Setting up a surround system

This is not a studio design, but I'm setting up a surround system at home for pleasure. Hope to get some advice and input along the way. My priority is surround movies, but I would also like it to work well for music in surround. Two channel music isnt a priority here, I've another setup for that. About the speaker system. I'm waiting for three Geddes Abbey speakers which will be used as fronts. These should arrive in January/February. http://gedlee.com/abbey.htm Geddes Harper will function as surrounds, but might be used some as fronts in the beginning. These can be seen with the first pictures I've added below. http://gedlee.com/Harper.htm Two Geddes bandbass subs will work as midbasses: http://gedlee.com/Subs.htm One Elemental Designs 18" sealed sub will take care of the deepest bass: http://www.edesignaudio.com/product_...roducts_id=640 Something worth noticing is that the Geddes Abbey and Harper speakers don't have any deep bass at all. That's why I'm using midbasses and they can be placed wherever they perform best in the room. To set up the three subs, I have a Behringer DCX2496 to help me. The Geddes speakers have waveguide and controlled directivity. They are suppose to be toed in so they cross the in front of the listener. They therefore need less absorption on sidewalls. In the beginning, till I receive the Abbeys, I will also use a pair of active Adam Classic Column MK3 speakers. Wish I has a projector, but I've to stick to my 50" Panasonic plasma for now. The dimensions of the room are as follows (pictures follows): - Width: Ca. 3,54 m - Length: Ca. 5,47 m mesasured from inside of the closet (I assume most of the bass will go through) - Height outside closet: Ca. 1,69 m - Height on other short side: Ca. 1,43 m - Height in the middle (heighest point): Ca. 2,47 m As one can see the room has sloping ceilings. And because of the closets built into the room, the height at each end differs. My plan is to fill the closets with low density porous material. Should be a nice basstrap. I do have general understanding of acoustics and treatments. I will create create a a RFZ for the frontspeakers and possibly also for the surrounds. Then I need to treat modes the best I can. I don't have experience with pressure based traps, but this might be an opportunity to get some. An even decay is also something I'll try to aim for.

I did try placing the Harper speakers at other positions in the room as can be seen below, but it didn't seem to measure any better. So I'll probably end up with the speakers at the position that was shown in the previous post. This is also the most practical placement and it will give me the opportunity to place some diffusors in the rear of the room. Started to fill the closet with low density fiberglass. I'll will fill it completely later, but this is a good start. A question about the fiberglass. It's practical to leave them in their packages. But they are packed quite dense and tightly that way. Would it be better to take them out and into something else so they become more loose? I don't know the density numbers of the fiberglass but this is stuff that is used for isolation, so it should have pretty low density.

I will have to wait till get my frontspeakers before I can finetune, but it's good to do some experience before to get to know the room and to make a plan for treatment. The surround speakers that I've used so far are also very similar to my coming Abbey frontspeakers. After basically deciding on placement of the front speakers, I went on with some treatment. Here's the ETC for each speaker before any treatment (besides the fiberglass in the closet): I then added two absorbents on the sidewalls. These were only placed by guessing where the first reflection points were. So better placement should yield a better result. These absorbents consist of rockwool 20 cm with 65 kg/m3 density and 2 cm of airgap. I'm wondering if the density is too high for this thickness. I could go down to rockwool of 40 kg/m3 if it will make any difference. Suggestions are welcome. I prefer using thick panels over a bigger airgap on sidewalls. This is how the ETC looked after this: It definetly took care of some early reflections at 7-9 ms.

I'm going to continue with ETC, and then I can post the development of frequency response and waterfall later. I added two superchunks (made out of 40 kg/m3 density rockwool) in the frontcorners and two absorbents (20 cm rockwool, 65 kg/m3 density) on the frontwall, but this didn't change the ETC. That might also indicate that I can make them reflectice og mids and highs. That's something I can try later. I then added two diffusors on the rear wall. I'm not sure what diffusors will do regarding movies. Movies as suppose to two channel music do have active surrounds. The same for concert DVDs, but I'm thinking it still would be good to have some diffusor in the rear of the room. It should deal with high gain later arriving reflections without making the space too dry and dead and also add some spaciousness. The box they're standing on is filled with rockwool. These are the diffusors I use: http://resonatorstockholm.com/optiffuser/ Mine have been painted with two strokes. I'm considering painting them now with the same color as the walls. This is the ETC after adding the diffusors: Reflections at 12-14 ms were minimized.

Continued by adding diffusors on the rear sidewalls. These boxes they're standing on are empty by the way. The ETC then gave this result:

Before I continue with other treatment, let's look at the frequency reponse with these changes. Again, these speakers don't have any deep bass. I've only measured from 90 Hz. This was with only the absorbents on sidewalls: Superchunks added in the frontcorners: Absorbents on frontwall added: Diffusors on rear wall added (on top of box stuffed with rockwool with 65 kg/m3 density): Diffusors added on rear sidewalls: The frequency response flatten with the superchunks and absorbents on frontwall but got a litte worse with the diffusors. Any comments to this?

Here's the waterfall before any treatment besides the fiberglass in closet and after all the treatment that I've showed pictures of. Let me know if I need to adjust the graph and how. Before: After treatment:

Except for the very early reflections, the ETC started to look pretty good. Obviously there are reflections coming from the floor and ceiling. First I placed the couch in the room. Added a carpet and the TV rack. The red carpet was too big (wasn't able to shut the door), so changed it to a bluish one. The ETC with the blue carpet: Table in front of the couch is nice to have. The ETC suffered though with some high spikes: Therefore put a cloth of wool on the table. New measurement: The cloth helped some, but not enough. I need to get something that covers the table better, or it might be solved when I treat the ceiling. I'm also interested in finding something to treat the floor bounce cancellation. I could put some Auralex foam pads under or on top of the carpet, but it's not going to look good. I think a better solution would to have a very thick (or two) carpets. I would like a new carpet anyway, I don't fancy the blue one. Any suggestions what I could buy?

I've started to take out the fiberglass bags of the closet and put the fiberglass more loosely in bags. This should work better. Finished one.

I'm not sure what you are showing in those final few graphs, but it certainly isn't frequency response! They look more like impulse response graphs to me. If so, you seem to have an RT-60 of less than 10 ms! :shock: Maybe you could explain what we are looking at in those graphs? From the frequency response graphs and waterfalls, you seem to be lacking bass in the system in general. I saw you said that your speakers don't have much low bass, and roll off below 90, but there doesn't to be much in the low end at all, and as a result you seem to have a mid-range boost, in the region 200 Hz to 2 kHz. Are you using some type of equalization in your signal chain? If so, then remove that from the circuit and repeat the tests. If not, then you might want to figure out what is wrong with your speakers, amp, measurement mic, or interface. Something just doesn't look right there. - Stuart -

Soundman2020 wrote:

I'm not sure what you are showing in those final few graphs, but it certainly isn't frequency response! They look more like impulse response graphs to me. If so, you seem to have an RT-60 of less than 10 ms! :shock: Maybe you could explain what we are looking at in those graphs?

The graphs are ETC- Energy Time Curve. They show the direct signal and the arriving reflections of the room. http://www.aes.org/e-lib/browse.cfm?elib=3735 http://www.daytonaudio.com/OmniMicV2/hs30.htm

Soundman2020 wrote:

From the frequency response graphs and waterfalls, you seem to be lacking bass in the system in general. I saw you said that your speakers don't have much low bass, and roll off below 90, but there doesn't to be much in the low end at all, and as a result you seem to have a mid-range boost, in the region 200 Hz to 2 kHz. Are you using some type of equalization in your signal chain? If so, then remove that from the circuit and repeat the tests. If not, then you might want to figure out what is wrong with your speakers, amp, measurement mic, or interface. Something just doesn't look right there. - Stuart -

First of all the speakers that I've used to measure with so far are surround speakers. I'm waiting for the front speakers to arrive. Secondly you need to look at this type of speakers as a system. They require multiple midbasses/subs. So far I havent' mesured with these. I hope do get that started soon, but I'm waiting for some electronics. But you still make a point. The graph isn't very flat. There are no EQ in the picture here. Some of it is caused by the room and possible placement. But some of it may be a bit of weakness with the speakers. I've already discussed this with the speaker builder, and he has adviced me to do some changes with the crossover which will lower a region. Not sure how important it is considering it's surround speaker, but it might be worth doing.

The graphs are ETC- Energy Time Curve. They show the direct signal and the arriving reflections of the room.

I guess you didn't read what I said in my comment: I did say that they look like impulse response curves ( ETC and IR are basically different ways of looking at the same data), and I did point out that if that's what they are, then you have a problem since your graphs are showing RT-60 times of less than 10 ms (which also does not jibe with the decay evident in your waterfall plots, showing times far longer and more plausible). On the other hand, if those scale markings really are "m" not "ms", then why are you showing graphs for a 60 m room when yours is not even 6? You might want to do some more research on this matter, since it looks like you aren't understanding it, or you aren't measuring it correctly. In any event, regardless of how it was derived (IR or ETC), RT-60 is not a valid measurement for such a small room. Small rooms do not have the ability to produce statistically valid reverberant fields. That can only be achieved in rooms whose dimensions are at least 6 or 7 times the lowest wavelength of interest. You also seem to have your room set up sideways! The ceiling can rise from the speakers going backwards, but should never descend again: that will create serious reflection issues. You should have set it up with the ceiling rising and falling from left to right, not front to back. That first very strong spike is most likely the rear wall reflection, and the following spikes are most likely coming from the angled ceiling (which isn't even treated, based on the photos).

First of all the speakers that I've used to measure with so far are surround speakers.

So you are running acoustic tests of your room with speakers that do not meet the requirements for running acoustic tests of rooms? :shock: Why would you do that? The speakers are coloring the sound, so all your measurements are invalid (unless you are compensating for the flawed speakers in software, but that clearly isn't the case, since the flaws are showing up in the graphs...) But I also just noticed that you are planing to use DIFFERENT speakers for the L-C-R speakers and the rear surrounds? That is not correct. All satellites should be identical, according to both ITU and EBU recommendations. All five speakers in any 5.1 system should be the same, located at the same height, and at the same distance from the listening position, or time-adjusted if the physical path length cannot be attained in the room. Which brings up the next point: your listening position and speaker orientation is incorrect for a 5.1 listening room.

I've already discussed this with the speaker builder, and he has adviced me to do some changes with the crossover which will lower a region. Not sure how important it is considering it's surround speaker, but it might be worth doing.

Flat response is ALWAYS important, always worth pursuing, regardless of what type of speaker you are using. The response should be as flat as possible across the entire published range for that speaker, and should roll off smoothly beyond that range.

Secondly you need to look at this type of speakers as a system.

No, actually you don't. When running room tests, you need to use one full-range speaker, or a good combination of a single woofer-tweet co-located with a matched sub-woofer. And when running speaker tests (after the room tests are completed), then you test ONE speaker at a time, calibrating that single speaker correctly for level and timing, as needed. Only once all the individual speakers are correctly calibrated can you run full tests on the complete system.

Some of it is caused by the room and possible placement. But some of it may be a bit of weakness with the speakers.

I beg to differ! Most of the uneven frequency response you see in those graphs is speaker related. There is some room related stuff going on too, obviously, but you'll never know which is which for sure until you use the right speakers for the job. One more comment here: You are conducting your tests at a level that is far too low. All your graphs show the same thing: you are using only a tiny fraction of the power that you should be using. Testing and calibration for each speaker should be done at the standard level of 85dB. Lower levels are not stressing the system at realistic levels, and are most likely not fully exciting many of the room modes that could cause you problems later. I'm also curious about your measurement mic: What mic are you using for these measurements? Hard to tell form the photos. Is that a Behringer ECM 8000? Also, what interface are you using, and did you test it for flat response with a simple loopback? Your mic setup is also incorrect. For ALL measurements, the mic should be centered at the listening position, aimed directly forward and lightly upwards at an angle of around 30 to 45 degrees. You need to measure and mark that position very, very accurately, so you can do future tests from the exact same position. Even a difference of a few mm can affect the response. If you don't get the mic back to the exact same location every time, then comparisons are invalid. The results themselves from each test will still be valid, but you cannot compare the changes in the results over time if the mic is placed in a different location for each measurement. However, your photos seem to show the mic in multiple random positions around the room, thus invalidating all comparison between measurements not take at the exact same position. Finally, you don't seem to have anywhere near enough treatment in that room yet. All of the photos so far show large bare walls, with only minimal bass trapping, very insufficient for such a small room. What I'd suggest is that you first rotate your room to orient it correctly, then get a good speaker setup with a single full-range speaker or combination, then correct your mic placement, and repeat the tests without and with the small amount of treatment that you have so far, but make sure that you are using the software correctly, since I find it highly improbable that your room has an RT-60 of under 10ms, as shown by some of the graphs (but not others). That would be my approach. - Stuart -

I can't see anything wrong with the ETCs I've posted and I have no idea why you're talking about RT60. I haven't posted RT60 and I don't plan to either, this being a small room without a diffuse field. The ETC shows the first 55 ms. Some of what you're saying is valid, but a lot is based on the fact that you haven't read my posts. First of all. I haven't moved the mic, but I've showed pictures of different possibly placements of speakers. Secondly, I have showed measurements from scratch without any treatment and was planning to show the progress. That's why you don't see much treatment yet. Give me some time. Having surrounds that are 100% similar to the fronts and center would be nice, but it would cost way more then I'm willing to pay for a surround setup. The important thing is that they have the same tonality, and they do to a large degree. I disagree with what you're saying about placement of surrounds. That would be correct for multichannel music, but for movies, which is my number one priority, they shall be above ear height and pointing in front of the listener to create a diffuse/ambient soundfield. 60-90 cm above ear height is ideal. Anyway, I'll continue later. Need to measure with the subs/midbasses. I will stick with the placement I've now and address the ceiling reflections. Playing down the long axis of the room would avoid this, but it's very unpractical and I can live with some compromises in this case. Painted my diffusors the same color as the walls. Looks better I think. :)

I can't see anything wrong with the ETCs I've posted

Well, I guess that's because you don't really understand what you are looking at! :)

I have no idea why you're talking about RT60.

I'm talking about RT-60 because that's one of the major things that both IR and ETC are useful for! :shock:

I haven't posted RT60 and I don't plan to either,

Ummm.... yes you have! You just didn't realize it. Your RT-60 is clearly evident from either an ETC graph or an IR graph (which are, as I already pointed out, two different aspects of the same data, and can be transformed into each other).

Some of what you're saying is valid,

No, all of what I said is valid. :)

Having surrounds that are 100% similar to the fronts and center would be nice, but it would cost way more then I'm willing to pay for a surround setup.

Fair enough, as long as you realize the shortcomings of such a setup.

The important thing is that they have the same tonality, and they do to a large degree.

"Tonality" is not an objective, measurable quality of speakers. It is a subjective invention of the audiophile world that has no place in serious acoustic measurement. "Frequency response" is an objective, measurable quality, and is the one actually used. What matters is that all the speakers should have flat frequency response across the entire range.

I disagree with what you're saying about placement of surrounds. That would be correct for multichannel music, but for movies, which is my number one priority, they shall be above ear height and pointing in front of the listener to create a diffuse/ambient soundfield. 60-90 cm above ear height is ideal.

Well, you can disagree all you want, but that won't change the facts. Both ITU and and EBU recommendations for critical listening rooms are crystal clear here. You might want to check into that before you continue to disagree with some of the foremost engineers from the two world leading organizations in this area... Disagreeing with my own personal opinion is one thing, but disagreeing with their findings is a bit different. Yeah, I've heard the same questionable tripe repeated endlessly on some home theater and audiophile sites on the web, but repeating a myth endlessly does not make it true. Scientific measurement of real, controlled situations in real, certified acoustic laboratories is what matters. The type of web site or person that repeats such myths is the same type that used to swear by carpets on the walls and egg crates on the ceiling! And the same type of place that insists that home theaters should attempt to emulate full-size theaters in terms of speaker placement and acoustic treatment. Maybe you've been spending too much time listening to such people... Simple scientifically verifiable fact: Placing speakers above the listening plane and tilting them down more than about 10 degrees creates false positioning interference patterns in the human pinna, causing misinterpretation of the vertical placement of sound sources, and mismatch with the visual placement. This psycho-acoustic effect is one of the causes of listening fatigue. You might want to check into that. In any event, your words are contradicting your photos: your speakers are most certainly not located 60-90 cm above the listening plain! Not unless you plan to listen while lying flat on the floor! :) Not only that, but it is physically impossible to set up your speakers like that. In fact, I doubt that you'll even be able to get them to the correct height with the current orientation: It seems to me that they'll be below the listening plane.

they shall be above ear height and pointing in front of the listener to create a diffuse/ambient soundfield.

Didn't you just tell me that it is impossible to have a diffuse field in a small room? So why are you trying to create the impossible? :shock: The ONLY reason you'd place speakers above or below the listening plane is if it is physically impossible to place them at the CORRECT height, due to intervening barriers in the room, or furniture, or some structural part of the room that prevents mounting them in the correct location.

I will stick with the placement I've now and address the ceiling reflections.

Sorry, but you don't have enough space to do that effectively, especially since we both agree that what you are seeing on your graphs is modal behavior, not just reflections. The space required for the type of treatment you'd need exceeds the space available, and would force your speakers even lower down, or out into the room.

Playing down the long axis of the room would avoid this, but it's very unpractical and I can live with some compromises in this case.

Once again, fair enough, as long as you are aware of the limitations.

Painted my diffusors the same color as the walls. Looks better I think.

QRD's in a such a small room? :shock: Surely you jest! Are you not aware of the lobing artifacts that you'll be getting from those? - Stuart -

Remember that nothing is definite from the pictures showed so far. I still haven't decided 100% on which shortside to place the speakers. The treatment in the room you see so far is basically just been thrown in. This is a progress. I'm well aware of the distance that is required from diffusors and I don't plan to place them so they will give me serious lobing issues. You can see measurements of the front and surround speakers here: http://gedlee.com/abbey.htm http://gedlee.com/Harper.htm They are not 100% similar of course, but this is good enough for me. There are not many who would buy big speakers for surrounds in a home setting. About ETCs that have posted. Can you please ellaborate on what you're seeing is wrong? I'm using this mic and premic (with a calibration file): http://www.content.ibf-acoustic.com/cat ... ucts_id=35 My soundcard is Lynx Two-B and I've run a loopback. It looks good. I'll make sure to measure at 85 dB from now on. I've read some papers on placing surrounds and I'm having difficulty seeing that we've clear standards. You also see DD and DTS recommend 60-90 cm above ear height and pointing them in front of the listener (not directly). Here's what Genelec writes on their page:

The rear channel angles can be raised by more than the front monitors as humans have poor poisitional perception of sounds from the rear but the standards do not agree with one another on this: 15º is perhaps sensible maximum here.

http://www.genelec-ht.com/learning-cent ... sitioning/ I've in the past tried placing surrounds at ear height, but I seemed to prefer having them higher which I think made a more ambient and surrounding soundfield. Many who have tried both have come to the same conclusion. But this isn't important right now and I can try both options later. We've to remember that movies and surround aren't accurate at all. It's more about creating something fun and artificial it seems like. But I'm open to information and knowledge that may have passed me. :) I'm now done filling the closets where I put the low density glassfiber more loosely in plastic bags. A lot of cutting to fit the sloping ceiling. Next step is to start measuring with midbasses and sub. This should give a clue about the best placement. I've two 12" midbasses which will operate in the 125-50 Hz region and one 18" to care of the bass below 50 Hz. I'm little uncertain which I should start with. Geddes (the speaker designer) recommend having at least on of the midbasses close to the fronts due to integration. Kind of wish now that I had bought a smaller sub. 18" is a little too big for this room.

Ok. Done some measurements of the midbasses. On both short sides of the room. There's some smoothing here (forgot to change the settings), but the graph doesn't look any different without smoothing. First started with one at the left side. Tried different placements and ended up about here for the most even response: Here's how the freq. response and waterfall looks like: I then started with the second midbass on the right side. Here I had more problems finding a good spot. Tried a lot, but the better one was probably about where the speaker needs to be placed. Measurement from the second one: Placing it on the right sidewall gave a response like this: As can be seen I'm dealing with a resonance around 62 Hz. I was able to move the midbasses to avoid a strong peak though. Several placements had a pretty strong peak at this frequency. Will continue with other side of the room in next post.

Let's see what happens at the other shortside of the room. I just left the rack and TV and only moved the midbasses. Started on the left side and also here I saw strong peaks at 62 Hz. I ended up having it on the sidewall out from the frontwall some for the best response. This gave me the following measurements: There's a cancellation at about 110 Hz. Other placements gave me less of it, but with a peak at 62 Hz. On the right side, I also found the placing the midbass on the sidewall yielded the best result but again with a cancellation. I did try moving the mic around some with this second midbass at the right side. Moving it forward and backwards didn't really change anything. Moving it to left and especially right side altered the response some. Mic to the left: Mic in the right seat: Ok. Any comments to what seems to be the better side based on this? I've not done any measurements with the large sub for the deepest bass. I'm pretty much thinking that if it gives me any serious peaks at low frequencies I can remove with EQ. Not ideal, but I'm not sure if I'll built resonators in this room. I do have a RPG Modex Plate type 1 which I can place here, but I'm uncertain how much only one of these will help. http://www.rpginc.com/products/modexplate/index.htm

Hello again. :D Still trying to figure out what is the best short side to place the speakers. Measuring the midbasses didn't give a definite answer. At least not to me. So I decided to measure the satelite speakers again. These are the surround speaker, but since they measure very similar to my fronts I believe they will give a good enough answer. Beware that there's no carpet in front of the speakers with these measurements. Don't look at the very early reflections. I will deal with that when I've found the best side. So I tried placing the satelites in front of the closets again. With treatment in the rear of the room (combination of basstrapping and diffusors) So how did this measure? Let's see. Freq. response and ETC. I then turned the setup to the other side again. The first picture show the satelites placed closer to frontwall which I tried at the end. Placed on this side they measurements were as follows. First placed at about 1 m to the frontwall. Continuing in next post.

Tried to place the satelites closer to the frontwall (about the same as the on the other side) which the latest picture in the last post showed. The freq. response didn't change much. To me it seems like the satelites are giving me a more even response when they are place in front of the closet. There's a higher level in the area between 200-400 Hz. This is the side where there's more ceiling height. I don't really see much difference comparing the ETCs. Seems like I can get a pretty similar ETC on both sides.