Are three inexpensive subs better than one expensive sub-Dr. Earl Geddes Approach - Page 70 - AVS Forum
Forum Jump: 
Reply
 
Thread Tools
post #2071 of 2083 Old 06-07-2014, 06:40 PM
Member
 
cap'n's Avatar
 
Join Date: Jan 2009
Location: Los Angeles
Posts: 50
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 4 Post(s)
Liked: 12
Quote:
Originally Posted by gedlee View Post

Todd

In my usage of the terms "walls" I meant "the bonding surfaces" which would include the floor and ceiling. The point was that the absorption does not occur within the space but at the boundaries of the space. A "worst" case" is not very useful if it is "unrealistic".

OK. I had assumed even in the simpler model that the absorption occurred at the bounding surfaces. You are saying it is modeled as occurring IN the space? Then why would one be able to assign it to length/width/vertical surfaces? It seems that if it was occurring IN the space it would not matter. In any case I'm sure even the more complex model you are talking about has some assumptions/simplifications implicit. In the end, when I have compared to, for example, BEM model data, the (simpler) model is reasonably accurate. I always try to stress in my papers that results are very general, and only the most general observations and conclusions should be made. If you can model a room more accurately, more power to you.
cap'n is offline  
Sponsored Links
Advertisement
 
post #2072 of 2083 Old 06-07-2014, 08:19 PM
Senior Member
 
gedlee's Avatar
 
Join Date: Dec 2008
Posts: 342
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 1 Post(s)
Liked: 23
Todd

Your model "assumes" the absorption is uniform and hence occurs throughout the space, and yes, my model is more accurate, but even then there are simplifying assumptions - but they are fewer and less restrictive. It would be good to understand the differences.

Yes, I can have large values of absorption on just one surface and none of the others - the surfaces are completely independent. Its too bad that you can't run the software, it is quite unique and represents the ultimate that an analytic approach can do.

Earl R. Geddes
GedLee
gedlee is offline  
post #2073 of 2083 Old 06-07-2014, 09:09 PM
Member
 
cap'n's Avatar
 
Join Date: Jan 2009
Location: Los Angeles
Posts: 50
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 4 Post(s)
Liked: 12
Earl, I will try on a different browser or on my Win XP virtual machine. Again though, I'm not sure having large absorption on one surface accurately modeled is particularly useful. Maybe for modeling small rooms that open into much larger and absorptive rooms...? I would be curious to see if your model predicted similar or different results to mine, on a macro scale. That is, something like a like seat to seat variance metric calculated for a defined seating area and sub setup, for a large range of room dimensions - and plotted. Would you see the same types of patterns I get (like in my latest AES paper).
cap'n is offline  
post #2074 of 2083 Old 06-08-2014, 08:46 AM
Senior Member
 
gedlee's Avatar
 
Join Date: Dec 2008
Posts: 342
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 1 Post(s)
Liked: 23
If the damping in my model is set to be low then mathematically the calculations will be the same as what you do. They will differ only when the damping is high, but then your model is no longer accurate. Hence comparisons are not possible where the complex calculations matter.

Earl R. Geddes
GedLee
gedlee is offline  
post #2075 of 2083 Old 06-08-2014, 06:09 PM
Advanced Member
 
andyc56's Avatar
 
Join Date: Feb 2013
Posts: 576
Mentioned: 1 Post(s)
Tagged: 0 Thread(s)
Quoted: 17 Post(s)
Liked: 182
Quote:
Originally Posted by cap'n View Post

Well, a big chunk of SFM optimization can be done using location/gain/ an delay only. Those should not be equpiment specific. The biquad, yes, though I would think you could get pretty close using standard values.

Oh, I forgot to reply to this part.

I am not trying to duplicate the way SFM does it, because of the patents associated with SFM. The idea is to optimize the response flatness at multiple listening positions simultaneously via individual EQ of each sub, as opposed to the SFM approach of first optimizing mean spatial variance with individual EQ, then following with global EQ to achieve overall response flatness. The only global EQ I would use would be for a "house curve" to boost 20 Hz by a few dB relative to the upper bass as recommended by Earl. The mains are a part of this optimization, so it is "optimizing the splice" as they say around here. I got the idea for the software from Earl's multi-sub video and thought I'd give it a try.

This approach should reduce mean spatial variance, but only as a side effect, as it does not optimize mean spatial variance per se. If one listening position has a response dip at some frequency while another has a peak there, and vice versa, this could make the MSV larger than if MSV were optimized separately. But I am not all that concerned, as long as the response at all listening positions is as flat as it can be.
andyc56 is offline  
post #2076 of 2083 Old 06-08-2014, 11:31 PM
Member
 
cap'n's Avatar
 
Join Date: Jan 2009
Location: Los Angeles
Posts: 50
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 4 Post(s)
Liked: 12
Quote:
Originally Posted by gedlee View Post

If the damping in my model is set to be low then mathematically the calculations will be the same as what you do. They will differ only when the damping is high, but then your model is no longer accurate. Hence comparisons are not possible where the complex calculations matter.

Hi Earl, What do you mean by "low" and "high" absorption? Other than an opening, or some fairly deep absorptive material, high absorption at low frequency would be rare (depending on what you mean by "high" I guess). I have always focused on low absorption case, since that is where modal problems are worst.
cap'n is offline  
post #2077 of 2083 Old 06-09-2014, 08:59 AM
Member
 
neworder71's Avatar
 
Join Date: Mar 2007
Posts: 146
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 1 Post(s)
Liked: 10
Quote:
Originally Posted by gedlee View Post

Andy - good point.

My CLD is made on the walls themselves. They are not a "product". They are made with standard RC-1 resilient channel and sheets of dry-rock bonded together with a flexible mastic. How they are made is in the book. The goal is to dampen the room at LFs the best you can because no amount of subs or EQ can beat a well damped room. But LF damping is exceedingly difficult to do and has to be done IN the room construction. You can almost never "add" good LF absorption after the room is constructed. You would have to tear the room down to the studs and start over.

Earl, I've gotten through most of your HT paper, very interesting. In it you mention a Helmholtz resonator. I plan to build a riser (I can't see over my front row), so I thought this might work for me. Other threads in AVS claim a Helmholtz resonator is difficult to get right or not that effective. Those threads recommend opening the sides of a riser as much as possible and stuff with insulation to make a broadband absorber. I find my room a bit dead at higher frequencies already, so going broadband doesn't appeal to me. Your instructions for the Helmholtz resonator seemed straight forward. But I have a couple questions:
1. it will border one side wall but not the back wall, does that matter?
2. Does it matter which side the opening are place on? Front, open side or back?
3. Rough size 9'x7'x8" (still need to do precise calcs of size and volumes), anything I need to consider or look out far? Is there something I'm missing that make this more difficult?

I realize going CLD is the best option for LF damping, but my DW will not allow a "to the studs" tear down.
neworder71 is offline  
post #2078 of 2083 Old 06-09-2014, 10:21 AM
Senior Member
 
gedlee's Avatar
 
Join Date: Dec 2008
Posts: 342
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 1 Post(s)
Liked: 23
It should not matter that the riser does not border the back wall all long as it is enclosed. Otherwise it cannot be a Helmholtz resonator. It also should not make much difference which side the ports face. Remember that the effectiveness of the resonator depends on the area of the port so don't make it too small.

You can do CLD on existing walls without going to the studs, t just won't be as effective.

Todd. In my software the absorption is input as Alpha, the absorption coefficient. This would almost certainly never exceed about .5 at LFs and even that would be very difficult. That said, with floating CLD walls I suspect that a value of .5 is achievable. In my room strong modes are hardly evident al all except to the lowest one at about 25 Hz. Above that no strong modes are apparent at all. This is not typical of course, but it is doable and I prefer to work with the better environments, not the typical ones. The low damping simplification is probably violated at values much above about .05-.1, almost no damping at all.

Earl R. Geddes
GedLee
gedlee is offline  
post #2079 of 2083 Old 06-09-2014, 02:31 PM
Member
 
neworder71's Avatar
 
Join Date: Mar 2007
Posts: 146
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 1 Post(s)
Liked: 10
Quote:
Originally Posted by gedlee View Post

Remember that the effectiveness of the resonator depends on the area of the port so don't make it too small.

Thanks...I may be a bit rusty on my math...but I'm coming up a with a very small port:

S=7.23 square inches

given:
30hz tune
c=343 m/s (converted to english units)
l'=1.725 inches (1.5 inches thickness of 2x8 times 1.15)
V=21504 cubic inches

Of course I can change l' and the port size goes up. I just want to make sure I'm calculating this correctly before cutting wood. Is there a minimum recommended neck length?

This would be my 2nd biggest volume, there would be 3 others.
neworder71 is offline  
post #2080 of 2083 Old 06-09-2014, 02:42 PM
Member
 
neworder71's Avatar
 
Join Date: Mar 2007
Posts: 146
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 1 Post(s)
Liked: 10
Quote:
Originally Posted by neworder71 View Post

Thanks...I may be a bit rusty on my math...but I'm coming up a with a very small port:

S=7.23 square inches

given:
30hz tune
c=343 m/s (converted to english units)
l'=1.725 inches (1.5 inches thickness of 2x8 times 1.15)
V=21504 cubic inches

Of course I can change l' and the port size goes up. I just want to make sure I'm calculating this correctly before cutting wood. Is there a minimum recommended neck length?

This would be my 2nd biggest volume, there would be 3 others.

For the other 3 cavities...doubling or having volume and port area...I end up with the same 30hz keeping c and l' constant. Does that sound right?
neworder71 is offline  
post #2081 of 2083 Old 06-10-2014, 08:16 AM
Senior Member
 
gedlee's Avatar
 
Join Date: Dec 2008
Posts: 342
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 1 Post(s)
Liked: 23
Half the volume and half the port area will remain the same tuning, but half as effective. I would increase the port length to he maximum you can do as this will yield the maximum value of absorption.

Earl R. Geddes
GedLee
gedlee is offline  
post #2082 of 2083 Old 06-10-2014, 11:13 AM
Member
 
neworder71's Avatar
 
Join Date: Mar 2007
Posts: 146
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 1 Post(s)
Liked: 10
Quote:
Originally Posted by gedlee View Post

Half the volume and half the port area will remain the same tuning, but half as effective. I would increase the port length to he maximum you can do as this will yield the maximum value of absorption.

Thanks. I'm guessing the external volume of the port should be subtracted from the volume of the chamber? It will be come more significant as you increase length.
neworder71 is offline  
post #2083 of 2083 Old 06-10-2014, 01:17 PM
Senior Member
 
gedlee's Avatar
 
Join Date: Dec 2008
Posts: 342
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 1 Post(s)
Liked: 23
That is all correct. I usually target about 1/3 - 1/4 of the total volume for the port.

Earl R. Geddes
GedLee
gedlee is offline  
Reply Subwoofers, Bass, and Transducers

User Tag List

Thread Tools
Show Printable Version Show Printable Version
Email this Page Email this Page


Forum Jump: 

Posting Rules  
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts

BB code is On
Smilies are On
[IMG] code is On
HTML code is Off
Trackbacks are Off
Pingbacks are Off
Refbacks are Off