This was directed to Erik. But...
And I said:
Which wasn't stated in any word. So.
Phase and frequency response are inexorably linked, though we more easily perceive frequency response amplitude differences.
That said, our hearing uses both phase and amplitude to determine location, and reverberation cues in recorded sound can convey a sense of location in space independent from amplitude.
The off axis response of a speaker like the Revox with widely spaced mid/high drivers has peaks and nulls due to the phase relationship of the multiple driver's different arrival times. It's on-axis frequency and phase response is reasonably smooth, it's off axis response is not.
The room's boundaries will reflect those peaks and nulls (constructive and destructive interference patterns) which will sound quite different than it's on axis sound, obscuring fine details that are more easily heard with a speaker like the Wega with closer driver spacing, and therefore smoother off-axis frequency and phase response.
Anyway, that's why I think measuring the polar response differences between the two speakers would reveal the reason Mike perceives the Revox "eating" reverb cues the Wega doesn't.
Art
Last edited:
I think you’re right about the last. But in a reverberant (more or less diffuse) field phase relations like those in the direct field of course are non-existent. So you have to get it from the direct field part of the total field at the listening spot.
That simple balance between direct and reverberant field determines the localization. The more ‘dry’ the sound, the more precise the image. Apart from dipoles perhaps.
That simple balance between direct and reverberant field determines the localization. The more ‘dry’ the sound, the more precise the image. Apart from dipoles perhaps.
I highly appreciate your all effort and discussion - such a pack of knowledge and points of view always opens new perspectives. And so I decided to experiment with acoustic domain. I have limited opportunities to acoustically treat my listening room more than a few resonant bass traps and absorbent panels that I have. I doubt anyone would love to have all the walls and the ceiling to be filled with felt or foam in a normal living room. But this fact does not limit experimenting with narrowing the dispersion angle of the speakers and make them more directional. The logic is quite simple: can't silent reflections from the ceilings, screens and walls, so let's limit the sound directivity so far that it hardly touches these surfaces and goes almost straight to the listener's ears.
My prototypes are very primitive and there's nothing like breakthrough in them, but the results are so interesting that I decided to share the "project" with you. From the factory Revox BX-350 have a quite wide sounding tweeter and for impulse-aligning purposes it sits backwards to the mid-woofers. That means, there is some space between a grille mesh and the tweeter. Inspired by BX-4100 with a foam dumping around the tweeter, although it has a different functionality, I started to mess with my 350's 😉 I will now show you some photos and some graphs - that's all very primitive but works in some way that I'd like to share with you. As far as you can stand the primitivity of the "project" - enjoy 😉
Here are Revox BX-350's in their natural shape and the very proper, one and only valid color:
And this is in-room SPL and phase graph from REW captured with UMIK-1 in the listening position:
...the 500 Hz dip is an inevitable asset of BX-350 set, the roll-off over 15 kHz definitely not. Now with manual room EQ in APO and PEACE:
Prototype number one substrates: polyester wool, detailing tool that I use to clean up the vinyls from dust and an analog notebook:
And now the implementation of this sophisticated idea 😉
...and the notebook...
...and if this made you smile - enjoy 😉 but first look at the SPL and phase graphs without EQ:
...and with a manual room EQ:
The second prototype - cut from an open cell foam to trap any sound trying to go elsewhere than perpendicular to the driver. Not a waveguide, not parabolic, not hyperbolic, just a primitive dispersion trap:
The pattern of a front of the foam is generic - just because I had this particular piece of foam panel. The driver:
...and the whole system:
What a handsome 😉
With an original 50-year-old grille mesh:
The measurements of a prototype 2 were not as impressive as its more primitive number one sibling. No EQ:
...with room EQ:
Prototype number three - narrower HF trap slot with additional layer of felt. The right one:
The left one:
The inside element closer view:
...the whole "HF Trap System" on Revox BX-350 😉
...closer:
...will be continued with graphs and photos in a minute 😉
My prototypes are very primitive and there's nothing like breakthrough in them, but the results are so interesting that I decided to share the "project" with you. From the factory Revox BX-350 have a quite wide sounding tweeter and for impulse-aligning purposes it sits backwards to the mid-woofers. That means, there is some space between a grille mesh and the tweeter. Inspired by BX-4100 with a foam dumping around the tweeter, although it has a different functionality, I started to mess with my 350's 😉 I will now show you some photos and some graphs - that's all very primitive but works in some way that I'd like to share with you. As far as you can stand the primitivity of the "project" - enjoy 😉
Here are Revox BX-350's in their natural shape and the very proper, one and only valid color:
And this is in-room SPL and phase graph from REW captured with UMIK-1 in the listening position:
...the 500 Hz dip is an inevitable asset of BX-350 set, the roll-off over 15 kHz definitely not. Now with manual room EQ in APO and PEACE:
Prototype number one substrates: polyester wool, detailing tool that I use to clean up the vinyls from dust and an analog notebook:
And now the implementation of this sophisticated idea 😉
...and the notebook...
...and if this made you smile - enjoy 😉 but first look at the SPL and phase graphs without EQ:
...and with a manual room EQ:
The second prototype - cut from an open cell foam to trap any sound trying to go elsewhere than perpendicular to the driver. Not a waveguide, not parabolic, not hyperbolic, just a primitive dispersion trap:
The pattern of a front of the foam is generic - just because I had this particular piece of foam panel. The driver:
...and the whole system:
What a handsome 😉
With an original 50-year-old grille mesh:
The measurements of a prototype 2 were not as impressive as its more primitive number one sibling. No EQ:
...with room EQ:
Prototype number three - narrower HF trap slot with additional layer of felt. The right one:
The left one:
The inside element closer view:
...the whole "HF Trap System" on Revox BX-350 😉
...closer:
...will be continued with graphs and photos in a minute 😉
...continuing, the back side of the open cell foam panel with the place for the tweeter's grill:
...and implementation on Revox BX-350:
Yes, I know the angle of the lower cut line of the foam is not at the optimal angle but that's still a prototype. The driver in its damping room 🙂
...and the whole system prototype ready to be measured:
The SPL and phase without EQ with something new, a 5 kHz dip:
...and with a manual room EQ:
...the 5 kHz dip could easy be fully resolved if PEACE EQ had more parametric EQ slots. Despite the no-waveguide, no-parabolic and nonhyperbolic nature of the HF dispersion trap and despite its primitivity, the changes are evident in the HF section of the graph in both SPL and phase plots. A few important details to measurements:
And now the most important area - the listening experience. Believe me or not, but that was the most surprising change in the sound system character I have ever heard. The HF trap can catch above 3200 Hz crossover point and of course something a bit lower but is limited to the tweeter. That should sound unnatural as part of mids has free wide dispersion and higher mids and the rest of HF being limited by the primitive felt-foam trap. But that's not the case!
These speakers started to bring more intimate yet natural listening experience with highs attenuated a bit, almost zero perceived room echo and precise percussion instruments. You can hear the hi-hat that is less prominent than before, but at the same time you need less volume for your mind and ears to fully and perfectly "resolve" the timbre of the instrument. This is just amazing thing - like listening to lows and lower mids in a room with speakers and having access to the HF precision compared to (good and precise) headphones.
Well, trying to achieve the signal reverberation reproduction of the Wega's with Revox speakers I achieved a completely new listening experience. Moreover, I was not aware of how many perceived echoes and reverbs actually come from the room itself as far as the decay phase is concerned. With the felt-foam monster fronts the Revox speakers are now echo and reverberation free. They eat them both more than before. But the precision of non-reflected sound from the tweeter is extraordinary. In what way?
In many ways. For example, most badly mixed and mastered rock music has an unpleasant portion of the sound around 4 kHz that my ears treat as annoying and tiring noise. The "monster-front-foam" modification of Revox speakers eliminated this effect to (almost) none and no, it is not evident from the SPL or clarity, GD etc. graphs in REW. Before the "mod" I used to "withstand" the fuzzy-guitar around-4 kHz sound and HF noise, now it's a pleasure to listen. That tells me these had to be distortions caused by room reflections in the higher-mid and HF area.
I know I haven't discovered anything new, but maybe the described results will encourage someone to experiment with speakers with different dispersion angles. The primitive nature of my prototypes makes me not brave enough to propose anyone to repeat the same idea with foam and felt things 🙂
But once again: a clean HF in almost-non-treated room is worth trying. Of course lower frequencies from the mid-woofers remain less directional, but at the same time they are less reflective in nature for the human ears and the whole system works for some reason.
If you disagree, then at least enjoy the photos 😉
Best,
Mike
...and implementation on Revox BX-350:
Yes, I know the angle of the lower cut line of the foam is not at the optimal angle but that's still a prototype. The driver in its damping room 🙂
...and the whole system prototype ready to be measured:
The SPL and phase without EQ with something new, a 5 kHz dip:
...and with a manual room EQ:
...the 5 kHz dip could easy be fully resolved if PEACE EQ had more parametric EQ slots. Despite the no-waveguide, no-parabolic and nonhyperbolic nature of the HF dispersion trap and despite its primitivity, the changes are evident in the HF section of the graph in both SPL and phase plots. A few important details to measurements:
- measurements with EQ were made with some attenuation integrated in the PEACE EQ to prevent clipping in the signal chain,
- measurements of prototype number 3 were made with HF attenuation switch on BX-350 set to +2 dB
And now the most important area - the listening experience. Believe me or not, but that was the most surprising change in the sound system character I have ever heard. The HF trap can catch above 3200 Hz crossover point and of course something a bit lower but is limited to the tweeter. That should sound unnatural as part of mids has free wide dispersion and higher mids and the rest of HF being limited by the primitive felt-foam trap. But that's not the case!
These speakers started to bring more intimate yet natural listening experience with highs attenuated a bit, almost zero perceived room echo and precise percussion instruments. You can hear the hi-hat that is less prominent than before, but at the same time you need less volume for your mind and ears to fully and perfectly "resolve" the timbre of the instrument. This is just amazing thing - like listening to lows and lower mids in a room with speakers and having access to the HF precision compared to (good and precise) headphones.
Well, trying to achieve the signal reverberation reproduction of the Wega's with Revox speakers I achieved a completely new listening experience. Moreover, I was not aware of how many perceived echoes and reverbs actually come from the room itself as far as the decay phase is concerned. With the felt-foam monster fronts the Revox speakers are now echo and reverberation free. They eat them both more than before. But the precision of non-reflected sound from the tweeter is extraordinary. In what way?
In many ways. For example, most badly mixed and mastered rock music has an unpleasant portion of the sound around 4 kHz that my ears treat as annoying and tiring noise. The "monster-front-foam" modification of Revox speakers eliminated this effect to (almost) none and no, it is not evident from the SPL or clarity, GD etc. graphs in REW. Before the "mod" I used to "withstand" the fuzzy-guitar around-4 kHz sound and HF noise, now it's a pleasure to listen. That tells me these had to be distortions caused by room reflections in the higher-mid and HF area.
I know I haven't discovered anything new, but maybe the described results will encourage someone to experiment with speakers with different dispersion angles. The primitive nature of my prototypes makes me not brave enough to propose anyone to repeat the same idea with foam and felt things 🙂
But once again: a clean HF in almost-non-treated room is worth trying. Of course lower frequencies from the mid-woofers remain less directional, but at the same time they are less reflective in nature for the human ears and the whole system works for some reason.
If you disagree, then at least enjoy the photos 😉
Best,
Mike
And now...
The Revox BX-350 by nature have heavy attenuation on the SPL graph surrounding some 500 Hz - that comes from the every single mid-woofer factory 400 Hz Fs increased by the enclosure. I guess. But what if four such mid-woofers located in a square lower part of the baffle may act as a virtual 30+ single cone woofer? 500 Hz wave half-length is about 30+ cm so...
Why not to put a phase plug in the center of the factory mid-woofer pyramid?
That's as insane as it is logical - at least give it a try even when there is not a well known math for such a phase plug shaping.
Just give me a minute 😉
Best,
Mike
The Revox BX-350 by nature have heavy attenuation on the SPL graph surrounding some 500 Hz - that comes from the every single mid-woofer factory 400 Hz Fs increased by the enclosure. I guess. But what if four such mid-woofers located in a square lower part of the baffle may act as a virtual 30+ single cone woofer? 500 Hz wave half-length is about 30+ cm so...
Why not to put a phase plug in the center of the factory mid-woofer pyramid?
That's as insane as it is logical - at least give it a try even when there is not a well known math for such a phase plug shaping.
Just give me a minute 😉
Best,
Mike
Mike,
Guess again, the free-air resonant frequency (Fs) of the woofers is not 400Hz, it is below the Fc (speaker in sealed box resonance).
The dropping impedance above Fs/Fc is normal, the second impedance peak reaching higher than Fs box resonance is due to the crossover design.
Because a "phase plug" in the center of the pyramid won't reduce the center to center distance between the drivers responsible for the constructive and destructive interference patterns off axis, also known as "lobing" or "comb filtering".
Art
Last edited:
Art,
my guessing was rather of a rhetoric nature 🙂 as Revox puts the resonant frequency at the magnets of their speakers. In case of a Revox BX-350 mid-woofers it reads as following:
...four of such drivers are connected in parallel. I have such spare drivers on the shelf from BX-230 (with 8 Ohms and 400 Hz Fs), can measure their impedance peak outside the box. I bet it will be exactly 400 Hz. Studer barely made any mistakes and it's hardly possible that they added a zero to the Fs of their drivers label by chance.
At the very beginning of this thread, I gave a schematic of BX-350 crossover. If you can find any circuit there which could bring such a peak in 500 Hz impedance then please explain, how it works. What I can see there in mid-woofer line is one LC band trap parallel to the mid-woofers and one RLC band pass bypass parallel to the main LF inductor. Both solutions actually make impedance lower locally at the SPL graph, as far as I understand the electronics. Crossover frequency is 3200 Hz. I bet that 4,7 uF and its series inductor address some resonance and 47 uF with inductor and resistors try to increase the SPL in certain frequency range. Look (GDN are mid-woofers, GDW is a tweeter, R means raster of capacitors - not important):
As to the phase plug - in a single driver, phase plug also does not alter distance between cone outer edges by the cone diameter and despite that it works, pushing some unwanted frequencies about an octave higher when half wavelength reaches cone diameter. As you do, I also hardly believe it will work in the center of four mid-woofers, but it is not enough for my mind not to give it a try 😉
Best,
Mike
my guessing was rather of a rhetoric nature 🙂 as Revox puts the resonant frequency at the magnets of their speakers. In case of a Revox BX-350 mid-woofers it reads as following:
...four of such drivers are connected in parallel. I have such spare drivers on the shelf from BX-230 (with 8 Ohms and 400 Hz Fs), can measure their impedance peak outside the box. I bet it will be exactly 400 Hz. Studer barely made any mistakes and it's hardly possible that they added a zero to the Fs of their drivers label by chance.
At the very beginning of this thread, I gave a schematic of BX-350 crossover. If you can find any circuit there which could bring such a peak in 500 Hz impedance then please explain, how it works. What I can see there in mid-woofer line is one LC band trap parallel to the mid-woofers and one RLC band pass bypass parallel to the main LF inductor. Both solutions actually make impedance lower locally at the SPL graph, as far as I understand the electronics. Crossover frequency is 3200 Hz. I bet that 4,7 uF and its series inductor address some resonance and 47 uF with inductor and resistors try to increase the SPL in certain frequency range. Look (GDN are mid-woofers, GDW is a tweeter, R means raster of capacitors - not important):
As to the phase plug - in a single driver, phase plug also does not alter distance between cone outer edges by the cone diameter and despite that it works, pushing some unwanted frequencies about an octave higher when half wavelength reaches cone diameter. As you do, I also hardly believe it will work in the center of four mid-woofers, but it is not enough for my mind not to give it a try 😉
Best,
Mike
Quite correct.
This is true as well, but could be misconstrued. The ear is indeed sensitive to phase differences between the ears, but there is no connection of this fact with the phase that one would see on a loudspeaker sweep. If the two speakers in a stereo pair are the same then any localization cues would be due to the recording (if no room interaction) and have nothing to do with speakers phase response. Of course the speakers in a room can severely mess this up with their directivity, as you have noted below.
I'm not clear on this terminology. Higher signal damping?
I'll bet the driver's impedance will be ~50 ohms at it's Fs.
16 ohm 400Hz is on the back of the driver, but Hz is cycles per second, not Fs, free air resonance.
I would not bet against the driver's impedance being ~16 ohms at 400Hz, or that it's impedance minima is near 400Hz.
Without the raw impedance curve of the drivers and the inductance of the crossover coils I can only guess about the circuit.
That said, I can't read the frequencies on this chart recorder impedance graph with a magnifying glass, but if the crossover components are not responsible for the peak, then the individual drivers must be ~50 ohms at that frequency.
View attachment 1455211
Looking forward to seeing your measurements of the driver's impedance response!
I'm not clear which terminology you are unclear on 🙂
If it's the terminology Mike calls the Revox "eating" reverb cues the Wega doesn't, my take is it's due to the Wega having better off-axis response than the Revox.
I suppose you could say the (yet to be measured) dips in the Revox off axis response are "selectively higher signal damping", but that sounds more like "marketing speak" than you or I would be comfortable with 😉
Art