Felt around tweeters

Larry Chanin

Hi All,

I started experimenting his afternoon with these cheap neoprene diffraction rings on my SDA-CRS+ which I'm using as a center channel speaker.

I've already upgraded the tweeters to the silk dome replacements. Here's a graph showing the before and after frequency response.

This seems to work in theory. I'm considering the more serious thick wool felt approach.

Larry

ttrentt

F1nut wrote:

JBL made a famous speaker!?!

Haha. Sorry, I am a big JBL fan - I know they seem to be frowned upon in this Polk community. Considering some of the prices JBL products pull on eBay, I would say they made some famous speakers. I am still looking for a set of 4311s

VR3

I am seriously thinking about trying these...

Larry did you notice a difference in sound, beyond the measurements?

Larry Chanin

Vr3MxStyler2k3 wrote:

I am seriously thinking about trying these...

Larry did you notice a difference in sound, beyond the measurements?

Hi Sid,

I haven't gotten that far yet. I've got five more tweeters in my setup and I want to measure what's going on before I get distracted by critical listening.

I don't have a trained ear and I don't want the placebo effect to unduely influence me. Frankly, I'm guessing that I won't be able to discern a difference with these rather basic rings, especially since its not feasible for me to do a blind A/B test where I can instantaneously switch between with and without the diffusions rings. Even if it turns out that I can't hear a difference, at least the measurements will allay any fears that I might be hurting my setup.

The measurements suggest that this is not entirely one of those worthless tweaks and it may be worth the additional effort to experiment with an AR Acoustic Blanket type approach. As you know the vintage speakers have sharp edges and may be more susceptible to diffusion effects than the modern Polks that have carefully designed grills. So applying thick absorption to the baffles of vintage speakers might give us more bang for the buck than if they were applied to a new Polk speaker.

Larry

VR3

I actually just took some foam tape I had laying around (1/8" thick, 1/2" wide) -- and put it around my tweeters...I cant really say I hear a difference, possibly less sibilance in some songs, not sure - still listening.

Larry Chanin

Vr3MxStyler2k3 wrote:

I actually just took some foam tape I had laying around (1/8" thick, 1/2" wide) -- and put it around my tweeters...I cant really say I hear a difference, possibly less sibilance in some songs, not sure - still listening.

Hi,

I'll be traveling for the next several days and won't be able to get to this for a while. I was thinking that a crude way to do an A/B test would be to install the diffraction devices around the tweeters on say just your left main speaker and leave them off your right. Then listening to a recording centered on the soundfield, change the balance from all the way left to all the way right and back. Then see if you can hear a difference between the main speakers.

Larry

George Grand

Famous speakers and good speakers are two different things. Blose are pretty famous too wouldn't you say?

The AR Blanket extends 3/4" past the tip of the domes. The tips are not at all flush with the outer edge of the blanket.

I told you it worked Dr. Chanin.

Larry Chanin

George Grand wrote:

Famous speakers and good speakers are two different things. Blose are pretty famous too wouldn't you say?

Hi George,

Perhaps infamous would be the more accurate term.

George Grand wrote:

I told you it worked Dr. Chanin.

I'd say the guys to convince would be Max, and to a lesser degree Mark.

Seriously, I think measurements might still be important to make sure the absorber is being placed correctly.

Larry

ttrentt

George Grand wrote:

Famous speakers and good speakers are two different things. Blose are pretty famous too wouldn't you say?

haha. You got me there. Will always be a huge fan of older JBL though. The big cast frame woofers were always beautiful to me

scottnbnj

Larry Chanin wrote:

...I was thinking that a crude way to do an A/B test would be to install the diffraction devices around the tweeters on say just your left main speaker and leave them off your right. Then listening to a recording centered on the soundfield, change the balance from all the way left to all the way right and back. Then see if you can hear a difference between the main speakers.

maybe another step might be to listen in mono. if you have a solid, stable image to start with (before adding felt to one side) and the changes hold true to your graph, this should be revealing.

)

Larry Chanin

scottnbnj wrote:

Larry Chanin wrote:

...I was thinking that a crude way to do an A/B test would be to install the diffraction devices around the tweeters on say just your left main speaker and leave them off your right. Then listening to a recording centered on the soundfield, change the balance from all the way left to all the way right and back. Then see if you can hear a difference between the main speakers.

maybe another step might be to listen in mono. if you have a solid, stable image to start with (before adding felt to one side) and the changes hold true to your graph, this should be revealing.

)

Hi Scott,

I'm afraid I don't follow. Can you elaborate?

The reason I suggested the method above is because we have very poor memories when it comes to compairing sounds. If more than a few minutes go by it is extremely difficult to make correct critical comparisons. That's why the switching in A/B test must be done almost instantaneously. Twirling the balance dial is almost instantaneous and it might be fast enough to let us make a meaningful comparison. Of course without being a blind A/B test this method is not scientific and is still subject to placebo effect.

By the way, I just attempted an other measurement using one of my main speakers SDA-1C's. There was almost no measurable difference. So small that they should be inaudible to human hearing, less than 1/4 dB. Refer to a closeup of the frequency responses.

Larry

scottnbnj

nt, sorry i'm gonna try again

)

scottnbnj

sorry larry, i'm not a very technical kind of guy, but, i played with pink noise and real time analyzers a long time ago and have been trying to spend more time listening than thinking about listening for quite a while, so, i'm not really schooled or fluid in all this stuff,.. and elaborating on anything often makes my head hurt, i hope it doesn't make yours hurt too.

anyway. i agree with what you said about a/b'ing, and i wasn't at all trying to say don't do that or it won't work. looping a short revealing clip of music might make it a bit easier. listening to both channels in mono just lets you shift your attention more quickly, focus more acutely and have a real time comparison or control reference for some things.

i think we agree that measurements like your graphs could mean alot of different things. for where i was going earlier, if you're familiar with the image in mono with a variety of tunes and the mains showed the large difference (as in graph one) with felt added to one side, i would think, with that magnitude of change, it would reveal distinct image shifts or instability. the types of tunes that it showed up on and how might make the differences easier to hear and make sense of than test tones and a/b'ing alone.

i mean, in a case where there's a db or two of gain or cancellation on one side, the image is shifting, and it's audible if you pick tunes that reveal it.

i guess you measured with a continuous tone. you've thought about this alot so you know, but for others, common spl measurements from continuous tones don't differentiate between direct, reflected or any other sort of sound, distortion or cancellation, much less how much later it arrives than the direct sound from the drivers. so, if the gain was from reflections that travelled a much longer path than driver > baffle > listening position or resonances that, say, built up in the room or enclosure surrounding the speak, it would take a longer sustained note for real world (music) gain to approach what the test tones showed.

so, say the mains' mono image is normally centered, solid and stable over a wide frequency range. but, when you add thick felt to one speak, the image in narrow frequency ranges stuck to or shifted or fluttered mostly towards the untreated speak on only long sustained notes or sounds. my first guess would be that the dbs of difference were later reflections or resonances that the felt somehow lessened the impact of on the treated speaker side of the room rather than merely diffraction off of some part of the baffle bouncing straight to the listening position. that sort of gain would be much easier and quicker to identify and isolate with the right tunes and a pair of speaks in mono than with spl measurements and/or a/b'ing.

another quick example might be if the gain effects clarity. it's not uncommon for a strong center image to draw towards the problem speaker side of the room too. listening in mono would just quickly isolate the problem further. treat the image to where it is stable and well centered, clarity is regained, and peace is restored relatively painlessly.

that the difference didn't hold true on the mains, as you said, obviously means the difference will be much more subtle, like soundstage layering and/or shades of blackness (more what i'd expect from controlling just baffle diffraction). being that you're using sda's might make things more complicated too. nonetheless, listening in mono still might help you tell a little more about what's going on and how much difference you could expect to hear with both speaks treated. like lots of other tools, the more familiar you are with it, the deeper it lets you hear and read into what's happening.

sorry i went on, i hope it made some sense.

)

Larry Chanin

scottnbnj wrote:

i guess you measured with a continuous tone. you've thought about this alot so you know, but for others, common spl measurements from continuous tones don't differentiate between direct, reflected or any other sort of sound, distortion or cancellation, much less how much later it arrives than the direct sound from the drivers. so, if the gain was from reflections that travelled a much longer path than driver > baffle > listening position or resonances that, say, built up in the room or enclosure surrounding the speak, it would take a longer sustained note for real world (music) gain to approach what the test tones showed.

Hi Scott,

Sorry for the delayed reply, as I said I was traveling.

Actually I was using a short MLS (Maximum Length Sequence) test tone. This is a sort of white noise.

As I mentioned earlier, we are dealing with two phenomena associated with Early Reflections:

1. A psychoacoustic effect whereby our brains blend the delayed reflected sound and the direct sound via Precedence Effect, thereby smearing the sonic image.
2. Comb Filtering whereby the reflected sounds and the direct sound constructively and destructively interfer thereby causing nulls and peaks in the frequency response, and causing lobing in a speaker's dispersion pattern.

The first effect has absolutely no bearing on measurement method. It is a function of how our brains work when interpreting sounds spaced short intervals apart.

The second, comb filtering, is a transient effect that occurs when the first reflections interact with the direct sound. A resonance is a steady-state effect caused by the formation of standing waves. This requires sustained reflections and takes some time to build after the sound is applied, and some time to dissipate after the sound ceases.

The changes in frequency response that I, and the other folks were measuring, is not likely to have anything to do with resonance effects. In order to have a measurable effect on resonances (the modal response of a room) requires the addition of a significant amount of aborption, i.e. a significant amount of the surfaces of the room must be covered with absorbers. These felt strips and rings were so small and thin as to have no measurable effect on the formation of standing waves. In fact, the addition of a person into the room had many times the absorption of the diffusion devices and therefore would have had more effect on resonance than the diffusion devices.

What the small diffusion devices were doing was reducing the amount and level of the first reflection reaching the listeners. By removing the effect of these early reflections the amount of comb filtering was reduced and the frequency response at the listening position was improved.

Larry

scottnbnj

Larry Chanin wrote:

Actually I was using a short MLS (Maximum Length Sequence) test tone. This is a sort of white noise.

very cool. mlssa was still too new when i was hacking around with this stuff for me to get my hands on it. but, i keep stumbling across software on the net...

i went to the acoustisoft website. do you have any thoughts about the add-ons and how etf compares to other similar programs?

)

Larry Chanin

scottnbnj wrote:

very cool. mlssa was still too new when i was hacking around with this stuff for me to get my hands on it. but, i keep stumbling across software on the net...

i went to the acoustisoft website. do you have any thoughts about the add-ons and how etf compares to other similar programs?

)

Hi Scott,

No, I'm afraid that ETF is more than a handfull for me and I haven't purchased any other comparable products.

I have purchased their R+D add-on. It's really a work in progress. It's supposed to help you identify the resonances in your room and set up your equalizer to deal with them. Its a rather time consumming process that requires taking up to 32 measurements throughout the room. It takes a pretty decent PC processor. With my 1 GHz cpu it takes quite a long time to calculate results even when its set for "fast compute".

Larry

Larry Chanin

Hi,

For those folks considering using felt treatments, here's a response I received from David Ralph to some questions I had.

Larry Chanin wrote:

I just ran across your excellent article on felt diffusors. It was very helpful in getting an appreciation for the pros and cons of this approach.

In your article you state:

"My recommendation is to not have any sort of treatment closer then the edge of the driver mounting plate, or at least not much closer."

I was wondering whether you might amplify on this statement. I know that some speaker manufacturers avoid the mounting plate with blankets of felt, but I alway thought this had more to do with maintaining access to the drivers for possible maintenance. Did you do any measurements of mounting closer to the drivers that showed poor performance? Do you have any other tips for placement?

[Dave]
Access to the units is good to maintain, but felt can have 2-sided tape added if removal/replacement is needed. Placing a 1/2" thick piece of felt too close to the tweeter dome can create more problems than it helps. This is because felt does reflect to some degree. It also can actually create something of a "chamber", in that the dome can be "boxed in" if felt on several sides is very close. But I do place it a bit over the faceplate. My latest system uses two smaller, 1/4" thick strips that overlap even more. The thinner felt has less of a reflection component due to less surface area on the side.

Larry Chanin wrote:

Do you have any opinions regarding a more aggressive approach such as using blankets of thick wool felt that covers large amounts of the baffle?

[Dave]
Note that John Dunlavy (of Duntech and Dunlavy Audio) used very thick felt blankets on the entire front of his very large designs. This has more to do with effectiveness in the midrange, as what's important is to have a significant distance through which the sound passes. The lower the frequency, the more is needed to be effective.

Regards,

Dave

Larry Chanin

Hi,

Here's an other response I received from Thomas over at Audio-worx who wrote the DIY article.

Larry Chanin wrote:

In your article you state:
"In this example a Masonite template was made to elevate the tweeter, to avoid destructive interactions with the grill cloth frame."

And you reduce the thickness of the felt.

I was wondering whether in an application where the speaker grill was not going to be mounted, whether you think the tweeter would have to be elevated or the felt thickness reduced?

You also mention that the cutout around the tweeter should be square rather than round. I was wondering if you could elaborate on the reasoning for this?

I also read the excellent article by David Ralph that you referenced. Have you done any similar measurements of the affects of this treatment on frequency response?

Do you have any additional tips on felt placement?

Thanks very much.

Larry Chanin

Hi,

Elevating the tweeter was specific to that design. I was using a premade enclosure with a factory grill frame made from 3/4" material. As a result I wanted to get the tweeter as high as possible to avoid diffraction interference with the grill frame.

Jon and I have been using felt on all our designs for more than a decade. There are probably some measurements around, but they were done long before we started posting on the net. I linked to Dave's testing because it's representative of what happens when one uses the felt.

If you look at the designs from Avalon acoustics, their entire grill frames are completely lined with wool felt with cutouts sized to the drivers.

One shouldn't use round cutouts close to a tweeter. Doing that would impact specific frequencies depending on the diameter of the hole.

Buy some felt and experiement. Use masking tape to place pieces around the baffle and listen to the effect. I'm sure you'll like what you hear.

Regards
Thomas

Flash21

Good info Larry, addresses some of the questions I had...thanks for the legwork.

Larry Chanin

Flash21 wrote:

Good info Larry, addresses some of the questions I had...thanks for the legwork.

Hi Steve,

You're welcome.

At some point I think I'll experiment with large amounts of the thick wool felt.

Right now my measurements show with the exception of one borderline case, for the most part the cheap neoprene rings yielded results that should be impossible to hear. That is, most of the changes in measured dB was below even the most acute human hearing. Where there were a few dB difference which might be heard by someone with a good ear, I found that the rings improved some frequencies and hurt others.

I ended up leaving one ring on my lower center channel and leaving the rest off of all my other speakers.

Larry

Larry Chanin

Hi,

For those interested in this topic I found two more revelant articles:

Understanding Cabinet Edge Diffraction by Andy Unruh

Felt Effects on Baffle Diffraction by David Ralph

Larry

nathanso

Nice trivia, Doro! I've got a pair of black velvet fronted Dahlquist DQM-3's I've had for 20-years, and a set of acoustic blanketed AR90 towers that I've had for a month or so. The AR's blanket is far thicker than any velvet. NHT also used foam bits (that always remind me of bike helmet pads) stuck to the fronts of their speakers in the area of the tweeters.

Larry Chanin

Hi All,

I bought a 1'x6'x1/2" chuck of the F13 wool felt. It is the less expensive stuff at $33.55 with shipping.

Attached is a photo of my SDA-CRS+ center channel with the baffle covered with the felt.

Also attached is the before and after frequency response.

Larry

scottnbnj

when you have everything put together, do you normally have something covering the framing around the speaks?

)

Larry Chanin

scottnbnj wrote:

when you have everything put together, do you normally have something covering the framing around the speaks?

)

Hi Scott,

Do you mean like this?



Its an acoustically transparent shadowbox.


You can see more of my setup by clicking on the link in my signature.

Larry

dorokusai

nathanso wrote:

Nice trivia, Doro! I've got a pair of black velvet fronted Dahlquist DQM-3's I've had for 20-years, and a set of acoustic blanketed AR90 towers that I've had for a month or so. The AR's blanket is far thicker than any velvet. NHT also used foam bits (that always remind me of bike helmet pads) stuck to the fronts of their speakers in the area of the tweeters.

Thanks Nathan

Larry - Great job with the extra information, it's been a nice read. I've always believed in the benefits of diffraction treatments but it's a tricky operation. I've flattened out the freq response in some older speakers with wool/felt, but it also degraded the midrange quite noticeably in almost all of my non-scientific tests. It's also so speaker dependent that while fun, can be a real PITA to tweak over time. I swept alot but the figures, while sweet on paper, didn't move me audibly. That foam ring stuff is bonafide crap....it's much more worth the time and effort to do exactly what you are doing now.

I plan on using it on my DQ10's when I get off my **** to restore them as it's already a known improvement for that particular speaker.

It's was fun to play around with, but eventually the ADD kicked in and I forgot where I was.

scottnbnj

Larry Chanin wrote:

Do you mean like this?

sorry larry, i didn't mean 'covering' visually (those grille covers look fickin a sweet btw). i meant acoustic treatments on the wood framing members extending out beyond the face of the baffle.

)

Larry Chanin

dorokusai wrote:

Larry - Great job with the extra information, it's been a nice read. I've always believed in the benefits of diffraction treatments but it's a tricky operation. I've flattened out the freq response in some older speakers with wool/felt, but it also degraded the midrange quite noticeably in almost all of my non-scientific tests. It's also so speaker dependent that while fun, can be a real PITA to tweak over time. I swept alot but the figures, while sweet on paper, didn't move me audibly. That foam ring stuff is bonafide crap....it's much more worth the time and effort to do exactly what you are doing now.

Hi Mark,

Thanks.

Assumming that the felt still introduces diffraction off the edges of the felt cutout, the article by Andy Unruh helped me understand Thomas' comment about why its better to cut square or rectangular holes in the felt. If the ring was round the distance from the tweeter to the ring would be the same so all the delays between the reflections and the direct sound would be the same. This would result in the biggest peaks and nulls in the frequency response. By using square or rectangular holes the delay distances vary, therefore the diffusion is spread out and the resulting peaks and nulls in the response are lessened.

Unfortunately, getting a better feel for the theory still doesn't help me figure out the optimum dimensions of those cutouts.

In my last measurement it would appear the felt degraded the midrange. It introduced a maximum difference in frequency response of about 3 dB, which would be audible, but just barely to a trained listener.

Most of the differences in the frequency response from the neoprene oval ring resulted in changes which never would be audible, and any listening tests would be futile.

Larry

Larry Chanin

scottnbnj wrote:

sorry larry, i didn't mean 'covering' visually (those grille covers look fickin a sweet btw). i meant acoustic treatments on the wood framing members extending out beyond the face of the baffle.

)

Hi Scott,

Thanks.

No I haven't tried to treat the frame, but it did occur to me that they might be the source of diffraction.

To try to see if there was a measurable effect I pulled the center channel speaker out further. I figured that this should yield less reflections off the frame. Surprisingly the frequency response got worse.

Larry

scottnbnj

Larry Chanin wrote:

I figured that this should yield less reflections off the frame. Surprisingly the frequency response got worse.

go with the force larry. your instincts were right.

)