Later I saw on the uncorrected frequency response that it was one room.
In fact, your speaker correction is done in a very specific way and it confused me.
I just went into the chart "comparison mode", if you press this button then you will also have this button.
This speaker system had a resonant tail at 109 Hz after correction. This confused me, since they usually try not to add resonant tails during correction. And that's why I thought that the measurements were simply taken in different rooms.
I already understood why this happened, but that's not what I was thinking about then.
This is the spectrogram of the speaker system with correction.
This is the spectrogram of the speaker system without correction.
The Y axis in the pictures shows time in milliseconds. As you can see from the picture, the correction added a resonant tail at 109 Hz.
I already understood why this happened, but that's not what I was thinking about then.
This is the spectrogram of the speaker system with correction.
This is the spectrogram of the speaker system without correction.
The Y axis in the pictures shows time in milliseconds. As you can see from the picture, the correction added a resonant tail at 109 Hz.
You are perfectly right - I sacrificed some resonance in favor of a flat FR which is not a perfect solution. And that was partly because I usually focus on FR when applying room EQ which I apply always manually, not using any auto-eq and it is a very time consuming process. If I understand well what you pasted here, you applied some time gating to this spectrogram:
...but still this 109 Hz resonance remains a question to solve. Should I leave it and maintain maximizing FR flatness or EQ a dip in FR to cancel the resonance? And this resonance at 109 Hz is added by the parametric EQ, as room has almost a null here. The 109 Hz EQ was added to even the bass notes SPL in clearly recorded and mixed genres of music, with separated instruments. In heavier mixes of rock and metal this 109 Hz makes an uncomfortable mess.
...but still this 109 Hz resonance remains a question to solve. Should I leave it and maintain maximizing FR flatness or EQ a dip in FR to cancel the resonance? And this resonance at 109 Hz is added by the parametric EQ, as room has almost a null here. The 109 Hz EQ was added to even the bass notes SPL in clearly recorded and mixed genres of music, with separated instruments. In heavier mixes of rock and metal this 109 Hz makes an uncomfortable mess.
It's up to you, not me. Usually, they try not to compensate for the pits, so as not to overload the speakers and not get tails, but this is not the law. Sometimes I partially compensate for the pits, but I boost the desired frequency no more than +3 dB. Very rarely no more than +6 dB, but all this is controlled by ear and the final decision is made after listening to musical material of different genres.
I applied these settings.
I suspect the difference in presentation probably comes down to the center to center distance difference between the Wego D200 and the Revox BX-350:
The Wega has a much more compact center to center distance, so it's off-axis response should be more consistent throughout the room. Since you are hearing a combination of direct and reflected sound past the near field of the speakers, the more compact virtual source can reveal details that may be obscured by the larger spacing.
The woofers resonant free air frequency (Fs) impedance peak is lower in frequency than the lower peak ~13 ohms of the woofers in the box (Fb), which is far higher than the four 16 ohm woofers in parallel= ~4.5 ohms minimum.
The stamp on the woofer is unusual, probably means it is ~16 ohms at 400Hz, but that's definitely not it's Fs.
Always a balancing act between "room boom", power response, and flat frequency response at the listening position.
Art
It’s probably the other way around. I bet that, in a normal room, the directivity of the systems apprehended will likely tell all the answers to the original question. That is, speakers should be EQ’ed flat at listening spot and pretty free of resonances. No news here, Amar Bose has made his living out of it.
Hi Art,
about the center to center distance - that's even worse that it seems from the first sight. That's because Revox has to deal not only with the distance you mentioned on the photos around the crossover frequency, which is 3200 Hz. Those four woofers play exactly the same signal, always and when we check out the 1/2 wavelength driver placing rule, these tiny woofers are much too far from each other to prevent lobing both vertically and longitudinally. Probably that is the reason why the baffle is not even but has four sloping surfaces meeting in the center point between woofers - their designers were aware of the challenge. But that's why I measured SPL vs amp volume with pink and white noise graph from the near field to have a clear picture how these drivers perform on alone. Almost on alone, of course, as full acoustic separation in such measurements of the whole speaker cabinet is not possible.
I still believe that a speaker that is not so good in echo/reverb reproduction should be louder at higher input signal and quieter at lower input, those losing the nuances as reverberations and hiding them from the ears of the listener. From my point of view the higher SPL of the speaker does not count as the listener always adjust the volume to feel comfortable at the peaks not the quietest parts of a source material. So if we have speakers with higher SPL we listen to them at lower amp volume, and those with lower overall SPL - with higher volume and if both speakers had similar SPL vs input signal graph slope, the effect should be pretty similar. But it is not 😉
Another thing to consider is that Revox woofers act as mid-woofers. Wega woofer is crossed over at some 900 Hz, Revox - at 3200 Hz. Wega has a sealed midtone dome which I prefer very much because it does not use a support of air volume from the back, as midtone cone drivers do (they have a separate closed volume for their exclusive use). Revox woofers always use the whole cabinet volume, even when playing around 3000 Hz and that can be evidently heard. But should it "eat" the source signal echo/reverberation of instruments?
Sometimes such reverberation is several seconds long so it is not about impulse response od the driver/cabinet, I think, rather an SPL vs input signal matters the most. But I can be easily wrong with above statement. But look: the reverberation of a hi-hat occupies high-mids and highs of the spectrum. With a clean mix (e.g. jazz, acoustic music) and enough silence in the background, the decay phase is far too long to treat it as impulse response. It is just a sound, a hiss, at the certain SPL. I can not give any example of any tweeter or other kind of driver that could have a decay phase at several kHz a few seconds long. So we have a hiss - say it is an equalized white noise ant two speakers. One plays this simulated reverberation hiss/noise quieter when compared to the overall perceived volume of music, another - louder and makes reverberation and echo more pronounced. That would be logical. But the graphs from the very beginning of this post this is not the case. Both speakers and their individual drivers, measured from near field, play the simulated reverberation (equalized white noise) with similar characteristic slope.
But what if we adjust comfortable volume to attack phase of the instruments? That would change the picture.
Regards,
Mike
about the center to center distance - that's even worse that it seems from the first sight. That's because Revox has to deal not only with the distance you mentioned on the photos around the crossover frequency, which is 3200 Hz. Those four woofers play exactly the same signal, always and when we check out the 1/2 wavelength driver placing rule, these tiny woofers are much too far from each other to prevent lobing both vertically and longitudinally. Probably that is the reason why the baffle is not even but has four sloping surfaces meeting in the center point between woofers - their designers were aware of the challenge. But that's why I measured SPL vs amp volume with pink and white noise graph from the near field to have a clear picture how these drivers perform on alone. Almost on alone, of course, as full acoustic separation in such measurements of the whole speaker cabinet is not possible.
I still believe that a speaker that is not so good in echo/reverb reproduction should be louder at higher input signal and quieter at lower input, those losing the nuances as reverberations and hiding them from the ears of the listener. From my point of view the higher SPL of the speaker does not count as the listener always adjust the volume to feel comfortable at the peaks not the quietest parts of a source material. So if we have speakers with higher SPL we listen to them at lower amp volume, and those with lower overall SPL - with higher volume and if both speakers had similar SPL vs input signal graph slope, the effect should be pretty similar. But it is not 😉
Another thing to consider is that Revox woofers act as mid-woofers. Wega woofer is crossed over at some 900 Hz, Revox - at 3200 Hz. Wega has a sealed midtone dome which I prefer very much because it does not use a support of air volume from the back, as midtone cone drivers do (they have a separate closed volume for their exclusive use). Revox woofers always use the whole cabinet volume, even when playing around 3000 Hz and that can be evidently heard. But should it "eat" the source signal echo/reverberation of instruments?
Sometimes such reverberation is several seconds long so it is not about impulse response od the driver/cabinet, I think, rather an SPL vs input signal matters the most. But I can be easily wrong with above statement. But look: the reverberation of a hi-hat occupies high-mids and highs of the spectrum. With a clean mix (e.g. jazz, acoustic music) and enough silence in the background, the decay phase is far too long to treat it as impulse response. It is just a sound, a hiss, at the certain SPL. I can not give any example of any tweeter or other kind of driver that could have a decay phase at several kHz a few seconds long. So we have a hiss - say it is an equalized white noise ant two speakers. One plays this simulated reverberation hiss/noise quieter when compared to the overall perceived volume of music, another - louder and makes reverberation and echo more pronounced. That would be logical. But the graphs from the very beginning of this post this is not the case. Both speakers and their individual drivers, measured from near field, play the simulated reverberation (equalized white noise) with similar characteristic slope.
But what if we adjust comfortable volume to attack phase of the instruments? That would change the picture.
Regards,
Mike
But that's not that simple as we can not in any way fix room decay issues at the room modes with EQ. Of course EQ alters FR but not room decay slope at room modes. Bass resonant traps (a lot of them), 30 cm thick membrane-like walls with tons of rockwool... At higher frequencies it seems to be easier but it actually is not because of the number of reflections. Differences in the source signal reverb/echo interpretation by the speakers can be heard from the near field by ears, as it can be from the listening positions. Those are very different situations, but the challenge exists in both.
Best,
Mike
@Art,
just wondering: what if the source signal reverberation/echo sits much above x-over frequency? Say it is some percussion instrument with decay phase echo in a 5-15 kHz range? Because if so, the tweeter is acting almost on-alone. Both in Revox and Wega. If you look at the Revox crossover schematic, which I provided at the beginning of this thread, the HF section has no dumping (LF section has a bit strange LC circuit for that). But even when listening from the near field by ears the difference in reproducing reverb from the source is evident - in favor of Wega. Wega has 19 mm dome tweeter which is or may be lighter to move by the motor magnet than the Revox 25 mm dome but that is something we do not know so far. I plan to do more experimenting with equalized white noise at different frequencies and volume levels hoping to finally find a difference in slope (volume dependent) of some characteristic of these speakers.
Mike
just wondering: what if the source signal reverberation/echo sits much above x-over frequency? Say it is some percussion instrument with decay phase echo in a 5-15 kHz range? Because if so, the tweeter is acting almost on-alone. Both in Revox and Wega. If you look at the Revox crossover schematic, which I provided at the beginning of this thread, the HF section has no dumping (LF section has a bit strange LC circuit for that). But even when listening from the near field by ears the difference in reproducing reverb from the source is evident - in favor of Wega. Wega has 19 mm dome tweeter which is or may be lighter to move by the motor magnet than the Revox 25 mm dome but that is something we do not know so far. I plan to do more experimenting with equalized white noise at different frequencies and volume levels hoping to finally find a difference in slope (volume dependent) of some characteristic of these speakers.
Mike
Unfortunately, the inverted pyramid focus can only correct the spacing and lobing problem at one point in space, leaving the "reverb eating" mess every where else.
I'd bet if you listened to the Revox outdoors (a semi-anechoic environment) at that apex point perpendicular to the four pyramid facets, you would find it's resolution to be as good or better than the Wega.
Reverberation covers the entire audio band width, and it's phase response varies both with time and frequency compared to it's source. "Natural sounding" reverb rolls off in the high frequency to mimic HF air absorption.
https://sengpielaudio.com/calculator-air.htm
It's hard to hear those subtle differences when a speaker's off axis reflected sound does not closely match it's on axis sound in phase and frequency response.
Polar response tests would highlight the obvious differences between the different speakers not apparent from their on-axis frequency and phase response.
Other than some "stiction" effects that may be evident in large woofer suspensions, loudspeakers are quite linear in amplitude response from sub-millivolt levels up to the power range that voice coil heating increases it's impedance, resulting in less SPL per given voltage, "power compression".
The less mass the voice coil has, the faster it can heat up.
The Wega, having smaller, lighter midrange and HF voice coils should exhibit "power compression" at a lower SPL than the Revox, less linear dynamic range.
As a test signal to determine the dynamic range at what level power compression sets in, pink noise would be a better choice than white noise, and M-noise ("Music Noise") would be better still at simulating musical power distribution.
https://meyersound.com/news/m-noise-test-signal/
That said, I think polar response, rather than on axis power compression tests will reveal the source of the Revox "eating reverb" cues.
Art
It’s not about isolated room modes at relative low frequencies. It’s about those higher up, which we refer to as reverberation. You can change those with acoustic measures, but you can also change the ratio direct field/reverberant field by adjusting the DI of the speaker. That way you experience the reverberations more or less as a part of the total signal, which again leads to the impression the one speaker sounds ‘dry’ and the other ‘lively’. I thought I had made that clear earlier and it’s widely known. See the rise of cardioid and dipole loudspeakers.
That is likely more about room acoustics. Which is a trade in itself. But could you or anyone explain what the relevance of phase is in the reverberant field? People claim a lot of things…
Even easier than changing speaker radiation pattern is just to reduce listening distance.
OP could setup these systems for very small listening triangle, say 1.5m or so, 5ft. If there is still differences in sound it's more likely result from the speaker itself (the direct sound) rather than due to directivity and room acoustics.
I do not know at what distance he listens, or at which toe-in and so on, but in general when one listens farther away it's more of the sum of all early reflections included which together way overwhelm direct sound except at highs. Listening at close proximity direct sound dominates for far wider bandwidth. These are very different perceptually, and while great speakers work for both a poor speaker could work for either quite nice while not so well on the other. On the close proximity listening also edge diffraction and phase related issues are more audible.
All this also raise question what's the listening skill? I don't even understand how it actually sounds like what the OP states, but now that he hears it it's only matter of him finding AB test that reveals why it is. Most of the time there is no single reason something sounds obviously different than something else, but multiple things combined.
Yeah, less so than with dome tweeters and mids I would assume.
On my ~constant directivity system with cardioid mid + waveguide, in a rather typical 5x6m family living room, distance where perceived sound dramatically changes is about 2.2m listening triangle, where effect of early reflections seem to toggle in/out from perception. Quite much closer than where the practical positioning puts the speakers and sofa, and it's smaller than I'd like. I'd like clear sound to reach all the way to sofa, about ~3m listening triangle.
This could be some "flaw" in my system as it's diy one, but I think that very narrow coverage system is needed to extend further. Or some thought on room acoustics. My friend has huge livingroom, 10x10 or so and there the listening triangle can be much bigger with similar DI system, before perception changes, before the sound seems to be "consumed" by early reflections.
Effect of early reflections to perception changes as one changes listening distance simply by walking closer / further from speakers staying equidistant to both. Even with narrow coverage system one could setup temporary toe-in to try and increase close side wall reflections while reducing direct sound to experiment with this.
This is the key thing with listening distance, it's a method to AB how "room sound" affects perception. Simplest test to this that I know is to listen phantom center of mono pink noise. Listening at close enough distance the phantom center collapses to a small bundle center between speakers, the speakers seem completely mute as if there is no sound in the room other than in front of you, well localized. When one starts to back out further from speakers the phantom center expands and gets "hazy" with lack of better terms, clarity and localization is lost with increasing early reflections and the phantom center could span even beyond speakers, seem to cover the whole "front wall" if you will. You could try and find if there is a particular distance where your perception changes significantly, some critical combination of things with early reflections and possibly other factors happen and auditory system changes state. I have it with one step basically, very weird thing.
Now, if you have these two perceptions, try any listening test on both. For example, whether the OP problem with decay is audible only on either distance, or with both? Now you can use logic to learn something from it, and have a method to toggle it on / off if its' only at the other listening distance, just walk closer / farther to add / remove the perception.
On my ~constant directivity system with cardioid mid + waveguide, in a rather typical 5x6m family living room, distance where perceived sound dramatically changes is about 2.2m listening triangle, where effect of early reflections seem to toggle in/out from perception. Quite much closer than where the practical positioning puts the speakers and sofa, and it's smaller than I'd like. I'd like clear sound to reach all the way to sofa, about ~3m listening triangle.
This could be some "flaw" in my system as it's diy one, but I think that very narrow coverage system is needed to extend further. Or some thought on room acoustics. My friend has huge livingroom, 10x10 or so and there the listening triangle can be much bigger with similar DI system, before perception changes, before the sound seems to be "consumed" by early reflections.
Effect of early reflections to perception changes as one changes listening distance simply by walking closer / further from speakers staying equidistant to both. Even with narrow coverage system one could setup temporary toe-in to try and increase close side wall reflections while reducing direct sound to experiment with this.
This is the key thing with listening distance, it's a method to AB how "room sound" affects perception. Simplest test to this that I know is to listen phantom center of mono pink noise. Listening at close enough distance the phantom center collapses to a small bundle center between speakers, the speakers seem completely mute as if there is no sound in the room other than in front of you, well localized. When one starts to back out further from speakers the phantom center expands and gets "hazy" with lack of better terms, clarity and localization is lost with increasing early reflections and the phantom center could span even beyond speakers, seem to cover the whole "front wall" if you will. You could try and find if there is a particular distance where your perception changes significantly, some critical combination of things with early reflections and possibly other factors happen and auditory system changes state. I have it with one step basically, very weird thing.
Now, if you have these two perceptions, try any listening test on both. For example, whether the OP problem with decay is audible only on either distance, or with both? Now you can use logic to learn something from it, and have a method to toggle it on / off if its' only at the other listening distance, just walk closer / farther to add / remove the perception.
Last edited:
It's interesting that you'd assume a DIY system to be flawed, and a non-DIY one to not be flawed.
This makes no sense to me.
Yeah in sense that there can be some mistake in there, I do not know "absolute" quality of my system other than it sounds better than other systems I have access to. But since it's subjective opinion I just don't know what it is for sure. Best explanation for which I could connect my perception to was when I found David Griesinger papers about Auditory Proximity and Limit of localization distance. So, just insecurity of being a hobbyist 😉 I've been trying to promote this listening distance stuff on forums a lot for people to try it out and comment if it's something of relevance for others, with their systems in their rooms. This is phenomenon with auditory system.
Hypothetize you had ideal 1deg coverage system pointed toward your ears, it would not radiate sound toward first specular reflections at all, so early reflections would not have any effect on perceived sound. Contrary, if you have typical speaker that radiates sound to all directions early reflections and sound of room contribute to perception some, which is evident with the listening distance changing test. For this reason, if you have some particular distance you'd like to get good clear sound without too much coloration from room sound, a suitably narrow coverage system should work.
What is suitably narrow coverage then? With the listening distance test, a lot of things regarding first specular reflections change. Delay, amplitude and angle for both speaker and HRTF changes. If I put narrower coverage speakers to same positions I have these ones, sit at the sofa, only amplitude of early reflections would change. Delays and angles would stay the same. Since I do not know what actually makes the perception change, I assume it's combination of any/all of these features. Since narrower coverage system affects only few of the changing parameters, I'd assume it would need to be very narrow coverage system to be effective, to overwhelm all of the remaining variables in this sense. But, this is just assumption. It could that vertical early reflections contribute more than horizontal, for example. It could be only amplitude matters, so then perhaps it needs to be just tad more narrower coverage.
Last edited: