Contour Filters

FredT300B · « **on:** 12 Sep 2007, 09:55 am »

I prefer the sound of a Fostex based single driver speaker with a contour filter. With a filter you lose some sensitivity, but it warms up the sound nicely by compensting for the rising sensitivity of the Fostex and Audio Nirvana speakers I have owned. Maybe I'm less tolerant of high frequency peaks than most, but without a filter I find the sound of these speakers too bright for my taste, especially the Audio Nirvana 8" driver.

What are some other single driver fans' listening experiences with and without contour filters in their Fostex or Lowther based speakers? I'm not looking for a theoretical discussion so much as the subjective experience of people who have listened to their own single driver speakers with and without a filter.

Scott F. · « **Reply #1 on:** 12 Sep 2007, 02:24 pm »

Hiya Fred,

I agree, some of the single driver do need notch filters to be listenable. Depending on which manufacturer is your favorite, the depth of the notch varies. A couple of local guys use the Fostex drivers in QW pipes. One uses the passive, speaker level version that makes it sound pretty darned good even when driven with low powered SET's (45, 2A3, 300b). The other guy does a passive line level (essentially a notch based tone control, the Paul Joppa version). It doesn't sound half bad either. I tried building one of those for myself (just for fun) but I never cared for the sound. Who knows, maybe it was the quality of the parts I used since I built it out of spares in my parts bins.

The speaker level notches I've built can sound fairly decent providing you use decent quality parts. I used one when I was first breaking in my PM2A's in Martin's MLTL cabinets. They sounded darn fine without too much sensitivity loss. If I hadn't gone with an active XO and vintage 15" woofers, I'd probably still be using those cabinets and filter.

I've also found that the lessor priced Lowthers which "shout" a bit, can be tweaked into submission. A combination of the cotton tweak and the foam tweak on the whizzers can do wonders without having to install a notch filter. I recently did this to a pair of DX4's using both tweaks at the same time. I lost some of the openness and presence that Lowthers provide but it sure made them infinitely more listenable.

That said, I've found that the PM2A and the PM6A really don't need notches after they are run in. I'm also playing with a PM2T (ticonal magnet) right now in a heavily modified QW pipe (the Teresonic Ingenium). I've got some PM2C's here that I'm getting ready to play with. I'll be interested to see how they sound. Eventually I'll have a complete database of the REAL Theile Small parameters for most all of the Lowther line. The Lowther website's T-S parameters are unreliable at best. It's going to take a while though. It takes forever to break in these drivers. I'll roll all of it into an article and put it up on the ETM website so everybody has reliable T-S's to refer to in the future.

MJK · « **Reply #2 on:** 12 Sep 2007, 02:56 pm »

Quote from: Scott F. on 12 Sep 2007, 02:24 pm

Eventually I'll have a complete database of the REAL Theile Small parameters for most all of the Lowther line. The Lowther website's T-S parameters are unreliable at best. It's going to take a while though. It takes forever to break in these drivers. I'll roll all of it into an article and put it up on the ETM website so everybody has reliable T-S's to refer to in the future.

http://www.quarter-wave.com/General/Lowthers.pdf

Scott F. · « **Reply #3 on:** 12 Sep 2007, 03:01 pm »

Dammit Martin, you beat me to it

Thats what I get for not checking your website......that'll learn me

Quote from: MJK on 12 Sep 2007, 02:56 pm

Quote from: Scott F. on 12 Sep 2007, 02:24 pm
Eventually I'll have a complete database of the REAL Theile Small parameters for most all of the Lowther line. The Lowther website's T-S parameters are unreliable at best. It's going to take a while though. It takes forever to break in these drivers. I'll roll all of it into an article and put it up on the ETM website so everybody has reliable T-S's to refer to in the future.

http://www.quarter-wave.com/General/Lowthers.pdf

miklorsmith · « **Reply #4 on:** 12 Sep 2007, 03:12 pm »

For anybody using a Squeezebox - download Inguz and use his semi-parametric EQ. It's instantly tuneable to taste, doesn't reduce sensitivity, doesn't add anything to the signal path, and has up to 9 bands so bass can also be bumped up a bit whereever ya like. It's the bee's knees as our trans-pond mates would say.

jrebman · « **Reply #5 on:** 12 Sep 2007, 03:32 pm »

I'm really looking forward to giving Roger Mojeski's passive line level contour filter a try. He measured his model 103 speaker, as well as a number of others, includoing fostex, Visaton, etc. and found that by applying a little passive contouring at the line level (either preamp outs, or power amp inputs), he could get them to sound much nicer. And no efficiency hit either.

No, I haven't heard them, but I am eager to.

-- Jim

JLM · « **Reply #6 on:** 13 Sep 2007, 10:02 am »

I'm planning on trying a Berhinger DEQ2496 to replace my baffle step/zobel circuit and put the finishing touches on room EQ (in my designed/dedicated listening room that already has six GIK 244 panels).

This will hopefully allow for direct connection of speaker cable to driver. I'm old enough to have started with zip cord, and have long since heard the difference between $1000/pair cables. But still drives me nuts that money is spent sending the signal through expensive/fancy speaker cables just to end up going through cheapy resistor/capacitors and their very ordinary leads in these sorts of circuits.

I hope to add a DAC back into the system and can then keep the DEQ2496 in the digital realm. With CIA VMB-1 chip based monoblocks and Bob Brines FTA-2000 single driver speakers, this should make for a sweet purist system.

planet10 · « **Reply #7 on:** 19 Sep 2007, 11:22 pm »

Quote from: miklorsmith on 12 Sep 2007, 03:12 pm

semi-parametric EQ...doesn't reduce sensitivity, doesn't add anything to the signal path

It may not reduce sensitivity, but it does ask the amp for more juice so it is indistinguishable from a loss in sensitivity (granted the amp does not have to drive the passive circuitry). And it certainly does add something to the signal path.

dave

planet10 · « **Reply #8 on:** 19 Sep 2007, 11:27 pm »

Quote from: FredT300B on 12 Sep 2007, 09:55 am

I prefer the sound of a Fostex based single driver speaker with a contour filter.

I always work towards fixing the problem at the cause instead of band-aiding it with a filter (sometimes you do need a band-aid). Hence the modified drivers that i use, the kinds of boxes i build, and the amps i use (the latter in a sense EQ as the impedance of the loudspeaker interacts with the output impedance of the speakers)

dave

JeffB · « **Reply #9 on:** 15 Oct 2007, 11:12 pm »

Does anybody have any thoughts as to how a contour filter compares to a parametric equalizer for fixing the rising response of many single drivers?
I would think that a parametric eq would be best for treatment of isolated frequency peaks. I think the parametric eq. would have a hard time fixing an extended rising response. I have no experience with a parametric eq so I am just asking. On the other hand, the parametric eqs typically offer some sort of graphing of the response from a plugged in microphone which seems useful for dialing in the right sound.

Ed Schilling · « **Reply #10 on:** 16 Oct 2007, 12:18 am »

A loudspeaker is a system not a collection of parts. There is not right or wrong answer here. If one designs a system for a particular set of goals and lives with the compromises that happen then all is well. I personally have built many systems that used all sorts of "compensation". I prefer not to design a system where they are needed.

If one simply wants to force a particular driver to work in a particular enclosure that has no business being done in the first place, then one can very easily modify the response through the use of "compensation" and the system might very well sound decent. This would be a "success" to many. Assuming the compromises taken by this route were not a deal breaker. And this is what most are doing in the single driver field as fas as I am concerned.

Bottom line.....there is no right or wrong and either approach works. As I stated, I prefer NOT to add stuff in the signal path. None of this rules out the fact that many ADD filters and such to their speakers and are very happy with the results. That's OK too.

Ed

miklorsmith · « **Reply #11 on:** 16 Oct 2007, 03:15 pm »

A parametric EQ is a contour filter, just a highly customizable one. It lets you set frequency "center", width of the bell curve on either side of that center, and amplitude. It can have a "shelf" which cuts off one side of the bell curve. There are other various features they can have as well. They work for cut and boost equally well.

Of course, adding an equalizer as a component has potential downsides as well, i.e. adding wires and connections to the signal path.

The reason I like the Inguz solution so much and have been trying to keep the idea alive as a fundamental function to Slimserver is that it works in the digital realm on digital files within their native environment (your computer) before going anywhere. It's the most transparent solution possible, and I can't hear it in the path. The Inguz version is more like a graphic EQ but you can set the centers. I would like a full-blown feature set but for now the Inguz solution is effective.

Ed - the idea is not one just compensating for loudspeakers, but of the loudspeaker/room interface and also for personal taste.

Kevin Haskins · « **Reply #12 on:** 16 Oct 2007, 03:20 pm »

I like filters.... the more the better.

JeffB · « **Reply #13 on:** 16 Oct 2007, 05:49 pm »

What is responsible for the rising response of so many wide range drivers?
Is there some non-linearity with the motor, or sound waves reflecting off the cone, or designing for a particular box geometry or something else?

From the graphs I have seen the problem seems to get worse with larger drivers.

Kevin Haskins · « **Reply #14 on:** 16 Oct 2007, 07:14 pm »

Quote from: JeffB on 16 Oct 2007, 05:49 pm

What is responsible for the rising response of so many wide range drivers?
Is there some non-linearity with the motor, or sound waves reflecting off the cone, or designing for a particular box geometry or something else?

From the graphs I have seen the problem seems to get worse with larger drivers.

There isn't anything wrong with the drivers. The rising response is there on purpose to counteract for the falling response off-axis. A big part of the "sound" of a single driver loudspeaker is due to their power response. Audiophiles tend to prefer a rolled-off top-end. Single driver loudspeakers provide that with their narrowing dispersion pattern. The larger the driver though, the more you have to boost on-axis response to account for the drooping high-end power response.

Ed Schilling · « **Reply #15 on:** 16 Oct 2007, 08:18 pm »

miklorsmith,

I think you missed my point. That aside I don't think we are talking about room/speaker interface or personal taste. Jeff is asking about filters as they pertain to single driver speaker design. I think.

Kevin is correct to a degree but again, this is NOT a separate issue. It is just part of designing for a given driver. Nothing is wrong and nothing is right. It is what it is on an individual basis.

I think it is common practice when designing any loudspeaker to take the individual drivers' response and characteristics in to account.

All single drivers are not alike and the stated responses are often "close at best". Not only that, things don't always sound like they look on paper.

The end result is what counts. If a fellow wants to design a speaker and thinks it needs comp then great. If a fellow thinks he needs to add it to an existing speaker for his "taste" then by all means he should do so but this would be in direct conflict with the designer but that's OK too.

Rising response of a driver is no more or less of a concern than "BSC". It all matters and speaker design is an art of compromise for the most part.

Ed

Kevin Haskins · « **Reply #16 on:** 16 Oct 2007, 08:27 pm »

Quote from: Ed Schilling on 16 Oct 2007, 08:18 pm

miklorsmith,

All single drivers are not alike and the stated responses are often "close at best". Not only that, things don't always sound like they look on paper.

The end result is what counts. If a fellow wants to design a speaker and thinks it needs comp then great. If a fellow thinks he needs to add it to an existing speaker for his "taste" then by all means he should do so but this would be in direct conflict with the designer but that's OK too.

Rising response of a driver is no more or less of a concern than "BSC". It all matters and speaker design is an art of compromise for the most part.

Ed

The thing about baffle step "loss" is that it isn't lost. You don't really loose the output even though it appears that you do looking at the anechoic on-axis measurement. Its still in the room with you... just arriving at a later time. I don't do a full baffle step because in my experience, people place loudspeakers in-room where you get re-enforcement from other boundaries and a full 6db is often WAY too much. A couple db is nice though.... makes it sound a little fuller and helps compensate for Fletcher-Munson type effects at lower volumes.

JeffB · « **Reply #17 on:** 16 Oct 2007, 08:39 pm »

Kevin,

Are you saying that the response at the sweet spot should be roughly flat because the rising response counters the off-axis drop-off at the angle towards the listening position?

I have a Visaton B-200 on an open baffle currently without any kind of filter.
It seems to do some things very well, but there is a character to its sound that I don't like.
I am pretty sure that this is due to a rising frequency response. However, the objectionable sonic signature
is present in the sweet spot, off-axis, in another room, and even outside on the patio.
I have a hard time believing that the rising response of this driver is designed to match the off-axis drop-off.
In fact many have commented that a 1mh inductor + 10Ohm resistor correction circuit helps greatly.

My baffle is about 18" wide.

So I started looking around for similar drivers hoping to find a flatter response and I am not seeing them.
This has led me to the thought that large wide-range drivers simply have a flaw in their response that requires correction on an open-baffle.

I am hopeful that a countour filter will fix the problems that I am hearing with this driver.

On the other hand, I love the open baffle bass I am getting out of a cheap Warrior 15".

miklorsmith · « **Reply #18 on:** 16 Oct 2007, 08:42 pm »

Oooo, Fletcher-Munson! The TacT 2.2XP includes a dynamic compensation for this. It's definitely a real phenomenon and it's Really Tough to have any speaker/room sound good at all volumes.

Of course, if you have a good filter/EQ you can create different settings for different volumes and music.

The individual taste comment was just that - some folks like more energy in the presence band and others don't. I know my single-driver bass reflex boxes like some boundary reinforcement for bass but then imaging suffers somewhat. If I could pull them out a bit and bump the lows though . . .

JeffB - I think Roger Modjeski has just the thing you need - the EQ in a Box. I'll be getting one here fairly soon and will report.

jrebman · « **Reply #19 on:** 16 Oct 2007, 08:54 pm »

Mike,

Have you taken a close look at the EQ in a box? It looks interesting to me as well, but IMO, to be truly useful for a wider range of SD speakers and boxes, it needs to have a couple more low cutoff points -- 80, 160, and 320 are useful but in my experience 640 and 1280 would be more desirable for a lot of today's drivers. My B-200s with phase plugs probably could have started somewhere between 1500 and 2000 for best results.

I understand what Jeff is saying as the B-200s are pretty capable speakers that do a lot right, but they are quite hot and need some taming in the upper middle to the highest octaves... IMO, but I should note that I am extremely sensitive to this.

-- Jim