Comb Filtering Confusion

**Dain Bramaged** · **10-06-06, 09:32 PM**

I was lead to this site for the first time less than an hour ago (following a download lead). It said I was free to download what ever I want for free (money wise), but the fee was to do some posting at this site. No problem. I like free ! So here goes. I started a thread at another forum yesterday but have only gotten one response so far so I thought I'd post it here and see what happens.

In my old house, I had a full 8.1 set-up with a Yamaha RX-V3300 (rated at 130wpc, but it doesn’t sound like that much) controlling everything. I had some older stereo components laying around not being used, so one day I got the itch to try and integrate them into the system. I ran the pre-outs from the receiver over to the Yamaha power amp (M85 - which has 2 separate mono-blocks and power supplies, one for each channel, 270wpc) which in turn powered another pair of identical speakers up front (same manufacturer, same model, etc.), so I had one pair being powered by the receiver and one pair being powered by the M85. The sound quality improved noticeably. I know it wasn’t just a perceived thing as everyone (not one exception) that heard it commented on how good it sounded. So, following the typical male logic of “if one is good, two is better…..” proceeded to buy more M85’s and more identical speakers to match their respective brothers in each channel. So by the time I was done, I had two full systems all being controlled by the one receiver. One system being powered by the receiver, the other system being powered by 4 M85s. The sound was amazing in that now horrendously over crowded small living room. I had two people tell me that they’ve never heard anything better, ever. (Although they’re not audiophiles, it’s still nice to hear.)

Then I sold my house and I now have everything packed away, waiting while I finish my new dedicated home theater room. It will be approx 15 x 45. In the mean time I’ve discovered some articles and discussions about comb filtering, which has me confused, and has led me to starting this thread. There seems to be three sides to this issue. One side says absolutely under no circumstances should you use two speakers per channel. Another side says that in real world applications comb filtering, if not impossible, is next to impossible to hear. And the third side says that it’s strictly luck of the draw – that there’s just too many variables to predict if your system will be compromised or not – you just have to set up your system and find out for yourself to see how the variables all come into play.

Well, with a substantial amount of money already invested, I’m going to be biased towards the “impossible to hear” side. Although with what I’ve read, I can certainly see where the “impossible to predict” side would have a legitimate argument. I have a strong tendency to shy away from the “absolutely don’t use multiple speakers” side because their argument seems to be based mostly on pure sine waves, which I don’t listen to except during calibrations.

This is where you come in. Is the big change in room size and layout going to act as a detriment ? Is the only reason why it sounded good was because I was pouring a large amount of power into a confined space ? I’m wondering if you think it’s going to be worth the extra expense and the extra time of basically setting up two separate audio systems in my new home theater. Which side do you adhere to, and why ?

**Sonnie** · **10-06-06, 11:14 PM**

Hello Dain and welcome to the Shack!

Quote:

It said I was free to download what ever I want for free (money wise), but the fee was to do some posting at this site.

Who told you this? We charge 100 bucks for all our downloads.

No, really, just kidding. While we don't require posting for the downloads we do encourage it. After all, posting is what helps us keep the site going.

I think we've got some guys who understand your issues and can address them.

**Danny** · **10-06-06, 11:32 PM**

Comb Filters occur when an identical sound emerges from 2 (or more) loudspeakers in different positions. It causes cancellation of some frequency's while amplifing others. To hear the difference wire two speakers in Mono and carefully align them so that they are in the same position (see picture below), then while you listen get a friend to move one backwards so that they are no longer aligned, you should hear a difference in sound quality. Comb filters are pretty much inevitable in any sound system however their effects can be minimised by careful placement of the speakers

**Dain Bramaged** · **10-08-06, 10:06 PM**

Good practical suggestion. I’ll definitely try that when I get everything set up.

Well, between the two forums it sounds like the general consensus so far is – “it’s impossible to predict”. I guess I’m starting to take on that attitude too. I’m thinking that with the room being 45 feet long, and the front and rear speakers being so far apart, I can’t help but wonder - just what will it sound like ? And now that I think about it some more, I’m really wondering how it will sound using so many bi-polar speakers in a room that large. Hmmmm. Not sure what to think now. Any more thoughts ?

**Wayne A. Pflughaupt** · **10-09-06, 12:34 PM**

Hi Dain,

You might want to check the last two posts at this thread (by Mitch and John). It’s more than just non-time-aligned speakers. John says that room reflections cause comb filtering, too. So, essentially there is no escaping it.

Even without the room, comb filtering is probably inevitable with any multi-channel system. Even if you minimize it by careful placement, anyone in your HT sitting outside the “sweet spot” will be “affected.” Not that that’s all bad. I’m sure you’ve sat outside the sweet spot in your system at one time or another. How did it sound? Aside from the obvious imaging penalty, I’ll bet it sounded just fine.

I recall experimenting with this a little back in the 80s, and I remember being pleased with the results as well. I’m guessing that the perceived improvement is caused by the slight delay, which is probably adding a sense of “spaciousness.” However, I wasn’t too enamoured with the prospects of having so many big speakers in my living room – which back then only meant two stereo pairs.

You might want to look into Yamaha’s Digital Soundfield Processing. It comes standard with most of their receivers. The processing mimics actual acoustical venues and spaces, so you can get the spaciousness without all the extra hardware and speakers you’re using now. You’ll want one of their upper-line models that utilize a second set of front speakers, which greatly enhances the effect.

Regards,
Wayne

**Ilpo** · **10-09-06, 09:39 PM**

Comb filtering is a common term to a measurable amplitude performance. Two identical signals summed with delay result in cancellation at frequencies where the delay is an odd multiple of half wavelengths. Even multiples add constructively and result in 6 dB addition of amplitude. Resulting amplitude response has deep dips, hence it visually resembles a comb. Audibility depends on many issues, frequency, delay etc. Ear does not integrate the two signals in a same way as microphone and measuring gear, but tends to localize the sound to the closest source.
The presented issue of having speaker pair for a single signal will mean that the directivity of each channel's combined speaker increases at all but lowest frequencies. Listening even slightly off axis of the pair (=channel) will have audible effect due to the directivity changes, and even more so if the speakers are not identical. So, although the results may be "impressive", they will not be truthful to the original signal.

**Dain Bramaged** · **10-10-06, 01:00 AM**

Ilpo,

"will mean that the directivity of each channel's combined speaker increases at all but"....

I haven't heard that before. Food for thought to be sure. I'd be genuinely interested in seeing a graphic representation that would reinforce that statement, if you feel like supplying one.

Wayne,

I hadn't given much thought to the room adding to the effect, but it does make sense if you think about it. You're right in that every system out there suffers from it to a certain degree. It's just a matter of to what extent, and to what extent each person does to try and control it.

**Ilpo** · **10-10-06, 06:42 AM**

About directivity; imagine you have two radiators for the same signal. Given that the radiators are identical, on the axis of the pair the radiated signals are in phase. Move your observation point sufficiently off axis so that the distance to the radiators is no more the same. At some frequency the path length difference is half wavelength and the signals cancel. You have first null in your radiation pattern. At very low frequencies the wavelength is so long that a small distance between radiators means practically nothing in this respect. (Example 100 Hz, 3,4 m wavelength, speakers are 0,5 m apart, at 90 degrees off axis the difference is still 0,5 m which is about 1/7 of wavelength.) Imagine you have only one speaker and do the same thinking. As the size of your radiator now is the size of your drive unit, this phenomenon does not happen (or to be exact, it happens but at much higher frequency, and then we are observing radiation from the opposite edges of the drive unit).
What if your speakers are not identical? The differences in the phase response mean that even on axis of the pair the signals are not in phase and the main lobe of the radiation pattern will depend on frequency.
So, you will hear interesting responses.
You will find the graphics on any acoustics textbook.

BR,

Ilpo

**Sir Terrence** · **10-10-06, 02:05 PM**

Quote:

Ilpo wrote:

About directivity; imagine you have two radiators for the same signal. Given that the radiators are identical, on the axis of the pair the radiated signals are in phase. Move your observation point sufficiently off axis so that the distance to the radiators is no more the same. At some frequency the path length difference is half wavelength and the signals cancel. You have first null in your radiation pattern. At very low frequencies the wavelength is so long that a small distance between radiators means practically nothing in this respect. (Example 100 Hz, 3,4 m wavelength, speakers are 0,5 m apart, at 90 degrees off axis the difference is still 0,5 m which is about 1/7 of wavelength.) Imagine you have only one speaker and do the same thinking. As the size of your radiator now is the size of your drive unit, this phenomenon does not happen (or to be exact, it happens but at much higher frequency, and then we are observing radiation from the opposite edges of the drive unit).
What if your speakers are not identical? The differences in the phase response mean that even on axis of the pair the signals are not in phase and the main lobe of the radiation pattern will depend on frequency.
So, you will hear interesting responses.
You will find the graphics on any acoustics textbook.

BR,

Ilpo

This is correct. I however would not use two identical speakers for a single channel because the sound will not be even over a number of seats. I certainly would not use two speakers for the center channel where dialog intelligibility is of the upmost importance. Comb filtering does add diffusion or the sense of spaciousness, but this works at the peril of clarity. You would be far better off just using a single speaker per channel IMO

**Ilpo** · **10-10-06, 09:56 PM**

Well, this is the only realistic conclusion. I just tried to explain the reasons why only one speaker per channel should be used.

Ilpo

**Dain Bramaged** · **10-12-06, 12:23 AM**

Ok guys. Don’t take this the wrong way – I’m not arguing with you, I’m just trying to understand something here. My towers are quite narrow. From the centerline of one speaker to the centerline of the next speaker is approximately 8”. I’ve seen speakers that have numerous drivers, all the same size, all in a single column in the enclosure, where the bottom driver might be as much as 18” or more away from the driver on top. Are you saying that a single speaker of that configuration would be better than 2 narrow towers side by side, and if so, why ? I'm sure some of the drivers are playing different frequencies, but we're still talking a sizeable spacing of like drivers.

**Dain Bramaged** · **10-16-06, 10:12 PM**

Where did everyone go?

**Ayreonaut** · **10-17-06, 12:02 AM**

Quote:

Dain Bramaged wrote:

I’ve seen speakers that have numerous drivers, all the same size, all in a single column in the enclosure, where the bottom driver might be as much as 18” or more away from the driver on top. Are you saying that a single speaker of that configuration would be better than 2 narrow towers side by side, and if so, why?

A vertical driver configuration is designed to integrate multiple drivers and works best for a certain listening height. Listeners in various places will be seated at similar heights and there is no problem.

A horizontal driver configuration works fine for the center listening position, but as you move to the side you get comb filtering effects. For example, center channel speakers with horizontal configurations can be problematic, and a center channel with the tweeter over the midrange is often recommended. This yields a more uniform horizontal distribution, and sounds clearer from the sides.

**Dain Bramaged** · **10-17-06, 09:12 PM**

I think I understand where you’re coming from as far as the dispersion pattern is concerned. However, with multiple drivers aren’t you still going to have comb filtering ? The sound waves are still going to overlap each other regardless of driver orientation, right ? Or are you saying that it’s just not as noticeable in a vertical array ?

**Ayreonaut** · **10-18-06, 07:33 AM**

A loudspeaker should be designed physically and electronically to integrate the drivers smoothly. Sometimes a lot of effort goes into "time aligning" the output from the multiple drivers.

**Bruce** · **11-26-06, 09:43 AM**

A single integrated speaker (multiple drivers) has the crossovers custom designed for the driver location/geometry to minimize many undesirable effects, including baffle step response.

Your two (2) speakers side by side playing the same signal has not been crossover optimized for this specific geometry, in fact, it probably compromizes some of the original response curve, because the baffle design is radically changed from the original design.

Everybody has their own preferences about what they like, so what you like is all that really counts.

In that small room of yours, you might have been listening to what we call the near-field speaker response. In a much larger room it will be be much different for a far-field speaker response.

Some of us may even be hamstrung by what we know (ignorance is sometimes bliss).

**Dain Bramaged** · **11-26-06, 10:33 PM**

Thanks for the terminology/clarification. I would have to agree with you on the near-field effect, especially after learning more about accoustics from the posters above.

"ignorance is sometimes bliss" Man do I hear you loud and clear on that one !

**drf** · **11-27-06, 03:39 AM**

And HomeTheatreShack takes first prize for the best well informed and knowledgable members.

I have seen combfiltering arise on other forums and the responses have been, well to be honest, terrible.
I have seen people who think it only applies to single speaker enclosures, that it is only an accostical phenominum and that it has nothing to do with reflections or delay

. But here it it has been encapsulated and explained clearly and also with some rather indepth examples of how it can be experienced/dealtwith in situ'.

I am so glad I found this forum.

To add to the topic heres a link to an article by Rod Elliot, concerning phase/time delay and how it relates to comb filtering. If you get a chance to read other articles on his site I highly recommend it as they are both educational and eye openers.

http://sound.westhost.com/ptd.htm

**Dain Bramaged** · **11-27-06, 07:59 AM**

Thanks for the link. It gets a little too far into the electrical side for my abilities (still learning), but I found some of it iteresting. I've added the website to my list of audio websites and will look at the other articles as I get time. Thanks again.

**brucek** · **11-27-06, 10:49 AM**

Quote:

Is the big change in room size and layout going to act as a detriment ?

So when will you hook all this back up and experience how it sounds in the big room?

brucek

**Dain Bramaged** · **11-27-06, 08:07 PM**

Unfortunately I can't answer that yet. About two weeks prior to starting this thread, I had put in a contingent offer on another house. I had thought that by now I'd at least have some nibbles on this house. No such luck. The housing market in this area has all but crawled to a total stop. So depending on how long it takes to sell this house, will determine when I can start on the home theater in the new house. Fortunately for me, the owners of the other house haven't gotten any other offers. Even though the housing market is really slow, I don't understand why people aren't interested in this one. It's more than fairly priced, plus we just sunk about $20k worth of upgrades into it without raising the price to make it more attractive. Still nothing. Go figure. I just hope that the other house doesn't get sold out from underneath me. At this point, all I can do is wait and see what happens. I'd appreciate any volunteers for good luck finger crossing.

**Dain Bramaged** · **05-11-07, 09:53 PM**

WELL IT'S ABOUT TIME !!!! GOOD GRIEF !! I was beginning to think we were never going to sell this place ! FINALLY !! And the house we wanted was still for sale so we bought it. Yea !

I deliberately started staying away from HT forums because it was getting to be depressing not being able to set-up my stuff. But now there's light at the end of the tunnel. There's a lot of good info in this thread and I plan on referring back to it when I start getting the theater set up. Thanks for the help. So don't be surprised if I start asking more questions. But that will be several months away yet.

**Wayne A. Pflughaupt** · **05-13-07, 09:25 PM**

You’re not gong to stay away until then, are you?

Regards,
Wayne

**Dain Bramaged** · **05-13-07, 09:43 PM**

Advertisement