Home Theater Forum and Systems banner
1 - 18 of 18 Posts

·
Registered
Joined
·
24 Posts
Discussion Starter · #1 · (Edited)
Hey guys,

I read a lot of your post and great ideas about optimizing the sound. I tried a bit, but now I need your help.

I want to mix my projects (jazz/world-pop/classical) in my living-room/studio. And this room is totally not symmetric and small and there is already some furniture that has no other place in our flat. That means, most things I read about acoustics I cannot use. So I think it cannot become perfect but I want to get the best possible.

I have already been doing a few mixes this way for a year and also released something on CD and some mixes were quite good. However, sometimes it was so hard to get a mix that translated well to other hifi-systems in other rooms or cars and it became really frustrating. I started to study room acoustics and REW and tried to used DRC (drc-fir.sourceforge.net), but it was not satisfying yet.

I moved the speakers to 3 different places in the room and changed the position slightly on each position and judged by ear. What I have now seems to sound best, but it's really hard to compare.

Problems still:
- bass and low mids - mixes don't translate -> probably because of the peaks and dips in freq (modal? SBIR?).
- stereoimage: weak phantom center
- hard to adjust reverb (fx) settings

Hope some of you experts can help me.

My setup is:

2 Yamaha MSP 5 - nearfield, on simple speaker stands
L->R distance: 114cm (tweeter center) ~ 45"
speaker -> Listener : 94cm ~ 37"
listening height:118cm ~ 47"
distance speaker->wall: ca 22cm (center tweeter) ~ 9"

1 sub (<100$ of a PC-2.1.) no real sub but gives at least a little bit 40Hz and up where the Yamahas are lacking.

Any suggestions, positions, physical room treatment, electronic correction etc welcome.

Regards,
Florian

Attached:
- sketch of the room
- REW measurement without room treatment
- REW measurement with three 47"x24" DIY hemp-absorber panels, 4"thick, wooden frame (one on the front wall slightly above the speakers)
- picture of my DAW working place
- REW *.mdat file
 

Attachments

·
Banned
Joined
·
516 Posts
Not sure what to tell you, as you state constraints and then want suggestions.

Modal behavior and symmetry are important!

The frequency response measurements simply confirm that you have significant comb filtering. No surprises there. And RT60s are not very useful considering one does not have a homogenous statistically well-mixed diffuse soundfield in a small acoustical space that the calculation assumes but rather a locally variable modal and specular soundfield.

A better tool to identify the causal factors creating the comb filtering would be full range ETC responses for each full range source. These would provide insight into the specific behavior that is causal to the locally variable destructive interaction displayed in the frequency response.

One problem that you could potentially correct is the speaker mounting and the early reflections/diffraction from the speaker mount and the work surface.

First you have something under the speakers (MoPads?) as well as the base of the speaker stand that both protrude in the front and act as diffractive and reflective sources.

If you need to isolate the speakers get rid of the diffractive source and get sheet or two of Sorbothane from Amazon.com for about $20 each. And change the orientation of the stand base so that the edge is parallel with the front of the speaker.

Also, you will have a strong detrimental reflection off the desk. Either get rid of the desk and replace it with a mobile computer stand large enough to just hold the keyboard, or move the desk a few feet out from the wall sufficient to cause reflections to be directed 'below' your head.
 

·
Registered
Joined
·
24 Posts
Discussion Starter · #3 ·
Hi SAC,

I read a few of your posts and I am happy that you share your expertise.
Maybe some of the constraints can be loosend if there is no other way, let's see what is necessary.
And I see that I forgot the room sketch.

I was already thinking about if the desk might be a problem. The mobile computer stand is an option.
What about a desk with an angled tabletop and smaller width?

I also worried if the distance speaker to speaker is to wide!? This would be resolved with another desk.
Maybe I do a measurement without the desk.

The speakers are set on EQ Acoustics AirSpace MonPads. So I will change the orientation of the stand and I can turn the pads around to get rid of the edge at the bottom of the speaker.
I could also make a measurement with pad and without, if you tell me what to measure (single speaker or full setup, freq. range, which REW settings).

I found Sorbothane sheets in Amazon US but could not find it in Germany. What is this Sorbothane and is it better than my "pro" MonPads?

Can you see any trouble in measurement coming from the closeness of the speakers to the front wall?

And can you tell me how to do that full range ETC response? I just uploaded the *.mdat file now, somehow it did not work with my last post.
 

Attachments

·
Banned
Joined
·
516 Posts
Just a few comments...

It is rather easy to see the early grunge aggravated by the speaker mount and the base as well as what is most likely the worksurface.

The asymmetry of the room topology results in dissimilar later prominent reflections.

Its hard to say exactly what to do as we cannot separate the base room response with out treatment from the treated response.

While treatment will not overcome the total effects of the asymmetry, you can mitigate the prominent returns so that they are more equivalent. But again, one would not want to over damp. Diffusion might be a better method to render the lower gain returns a bit more 'equal' as far as treatment might be of use...

As you are not using a hardware propagation loopback and the direct signal is set to T=0ms, we do not have the benefit of being provided with an actual time of flight for each reflection. Thus the blocking method could be employed to resolve each of the significant reflection into their vector paths and points of boundary incidence. Being able to envision the spatial distribution might add quite a bit of understanding into the actual dynamics of the space allowing further adjustments of the siting and listening position so as to mitigate asymmetrical effects.

Yes, Sorbothane is 'special' dissipative elastomer that is specifically recommended. Three simple 1/4-3/8" thick feet are more than adequate to address the mechanical coupling of each of the speakers to the stands. You might check Amazon.com.uk or the Sorbothane website for international distributors..
 

Attachments

·
Registered
Joined
·
24 Posts
Discussion Starter · #5 ·
Hey that's great. Thanks for the effort and the pdf. It helps to understand. I did not know it is possible to analyse the problem from a simple measurement that deeply. I see the difference, but I have to admit that I still don't get what that ETC is and how it works. Y is dB but what is X? m for meter? But what meter? No clue yet. I tried to google it. If you have a link where I can learn it, I'd appreciate it.

I think I also got what you meant by
responses for each full range source.
And:
While treatment will not overcome the total effects of the asymmetry, you can mitigate the prominent returns so that they are more equivalent. But again, one would not want to over damp. Diffusion might be a better method to render the lower gain returns a bit more 'equal' as far as treatment might be of use...
I was thinking about diffusion, too. What kind of diffusors do you recommend and where should I put them?

Today I was examining the dips at 102Hz and 129Hz. I played them as sine wave separately and looked where in the room I could hear them. I found them very loud in lowest and highest shelf of the wardrobe and above.
Can I treat them effectively by putting something like rockwool in there?

Now that you checked the measurement and saw the sketch of the room, should I try to change something about the listening- and speaker-postion?
 

·
Registered
Joined
·
24 Posts
Discussion Starter · #6 ·
ha ... ETC ... I come closer. :ponder: X-Axis is for time... and m was just millisecond!?
I made another measurement and went to "Filtered IR" clicked ETC envelope and hid the main curve.

And to me it looks like you were right. Without desk the desk peak is gone. :clap:
Your PDF helped me a lot.

However how does this effect that the measurement shows "sound". I tried to figure it out by my ears, and tried to hear the difference, but it took some time to remove all the equipment and I felt so different then sitting in front of an almost naked wall.
Which frequencies are affected?
And how does early grundge and mount diffraction sound like?

I attached the measurements without the desk and also with the MonPad turned around and aligned to the speaker and stand. In general, is it better in this stage to measure with the sub or without?

And I attached a pictures of how the stands look like now and how I arranged the absorbers, with nothing in mind.

From Sorbothane I found those feet on ebay is this what you mean? http://www.ebay.co.uk/itm/4-SORBOTH...mponents_RL&hash=item43a92e9feb#ht_1378wt_698
Do I put them between speaker and stand or also between stand and ground?
 

Attachments

·
Banned
Joined
·
516 Posts
For the ETC we measure each source INDIVIDUALLY as earlier stated.
Otherwise we have two sources whose signals arrive at differencing times.

Generally, measurements are made at the same precise point at the listening position.

Nulls below ~200 Hz are normally modal nulls, not technically a result of he cancellation of specular reflections where the sound behaves more similarly to focused rays of light - at least for conceptual purposes, as the wavelengths are smaller than the incident surfaces they encounter.
 

·
Registered
Joined
·
24 Posts
Discussion Starter · #8 ·
Thanks SAC for the explanation.
Would be great if you could answer me
...how does this effect, the reflexions, that the ETC measurement shows "sound".
...
Can I find out which frequencies are affected?
And how does early grundge and mount diffraction sound like?
I understood that from the time(ms) I can calculate the distance of the reflecting object. That's cool.

From Sorbothane I found those feet on ebay is this what you mean? http://www.ebay.co.uk/itm/4-SORBOTH...mponents_RL&hash=item43a92e9feb#ht_1378wt_698
Do I put them between speaker and stand or also between stand and ground?
How big is the difference in sound compared to my MonPads?

I changed the setting of my existing absorbers now and did a new measurement (attached). What do you think?

And I am thinking about building a few more absorbers
- 1-2 panel for the rear wall
-1-2 panels on the ceiling
- bass traps
- corner traps
Any recommendations for my specific room setting?
How many, how big?


And:
Today I was examining the dips at 102Hz and 129Hz. I played them as sine wave separately and looked where in the room I could hear them. I found them very loud in lowest and highest shelf of the wardrobe and above.
Can I treat them effectively by putting something like rockwool in there?
 

Attachments

·
Registered
Joined
·
19 Posts
And I am thinking about building a few more absorbers
- 1-2 panel for the rear wall
-1-2 panels on the ceiling
- bass traps
- corner traps
Any recommendations for my specific room setting?
How many, how big?
Just a quick glance.

Seeing that you can't move your equipment and listening position to make left and right symmetrical, but you can make measurements, then first you make baseline measurement with no treatments in the room. Measure and post L and R driven separately - you cant separate them from the ETC display. Don't forget that your speakers are too wide apart (114cm) for your 94cm speaker-listener distance. Make them equal and don't move the mic between different measurements! Mark the spot exactly with a plumb hanging from the ceiling.

Then you resolve all spikes that are up to 10ms (and above -20dB) after the direct signal to their corresponding reflection location using the 'blocking method'. Search for SAC-s excellent tutorial how to do this.
After you identified their exact location, put there a 10cm absorber, at least 60x60cm, at each identified spot with a 10cm gap and re-measure.
This will clean up your stereo image and hopefully leave your room 'live'. Then you still have to deal with modal issues and L/R asymmetry.
 

·
Registered
Joined
·
24 Posts
Discussion Starter · #10 ·
Hi dannut,

thanks for this great answer. This step by step instruction helps me a lot in my situation.
Now I tried to find SAC-s tutorial. However, "ETC" is not an allowed keyword for the search and "blocking method" gives no result. I was searching for a while now.
Could you please post the link for me?

Regards,
Florian
 

·
Registered
Joined
·
19 Posts
No, I can't find it, it could have been posted in an alternate universe.:scratch:

You identify an offending reflection in a ETC display, then put a piece of absorbing material near the mic and 'search' the direction of the reflection while updating the measurement.
When you found the direction, you go farther from the mic with the absorber while updating the measurement until you are at the offending surface. Thats it.
 

·
Registered
Joined
·
24 Posts
Discussion Starter · #12 · (Edited)
Okay. I started the procedure. It's hard to find those reflecting objects corresponding t 1ms to 5-6ms.
Can it be, that a reflection after 3.84 ms comes from the front wall behind the speaker?
I put 20cm foam there and it disappeared.
But what frequency does the speaker send to the front wall?
Can I find it out from the ETC?

And in general. Am I on the the right track!?

*.mdat file attached.

PS:
Why shows SAC to be banned?
 

Attachments

·
Registered
Joined
·
24 Posts
Discussion Starter · #13 ·
Okay slowly but steady...
I fixed a few peaks:
3,84ms : somewhere at the front wall close to the right speaker, or the right speaker itself!? But strangely this peak is not on the ETC of the right speaker!?

4,65ms: from the floor
5,53ms: from the ceiling
This one from the ceiling disappeared when I put an 0.6mx 1m absorber there.
I even could cancel most of the peaks by just building a 1,2mx 1m cloud on the ceiling above the listening position, and absorption on the front wall and the rear wall but that way it become to dry for me.

Is there a way to deflect those first reflections instead of absorbing them?
A wooden panel should have a size bigger then the wavelength of the reflected wave.
However, what frequencies are reflected from ceiling, floor, frontwall?
A 0.6mx1m deflector could change the direction of frequencies bigger than 571Hz and partially bigger than 343Hz, right!?

What frequencies make trouble for stereo-imaging and clarity?
 

·
Registered
Joined
·
19 Posts
You only posted the L speaker measurement, please post L and R together.
Do you have the speakers asymmetrically toed? Have you tried them without toe-in? If you have trouble finding reflections, then more intuitive could be the string-method. But you need to take your measurements with a loopback connected for timing reference.
 

·
Registered
Joined
·
24 Posts
Discussion Starter · #15 ·
What does "asymmetrical toed" and to "toe in" mean?
The string method worked finally. Thanks a lot.
A used the loopback connection in "preferences". Don't know how to get it in the measurement.

Now I made a few measurements before and after treating the possible reflection points. Can you tell me if I am making progress.

All (always single L, single R and one time combined) measurements are here : http://dl.dropbox.com/u/63267849/serveral measurements for hometheatershack.mdat
"after ETC" is the recent one.
 

·
Registered
Joined
·
19 Posts
edit: just saw your post, try to look at the measurements sometimes tomorrow?

no toe in - speakers face forward. Toe-in, speaker face you - is it clear(er)?
 

·
Registered
Joined
·
24 Posts
Discussion Starter · #17 ·
cool.
so my speakers are not asymmetrically toed, they are pretty symmetrical toed in.(measured with measuring tape)
I have not tried it sometimes, but I am not sure.
Why do you ask, what could this be helpful for?
 

·
Registered
Joined
·
19 Posts
First, look at the overlays display, select 'R no reflection' and 'R with cloud'. You introduced a reflection at 3,8ms and eliminated 3 different reflections after that. Is it a reflection from the 'cloud'? How high did you put it? If you can trace it back to it, then you see that so called 'absorbers' are not ideal devices and you should seek to eliminate the reflective property of the absorber by different/no covering, angle etc. Or you have some inconsistency between different measurements, which makes the whole process that more difficult. Also you have the same 3,8ms reflection on all your L speaker measurements.

If you want a 'lazy' compromise, then a cloud absorber and something (absorber/diffuser/reflector) for the frontwall/backwall flutter and take measurements to see what anomaly remains.

Have you thought about how you going to tame the modal region? You could build pressure-based absorbers targeting specific response anomalies, iterative process, not trivial. Or just put thick, 'fluffy' 'super-chunk' porous absorbtion everywhere you can. First test where the modal response is most effectively mitigated, then for the final version cover it with 6-10mil (0,2mm) plastic, to reflect the high frequencies.
 
1 - 18 of 18 Posts
Top