Acoustic Simulation for Michael

Tenson

Tenson

Moderator
Joined
Nov 12, 2003
Messages
5,947
Reaction score
1
Location
Kent, UK
I did this simulation of Michael's room acoustics so thought I might as well post it up here for anyone who could be interested to see CARA at work.

The system uses Wilson Sofia speakers, so I made a similar design in the CARA loudspeaker editor - Same size box with a 10" woofer, 7" midrange and 1" tweeter. Crossover at 110Hz and 2800Hz.

Here is the original room layout.
Original-Layout.jpg


Here are the simulated results for frequency response and sound arrival time. Ideally, the sound arrival time should be a flat line at the top of the graph, but as you can see there are frequency ranges where sound arrives later, which is due to reflections.

Original-Results.jpg



So now we can try a couple of different speaker and listener positions. I let the computer run an optimisation process within some position limits for the speakers and listener. Two options came up as seeming to work better.

1-Layout.jpg

1-Results.jpg



2-Layout.jpg


2-Results.jpg


As you can see, both options have a smoother frequency response than the original layout, however option-2 has a lot of spikes in the sound arrival time, meaning it might be a more coloured sound than option-1.

So what if we keep the original layout but add some acoustic panels? Below is what happens when we add a 1m square 75mm thick foam absorber on the wall behind the listener.

Original-Layout+BkAbsorb.jpg


If we compare that to the first set of results there are clearly far fewer late arriving sounds in the mid-high range. Great!

How about also adding some absorbers on the pillars either side of the speakers? Well I tried it on all faces of the pillars, and the ones facing inwards towards each other seemed to help most.

Pillar-Absorber.jpg


Original-Results+BkAbsorb+PlrAbsorb.jpg


Looking very nice now in the mid and high range, but those foam panels don't improve the bass, as you can see the response is still a bit ragged.

Lets try optimising the positioning of the listener and the speakers now we have those panels in place. The layout not surprisingly looks a lot like option-1, but tweaked a little bit to make best use of the absorbers.

Option3-Layout.jpg


Option3-Results.jpg


Okay that does seem to be the best balance so far :) The mid-high range has few reflections and the bass is smooth.

Hope that helps Michael! Seems like adding a panel behind the listener is of great value, and if possible a little re-positioning will improve the bass too.
 
Thanks really appreciate those simulations. The background to the request was a concern that the pilliars may have a significant impact on the acoustics - and I also feel there's some lack of bass energy in the room (with both Proacs and to a lesser extent my current Wilson Sophias).

Some thoughts / comments :

1) it would appear that adding treatment to the pilliars doesnt have a significant effect - based on these results I'm unlikely to bother especially give the visual impact of adding them to the inside surface.

2) Not so with the back panel - I must try one - how high should I mount it ? (there's a long radiator behind the sofa)

3) I'm limited somewhat with speaker/ sofa placement given this is our lounge. Option 1 and 3 put the sofa too far into the middle of the room. Whilst I could bring the speakers forward I struggle to see how that could improve bass. Surely bass improves as you move nearer the back wall - seems counter intuitive ?
 
I thought option 1 and 3 might have the sofa too far into the room, hence I also posted option 2. This is no worse than your current arrangement in the mid-high and should give a smoother bass response. Worth trying?

Modal acoustics is more complex than simply getting more bass by moving speakers closer to the walls. This is true to a degree, if the speakers are very close, but with free-standing speakers that don't go inside the wall you will still get funny things going on. Give option 2 a go and see what you think. It should cure the bass lacking around 50Hz.

I've also attached a video of how the sound spreads out from one speaker with your original layout. You can see the pillar helps to reduce the first side wall reflection a little, although you still get some diffraction going on.


P.S. You can't see the pillars, but you can see the outline of them where the sound wave doesn't go.
 
Here is one more option you could try with the listener nearer the back wall. There is also an absorber panel behind.

Option4-Layout.jpg


Option4-Results.jpg
 
Simon, based on using this tool for several years - there is no use in extending the sim up to 20Khz at all.

In fact I strongly recommend you limit to c. 800Hz.

The reason is simple: everything above c. 500Hz is governed by surface finish far more than anything else*. However it appears CARA assigns a linear weighting to flat response - which means the intangibles of exact finish and the exact details of the speaker models dominate its recommended positioning - when it really should not. The real problem to be solved is achieving in-room a natural acoustic balance of the bass response 'maximally flat' w.r.t the room boundaries.

So i suggest : limit the modelling upper freq parameter to 5-800hz or so (up the recursion if you like because it's quicker over a narrower range) - and then result seems to solve properly for the positioning-dominated bass range. IMO of course.

Then, for domestic sake, select symmetry against one or more boundary - it makes less than buggerall difference to the final best-outcome prediction.

If you need a real-world example: I got corroboration between my 'speaker placement by 2yrs by ear and measurement' and the CARA model well within 10mm this way following this 'crude' technique!

- And if you really want quick'n'dirty - stay under 300Hz and you just do not need to model anything under 1000mm in any dimension - but the results are still very, very good! And you can get right on the nose in about 10mins flat... leaving time free to test by ear/mic.


*for everyone else - the half-power point in most music is 300Hz or below. That is, 50% or more of the acoustic power is firmly in the bass - below lowest midrange.
 
Last edited by a moderator:
Hi Martin,

Thanks for the advice. I didn't realise you could limit the upper FR, is that done in the 'SPL Target Function' dialogue?

Your conclusion does seem to make sense, only I think non-symmetric placement tends to help in both the modal and specular ranges.

Do you also find that if you start the optimisation with the speakers and listener in a different location it will find different optimised positions? I tend to run it 2 or 3 times with different start locations.
 
Following Martins advice for tweaking the settings to focus below 800Hz, the results do seem more consistent. It ends up with much the same positioning each time I run an optimisation process.

I hope you don't hurt your back moving these big speakers around a lot!

This is what it suggests.

5-Layout.jpg


5-Results.jpg
 
Hmm - that's very interesting - planning to have a play late afternoon - not sure what the GF will say having the speakers that close - but it's good the sofa is quite far back. Can you run Option 5 with the back panel ?

A bit strange having one speaker much closer to the side wall - I guess the asymmetric set up helps reduce the sound reflections ?
 
Last edited by a moderator:
Tried Option 2 and 5 and the good news is they both sound better than my original set up. I would say option 5 provides the cleanest sound but for some reason the option 2 seems to produce slightly better bass.

So I'm going to go for option 2 for now and live for it for a while. Theres one matter that I find confusing with the simulation. If I want to sit in the middle of the two speakers I measured a distance approx 2.9m from the LH wall. The diagram suggests it should be 2.24m which puts me closer to the left hand speaker - can that be right?
 
A big thank you to Simon. Been listening over the last couple of days and there's a definite improvement in bass response and overall sound quality with the new layout. Funny I thought you'd always get better bass closer to the back wall but thats clearly not always the case. Very pleased with the outcome have now got an acoustic panel on order for the rear wall :)
 
Update - been living with the new set up for a while now. I also added a 5'' deep 4' x 2' acoustic panel behind my head and also put one in the LHS corner along side the sofa for good measure.

Overall continue to be very impressed with the results. Best value upgrade Ive made with the room treatment.

One question Simon - would I benefit by moving the sofa further into the room (by a few inches) ? I wasn't clear whether the sitting position from the back wall was critical or not, and everything I read suggests you need to sit further into the room if possible. I cant do the 38% thing but I could move forward by a few inches. Would that potentially mess things up ?

I know I could experiment but any changes are likely to me marginal so may just do it (or not) depending on the simulation - please note the final simulation was based on slightly different dimensions to those above per our final email exchange.

Thanks
Mike
 
Hi Mike, the listening position is equally important as the speaker positions. Further in to the room isn't necessarily better but the last arrangement I sent you by PM should be correct for your room dims.
 
This is very interesting.

If you measure it for real with those speakers, would you really get that much SPL at 20Hz?

In my messing about, I have found measuring very useful, but it doesn't tell you much, if anything, about tone, clarity and dynamics of sound you hear.
 
The speakers are simulated based on the specs Wilson give, so probably not that accurate in exactly how low they go. More important is the physical position of the bass drivers.

As Martin said earlier, the tone and clarity (i.e. midrange and up) tends to be dominated by the small details of the room so this simulation is much more about finding the position that gives the best LF response.
 
It's not a problem to go low with a 10" driver provided the box is not too small. The Wilsons are not exactly bookshelf speakers!

The maximum output level and sensitivity will be sacrificed compared to a horn loaded design of course. That said, I personally think a 10" driver can push out a lot more level than most people need, even at 20Hz (in room).

The Advantage go to about 25Hz in most rooms with 2x 7" drivers in each cabinet. They go pretty loud and dynamic too. Only those who are very serious about bass would need more IMO.
 


Write your reply...

Latest posts

Back
Top