpg


I tried toeing the speakers in a bit more so that the tweeter beams crossed about a foot behind my head instead of several feet behind my head. This had the effect of increasing image focus and reducing lateral reflections.

02-L-R-Avg-spkrs-against-front-wall-tilted-GIK-Tri-Traps-reflective-side-out-usual-toe-in-vs-more.jpg


Compared to previous locations I don't think soundstage depth is as good when the speakers are this close to the front wall. Vocals in particular have a more 'forward' placement. Also, even though there is evidently significantly more bass with the speakers in this position, the presentation still sounded surprisingly lean after my initial attempt at EQ:

03-L-R-Avg-spkrs-against-front-wall-tilted-GIK-Tri-Traps-reflective-side-out-more-toe-in-with-vs.jpg


My second attempt knocked the level down a further -0.5dB and widened the Q to encompass two octaves instead of one. This sounds more balanced to my ears and also puts a little more distance between listener and singer. The peak at 870Hz intrigues me, I might try absorption again on the area of wall between the speakers and see what effect this has.

04-L-R-Avg-EQ-second-attempt.jpg
 


advertisement


Back
Top