Hi,
quick on this one: yeah cardioidish pattern has very little attenuation toward first ceiling and floor specular reflections. Behind the speaker the pattern is due to two sounds superimposing, disturb either of the two and the pattern breaks.
A carefully crafted pattern in ideal environment of the simulator using multiple meter observation distance works nicely as both of the two sounds that together make the destructive interference arrive to the observation point uninterrupted. But when implemented into reality, the wall behind speaker could be really close to the speaker, irregular shape/objects nearby, non-coaxial sources for the two sounds making the pattern which means different specular reflection angles and path lengths through the wall to listening spot and the pattern isn't likely playing out quite like in the simulator.
Another, I didn't check how much attenuation you get toward wall behind speaker, but you'd likely want -20dB attenuation to the specular reflection to significantly reduce it's effect on interference at listening spot. More over, it's just one of six first specular reflections in a room and as you've noticed it's not gonna fix all problems of a room. But, you could likely get some audible benefit from the pattern, if you keep on tweaking and figure out what to listen for. Since you already got sound that doesn't capture your attention thats already best for now! Have fun!🙂
Last edited:
My own OCD with the graphs has been driving me in making the kind of speakers that I have been trying out till now..
But with this latest one, I fought against that tendency and tried to do what a layman would do.. Buy the best quality drivers available and Merge them together into a speaker using a sensible (ish) crossover.
I myself am surprised at how good they are sounding compared to all my prior attempts. 😀
But if someone likes the looks of a traditional two way speaker and wants good enough vertical polars, then it would probably be hard to beat this speaker I guess..
https://www.audiosciencereview.com/forum/index.php?threads/ascilab-a6b-review.60101/
your experience confirms that vertical directivity is much less important than horizontal directivity. But they also can't be ignored, as you were moved to improve the vertical response, albeit not near the tweeter-mid crossover.
re' OCD
I keep telling myself that the lines on a horizontal line chart don't need to be perfectly flat, but I still keep trying to get them as flat as possible.
re' OCD
I keep telling myself that the lines on a horizontal line chart don't need to be perfectly flat, but I still keep trying to get them as flat as possible.
Wasn’t simplier before with the horn? And you still have the crazy low distortion and dynamics
@ciobi69: With a well designed big horn it is easier to get directivity control above 800Hz or so.. And the EXAR 400 horn I have does a decent job at that.. Probably the Gen2 horns do a much better job w.r.t pattern control but I haven't had any experience with those yet. So the above 800 Hz - 1kHz to 20kHz part is sorted.
It is the part of the spectrum that is handled by drivers below what the horn covers the bigger issue for me.
In my experience, in the placement-listening situation I have, subjectively, the sound from the horn is so clear and so focussed that it was not so easy (please remember that there was no serious issue I observed and all this maybe nitpicking and hugely influenced by personal preferences) to pay attention to the music content that lies below 1kHz. My hypothesis was that this was due to difference in the percentage of direct sound. That is why I tried to use a cardioidish pattern below the horn. So that the there wont be a lot of directivity mismatch. But somehow, it turned out to be not enough subjectively, probably, as there was still big directivity step and a bit too steep power response as shown in below pic.
Note that the below 500Hz ish range is still not having high enough directivity.
Maybe if it was a super/hyper cardioid ish pattern below the horn, it might have been better? I don't know...
Maybe just the 15inch woofer alone (with subwoofer support) with the horn would have been better? since the spin data and directivity transition looks better as shown below:
I don't know the answer.. But I doubt it would have been enough (for my taste) since there is stillquite a bit of a directivity step below 1kHz.
Maybe if the horn, instead of having that 10dB ish DI above 1kHz, had a 6dB to 8-9dB climb, I would have liked better?
There was so many questions I have for which I dont have clear answers to.. 🙂
But one thing I am discovering again with speakers having lower power response slope is that it is easy to listen to them 2+ m away from the speakers/more relaxed sound/able to focus on the whole spectrum easily than some parts appearing super clear.. 🙂
The super clear part is very exciting initially but over time it sort of builds up fatigue (at least for me)
The below is the data for the current 4 way speaker.
Another thing I keep finding is that the Purifi/SB15NBAC at lower volumes sounds much more alive/clear/(I don't know the word) in the midrange than the 15inch driver (15PR400). This could be due to some driver characteristic/directivity effect etc.. again I dont know..
So after all this experimentation, my preferences are clear to some extent to me. I prefer drivers like the Purifi for the midrange. I prefer big drivers for the low frequencies. I prefer the T34A kind of tweeter/ compression driver like the BMS 4550 or similar in a relatively low directivity waveguide for the highs.. 🙂
It is the part of the spectrum that is handled by drivers below what the horn covers the bigger issue for me.
In my experience, in the placement-listening situation I have, subjectively, the sound from the horn is so clear and so focussed that it was not so easy (please remember that there was no serious issue I observed and all this maybe nitpicking and hugely influenced by personal preferences) to pay attention to the music content that lies below 1kHz. My hypothesis was that this was due to difference in the percentage of direct sound. That is why I tried to use a cardioidish pattern below the horn. So that the there wont be a lot of directivity mismatch. But somehow, it turned out to be not enough subjectively, probably, as there was still big directivity step and a bit too steep power response as shown in below pic.
Note that the below 500Hz ish range is still not having high enough directivity.
Maybe if it was a super/hyper cardioid ish pattern below the horn, it might have been better? I don't know...
Maybe just the 15inch woofer alone (with subwoofer support) with the horn would have been better? since the spin data and directivity transition looks better as shown below:
I don't know the answer.. But I doubt it would have been enough (for my taste) since there is stillquite a bit of a directivity step below 1kHz.
Maybe if the horn, instead of having that 10dB ish DI above 1kHz, had a 6dB to 8-9dB climb, I would have liked better?
There was so many questions I have for which I dont have clear answers to.. 🙂
But one thing I am discovering again with speakers having lower power response slope is that it is easy to listen to them 2+ m away from the speakers/more relaxed sound/able to focus on the whole spectrum easily than some parts appearing super clear.. 🙂
The super clear part is very exciting initially but over time it sort of builds up fatigue (at least for me)
The below is the data for the current 4 way speaker.
Another thing I keep finding is that the Purifi/SB15NBAC at lower volumes sounds much more alive/clear/(I don't know the word) in the midrange than the 15inch driver (15PR400). This could be due to some driver characteristic/directivity effect etc.. again I dont know..
So after all this experimentation, my preferences are clear to some extent to me. I prefer drivers like the Purifi for the midrange. I prefer big drivers for the low frequencies. I prefer the T34A kind of tweeter/ compression driver like the BMS 4550 or similar in a relatively low directivity waveguide for the highs.. 🙂
Attachments
Thank you for continuing to report your findings, it remains very interesting to follow. While your experiments re. directivity preference (db ("how high," but slope/transition, at which frequencies), etc. have confounding variables (e.g. room acoustics), they nonetheless are interesting datapoints for us as we contemplate directivity options for our designs. And, of course, they are super valuable for you regardless!
Bill
Bill
Have you considered alternatives to cardioid and monopole for fixing specific issues with bass in the room? Like a dipole for reducing side patterns and increasing rear, ripole four-directional, or a totally dispersed set of subwoofers?
It's been really fun watching this project develop over time.
It's been really fun watching this project develop over time.
I think what you have been able to do through building and experimenting is to experience quite a few different types of good designs. The knowledge that you gain from this should not be underestimated and unfortunately can't really be gained any other way. We can all discuss endlessly the pros and cons of every approach but it is impossible to know ahead of time how they all come together in a design and whether the right tradeoffs were made for a particular situation.
There is plenty of reasearch from blind testing that shows most people in the same room with the same equipment rate speakers in much the same way.
What to me is more interesting came from Soren Bech where he tested different speakers in different rooms and positions and found that the preference scores changed. It is a shame that more funding wasn't put into pursuing that further to understand better how the different styles of speaker and room together affected the preferences.
I think it will also be interesting to see if you find making the directivity and overall response smoothess better with the Purifi/Bliesma top has any real impact on how you rate the combined speaker yourself.
Thanks @fluid.. 🙂
Now I am interested in reading what Soren Bech has done.. I will try to read his papers.
Regarding trying out the waveguide on the Bliesma, that is the next step..
It has been a very expensive tweeter and I had to pay double the original price of the tweeter to get it home after duties and taxes.. So I am just gathering the courage needed to separate the grill from that tweeter.. 😀
Now I am interested in reading what Soren Bech has done.. I will try to read his papers.
Regarding trying out the waveguide on the Bliesma, that is the next step..
It has been a very expensive tweeter and I had to pay double the original price of the tweeter to get it home after duties and taxes.. So I am just gathering the courage needed to separate the grill from that tweeter.. 😀
Thank you 🙂
I have just tried multiple sub-approaches with the previous 3-way speaker configuration.
4 subwoofers spatially separated out in the room (2 of those dual Satori WO24P boxes and 2 of those NERO SW800s). It really smoothed out the frequency response in the bass, but I haven't seen any significant change in the decay pattern. In fact, I have not done much in that direction, as I was subjectively happy with the low bass part. Since I have all these woofers anyway, I will try out DBA sometime.
The one I referred to is 'Perception of timbre of reproduced sound in small rooms", which is behind the AES paywall.
There is a figure from it in this paper
https://www.researchgate.net/public..._sound_quality_-_a_review_of_existing_studies
These two are also good, Timbral aspects of reproduced sound in small rooms I and II
https://orbit.dtu.dk/files/4415806/Bech.pdf
https://www.audiosciencereview.com/...eproduced-sound-in-small-rooms-ii-pdf.131490/
I have been listening to hundreds of songs on the new 4 way "PureBliss" system.. 😀
I love this system for everything around mids and higher..
Here is another hypothesis..
I think I now understand the impact of excess group delay in the bass region.
One thing I loved about my earlier dual woofer Satori system-based 2-way speakers (I never liked the same woofer system-based 3-way speakers as much in the bass) is the impact of the bass on the body. Every time it plays something with bass, it feels like getting punches here and there on the body.. 🙂
Even though my current 4-way system goes deeper in bass as per graphs, I miss that impact on the body even with those dual 15inch woofers per side.. I have been pondering about this for a couple of days now. And I then saw some posts by Kimmosto's recently, especially this one:
https://audiosciencereview.com/foru...ble-group-delay-subwoofers.60126/post-2232839
Let me compare the two systems now.
System-1 (2way with dual Satori WO24P-8 & BMS 4550 on)
System-2 (4way PureBliss system)
In videos above, the Satori system play the same song at the same position as my current speakers.
Satori system plays in stereo while the other is playing in mono. If someone cares to both watch videos with a good pair of headphones and focus on the bass alone, it can be found that it is hard (for me at least) to point out differences in the tonality of bass.
But on the sofa where I sit for listening, there is an easily identifiable difference in the impact of bass on the body.
The differences in excess group delay between the two systems can be seen in above graphs (the 4way system has more excess group delay below 200Hz due to the steep filters I used in and around the 200Hz region and below. This was because I didn't like LR2 due to the feeling of sound seeming to come from 2 drivers instead of as a whole in some songs).. 🙂
Now I think I need to do something to try and bring back that bass impact.. (linear phase crossovers maybe..?)
I love this system for everything around mids and higher..
Here is another hypothesis..
I think I now understand the impact of excess group delay in the bass region.
One thing I loved about my earlier dual woofer Satori system-based 2-way speakers (I never liked the same woofer system-based 3-way speakers as much in the bass) is the impact of the bass on the body. Every time it plays something with bass, it feels like getting punches here and there on the body.. 🙂
Even though my current 4-way system goes deeper in bass as per graphs, I miss that impact on the body even with those dual 15inch woofers per side.. I have been pondering about this for a couple of days now. And I then saw some posts by Kimmosto's recently, especially this one:
https://audiosciencereview.com/foru...ble-group-delay-subwoofers.60126/post-2232839
Let me compare the two systems now.
System-1 (2way with dual Satori WO24P-8 & BMS 4550 on)
System-2 (4way PureBliss system)
In videos above, the Satori system play the same song at the same position as my current speakers.
Satori system plays in stereo while the other is playing in mono. If someone cares to both watch videos with a good pair of headphones and focus on the bass alone, it can be found that it is hard (for me at least) to point out differences in the tonality of bass.
But on the sofa where I sit for listening, there is an easily identifiable difference in the impact of bass on the body.
The differences in excess group delay between the two systems can be seen in above graphs (the 4way system has more excess group delay below 200Hz due to the steep filters I used in and around the 200Hz region and below. This was because I didn't like LR2 due to the feeling of sound seeming to come from 2 drivers instead of as a whole in some songs).. 🙂
Now I think I need to do something to try and bring back that bass impact.. (linear phase crossovers maybe..?)
Attachments
Last edited:
I found a very similar phenomenon in my experiments. I found that in the 100 - 300 Hz crossover region, a 2nd order filter was superior to a 3rd or 4th order filter. Even when the frequency response was EQd to be identical, I could easily identify the LR4 vs the LR2. It was as you described, a lack of impact in the mid to upper bass.
I suggest you try LR2 again for all your filters below 500 Hz. You may find a way to eliminate "the feeling of sound seeming to come from 2 drivers instead of as a whole in some songs... " by careful EQ and gentle notch filtering.
I've been an enthusiastic for flat phase / reduced group delay, for a long time. Especially for lower frequencies. I think it sets a foundational timing stage for the rest of the musical spectrum to sit on. Totally agree that headphones cannot convey the experience. Our mind / body integration knows rythmn when it hears and feels it I think.
Anyway, if you try linear-phase crossovers, I sure do want to hear your take on them.
If you can tolerate a little latency, you can can have both higher order xovers to solve the sound coming from two drivers, and reduce the group delay below 200Hz.
I find higher order complementary linear-phase xovers do a better job managing out-of-band response summations than trying to use PEQs and/or notch filters, which tend to have high Q's and strong gains. High Q & strong gain (negative included) = bad juju imo.
Whereas high order xovers more smoothly squelch out of band anomalies into non-contribution.
And higher order also lets me use a wider range of crossover frequencies, because I don't have to concern so much with long out-of-band response tails. So the best xover frequency for good polar response is more likely to be useable.
And further with higher order, lobing is reduced due to narrowed frequency ranges of lobing potential.
All that said in favor of higher order linear-phase crossovers, they do need to achieve acoustic complementary response to negate pre-ring potential.
(I say negate potential, because other than heavily contrived electrical filter demos, of FIR filter misapplication, I've never heard pre-ring.)
You can also do low order lin-phase xovers of course. They just make it harder to get fully complementary acoustic response. (same as with IIR)
Anyway again, sure hope you try this. You communicate your experience very well with high credibility 🙂
Hi, cool, system develops with listening skill, right? 😀
@tmuikku: Yes.. It is definitely fun refining the system as per my requirements. Earlier I have been just focusing on getting the graphs right. Now I am trying to understand/correlate graphs with what I am hearing, trying to correlate with what I expect/like w.r.t music playback and adjust the whole system.. 🙂
For example the other day, I kimd of understood that the reason why my system sound right even from other rooms in the house could be because of smooth power response curves. This was sorely lacking with my QA 2020i speakers.. I could detect wierd shifts in tonalities when I moved from one room to the other..
For example the other day, I kimd of understood that the reason why my system sound right even from other rooms in the house could be because of smooth power response curves. This was sorely lacking with my QA 2020i speakers.. I could detect wierd shifts in tonalities when I moved from one room to the other..
@vineethkumar01 , check the concept of blauerts bander, https://de.m.wikipedia.org/wiki/Blauertsche_Bänder
It matches the freq response in your above post.
This in addition to your observation of group delay. ;-)
It matches the freq response in your above post.
This in addition to your observation of group delay. ;-)
Same with me, going for linear phase and spl, and assuring the smooth response curve at mlp matches ~ the predicted in-room response also implies smooth power response. I used latest version of Vituixcad to establish these curves. One of the observations was indeed the eveneens at different places in the living room or even on the patio.