Re: Re: Re: Re: Re: Re: Re: Re: Multiple Small Subs - Geddes Approach
The reason why human can distinguish tones of different frequency is because it ocurrs long enough. Therefore reverberation time has everything to do with listening impression, and does come down to personal preference. If you had the same measured SPL response in an anechoic chamber and in a room you will hear significant difference.markus76 said:
The preference to have huge level variations (?5dB) at low frequencies?
Multiple subs do nothing to the low frequency reverberation time, that's true. Only active (not available as a product) or passive absorption can help. But what multiple subs do is to smooth the frequency response in a way it can't be done with any other method (I know of). That IS the number one goal for low frequency reproduction.
What are you trying to say?
We are talking about sound reproduction in acoustically small spaces. There will never be an anechoic living room, especially not at low frequencies. A tone of 80 Hz has a wavelength of 4.30 m! There's no way to escape the immediate superposition of the room's response at low frequencies. Even before you get the chance to determine a low frequency sound's pitch the room already has taken over.
The soundfield in small rooms has to be divided into several frequency ranges to understand its audible effects. You can't apply geometric acoustics to the modal region.
We are talking about sound reproduction in acoustically small spaces. There will never be an anechoic living room, especially not at low frequencies. A tone of 80 Hz has a wavelength of 4.30 m! There's no way to escape the immediate superposition of the room's response at low frequencies. Even before you get the chance to determine a low frequency sound's pitch the room already has taken over.
The soundfield in small rooms has to be divided into several frequency ranges to understand its audible effects. You can't apply geometric acoustics to the modal region.
gedlee said:
That's the key difference - symmetry.
When I refer to a Welti arrangement, I am talking about four corners, four wall midpoints or two midpoints. This was the recommended placement at the conclusion of the original Welti study, which was put into a whitepaper on the Harman International website.
When I talk about a Geddes arrangement, I am talking about pseudo-random placement, specifically one in a corner, one at mid height and at least one in a random place.
Personally, I have always felt that measurements were key, always had a slight concern that neither set of rules could be considered absolute. I had modeled rooms of various shapes with listening areas in different places, each having both Geddes and Welti arrangements.
It became clear after a while that there was no clear winner. So my conclusion was start with an idea where you'd like the subs to be, keeping in mind what your goal is. It can be a Geddes arrangement or a Welti arrangement, try them both or neither. Just don't bunch the subs up together.
You might have to move them around a little, perhaps reverse connections on one or more subs. But with even just two subs, you'll find response is made much smoother if you put them in the right places. Three is better still and with four, you can put them practically anywhere and expect smooth bass.
Wayne Parham said:
Hey Wayne,
Actually its 1) in corner, a must 2) along an opposing wall someplace (lots of possibilities here), recommended 3) anywhere but a corner, away from the first two and if possible, above the center line.
This keeps getting misquoted all the time. Lets get it straight from now on OK.
To me there is a clear winner - three subs instead of four is a pretty clear advantage. You seem to want to hold out on that assesment, but I don't.
I know you like your pseudo-random three-sub placement. You've been adamant about it for years, saying it was fundamentally different than Welti's proposed placements. That's why I was surprised to see comments here that appeared to declare them as equivalent methods. I would suggest they are similar, probably equivalent in some situations, but still there are some differences.
As I recall, when you and I first talked about this, you used to simply call your placement "random". Your beef with Welti was the symmetry, and your point was to introduce some random element, to decorrelate the sources. Looking back through E-mails and forum posts, this is what I see.
I suggested that random included a lot of undesirable placements like all subs lumped together so you came up with a set of rules to enforce a breakup of the groupings. That's when you started saying they should be put in one corner, one above mid-height and one random. You made this very clear, because I wanted to do a comparitive study and I asked for actionable directions for a "Geddes placement" to compare with Welti's.
When did you refine your original placement suggestions to include a wall? Last time I talked to you, it was one in a corner, one above mid-level vertically and the third random. Now it is corner, along an opposing wall and above mid-height for the third? It's great to make refinements, and this suggestion is definitely new since I talked to you last. Didn't I also read the mid-height suggestion is optional?
By the way, as far as three subs being better than four, I want to make the point that sometimes two subs do pretty well too, especially if the mains go low and are in the right places. You can usually do better with three and almost can't go wrong with four but the point is sometimes a pair is enough.
As I recall, when you and I first talked about this, you used to simply call your placement "random". Your beef with Welti was the symmetry, and your point was to introduce some random element, to decorrelate the sources. Looking back through E-mails and forum posts, this is what I see.
I suggested that random included a lot of undesirable placements like all subs lumped together so you came up with a set of rules to enforce a breakup of the groupings. That's when you started saying they should be put in one corner, one above mid-height and one random. You made this very clear, because I wanted to do a comparitive study and I asked for actionable directions for a "Geddes placement" to compare with Welti's.
When did you refine your original placement suggestions to include a wall? Last time I talked to you, it was one in a corner, one above mid-level vertically and the third random. Now it is corner, along an opposing wall and above mid-height for the third? It's great to make refinements, and this suggestion is definitely new since I talked to you last. Didn't I also read the mid-height suggestion is optional?
By the way, as far as three subs being better than four, I want to make the point that sometimes two subs do pretty well too, especially if the mains go low and are in the right places. You can usually do better with three and almost can't go wrong with four but the point is sometimes a pair is enough.
My feeling is that you can have an excellent result with only two subs if you're completely free with their placement and have a lot of time (days). But that's virtually never the case. So 3 subs seems to be the best recommendation that will yield best results in a vast variety of domestic listening rooms.
Wayne
You are arguing semantics. Welti's and my approach are "fundamentally" the same - both use multiple subs. The details differ.
The original argument was your claim that Welti's four symmetrical subs was "optimum" and I disputed that saying that a more random choice was "just as good with three subs as Welti's with four". You asked for a specific arraingment to test and I gave you one. It's ridiculous to claim that "random" includes lumping the sources together.
From the begining I have suggested one in a corner, this has never changed. And I have suggested one mid-way along a wall but as far from the corner one as possible. This equates to along an opposing wall - where exactly is not as important. The third sub has always been a random choice, but again, it should be as far away from the other two as possible and above the midline appears beneficial. I have never changed on these recomendations and please don't suggest otherwise. But this is NOT completely random, but its not symmetric either.
These are not "fundamental" differences between Welti and myself. In the context of using multiple subs or not its a minor difference. In the context of "I'm going to use multiple subs, how many should I use and where should I place them?", its a bigger difference.
But there are no hard and fast rules here just as there is no one room which we are talking about. The optimum is as varied as there are rooms. There are "principles" that one can use and I have stated those consistantly all along.
I've often used two subs. It can work. Three virtually always works. One is usually a disappointment.
You are arguing semantics. Welti's and my approach are "fundamentally" the same - both use multiple subs. The details differ.
The original argument was your claim that Welti's four symmetrical subs was "optimum" and I disputed that saying that a more random choice was "just as good with three subs as Welti's with four". You asked for a specific arraingment to test and I gave you one. It's ridiculous to claim that "random" includes lumping the sources together.
From the begining I have suggested one in a corner, this has never changed. And I have suggested one mid-way along a wall but as far from the corner one as possible. This equates to along an opposing wall - where exactly is not as important. The third sub has always been a random choice, but again, it should be as far away from the other two as possible and above the midline appears beneficial. I have never changed on these recomendations and please don't suggest otherwise. But this is NOT completely random, but its not symmetric either.
These are not "fundamental" differences between Welti and myself. In the context of using multiple subs or not its a minor difference. In the context of "I'm going to use multiple subs, how many should I use and where should I place them?", its a bigger difference.
But there are no hard and fast rules here just as there is no one room which we are talking about. The optimum is as varied as there are rooms. There are "principles" that one can use and I have stated those consistantly all along.
I've often used two subs. It can work. Three virtually always works. One is usually a disappointment.
markus76 said:
Which is why it is nice to be able to model the room, to get a first look at your options. It's an easy way to do a lot of what-ifs, find the general areas that seem to work best, then setup and measure for fine tuning.
There may be other options around, but I've been using CARA for a few years and really like it. You can model your room to whatever level of detail suits you, then positioning subs within it is easy.
For a serious new design CARA looks quite reasonable. But somehow finding the ideal speaker locations and then being able to actually put them there is another thing. Usually the subs go where they can, not where they should. Then its how to set them up to get the optimum. This, to me, is the heart of the problem. I agree that some good PC software that did this for you, like the JBL box, would be very useful. But I'm not about to start writting code again.
For me there are two very important things that people need to remember when attempting this. Two of the three subs should be able to be located fairly easily within the listening environment; the corner and adjacent wall subwoofer. Don't forget the third on which is for non-corner above centerline height placement. I have not modeled any rooms but this makes complete sense to me because we are attempting to even out a frequency response that is room dominated in a 3-D space. To do this best makes me think that something above the centerline would be of great advantage.
The second item is that some people state that doing level setting by ear is fine. Having seen a setup performed confirmed the fact that measurements are a huge benefit. Having so many variables from placement to levels to phasing just leaves too many possibilities for ear tuning. I would guess that most people who do tuning by ear end up with a very bass emphasized playback.
The second item is that some people state that doing level setting by ear is fine. Having seen a setup performed confirmed the fact that measurements are a huge benefit. Having so many variables from placement to levels to phasing just leaves too many possibilities for ear tuning. I would guess that most people who do tuning by ear end up with a very bass emphasized playback.
Markus,
Thanks for your write-up. It is excellent.
It would be very helpful if you could provide a bit more information about the final XO points, XO slope and EQ you used in setting up your subs resulting in the 4 graphs shown in your write-up.
A number of questions regarding the XO points and XO slopes were previously raised by a few people but I have found no descriptions about those yet.
Regards,
Bill
Thanks for your write-up. It is excellent.
It would be very helpful if you could provide a bit more information about the final XO points, XO slope and EQ you used in setting up your subs resulting in the 4 graphs shown in your write-up.
A number of questions regarding the XO points and XO slopes were previously raised by a few people but I have found no descriptions about those yet.
Regards,
Bill
