An externally hosted image should be here but it was not working when we last tested it.
The guy here looks like someone from a Bond movie - or is this you Wesasyo photoshopped in?
Funny thing is... good projectors are almost cheaper than big screen TVs now... and at 110" or more... the word is... immersive!
It was an easy sell to my wife. She wasn't too sure about it when I was installing the huge screen on the wall, but after the first movie, I could see it would not be a problem!
It was an easy sell to my wife. She wasn't too sure about it when I was installing the huge screen on the wall, but after the first movie, I could see it would not be a problem!
I know this is an older discussion area but I wonder if you'd be willing to try a simple experiment. JBL is now controlling directivity in some of their systems via mechanical means with cone drivers and focused lenses covering part of the cone. A simple application of masking tape between the drivers vertically just might give you/us a window into controlling combing and reducing HF cancellations. I'm not entirely shure how the FR will be effected overall but I suspect a substantial decrease in the overall FR tilt.
Other than using smaller drivers such as the TC6, I don't see any other way to mitigate the effects even with extensive DSP or shading. I've used high pass filters on the central drivers with good results in the past.
Believe me I've spend quite a bit of time thinking about such measures. But I've also followed a different train of thought I deemed important. One that tries to eliminate diffraction as much as possible. Putting anything in front of the cones would not help with that particular goal.
The sound I get, even though comb filtering is measurable, is convincing enough to accept it as is. There have been many plans in my head to try and avoid comb filtering. One of them involved shading higher frequencies in the outer drivers with shelving filters. But in the end I am way too pleased with what I've got right now, despite comb filtering being a reality in measurements.
It (comb filtering) will always be a reality in stereo of any kind. And a reason for me to play with cross talk compensation. The intelligibility of my playback improved by creating some minor cuts at certain frequencies related to head size. Mostly noticeable in the phantom part of the playback.
The downward slope due to comb filtering of the speakers when measured raw was way less than anticipated at the start of this project. And with the FIR correction became a non issue with these floor to ceiling arrays. Any dropping response in my current measurements are there by choice.
Two things I value in music above anything else: first is tonality of the system. Second (and partly responsible for point one) is the timing. Imaging is wonderful and comes in at number 3 but it must not stand in the way of the first two points. And I'm glad to say it doesn't need to be the case.
So it's hard for me to introduce something that might create diffraction. As that's one important part I'm trying to avoid. As that will result in very small timing issues of it's own. Something JBL seems to do more often, judging the horns on their M2 flagship.
I did think, as stated above, about using reticulated foam. That's about the max I'd consider putting in front of the cones.
The lack of diffraction and early reflections lets the speakers completely disappear. I bet ra7 and other line array owners who paid enough attention to their environment can confirm that. But more importantly to me it creates less timing errors. The comb filtering (that I've never noticed in any of my listening) becomes less of an issue.
The sound I get, even though comb filtering is measurable, is convincing enough to accept it as is. There have been many plans in my head to try and avoid comb filtering. One of them involved shading higher frequencies in the outer drivers with shelving filters. But in the end I am way too pleased with what I've got right now, despite comb filtering being a reality in measurements.
It (comb filtering) will always be a reality in stereo of any kind. And a reason for me to play with cross talk compensation. The intelligibility of my playback improved by creating some minor cuts at certain frequencies related to head size. Mostly noticeable in the phantom part of the playback.
The downward slope due to comb filtering of the speakers when measured raw was way less than anticipated at the start of this project. And with the FIR correction became a non issue with these floor to ceiling arrays. Any dropping response in my current measurements are there by choice.
Two things I value in music above anything else: first is tonality of the system. Second (and partly responsible for point one) is the timing. Imaging is wonderful and comes in at number 3 but it must not stand in the way of the first two points. And I'm glad to say it doesn't need to be the case.
So it's hard for me to introduce something that might create diffraction. As that's one important part I'm trying to avoid. As that will result in very small timing issues of it's own. Something JBL seems to do more often, judging the horns on their M2 flagship.
I did think, as stated above, about using reticulated foam. That's about the max I'd consider putting in front of the cones.
The lack of diffraction and early reflections lets the speakers completely disappear. I bet ra7 and other line array owners who paid enough attention to their environment can confirm that. But more importantly to me it creates less timing errors. The comb filtering (that I've never noticed in any of my listening) becomes less of an issue.
Diffraction from baffle edges
Diffraction..........interesting topic as it arises. It's become the acoustic 'jitter' associated with digital audiophiles encompassing a whole host of anomolies. I'm pretty confident you've already read the above link but just in case..........
So after applying the Linkwitz study and known diffraction artifacts we can see that the destructive nature of opposing wavefronts across drivers is far worse than those that would be created by an obstacle to those wavefronts. Take it from Linkwitz......the man knows his stuff so every time he builds something that physically appears to be a 'diffraction' nightmare with hard and sharp edges, think again as he's taken the time to dispel the B.S.
Diffraction..........interesting topic as it arises. It's become the acoustic 'jitter' associated with digital audiophiles encompassing a whole host of anomolies. I'm pretty confident you've already read the above link but just in case..........
So after applying the Linkwitz study and known diffraction artifacts we can see that the destructive nature of opposing wavefronts across drivers is far worse than those that would be created by an obstacle to those wavefronts. Take it from Linkwitz......the man knows his stuff so every time he builds something that physically appears to be a 'diffraction' nightmare with hard and sharp edges, think again as he's taken the time to dispel the B.S.
Yes I've read Linkwitz, Geddes, Toole, Keele, etc. Have you also read Speaker Dave's (David Smith) contributions on this forum? His paper on audibility of comb filtering with floor to ceiling line arrays? It is linked in one of his posts somewhere from his McIntosh days.
I have huge respect for all of them. I've basically read all of their input I could find. But I am allowed to make up my own mind, right?
With a frequency response like this at the listening position:
Should I really be worried? Smoothed 1/12 octave.
As said before, I also followed John Dunlavy and took some of his advice. Most of the gentlemen mentioned above are not worried about phase. In stark contrast to them I am (and so was John Dunlavy)! That was one of my priorities with these arrays. And a reason why I build something they wouldn't think or dream of. Look up Gedlee on arrays on this forum and you'll see his view.
So, no, that does not change my mind. I like to make up my own mind on all of these issues. That's why this thread got kind of long in the first place
. My goal from the start was to create a time coherent speaker. All of the prior names mentioned look at quarter wave distances. I looked at the distance from each driver to my ears at the listening position and went from there.
Nearly losing my mind trying to solve the puzzle I followed my own instincts and took from each of the great minds above what I could use to create my own mix of thinking. Many questions were asked to the guys that finished arrays before I did to strengthen my view. And as seen I went all out with a set of my own priorities. Diffraction being one of them.
No need to agree with me, this is just my view on things. One I'm pretty satisfied with and leading to minimum phase behaviour trough the entire pass band of the speaker from 17 Hz to 17000 Hz.
Left, right and combined response at the listening position:
Gated with frequency dependant window of 4 cycles
Frequency response of left speaker at listening position:
Again gated with frequency dependant window at the listening position
Group delay at listening position:
Spectrum plot of the stereo pair:
Result shown down to -20 dB depth
The above was my personal goal, well some of the plots shown here were with complete linear processing, I did change my mind about that. Minimum phase has become my end goal. Validated trough listening.
Now why would I worry about diffraction so much? Together with first reflections diffraction is the main reason we would/could be able to find the source of the sound trough listening. Take those keys away (down enough in SPL) and you have completely disappearing speakers. Even these huge ones.
I see no reason to compromise that behaviour, no matter what Linkwitz would have to say about that particular subject.
Should I have listened to the wise men? I'm glad I'm stubborn enough to make up my own mind. You've got to be to work on something like this for over 4 years.
If I can come up with a solution I will test it. But it simply does not involve putting anything in front of the cones. Unless it's some sort of reticulated foam.
I'm in no way trying to get into an argument here. Just showing my point of view. The one I needed to hold on to, just to be able to try and finish this project.
Somewhere half way I even mentioned that new input can make you doubt your choices. Reading on this very forum has made me wonder about my ideas many times. But I needed to cling to my own train of thoughts to prove something to myself. And from where I stand it worked out really well.
Sorry for the long reply
.
Last edited:
Good start! Buckle up, you're in the Multi way forum
.There's little I would have dissuaded you from trying as you built these.......I'm a fan of the fullrange array concept as well. We've seen the concept proven before with the fiberglass white cone vifas and the execution was not nearly as precise as yours. Many of the for mentioned designers would cringe at the thought of what you built from a conceptual standpoint..........but that didn't deter you and that's commendable in itself.
I simply mentioned the tape as a very quick and temporary means of testing and interpreting data. We'll have to just agree to disagree on the topic of diffraction I suppose. Given the details of your design, we don't disagree on much else so I call that an acceptable compromise.
Narrow baffles are the key to imaging and a disappearing speaker IMO but there's little to prove it in the means of current measurements and thinking other than what's been to my ears the very obvious results when listening. If you'd like to discuss that topic sometime, shoot me a pm.
I simply mentioned the tape as a very quick and temporary means of testing and interpreting data. We'll have to just agree to disagree on the topic of diffraction I suppose. Given the details of your design, we don't disagree on much else so I call that an acceptable compromise.
Narrow baffles are the key to imaging and a disappearing speaker IMO but there's little to prove it in the means of current measurements and thinking other than what's been to my ears the very obvious results when listening. If you'd like to discuss that topic sometime, shoot me a pm.
Nothing wrong with that, and thanks for the thought.
Maybe in time I'll be up for the experiment. Need to try and keep my mind open after all. I would have no problem talking about the disappearing act right here though.
I'm still convinced we as a specie have a very well developed hearing system that can determine origin of sound trough very small time differences.
My experiment, that anyone could repeat on cross talk cancelation proved it to me once again. Mix your stereo by inserting an inverted copy back into the signal with only a very small time delay between them.
While this was not successful in improving what I wanted to improve, mix the signal in a bit louder (~ -15 dB instead of the mentioned -18 dB) and listen.
Now tell me it doesn't do anything to the origin of the sound heard. It definitely did over here. The actual delay might need tweaking, as it is based on my head size plus listening and speaker to speaker distance. You can also remove the low pass and high pass to make it stronger.
This is pretty powerful stuff. Even though we are talking about a very small amount of delay.
This experiment was followed by others, like this one:
Mixing in an almost mono copy of the phantom centre. More successful than the first, but in the end I was undecided. It did lead to a Few cuts with EQ to the phantom centre signal trough mid-side processing combined with the earlier mentioned S-curve. Still a work in progress though.
Maybe in time I'll be up for the experiment. Need to try and keep my mind open after all. I would have no problem talking about the disappearing act right here though.
I'm still convinced we as a specie have a very well developed hearing system that can determine origin of sound trough very small time differences.
My experiment, that anyone could repeat on cross talk cancelation proved it to me once again. Mix your stereo by inserting an inverted copy back into the signal with only a very small time delay between them.
While this was not successful in improving what I wanted to improve, mix the signal in a bit louder (~ -15 dB instead of the mentioned -18 dB) and listen.
Now tell me it doesn't do anything to the origin of the sound heard. It definitely did over here. The actual delay might need tweaking, as it is based on my head size plus listening and speaker to speaker distance. You can also remove the low pass and high pass to make it stronger.
This is pretty powerful stuff. Even though we are talking about a very small amount of delay.
This experiment was followed by others, like this one:
Mixing in an almost mono copy of the phantom centre. More successful than the first, but in the end I was undecided. It did lead to a Few cuts with EQ to the phantom centre signal trough mid-side processing combined with the earlier mentioned S-curve. Still a work in progress though.
Frequency response of left speaker at listening position:
Again gated with frequency dependant window at the listening position
This one should say: Frequency an phase response of the left and right speaker combined at listening position. Time window for edit has long passed before I noticed the slip-up.
To emphasize the result: actual measured results are shown here. Not some simulations.
Last edited: