DSP midrange directivity control aka kinda cardioid - Page 2 - AVS Forum
Forum Jump: 
 3Likes
Reply
 
Thread Tools
post #31 of 56 Old 08-13-2014, 06:32 AM - Thread Starter
Senior Member
 
Nate Hansen's Avatar
 
Join Date: Aug 2012
Location: Rochester, MN
Posts: 213
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 50 Post(s)
Liked: 52
Quote:
Originally Posted by Martycool007 View Post
Couple of quick questions, what does a High Shelf and/or Low Shelf filter do of accomplish? I keep seeing people talk about using a High or Low Shelf filter, just never way sure what that does, exactly?


If you want I can send you a file to play with in REW. I always learn best by getting hands-on.
Nate Hansen is online now  
Sponsored Links
Advertisement
 
post #32 of 56 Old 08-13-2014, 12:05 PM
Advanced Member
 
MBentz's Avatar
 
Join Date: Jun 2006
Location: Chicago
Posts: 949
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 21 Post(s)
Liked: 41
Nate, what are your thoughts about using the corners of the room to accomplish the directivity that you seek?

That ~2dB hit in sensitivity is a bit less elegant than the elegant polars you're getting. Have you experimented with moving the slots further back - maybe using the rear corners instead? I would think the frontal lobe attenuation would be reduced the further back you put the rear-facing attenuator.

Also, what about a variable width slot? What I'm thinking is a way of making the Q of the "port resonance" wider in such a way as to only allow low frequencies to come out the back - and maybe in the process you can get some extra LF boost (although that will be minimal with a low Q design).

I've been trying to write my own directivity prediction software, but I haven't found a free version of MATLAB that implements Jacobian functions properly. Sci-Lab has been the best one I've found thus far. You might be able to use Akabak to simulate the Q of a variable width port (you could just do it with multiple ports of different diameters to approximate the idea - or you could just build it that way too).

-Mike Bentz
~It's all about compromise~
MBentz is offline  
post #33 of 56 Old 08-13-2014, 12:35 PM - Thread Starter
Senior Member
 
Nate Hansen's Avatar
 
Join Date: Aug 2012
Location: Rochester, MN
Posts: 213
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 50 Post(s)
Liked: 52
If I had good room corners is build corner horns in heartbeat! It was something I was considering when I was house shopping last year.

Unfortunately the efficiency loss is a necessary evil of dipole/cardioid systems due to the way they get the directivity by cancellation. This is mitigated to a point by using higher eff drivers. No free lunch as usual. In my system I use flanking woofers up to 300 hz so I won't be applying the full 6dB of roll-off compensation.

I'm fairly sure that rear corner vents won't work due to the greater distance to the baffle plane. It would work to a certain point (my rear vent was only good to 200hz or so) then the pattern would widen before the piston diameter takes over. Varying the slot position on the side is something I ought to investigate so thanks for pointing that out.

I'll have to think on the variable width slot.
Nate Hansen is online now  
post #34 of 56 Old 08-13-2014, 04:27 PM - Thread Starter
Senior Member
 
Nate Hansen's Avatar
 
Join Date: Aug 2012
Location: Rochester, MN
Posts: 213
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 50 Post(s)
Liked: 52
Here's the plot of the rear vent cardioid to give you an idea of what I was talking about before. Really not very directive at all in the mid-bass and lower midrange. Basically similar to the monopole response but with greater rear attenuation. The vent itself is on the back of the cab and is about 15" square with a 19" path-length to the baffle plane:



Looks like rain this weekend, but if not I'll be out annoying my neighbors as usual . I'm going to rebuild my test box so it's narrower (22" wide right now for the corner vents). It'll be 19" wide and I'll have removable side panels to test different front to back slot positions. I may even drag a TD out and see how it does. My hope is this will be the final test and be close to the final box so I can start building. I'm thinking the final box will have 2" radius roundovers on the baffle.

One thing I haven't addressed, but have been thinking about, is directivity in the vertical plane. Everything I've done so far has basically ignored this. It would be interesting to see if the side vents have any effect. I'm sure it won't be as dramatic as what we see horizontally but I wonder if it changes things at all. At this point I can't really think of a good way to measure it without impeding the side vents. I might add a vent on the bottom of the cab (it'll be up off the floor about 12" so I'll have to build stands to accommodate). Maybe increased directivity in the vertical plane can help with the floor bounce...
Nate Hansen is online now  
post #35 of 56 Old 08-14-2014, 03:59 AM
AVS Special Member
 
Martycool007's Avatar
 
Join Date: Jun 2011
Posts: 3,692
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 280 Post(s)
Liked: 172
Quote:
Originally Posted by Nate Hansen View Post


If you want I can send you a file to play with in REW. I always learn best by getting hands-on.
Nate, thanks for the reply! It is my understanding that a Low Shelf Filter is used to either boost or cut frequencies below the Fc, and conversely, a High Shelf Filtr can boost or cut frequencies above the Fc, is this correct?

When is it beneficial to use a Low or High Shelf Filter? I mean, if you have access to a DSP/EQ-able setup like I have, iNuke1000dsp's for LCR's, can you not just use traditional boost or cuts to the frequencies in question? What does a Low or High Shelf filter offer over these methods?
Martycool007 is offline  
post #36 of 56 Old 08-14-2014, 12:33 PM
Advanced Member
 
MBentz's Avatar
 
Join Date: Jun 2006
Location: Chicago
Posts: 949
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 21 Post(s)
Liked: 41
1) What frequencies are you referring to when you say mid-bass? We have data here, it'd be easier to communicate if we use actual number. Or maybe I'm just being a stubborn engineer

2) What is that dip at ~220Hz showing up in every measurement? I would recommend adjusting the measurement configuration to guarantee you're not dealing with reflections at these frequencies. How do you know they're not affecting what you think the polars are doing?

3) I think minimizing front lobe attenuation should be a major focus of the approach - or at least a means of comparing the merit of different approaches. What is the relative weighting of the polar width versus distortion? I think it should be possible to avoid front lobe attenuation. I know I've seen subwoofer measurements getting as low as only 0.5dB on the front lobe - there was a small expense at the polar pattern, but I personally would prefer that versus 2dB on the front lobe with perfect polars. The power response in the room can be adjusted via other means, but distortion can't...

4) You'll improve the SNR of your measurements if you start by EQ'ing the on-axis response to be flat. In REW I'll take that mostly flat EQ'd response and then make it my calibration file - now everything is referenced to that on-axis response and I don't have to twist the response plots in my head to understand the attenuation for each axis.

5) I noticed that you're increasing the distance between angles as you get wider out. Clover-leafing usually happens at the wider angles, so I would recommend keeping a 10 degree step all the way through. I built myself a lazy susan to make this go faster (I cut holes for each angle and used a pin).


Anyways, just thinking out loud - as you've mentioned I've seen this pop up very infrequently, but it's something I think is worth investigating a little bit....especially if it's as simple as cutting holes in the box. I've decided on going the full-sized horn approach, but I may give this another look since the solution size should be smaller.

-Mike Bentz
~It's all about compromise~
MBentz is offline  
post #37 of 56 Old 08-14-2014, 03:04 PM - Thread Starter
Senior Member
 
Nate Hansen's Avatar
 
Join Date: Aug 2012
Location: Rochester, MN
Posts: 213
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 50 Post(s)
Liked: 52
1.) Mike - sorry, when I say midbass I'm speaking about 100-300hz. Might as well use numbers!

2.) I'm fairly sure that dip at 220 hz is a reflection due to the distance of the woofer to the ground. It shows up in every measurement where I was using the old modified box. If you look at the measurements of the front corner vent cab it's gone because the driver center is lower. Good point about it affecting the polars as it very well could, though I'm mostly concerned about 300hz and up.

3.) I'd like to have my cake and eat it too , but I just don't see how this is possible. The whole point of this is the increased directivity and the lower sensitivity is the trade off. I'm willing to trade that for the polar pattern to be honest. I really don't see how this can be improved with this kind of design....it's the nature of the beast. My version with the 8" side drivers did well to control the pattern from 300hz+ while leaving the region below that omni so there's little sensitivity lost where I believe it would matter most: at lower frequencies. The trade-off there is more $$ in drivers and amps. This was the direction I thought I'd go when I started all this but when I saw I could get close with the acoustic resistance box I switched gears.

I tend to listen at about 80-85dB, and you could figure 20-25dB for dynamic peaks on good recordings (to be honest the stuff I listen to isn't that good). Is that 2dB of attenuation really that huge of a deal with quality pro drivers used in the home? These will be high passed at 80hz.

4.) I didn't take the time to eq everything flat because I was working on 5-6 different iterations each time I was out. It would increase the SNR to be sure, but I felt that I was high enough above the noise floor that I could get a good picture of the polar pattern as is.

5.) I increase the angle as I get wider out for a few reasons. One - to spare my neighbors having to listen to 30 sweeps for each set of polars . Also it saves time. I don't think it's as critical to be as detailed at the wider angles since it's far out of the listening window, but it would be worthwhile to check it out once I get close to a final design. I have a thread about some diy waveguides (quiet now) going over on DIYA, and Geddes mentioned that this is the measurement scheme he uses. To me his reasoning is sound.

I have a lazy Susan turntable I built so doing this is all pretty easy once I have everything setup.
Nate Hansen is online now  
post #38 of 56 Old 08-14-2014, 04:21 PM - Thread Starter
Senior Member
 
Nate Hansen's Avatar
 
Join Date: Aug 2012
Location: Rochester, MN
Posts: 213
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 50 Post(s)
Liked: 52
Marty - a shelving filter is useful because it raises the entire spectrum above or below the corner, depending if it's a high or low shelf. A peaking filter only affects a small range of frequencies depending on the q of the filter which we talked about before. If you look at my graphic again I think you will see how it works with this in mind.

A low shelf of about -8dB centered around 800hz would be useful for the response shown in post 34.
Nate Hansen is online now  
post #39 of 56 Old 08-15-2014, 04:51 AM
AVS Special Member
 
Martycool007's Avatar
 
Join Date: Jun 2011
Posts: 3,692
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 280 Post(s)
Liked: 172
Quote:
Originally Posted by Nate Hansen View Post
Marty - a shelving filter is useful because it raises the entire spectrum above or below the corner, depending if it's a high or low shelf. A peaking filter only affects a small range of frequencies depending on the q of the filter which we talked about before. If you look at my graphic again I think you will see how it works with this in mind.

A low shelf of about -8dB centered around 800hz would be useful for the response shown in post 34.
So based on the plot in post #34 , a low shelf filter of around -8db at 800hz would take down everything below that 800hz and make the response more flat? It is my understanding that by using a Low or High Shelf Filter you are either cutting, ie: a low shelf filter, or boosting, ie: a High Shelf Filter over the entire range either above or below the number specified by the user? That makes sense, only thing I am unsure of is when to use a High or Low Shelf Filter, instead of a traditional filter, ie: boost or cut at "x" or "y" frequency with what ever Q you need? Are the shelf filters mainly beneficial when the entire response needs to he adjusted above or below a specific area?
Martycool007 is offline  
post #40 of 56 Old 08-15-2014, 09:46 AM
Advanced Member
 
Bassment's Avatar
 
Join Date: Aug 2013
Location: Ontario, Canada
Posts: 887
Mentioned: 1 Post(s)
Tagged: 0 Thread(s)
Quoted: 401 Post(s)
Liked: 161
Marty, a high shelf will increase from the chosen frequency and above. So say you want to boost the treble, you will high shelf at 2000 hz by 2 dB. If you want to lower the treble, high shelf at 2000 hz -2 dB. Low shelf goes the opposite way in the frequency spectrum. If you want to boost from 0-30 hz, you low shelf at 30 hz + 5 dB.

A traditional filter will just increase or decrease the SPL at a specific frequency or over a small range. A shelf can do it over a large range. You are correct in saying shelf filters are beneficial when the entire response needs to be adjusted above or below a specific frequency.
Bassment is offline  
post #41 of 56 Old 08-15-2014, 11:14 AM
AVS Addicted Member
 
LTD02's Avatar
 
Join Date: Oct 2003
Posts: 16,841
Mentioned: 23 Post(s)
Tagged: 0 Thread(s)
Quoted: 555 Post(s)
Liked: 1164
cool project.


just thought that I'd kick in that this method works well for controlling the rear wave for bass applications.
the front wave summation though isn't perfect because (obviously) the rear driver running on both a distance and electronic delay doesn't sum the leading edge of the front firing wave.
also, the summation is frequency dependent -- some frequencies will be running with more or less summation than others.
might be best to try a large physical horn that just doesn't go quite as low. not sure...

Listen. It's All Good.
LTD02 is offline  
post #42 of 56 Old 08-15-2014, 12:03 PM - Thread Starter
Senior Member
 
Nate Hansen's Avatar
 
Join Date: Aug 2012
Location: Rochester, MN
Posts: 213
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 50 Post(s)
Liked: 52
Ltd - my intention before I started building these waveguides and "cardioid" boxes was to build large Synergy horns. As you can see I got a little sidetracked.......at this point I intend to take the waveguide and directive 15 as far as I can. I'll likely tackle the big horn project next year.
Nate Hansen is online now  
post #43 of 56 Old 10-24-2014, 05:14 PM - Thread Starter
Senior Member
 
Nate Hansen's Avatar
 
Join Date: Aug 2012
Location: Rochester, MN
Posts: 213
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 50 Post(s)
Liked: 52
Curved cardioid cabs coming along.....
rhodesj likes this.
Nate Hansen is online now  
post #44 of 56 Old 10-25-2014, 05:21 AM
AVS Special Member
 
Martycool007's Avatar
 
Join Date: Jun 2011
Posts: 3,692
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 280 Post(s)
Liked: 172
Nate, I know that we have gone over your priorities for this build, several times, but, would this whole process and overall goal of having constant directivity from 300hz and up not be easier by simply building a 3-way with a mid- horn for the 300hz to 1,000hz area crossing to a compression driver and waveguide at say 900hz ( ie Seos-15 + DNA-360) as well as crossing at 300hz to a traditional woofer? Would this not provide constant directivity in the same manor from 300hz to 900hz? Doesn't Wayne's mid- horn hold pattern down to around that point? (I think he even offers it in flat pack form and uses it in his 7Pi and 8Pi speakers).
Martycool007 is offline  
post #45 of 56 Old 10-25-2014, 05:55 AM - Thread Starter
Senior Member
 
Nate Hansen's Avatar
 
Join Date: Aug 2012
Location: Rochester, MN
Posts: 213
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 50 Post(s)
Liked: 52
Depends on your definition of easier. A waveguide that controls directivity to 300hz would be about 45" wide. In that case I'd make a synergy horn.......and that's a project for next year.

The Pi midrange horn is 24" wide, which gets you control down to around 550hz. It's s straight conical horn with no secondary flare or round over, so I would think the pattern will narrow before it loses control if it's used down to 300hz. This is something you would see in a contour plot, and it's called waist-banding. The Pi midrange horn is meant to be used in a corner, where the walls of the room extend the horn.

My reason for going with cardioid is to be able to use my TD15 woofers and still get reasonable directivity control down to 300hz.
Nate Hansen is online now  
post #46 of 56 Old 10-25-2014, 09:18 AM
AVS Special Member
 
tuxedocivic's Avatar
 
Join Date: Feb 2010
Location: Ladysmith, BC
Posts: 5,096
Mentioned: 11 Post(s)
Tagged: 1 Thread(s)
Quoted: 567 Post(s)
Liked: 659
Quote:
Originally Posted by Nate Hansen View Post
Curved cardioid cabs coming along.....
Sick!

Those are some huge round overs
tuxedocivic is offline  
post #47 of 56 Old 10-25-2014, 12:06 PM - Thread Starter
Senior Member
 
Nate Hansen's Avatar
 
Join Date: Aug 2012
Location: Rochester, MN
Posts: 213
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 50 Post(s)
Liked: 52
Thanks. It's facing up in the pic. 3" radius on the baffle (only on the sides) and the back corners are 12" radius because it looks cool.....
Nate Hansen is online now  
post #48 of 56 Old 10-25-2014, 06:39 PM
AVS Special Member
 
DS-21's Avatar
 
Join Date: Jun 2005
Posts: 3,554
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 211 Post(s)
Liked: 247
That's really cool!

Where are the vents going to be?

--
"In many cases there aren’t two sides unless one side is 'reality' and the other is 'nonsense.'" - Phil Plait
Serious Audio Blog 
Multichannel music (and video) urban loft living room system 
DS-21 is offline  
post #49 of 56 Old 10-26-2014, 07:16 AM - Thread Starter
Senior Member
 
Nate Hansen's Avatar
 
Join Date: Aug 2012
Location: Rochester, MN
Posts: 213
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 50 Post(s)
Liked: 52
The vents will start where the baffle roundover meets the sides. Unfortunately I might have to shelve this project until next year. I'm going to look at it closely in a bit but I don't think some of my kerf cuts were fully penetrated by the glue, so it's a bit wobbly. Maybe that was just the fiberglass fumes .
Attached Thumbnails
Click image for larger version

Name:	eoswg and cardioid model.PNG
Views:	17
Size:	126.8 KB
ID:	331666  
Nate Hansen is online now  
post #50 of 56 Old 10-26-2014, 07:46 AM
Advanced Member
 
Face2's Avatar
 
Join Date: Mar 2007
Location: Long Island, NY
Posts: 774
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 62 Post(s)
Liked: 58
Are those areas still accessible? You can fill them with putty.

Mike
Face2 is offline  
post #51 of 56 Old 10-26-2014, 11:18 AM - Thread Starter
Senior Member
 
Nate Hansen's Avatar
 
Join Date: Aug 2012
Location: Rochester, MN
Posts: 213
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 50 Post(s)
Liked: 52
Well, there's actually some voids where there's glue on top, but not inside. That said they actually seem pretty solid and I'm going to move forward. I think I'll paint up the outside of the curves with fiberglass resin to strengthen it up a bit.
Nate Hansen is online now  
post #52 of 56 Old 11-02-2014, 04:08 PM - Thread Starter
Senior Member
 
Nate Hansen's Avatar
 
Join Date: Aug 2012
Location: Rochester, MN
Posts: 213
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 50 Post(s)
Liked: 52
I've got the curvy cabs to a point where they're fully functional, albeit ugly. They are stuffed with Ultra Touch recycled denim 3.5" thick insulation, and a TD15M. They were measured about 7' off the ground. GP would have been a better method for getting the low end but the idea was to work out a xo with the waveguide and see the directivity of the system. It was kinda breezy so my work with the wg was a flop and I didn't get the xo done. I did get some data on that but I'm not confident in it's accuracy. I do believe that I got good data on the cardioid cab so I figured I'd share. These are quite a bit different than the cabs I was experimenting with earlier in this thread because of the baffle roundovers......that changes the distance from the baffle plane to the rear vents which determines how high the rear radiation will work to control the directivity. With the 3" radius roundover that distance works out to about 4.5" which is 1/4wl of 750hz. The idea is to smoothly transition from the cardioid pattern to the piston diameter narrowing the dispersion. This seems to have worked out quite well I think. You can see from the full polar plot that (at least for the angles I can view in ARTA) these are not a perfect cardioid.......at all! They do control the directivity well though. From 300hz up there is about 10dB and greater rejection to the sides and rear. Not too shabby but more would be better.
Attached Thumbnails
Click image for larger version

Name:	full polar plot.PNG
Views:	20
Size:	56.7 KB
ID:	346153   Click image for larger version

Name:	cardioid contour plot.PNG
Views:	28
Size:	363.5 KB
ID:	346161   Click image for larger version

Name:	wg+cardioid.jpg
Views:	29
Size:	122.5 KB
ID:	346169  
Nate Hansen is online now  
post #53 of 56 Old 11-02-2014, 04:53 PM - Thread Starter
Senior Member
 
Nate Hansen's Avatar
 
Join Date: Aug 2012
Location: Rochester, MN
Posts: 213
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 50 Post(s)
Liked: 52
It took 9 filters to flatten out the response. It's pretty flat on its own from 300-1khz. There's a +5dB low shelf centered at 200hz to bring up the low end. The roll off is due to the rear wave cancellation and you see this with dipoles as well. I did the measurements at about 80-85dB and a distance of 8'. There's visually quite a bit more excursion on this cardioid cab than there is on a sealed cab at a similar level.....which is to be expected with that kind of boost. I'll be using these with mid bass modules, so I won't apply that much boost when I use them. Who knows, I may even set these on top of the mbms and high pass the cardioid cab at 200 or so. If one wanted to use these cardioids in a 2-way the TD15M is not the driver to use due to its limited x max but I knew that going in. Even with a more capable driver I'm not sure I'd want to use this type of setup in a dedicated 2-way. There's quite a bit of excursion for the relatively low levels I was testing at for a driver that's playing well up into the midrange.
Nate Hansen is online now  
post #54 of 56 Old 11-03-2014, 10:46 AM
Senior Member
 
pitviper33's Avatar
 
Join Date: Sep 2011
Location: Eastern TN
Posts: 423
Mentioned: 1 Post(s)
Tagged: 0 Thread(s)
Quoted: 66 Post(s)
Liked: 39
Very impressive Nate! The directivity plot looks great. So what's next? Wait for a calmer day and get the crossover dialed in?
pitviper33 is offline  
post #55 of 56 Old 11-03-2014, 11:02 AM - Thread Starter
Senior Member
 
Nate Hansen's Avatar
 
Join Date: Aug 2012
Location: Rochester, MN
Posts: 213
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 50 Post(s)
Liked: 52
Yeah, as long as it's not too cold out!

I should have tried it with a high pass, I'm sure that would have reduced excursion quite a bit.
Nate Hansen is online now  
post #56 of 56 Old 11-11-2014, 03:24 PM - Thread Starter
Senior Member
 
Nate Hansen's Avatar
 
Join Date: Aug 2012
Location: Rochester, MN
Posts: 213
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 50 Post(s)
Liked: 52
So I had twins!

These are functionally complete, though they're pretty fugly. At this point I'm thinking about bracing. It will be minimal, with a beefy brace tying the top and bottom panels together, and a brace from the back panel to the central brace. I'd like to try to use a 3/4" threaded rod that attaches to the back panel, ties into the vertical central brace, and then butts up against the driver magnet. Not quite sure how I'll tie all that together though when the driver is in and the box is stuffed. I may forego bracing altogether for now. My hope is to have these up and running with my waveguides by this weekend.
Attached Thumbnails
Click image for larger version

Name:	IMAG0173.jpg
Views:	18
Size:	87.3 KB
ID:	362305  
Nate Hansen is online now  
Reply DIY Speakers and Subs

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