AudioCircle
Industry Circles => GR Research => Topic started by: Danny Richie on 9 Jul 2019, 06:55 pm
-
More on the subject of comb filtering.
Check it out: https://www.youtube.com/watch?v=MGSfaKWcetQ&feature=youtu.be&fbclid=IwAR2OVNNj25koC10zi0bJhjxK2TxiTSb6pzlEJUtUUSsPnCJBzHrj-vUbApM
-
Danny, are you implying that any virtually every speaker with more than one tweeter will succumb to comb filtering?
Are there exceptions or fixes to this problem? I ask because there are highly regarded speaker designers on this forum, including yourself, who have designed speakers with multiple tweeters.
Aside from marketing, why would a speaker designer add a second tweeter?
Are tweeters added to the rear of the cabinet exempt from comb filtering?
And how are line arrays different?
-
Danny, are you implying that any virtually every speaker with more than one tweeter will succumb to comb filtering?
It is more than just an implication. It is a fact.
Are there exceptions or fixes to this problem? I ask because there are highly regarded speaker designers on this forum, including yourself, who have designed speakers with multiple tweeters.
the only speakers that I have designed with multiple tweeters have been line sources.
Aside from marketing, why would a speaker designer add a second tweeter?
I don't know. I can't think of any good reason.
Are tweeters added to the rear of the cabinet exempt from comb filtering?
Rear firing tweeters don't effect on the on axis response.
And how are line arrays different?
Line arrays are very different. They act more like a single continuous driver. If you pulled out all of the tweeters except for two of them then the response will be just like what I showed in the video. But while two of them may be causing a cancellation others that are playing are not. Their distances are different. What you get is +/-2db ripples that move around as height changes. Dome tweeters or drivers that are spaced too far apart still have a mess of a response though.
-
Wouldn't it only be the case if both tweeters are covering the same frequencies. The frequencies are too high for the tweeter separation.
With an array of tweeters such as on the Tekton such that only one tweeter is playing the high frequencies and the rest are low passed to form a midrange driver from the surrounding tweeters, you wouldn't have the same issue would you?
-
Another fine presentation
Gary
-
Great video Danny. The demo makes it clear how moving the mic down just two inches from dead center at a meter away causes a huge drop in output of the upper frequencies. Due to out of phase cancellation. Then much worse as the mike is moved down two more inches and finally two more. It seems very easy to understand to me.
Thanks for doing these,
Jeff
-
Danny, I think these videos are a great idea. Fact based information without the hyperbole and marketing slant that seems to drive the ever shrinking "audiophile" market.
Quite possibly a self-serving idea for future subject, but here goes. Since convenience seems to be a driving factor in music reproduction among many now, maybe some real world ways to integrate optimization into conventional rooms. By conventional, I mean non-dedicated listening environments that are shared with folks that aren't basement-dwelling audio trolls. (I consider myself to at least partially match that description).
Voices of reason such as yours, along with easily implemented ideas and perhaps in-room measuring methodolgy would be welcome. Some sort of rentable "kit" would be cool. For me, at least, the fun starts leaving the equation when the whole morass gets too technical and demanding of my time and energy. It's a hobby, but it's one of many for me and I suspect others.
I realize all that is sort of a meandering diatribe. Sufice to say, I appreciate the videos!
-
Quite possibly a self-serving idea for future subject, but here goes. Since convenience seems to be a driving factor in music reproduction among many now, maybe some real world ways to integrate optimization into conventional rooms. By conventional, I mean non-dedicated listening environments that are shared with folks that aren't basement-dwelling audio trolls. (I consider myself to at least partially match that description).
Voices of reason such as yours, along with easily implemented ideas and perhaps in-room measuring methodolgy would be welcome. Some sort of rentable "kit" would be cool. For me, at least, the fun starts leaving the equation when the whole morass gets too technical and demanding of my time and energy. It's a hobby, but it's one of many for me and I suspect others.
I agree wholeheartedly. Your talks should be practical. We rarely hear from experts about setup in a "typical" room, or optimizing bass by ear, or actual demonstrations of the differences in parts quality. Also, you should showcase your personal system. I'd also love to hear actual scenarios of visiting someone's home and showing them how to maximize their system.
-
Wouldn't it only be the case if both tweeters are covering the same frequencies. The frequencies are too high for the tweeter separation.
With an array of tweeters such as on the Tekton such that only one tweeter is playing the high frequencies and the rest are low passed to form a midrange driver from the surrounding tweeters, you wouldn't have the same issue would you?
Those would not suffer from comb filtering issues up top for sure. It is possible that they have some comb filtering in the off axis but it depends on how high the mid tweeters are allowed to play. For instance if the the mid-tweeters play up to 3kHz. That wavelength is about 4.5" long. The acoustic centers are about that far apart. So you might see some in the off axis. But if they only play up to 2kHz or so then that wavelength is about 6.5" long. So you might not see any issues.
Maybe I can get Eric to chime in on that. I'll send him a message.
-
Great informative video! The data presented makes it superior to other folks info audio videos.
Several of the questions asked at the YouTube page are relevant.
The one I thought of also, where one is playing a monaural record.. Does the fact I am using stereo speakers mean much? (for comb filtering. My thought is the DISTANCE between the speakers, being several feet instead of just inches, makes the comb filtering problem 'go away'.) but I do not know if this is true.
So the question about playing monaural source material... What does that do? is a good question. And is there SOME FREQUENCY at which the actual typical distance between stereo pair of speakers IS going to create the comb filter problem playing monaural source???
The second question, which seems to have more confusion, and 'alternate theory' kind of commenting is about LINE ARRAY. Seems some folks posting think a line array does the same thing. But from the actual video, it seems he is saying no. a line array (or a ribbon tweeter) is different.
And finally my question about what about a multi driver system, with say a five way design. (Like Infinity IRS)
Anyway, in general, HOW FAR APART do drivers need to be, for what frequency range...
Clearly the tweeter range, two suck. And with the demo the tweeters are a few inches apart.
So my general question boils down to what is the relationship of DISTANCE to frequency response matters to the comb filter problem... Which adds in the mono. dual mids all into one question.
One other thing that was interesting... How moving the mic changed the frequency of the suckout... amazing...
-
Girl, you ask a lot of good questions.
I'll try and take them as you gave them.
Great informative video! The data presented makes it superior to other folks info audio videos.
Several of the questions asked at the YouTube page are relevant.
The one I thought of also, where one is playing a monaural record.. Does the fact I am using stereo speakers mean much? (for comb filtering. My thought is the DISTANCE between the speakers, being several feet instead of just inches, makes the comb filtering problem 'go away'.) but I do not know if this is true.
I answered that question earlier today. The distance can push the cancellation much lower in range, but the longer wavelengths could all couple with each other. It all depends on the distance that they are apart and from you and if there is a difference in those distances.
Here is what I replied tot he poster on Youtube.
Whether you drop out a channel or not could depend on the speaker placement and distance that they are from each other compared to you. And the number of recordings made that way are limited. Go with what works best (or sounds best) for you and your application.
So the question about playing monaural source material... What does that do? is a good question. And is there SOME FREQUENCY at which the actual typical distance between stereo pair of speakers IS going to create the comb filter problem playing monaural source???
Lower wavelengths will couple as the distances (time delay of one speaker compared to another) are not that different. But a little time delay difference up top will knock them right out. But so can a wall reflection.
The second question, which seems to have more confusion, and 'alternate theory' kind of commenting is about LINE ARRAY. Seems some folks posting think a line array does the same thing. But from the actual video, it seems he is saying no. a line array (or a ribbon tweeter) is different.
Yeah, with a line array instead of two sources you have a lot of sources. So there are many very slight time delay differences. With a line array I can move the microphone up a few inches at a time and will see +/-2db ripples across the top end. As the mic is moved up or down you still have +/-2db ripples, but the peaks and dips will move around and swap places. The effect is actually still much less than room related reflections.
And finally my question about what about a multi driver system, with say a five way design. (Like Infinity IRS)
Anyway, in general, HOW FAR APART do drivers need to be, for what frequency range...
Something to keep in mind in that regard is the length or distance of a wavelength. See this chart: http://www.soundoctor.com/freq.htm
And then compare that to how far the drivers are apart. Then you have to consider the time arrival differential verses wavelength to know how much phase rotation is occurring and if there will be cancellation.
Clearly the tweeter range, two suck. And with the demo the tweeters are a few inches apart.
Yes, always.
So my general question boils down to what is the relationship of DISTANCE to frequency response matters to the comb filter problem... Which adds in the mono. dual mids all into one question.
One other thing that was interesting... How moving the mic changed the frequency of the suckout... amazing...
Again compare wavelengths verses distance differential. That website with the chart will really help.
And sharp questions...
-
I recall the Tekton Enzo, with a three tweeter line array, measured pretty well by Stereophile. Even though it is a 2.5 way with ten inch woofers the mids looked decent.
-
Would this have the same effect with two different types of tweeters? For instance I have a Linaeum dipole ribbon "super tweeter" and even though it is lower sensitivity than my other speakers it seems to just make every pair of speakers I have matched it with sound significantly better. I've never done measurements but just the overall enjoyment of sound and the naturalness of the music improves greatly as well as the sense of air/space. Does the different dispersion and output prevent comb filtering from happening? I should get some measuring stuff just out of curiosity..
-
Would this have the same effect with two different types of tweeters? For instance I have a Linaeum dipole ribbon "super tweeter" and even though it is lower sensitivity than my other speakers it seems to just make every pair of speakers I have matched it with sound significantly better. I've never done measurements but just the overall enjoyment of sound and the naturalness of the music improves greatly as well as the sense of air/space. Does the different dispersion and output prevent comb filtering from happening? I should get some measuring stuff just out of curiosity..
Stay tuned till next Tuesday. The topic involves adding a super tweeter and what it does to the response.
-
Magnepan ribbon tweeter has graphs showing even response up to 50,000Hz and still going..... :thumb:
-
Regarding my last comment: I tried adding the super tweeters to my newly acquired Super Vs last night. First time I’ve ever heard them detract from the sound quality, it was immediately apparent and I pulled them off. I guess when it’s done right it’s done right. Great speakers Danny 👍
-
I recall the Tekton Enzo, with a three tweeter line array, measured pretty well by Stereophile. Even though it is a 2.5 way with ten inch woofers the mids looked decent.
Hmmm. 'Pretty well' compared to one's expectations of such a design. It certainly isn't Zu awful.
But objectively both horizontal and vertical dispersion are problematic. Look at what happens at the crossover. This is fudged by the way Stereophile measure, but if you compare to a speaker that is sensibly designed like a Revel or KEF Reference it becomes obvious.
-
Regarding my last comment: I tried adding the super tweeters to my newly acquired Super Vs last night. First time I’ve ever heard them detract from the sound quality, it was immediately apparent and I pulled them off. I guess when it’s done right it’s done right. Great speakers Danny 👍
If you watch the new video https://www.youtube.com/watch?v=JTBAx9x5fUU&t=912s
Danny mentions putting the supertweeter ON TOP facing upwards... You might try that. But watch the video first.
-
Hmmm. 'Pretty well' compared to one's expectations of such a design. It certainly isn't Zu awful.
But objectively both horizontal and vertical dispersion are problematic. Look at what happens at the crossover. This is fudged by the way Stereophile measure, but if you compare to a speaker that is sensibly designed like a Revel or KEF Reference it becomes obvious.
i wouldn't expect them to compare well with either of those myself, never having heard them. 'Pretty well' is better than poor though. Although the mids have a rise, as you mentioned, subjectively that might not be too objectionable in the room.
-
I'm not sure this question was answered, but I may have missed it.
Does reducing the distance between the tweeters reduce the comb filtering effect?
Specifically I built a center channel with 4 tweeters stacked. The tweeters themselves had virtually no faceplate/waveguide and were similar to this - https://www.parts-express.com/dayton-audio-nd16fa-6-5-8-soft-dome-neodymium-tweeter--275-025 (https://www.parts-express.com/dayton-audio-nd16fa-6-5-8-soft-dome-neodymium-tweeter--275-025) When I stacked them i put them as close together as possible perhaps 1/16" separating them. I haven't tested them as you did, but would you predict much if any comb filtering?
Conceptually I envision them as a short line array due to how close the active surfaces of the tweeters are.
-
@Danny, what do you think about the supertweeter at 18 kHz (with only a cap). With the bass-reflex coincident with it.
Big woofer + tweeter + supertweeter.
-
And Big woofer + midwoofer + tweeter
Old 3-ways from Yamaha. Yamaha NS-1000.
(http://maty.galeon.com/WP-imagenes/hard/Yamaha-NS-1000.jpg)
Yamaha NS-1000, crossover
(http://maty.galeon.com/WP-imagenes/hard/Yamaha-NS-1000-crossover.png) (http://maty.galeon.com/WP-imagenes/hard/Yamaha-NS-1000-crossover.png)
Yamaha NS-1000, frequency
(http://maty.galeon.com/WP-imagenes/hard/Yamaha-NS-1000-frequency.png) (http://maty.galeon.com/WP-imagenes/hard/Yamaha-NS-1000-frequency.png)
Yamaha NS-1000, impedance and phase
(http://maty.galeon.com/WP-imagenes/hard/Yamaha-NS-1000-impedance-phase.png) (http://maty.galeon.com/WP-imagenes/hard/Yamaha-NS-1000-impedance-phase.png)
Closed box, without bass-reflex if I am not wrong.
-
https://www.hifiengine.com/manual_library/yamaha/ns-1000m.shtml
Specifications
Type: 3 way, 3 driver loudspeaker system
Frequency Response: 40Hz to 20kHz
Power Handling: 100W
Crossover Frequency: 500, 6000Hz
Impedance: 8Ω
Sensitivity: 90dB
Bass: 1 x 300mm cone
Midrange: 1 x 88mm dome
Tweeter: 1 x 30mm dome
Enclosure: infinite baffle
Dimensions: 675 x 375 x 326mm
Weight: 31kg
http://www.troelsgravesen.dk/Yamaha-NS1000.htm
-
More on the subject of comb filtering.
I guess I haven't hung around enough to see what everyone is doing :oops: Very cool danny :thumb:
I recall the Tekton Enzo, with a three tweeter line array, measured pretty well by Stereophile. Even though it is a 2.5 way with ten inch woofers the mids looked decent.
They do not work like a line source. A line source needs all the drive units to be center to center within the wave length of the lowest freq prominently produced . And, they should be as tall as the room to create a "line source" type wave front. Of course, there are watered down versions of this that have some success. Tekton (say the Impact) uses attenuation and outright phase rotation to blend the array. Rather clever.
I think for the most part, danny is talking about speakers like these:
(https://i.imgur.com/vfD84Tr.jpg)
-
(https://i.imgur.com/vfD84Tr.jpg)
Okay, that made me laugh.
-
Bub, that takes me back! :green:
-
Bub, that takes me back! :green:
Aspects of that go way back. Only you could have made that pic.
-
I'd also love to hear actual scenarios of visiting someone's home and showing them how to maximize their system.
That would be so cool! :D :D
-
How to add a super tweeter, or how not to by Danny Richie
https://www.audiocircle.com/index.php?topic=161633.0
LspCAD simulation of my very strange 3-ways:
https://www.audiocircle.com/index.php?topic=166472.msg1770575#msg1770575
To finish with the strange Big 3-ways, the SPL graph with and without super tweeter. Speakers simulated as 8.2 Ohms.
(http://maty.galeon.com/WP-imagenes/hum/cajas-grandotas-LspCAD-SPL-graph-with-8-Ohms.png) (http://maty.galeon.com/WP-imagenes/hum/cajas-grandotas-LspCAD-SPL-graph-with-8-Ohms.png)
(http://maty.galeon.com/WP-imagenes/hum/cajas-grandotas-LspCAD-SPL-graph-with-8-Ohms-without-supertweeter.png) (http://maty.galeon.com/WP-imagenes/hum/cajas-grandotas-LspCAD-SPL-graph-with-8-Ohms-without-supertweeter.png)
Without the super tweeter the sound is less airy. The simulation confirms it.
-
How many here can actually hear a super tweeter? (Most audiophiles are age 60+ and probably can no longer hear above 12,000 Hz.)
And how does one overcome comb effects at those frequencies? (18,000 Hz sound waves are 0.75 inches long.)
-
How many here can actually hear a super tweeter? (Most audiophiles are age 60+ and probably can no longer hear above 12,000 Hz.)
And how does one overcome comb effects at those frequencies? (18,000 Hz sound waves are 0.75 inches long.)
The problem is that the way most people try to add it causes huge amounts of comb filtering well down into an audible range.
-
Danny,
I watched your Fast Bass video last night. It was so informative and eye opening.
In that approx 20 min video you taught me everything I've been doing wrong over the last 40 years (building loudspeakers off and on) never getting the bass
as I wanted it to be, even with JBL 2245H 18" Drivers, 4-12 inch drivers, 3-box bi-amp systems, 4-box bi-amp systems, 8cf ported bass cabs, 4cf cubic ported
bass cabs, Electronic xovers, Mono bass, stereo bass, and now i've been working on two 4.5cf Push-Pull bass cabs off and on for the last coulpe years questioning
my designs if its worth continuing or not.
Where the hell have your video's been since I was 12, now 62?
You could have saved me a lot of time money and disapointment over the decades.
I've lost so many years doing it wrong.
I'm not complaining because I still love doing it after 40 years and will be building as long as I'm able.
Awesome work Danny, just awesome. Please continue.
Steve
-
Danny, Rant #2. Kidding.
Now I dream of about eight - 7" drivers per side, crossed over at about 200Hz with maybe 6db gain around 25Hz with 2-ways or 2.5 ways on top or on the sides
bi-amp and active x-over. See what you've done to me. I LOVE IT...............
-
No wait. Make that eight-7" drives per side and 4-ribbons per side. I've never heard ribbons before.
I think maybe finally I'm on to sometime.
You need to write a book and get it published.
Oh yeah. Can I borrow some cash.
:thumb:
-
Now I dream of about eight - 7" drivers per side, crossed over at about 200Hz with maybe 6db gain around 25Hz with 2-ways or 2.5 ways on top or on the sides
bi-amp and active x-over. See what you've done to me. I LOVE IT...............
Active crossover? Oh, hell naw!! Danny will take you to school for that one. :o
-
Sorry, but latest generation DSP crossovers can really compete in speaker designs. The past is gone.
-
Sorry, but latest generation DSP crossovers can really compete in speaker designs. The past is gone.
I don’t understand how anything on the digital side matters. I can see once the signal is analogue it can degrade but by definition a digital signal is on or off, there is no noise...
-
Because you have to have excellent D/A converters to keep the digital processing sounding good. Not many DSP processors out there do that correctly.
-
Active crossover? Oh, hell naw!! Danny will take you to school for that one. :o
I'm up for that :thumb:.
-
Active crossover? Oh, hell naw!! Danny will take you to school for that one. :o
Danny actually mentioned the DSP crossover I use in one Tech Talk. I demoed it for him on the line force speakers with fine results. It has gotten even better since that demo.
-
I've learned that the way to do it is to use servo subs for the bottom end. They then use their own amp and control system that will allow you to do virtually anything you want with them.
With low end duty handled with a separate amp then your amp has no current demand and sees an easy load. Then use your amp to drive the uppers using passive filters.
And for an all digital system Rich's system is the only one out there that I can recommend. He did his homework, and it is really good.
-
Thanks Danny! :D
-
And... much better with minimum phase PEQ, emulating passive RLC filters, speakers response and the nature. At least with acoustic and electrified instruments. Without Autotune, of course.
Modern commercial recordings - so badly recorded, with low or very low DR, usually with only synthesized instruments are other world.