Here is more data:
This is clone of Linkwitz's Pluto speaker; a fully active DSP controlled implementation:
Tweeter is 2" Peerless full range mounted in end of 12" long pipe that is filled with damping material and left open at the back. Tweeter is about 3" above edge of SEAS L16RN-SL, manufactured for Linkwitz. Crossover point is set at about 1kHz.
Tweeter is set to ear height. Tweeter axis is tilted about 5° up from horizontal.
Measurements for correction were done at 9" horizontally from lower edge of tweeter cone. At this distance first room reflection is from floor and is about -42dB from direct sound. Correction is done using entire IR employing Kirkeby inverse.
Measurement of corrected system was made from 9". Additional measurements of corrected response were made 15° below reference position, at 30° horizontal by 30° vertical, 45° horizontal, and 90° horizontal; all at 9" from tweeter center. Measurement was also taken at ear height from preferred listening distance of 42".
Windowing is Blackman-Harris 7 term 12ms for left and right of window center.
Here are overlays of raw responses of woofer, tweeter, and corrected system:
Overlays of corrected system and off axis measurements:
Ripple off axis is primarily due to diffraction of tweeter and is superimposed on overall directivity of tweeter.
In above plots no smoothing is applied; 1/3 octave smoothing is revealing of tweeter directivity:
The 30° vertical component of 30° horizontal by 30° vertical measurement reveals minor lobing behavior at 1kHz.
Overlays of corrected response at 9" v 42":
The above plots reveal combination of room reflections, and differences in tweeter diffraction at 9" v 42". This is clarified by addition of 4ms gating to 42" measurement, thus removing most of room reflections.
Above 3kHz 42" gated v no gate are virtually identical. Is this ripple audible? Not to my ears. Are early room reflection audible? Not to my ears. Isolated broadband burst signals do produce audible reflections, but are heard only from walls with reflected delays >12ms.
Speaker may be considered as minimal baffle/waveguide design.
The results are consistent with soongsc's posted results
This is clone of Linkwitz's Pluto speaker; a fully active DSP controlled implementation:
An externally hosted image should be here but it was not working when we last tested it.
Tweeter is 2" Peerless full range mounted in end of 12" long pipe that is filled with damping material and left open at the back. Tweeter is about 3" above edge of SEAS L16RN-SL, manufactured for Linkwitz. Crossover point is set at about 1kHz.
Tweeter is set to ear height. Tweeter axis is tilted about 5° up from horizontal.
Measurements for correction were done at 9" horizontally from lower edge of tweeter cone. At this distance first room reflection is from floor and is about -42dB from direct sound. Correction is done using entire IR employing Kirkeby inverse.
Measurement of corrected system was made from 9". Additional measurements of corrected response were made 15° below reference position, at 30° horizontal by 30° vertical, 45° horizontal, and 90° horizontal; all at 9" from tweeter center. Measurement was also taken at ear height from preferred listening distance of 42".
Windowing is Blackman-Harris 7 term 12ms for left and right of window center.
Here are overlays of raw responses of woofer, tweeter, and corrected system:
Overlays of corrected system and off axis measurements:
Ripple off axis is primarily due to diffraction of tweeter and is superimposed on overall directivity of tweeter.
In above plots no smoothing is applied; 1/3 octave smoothing is revealing of tweeter directivity:
The 30° vertical component of 30° horizontal by 30° vertical measurement reveals minor lobing behavior at 1kHz.
Overlays of corrected response at 9" v 42":
The above plots reveal combination of room reflections, and differences in tweeter diffraction at 9" v 42". This is clarified by addition of 4ms gating to 42" measurement, thus removing most of room reflections.
Above 3kHz 42" gated v no gate are virtually identical. Is this ripple audible? Not to my ears. Are early room reflection audible? Not to my ears. Isolated broadband burst signals do produce audible reflections, but are heard only from walls with reflected delays >12ms.
Speaker may be considered as minimal baffle/waveguide design.
The results are consistent with soongsc's posted results
How is it a minimal baffle / waveguide design ?
The problem with this "minimal baffle" design is that you still have a flat baffle which just happens to be slightly bigger than the tweeters own baffle, is round, (bad) and has a sharp 90 degree bend all the way around, (also bad) pretty much a worst case scenario for diffraction, and it shows in the raw measurements of your first graph, both the steep rise from 1.5-5Khz and what looks like 2dB of continuous ripple from 4Khz up.
To do this sort of design properly there needs to be a smooth continuous half sphere curve from the dome right around to the sides of the tube, a bit like the tweeter pod in a B&W Nautilus 802 etc...as it is you just have a flat round baffle, as small as it might be.
If you now try to EQ all that ripple on axis you will have massive amounts of ripple off axis through the entire treble range, this shows clearly in your second response graph. If anything this is a poster child for why on axis EQ in the presence of large amounts of diffraction is a bad idea! 😉
You might say it sounds fine to you, but I think you'd find if you eliminated that diffraction from the sharp edges at the end of the tweeter tube that it would sound a whole lot better and need far less EQ! (The rise from 1.5-5Khz would still be there but the ripple should be gone)
Last edited:
Listening to Pluto Clone at 9" v 42" with music program is virtually identical. Sound stage moves closer, and SPL goes up. Two unavoidable outcomes.
Sound is better than reference quality headphones in terms of tonal balance, clarity, and of course fully externalized sound stage.
Up close they sound better than any B&W speakers that I have listened to.
Tight diffraction is only audible using long swept sine signal.
Content above 2.5kHz provides primarily intensity cues in sound stage formation.
I've also implemented speakers with EQ using standard filters. Simple broad notch filter for 1.5kHz-5kHz peak and proper level and delay at crossover yield extremely good listening, but lack imaging specificity and clarity, especially with metallic percussion.
Big B&W midrange with high cross has much more clarity wrecking performance than Pluto implementations.
Sound is better than reference quality headphones in terms of tonal balance, clarity, and of course fully externalized sound stage.
Up close they sound better than any B&W speakers that I have listened to.
Tight diffraction is only audible using long swept sine signal.
Content above 2.5kHz provides primarily intensity cues in sound stage formation.
I've also implemented speakers with EQ using standard filters. Simple broad notch filter for 1.5kHz-5kHz peak and proper level and delay at crossover yield extremely good listening, but lack imaging specificity and clarity, especially with metallic percussion.
Big B&W midrange with high cross has much more clarity wrecking performance than Pluto implementations.
I knew that this absurd conversation would all come down to "Yea, but it sound good to me." as it has. The classic argument to end all arguments. 🙄
So now that we have reached this absurdity can we get back on topic?
Is accuracy in response to ±30° topic addressed by waveguides/baffles? This appears to be gist of early posts in thread.
DSP based Pluto Cone produces highly recognizable square waves on axis, off axis, and at listening position from 70Hz-4kHz.
Accuracy extends to much less demanding waveforms such as percussion in music as well. Here is percussion of drum being struck. It is about 94ms of recording from CD. Top track is recording on CD, lower track is recording of speaker playing CD:
This is 10ms zoom to cursor location in above waveform:
Measured performance of accuracy is exceptional. Am I hearing what is on CD?
Sure looks like it. Does it sound good? Yes.
DSP based Pluto Cone produces highly recognizable square waves on axis, off axis, and at listening position from 70Hz-4kHz.
Accuracy extends to much less demanding waveforms such as percussion in music as well. Here is percussion of drum being struck. It is about 94ms of recording from CD. Top track is recording on CD, lower track is recording of speaker playing CD:
An externally hosted image should be here but it was not working when we last tested it.
This is 10ms zoom to cursor location in above waveform:
An externally hosted image should be here but it was not working when we last tested it.
Measured performance of accuracy is exceptional. Am I hearing what is on CD?
Sure looks like it. Does it sound good? Yes.
Last edited:
It's hard to extract more information from CD as matching of waveforms.
Which colors are the same?
What is shape of wavefront exiting B&C DE250 and entering waveguide?
Which colors are the same?
What is shape of wavefront exiting B&C DE250 and entering waveguide?
Hello Barleywater
They don't match,😱 there are obvious differences so what does that mean??
Rob🙂
Last edited:
Have you ever recorded a music signal through a speaker and compared it to the music signal? Or square waves? What level of accuracy have you achieved?
I superimposed both 10 ms graphs:
There is a periodical error of about 4 ms length. You've got something wrong in the 200-300 Hz range. 😛
But getting serious: The reproduction is impressive. 🙂
There is a periodical error of about 4 ms length. You've got something wrong in the 200-300 Hz range. 😛
But getting serious: The reproduction is impressive. 🙂
That was just a bad joke
It is indeed quite impressive.
Once the phase is linear such comparisons are possible. Have you tired subtracting the two tracks to listen to residual differences?
If the phase and amplitude is linear all that should remain will be non linear distortions (IMD, harmonic, thermal, ...). That would be interesting to really "hear" these isolated from the original signal.
How do you know what is "impressive" and what isn't? It could be that the differences that you see might sound terrible or they could be completely inaudible. I know of no definition of audible quality in the time domain.
Like a showed a couple pages back, the minimal baffle approach does work, even for tweeters. Below you can see the raw response of the Dayton ND28F tweeter (measurements at 0, 15, 30, 45 an 60 degrees).
In my opinion a pretty good result up to 10k. I personally don't care much about dispersion above 10k.
An externally hosted image should be here but it was not working when we last tested it.
In my opinion a pretty good result up to 10k. I personally don't care much about dispersion above 10k.
Come on Earl, it is impressive, as a technical achievement if nothing else, when a loudspeaker creates a near perfect replica of a complex input signal.
Sure, we know that phase accuracy is not a necessary criteria, and that some other speaker may have less phase accuracy, better frequency accuracy and surpass this one sonically. Sure it is a one dimensional look at a multidimensional object. Still, such waveform accuracy is rare in a loudspeaker and indicates a design that does some things very right.
David
Is it rare? I don't know if it is impressive or not, I have never done this test on my own system. How was it done? Clearly there are no reflections in the data or it would never look like that. Near field? I don't think that waveform matching like that would be hard at all in the near field.
Were these done during the same test session as the frequency response curves without moving the setup?
I agree.
However, this has a limited bandwidth representation. It would be interesting to see a more complex music passage, and be consistent within 15 degrees off axis, that would be IMPRESSIVE!
Last edited:
Earl,
I am wondering whether you have tested you waveguide without the rounded lip termination? How do measurements look and how it sounds?
I am wondering whether you have tested you waveguide without the rounded lip termination? How do measurements look and how it sounds?
On purpose I avoided any term like "sound" or "audible". Perhaps I should have written "The reproduction looks impressive".
My superposition wants to show, that already the most primitive visual comparison reveals differences in the acoustical domain.
Rudolf