Aha - you tried power shading.
A CBT is actually a delay shading device and that can be done with multichannel dsp easily too. Virtual curvature can be concave or convex as CBT. A tall full-height tower might be better still.
Power shading Vs time delay shading?
Hi Juhazi,
Yes, I have never tried time delayed driver shading because I am trying to maintain the most even power response possible.
In my view any attempt to delay the first arrival times of different drivers is
fundamentally flawed and impossible in real rooms.
If one looks at the theory, time delayed driver shading, requires all drivers to be time aligned for one microphone point....Not a pair of mics separated by 22cm or so, the average distance between adult human ears.
A 22cm wavelength corresponds to 1,650 Hz, if one uses a half wavelength error as unacceptable phase error, all music above 825 Hz is phase distorted.....
Then multiply these errors by room reflections....
In addition any movement from the "head in a vice" listening position and the theory collapses....Along with the sound stage!
Power shading is bad....Time delayed driver shading is a nightmare!
Hope this helps and all the best
Derek.
Hi Juhazi,
Yes, I have never tried time delayed driver shading because I am trying to maintain the most even power response possible.
In my view any attempt to delay the first arrival times of different drivers is
fundamentally flawed and impossible in real rooms.
If one looks at the theory, time delayed driver shading, requires all drivers to be time aligned for one microphone point....Not a pair of mics separated by 22cm or so, the average distance between adult human ears.
A 22cm wavelength corresponds to 1,650 Hz, if one uses a half wavelength error as unacceptable phase error, all music above 825 Hz is phase distorted.....
Then multiply these errors by room reflections....
In addition any movement from the "head in a vice" listening position and the theory collapses....Along with the sound stage!
Power shading is bad....Time delayed driver shading is a nightmare!
Hope this helps and all the best
Derek.
Mondo, if I may, picture a floor to ceiling array of full range drivers without any shading.
Next step, imagine what the floor and ceiling reflections do to the above image of the array.
Now picture that same floor to ceiling array with shading applied top and bottom.
Can you see how adding the floor and ceiling reflection of that one (with bigger movements of the cones in the centre) could be viewed as a problem?
At least in my head it works that way. I picture the top one as one big continuous array with some shading due to attenuated floor and ceiling reflections.
The second proposal would form a wavy shaped array, again, all in my head.
Does that make sense? That's how I view the difference and why I chose not to shade my arrays. Well that, and some other reasons.
I agree Wesayo, that is indeed the way I visualise the operation of a full height array (and also the way is explained in all the literature ive read).
I simply don't like using a boundary to 'improve' a concept.
How then does the same array perform in the absence of boundaries? How does one correlate the psuedo anechoic measurements some take outside with reality?
I shall be cautious (I have not tried anything like the number of concepts as yourself or others - and I am familiar with your relative success with BMR, but here I'm afraid I also disagree. Ive not heard a single BMR that sounded better than...a transistor radio.)
However, I agree with shading the downside is the loss of a major advantage of non shaded arrays, power handling, linearity and a point source effect.
But I disagree about the audibility of combing. The phasing is really very obvious to me in the few arrays I have listened to. (mostly j arrays) I also dislike the magnified image they produce, 8ft tall singers and the like. Although I am.sure there is media/genres where this isn't a problem at all for the listener.
Another good point, I had forgotten about Bessel arrays. That concept interested me but I have never gotten around to trying it out. Certainly it interests me more than the CBT and to a lesser extent the full height monsters!
I think.i have that link bookmarked somewhere, as it seemed to answer all my misgivings with 'conventional' arrays. It does make me wonder how much more difficult time alignment may be (though its pretty much a futile effort to do so in more than one seat position in one room anyway)
Last edited:
Really? Think again. Line arrays solve a lot of important issues in home sound reproduction. Floor and ceiling reflections -- GONE! Efficiency -- high. Distortion -- low.
A lot of emphasis is on horizontal directivity, which in my opinion is misplaced. Our two-ears and brain are very good at separating sound sources in the horizontal axis. Toole has shown that in fact side reflections add a sense of spaciousness. On the other hand, our ears and brain cannot separate out floor bounce and ceiling bounce. It changes the perceived frequency response and also localizes the speaker reducing the depth illusion.
Think about the issues carefully. There are many positives to line arrays. Toole pretty much says that this is the ideal speaker for home sound reproduction. Regarding the comb filtering. It is pretty high up in frequency and if you sit far enough from the array, is a non-audible issue.
Both problems have not been an issue in my particular experience. The tall singer phenomenon I could only replicate on some material, but not at all with my current DSP settings (*). I did try as many genres as possible. It gets mentioned often though.
If you experienced it on tall arrays, may I ask what the arrays were like? Full range drivers in a row like ra7 and myself? Or multi way arrays?
And also what kind of EQ was used? Not arguing here, just trying to find out why arrays get mentioned with these symptoms.
Have you noticed this ra7?
(*)= I've heard it, on some pop material, but it was a while back. I think it was a Tina Turner track from the 80's. I have to look at what settings I was using at the time as this has not been a problem at all in a long, long time. It could very well be when I was still running IIR PEQ filters. I do remember posting how my first impression of using the FIR filters was as if the arrays had been tamed pretty good.
From the early days I started using it, quoting myself from my thread:
The "lost" ambience is back though, but the raw arrays with global EQ did have a charm of their own to knock you over with big orchestral music. It was pretty wild!First remark about DRC: it manages to take these huge line arrays down to really small size speakers.
Last edited:
^ I totally agree, vertical directivity is very important. This may have something to do with my fondness of horn loaded ribbon tweeters (well the only ones ive heard and own Fountek cd3.5h)
My trouble with arrays, I guess stems from the large listening distance required - sadly on our poxy little island, sufficient room for that kind of experience is reserved for the wealthy.
I also enjoy nearfield listening very much and I guess that precludes me from enjoying an array? My own system.is flawed, polar bloom @ 3khz from.a dome tweeter (something i plan to rectify with ribbons) and floor bounce (and ceiling) from the standmount design.
But again some tracks always make the hair on my neck and arms stand up (Temple of the Dog - Your saviour esp. 3:10 onwards, indeed much of the album)
Wesayo: I have a limited listening time on one pair of arrays. Multiway, like a straight CBT, small 4" woofers and 20mm domes. They were diy made by a colleague from uni who was addicted to them. I found close mic'd vocals were the worst ('Unbreak my heart' - cant recall the singers name....)
Orchestral music was far better, rock both studio and live seemed ok, somewhere in the middle of the others. Pity I'm not a huge classical music fan really.
However, in surround with movies the effect was rather good, I'm just not sure I felt the same about music (just as music in the cinema always makes me balk, but the movie delights)
I don't think EQ was used, other than to correct the HF roll off, and no DSP or convolution IIRC.
That was when I stumbled upon the Bessel array, thought it was perfect, read about more and realised there is always a downside, so again didnt persue it. I guess I'm settled with point source and nearfield, until I can afford the space to let an array or large speaker system shine
My trouble with arrays, I guess stems from the large listening distance required - sadly on our poxy little island, sufficient room for that kind of experience is reserved for the wealthy.
I also enjoy nearfield listening very much and I guess that precludes me from enjoying an array? My own system.is flawed, polar bloom @ 3khz from.a dome tweeter (something i plan to rectify with ribbons) and floor bounce (and ceiling) from the standmount design.
But again some tracks always make the hair on my neck and arms stand up (Temple of the Dog - Your saviour esp. 3:10 onwards, indeed much of the album)
Wesayo: I have a limited listening time on one pair of arrays. Multiway, like a straight CBT, small 4" woofers and 20mm domes. They were diy made by a colleague from uni who was addicted to them. I found close mic'd vocals were the worst ('Unbreak my heart' - cant recall the singers name....)
Orchestral music was far better, rock both studio and live seemed ok, somewhere in the middle of the others. Pity I'm not a huge classical music fan really.
However, in surround with movies the effect was rather good, I'm just not sure I felt the same about music (just as music in the cinema always makes me balk, but the movie delights)
I don't think EQ was used, other than to correct the HF roll off, and no DSP or convolution IIRC.
That was when I stumbled upon the Bessel array, thought it was perfect, read about more and realised there is always a downside, so again didnt persue it. I guess I'm settled with point source and nearfield, until I can afford the space to let an array or large speaker system shine
Last edited:
I agree with Wesayso. I never experienced the really tall singers. It basically sounds exactly like I'm listening to a point source. But if I stand up, the image is again at my ear level. On some recordings, it sounds like they're on the stage. But a cello sounds about the size of a cello and same with other instruments. The "tallness" is heard at around 7 kHz, I read somewhere. Boosting and lowering 7 kHz changes the height (I haven't tried it).
A lot has to do with the room as well. I am finding that I damp the opposite wall reflections, the sound collapses between the speakers, whereas if I let the opposite reflections go, then I get more spaciousness outside the speakers (to be expected). But in the damped case, on very dry studio recordings, the sound becomes big and there is an almost in head localization. I have experienced this with point sources also. The positive of the damped scenario is that recordings which have real reverberation, for example, recordings in a church, sound very life-like and give the impression of really being there.
A lot has to do with the room as well. I am finding that I damp the opposite wall reflections, the sound collapses between the speakers, whereas if I let the opposite reflections go, then I get more spaciousness outside the speakers (to be expected). But in the damped case, on very dry studio recordings, the sound becomes big and there is an almost in head localization. I have experienced this with point sources also. The positive of the damped scenario is that recordings which have real reverberation, for example, recordings in a church, sound very life-like and give the impression of really being there.
Last edited:
With all due respect....
This is funny
Currently there are no high end BMR's available to the public....!
Parts Express only sell the TV derived drivers as do the UK distributors.
Your comment above is why CSS in the UK wont deal with DIY guys...
I'll leave you ponder the laws of cause and effect
A few years ago I might have gone with a reply along the lines of :
" How interesting, please can you post some pics of your designs and.....
which software programs and microphones you are using.....
maybe you could shed some light on how the different drivers Sd's and motors affect the power response....
how they integrate as point sources Vs arrays....
Sd Vs crossover Vs SPL....Whats your thinking on the balance.....
Mms over Bl, whats your target sweet spot for each driver Sd...
how does that compare with the cone drivers you use....
How do you compensate for the inherent driver resonance.....
Plus 100 other questions.....
But hey, you got all that covered and in your "experience"....All BMR's suck.....What do I know, eh?
This is funny
Currently there are no high end BMR's available to the public....!
Parts Express only sell the TV derived drivers as do the UK distributors.
Your comment above is why CSS in the UK wont deal with DIY guys...
I'll leave you ponder the laws of cause and effect
A few years ago I might have gone with a reply along the lines of :
" How interesting, please can you post some pics of your designs and.....
which software programs and microphones you are using.....
maybe you could shed some light on how the different drivers Sd's and motors affect the power response....
how they integrate as point sources Vs arrays....
Sd Vs crossover Vs SPL....Whats your thinking on the balance.....
Mms over Bl, whats your target sweet spot for each driver Sd...
how does that compare with the cone drivers you use....
How do you compensate for the inherent driver resonance.....
Plus 100 other questions.....
But hey, you got all that covered and in your "experience"....All BMR's suck.....What do I know, eh?
The TC9 also upper frequency range directivity is so narrow that it is not hitting your side walls- little "spaciousness" above 5kHz.
The TC9 has only 30 degree dispersion (-6dB) at 16 kHz, it's -17 dB at 45 degrees off axis.
The narrowing dispersion results in less HF boost required for flat on axis response, and makes your baffle/90 degree corner waveguide diffraction "issue" not a concern.
The polars below were with the TC9 driver mounted in a 1 liter bottle, basically no baffle other than the speaker mounting flange.
Unfortunately, just noticed that I repeated the 10 degree test, so my statement above is based on the 15 degree off axis result being between the 10 and 20 degree graphs
.Art
Attachments
Last edited:
This is funny
Currently there are no high end BMR's available to the public....!
?
Funny? Yeah, this overkill overreaction is hilarious!
Well done, Quite correct. I never mentioned High Quality or Hi Fi BMR. So please refrain from putting words in my mouth.
Human nature that "with all due respect" usually precedes a response showing the opposite traits, and ironic that one should latch on to an erroneous interpretation of words I never wrote - as opposed to criticism of the plethora of things I do write (which ARE probably due some criticism)
Last edited:
If I may be pedantic for a moment? (apologies, otherwise)
"I believe the first wave front to be the most important (due to that precedence effect)."
The precendence effect doesn't mean that early reflections won't matter, or can be ignored. Just that they blend into (but still affect) the sound of the first arrival (which does by itself determine the perceived direction of the overall sound).
And, back a few pages, I think the reference to the singer on "Let it Be" is to Paul McCartney? Lennon only sings background "ahhs" on that cut, I'm pretty sure.
And the issues with the UCA202 and 222 are that different sample clocks are used for the recording and playing streams (same nominal rate, but not in synch with each other, on is from a local crystal, the other is via a phaselock multiple on the USB clock). So you can't get a consistently precise timing using them, and you can NEVER do any signal averaging in a measurement setup with them without screwing up the results.
(apologies, otherwise)"I believe the first wave front to be the most important (due to that precedence effect)."
The precendence effect doesn't mean that early reflections won't matter, or can be ignored. Just that they blend into (but still affect) the sound of the first arrival (which does by itself determine the perceived direction of the overall sound).
And, back a few pages, I think the reference to the singer on "Let it Be" is to Paul McCartney? Lennon only sings background "ahhs" on that cut, I'm pretty sure.
And the issues with the UCA202 and 222 are that different sample clocks are used for the recording and playing streams (same nominal rate, but not in synch with each other, on is from a local crystal, the other is via a phaselock multiple on the USB clock). So you can't get a consistently precise timing using them, and you can NEVER do any signal averaging in a measurement setup with them without screwing up the results.
If I may be pedantic for a moment?
"I believe the first wave front to be the most important (due to that precedence effect)."
The precendence effect doesn't mean that early reflections won't matter, or can be ignored. Just that they blend into (but still affect) the sound of the first arrival (which does by itself determine the perceived direction of the overall sound).
That part looks like a quote from something I said
.And that's why I try to damp or prevent those early reflections from happening as much as possible. A horn would be far more successful to prevent them, we have to do a little more to succeed there. I totally agree with your statement. That view also meant for me to try and make that first wave front as time coherent as I could get it. Hard to get a true more than 20 dB down after the main pulse in a living room though.
Ra7's impressions where he damped the opposite walls show that it gets increasingly difficult to determine distance to the speaker once the early reflections are sufficiently damped.
Last edited:
Let's try a visual presentation of that.
Here's a wavelet at the listening position of the combined left and right array down to -15 dB in my room:
Evidently I still have some early reflections, the most severe at ~7 ms (I.m.h.o. from my back wall behind the listening position). This result is due to a combination of damping panels at (obvious) first reflection points and DRC processing.
When looking at -10 dB down on the same plot everything is a lot more clear:
Even some horns have problems achieving that though, as seen in a paper from the late Jean-Michel Le Cléac'h (may he rest in peace).
JMLC's Horns_measurements_ETF2010d.pdf
A lot of that seems to depend on horn termination, but the wavelets do show early reflection or diffraction effects on most of these horns. The severity and level depends a lot on the design of that horn. A lot of them are louder than -15 dB down.
With a bit of work and processing I think ra7's corner horn will do better than me in my living room. Not having to worry as much about floor and ceiling reflections helps the arrays compared to more conventional speakers.
For me that was a big enough reason to try arrays. I couldn't sell the idea of a horn speaker to my girl. Strangely enough she had less difficulty with arrays, lucky me.
As for me, this is it, I can't do more without getting into trouble here. I've been told I cannot place more damping or diffraction panels in our living room. But I'm quite happy with the results I was able to obtain. There was a lot of help from the DRC processing. Yet, due to using the damping panels I could get away with using short frequency dependant windows to correct mostly speaker and a little "room".
Here's a wavelet at the listening position of the combined left and right array down to -15 dB in my room:
Evidently I still have some early reflections, the most severe at ~7 ms (I.m.h.o. from my back wall behind the listening position). This result is due to a combination of damping panels at (obvious) first reflection points and DRC processing.
When looking at -10 dB down on the same plot everything is a lot more clear:
Even some horns have problems achieving that though, as seen in a paper from the late Jean-Michel Le Cléac'h (may he rest in peace).
JMLC's Horns_measurements_ETF2010d.pdf
A lot of that seems to depend on horn termination, but the wavelets do show early reflection or diffraction effects on most of these horns. The severity and level depends a lot on the design of that horn. A lot of them are louder than -15 dB down.
With a bit of work and processing I think ra7's corner horn will do better than me in my living room. Not having to worry as much about floor and ceiling reflections helps the arrays compared to more conventional speakers.
For me that was a big enough reason to try arrays. I couldn't sell the idea of a horn speaker to my girl. Strangely enough she had less difficulty with arrays, lucky me.
As for me, this is it, I can't do more without getting into trouble here. I've been told I cannot place more damping or diffraction panels in our living room. But I'm quite happy with the results I was able to obtain. There was a lot of help from the DRC processing. Yet, due to using the damping panels I could get away with using short frequency dependant windows to correct mostly speaker and a little "room".
Last edited:
If I may be pedantic for a moment?
"I believe the first wave front to be the most important (due to that precedence effect)."
The precendence effect doesn't mean that early reflections won't matter, or can be ignored. Just that they blend into (but still affect) the sound of the first arrival (which does by itself determine the perceived direction of the overall sound).
And, back a few pages, I think the reference to the singer on "Let it Be" is to Paul McCartney? Lennon only sings background "ahhs" on that cut, I'm pretty sure.
And the issues with the UCA202 and 222 are that different sample clocks are used for the recording and playing streams (same nominal rate, but not in synch with each other, on is from a local crystal, the other is via a phaselock multiple on the USB clock). So you can't get a consistently precise timing using them, and you can NEVER do any signal averaging in a measurement setup with them without screwing up the results.
Nwavguy did a thorough review and test of the UCA202 and found it to be quite good for a $30. He did not point out any clock topology flaw like this and his jitter tests did show very low side bands due to jitter at -120dB.
NwAvGuy: Behringer UCA202 Review
I have one and have looped back tests on it with no apparent phase anomalies. Would not the two unsync'd clocks show up there? It's actually a very decent sounding headphone amp if you have 60R headphones.
Last edited:
Hi Allen, yes, there is definitely a smoothness and tonal accuracy about the lower midrange, which has always sounded colored with point sources. Example is the helicopter sound on "The Wall." The helicopter chops have a fatness and speed that sounds so real.
But I am not sure I would associate that with the corner loading. I would say it is more because of the avoidance of floor and ceiling reflections that typically affect the lower midrange.
That Floyd track is the perfect way to explain to subjectively explain the lower midrange coherency you're hearing - I know what you mean. I miss a larger listening space and the awful 3 ways I picked up as a teenager. Although my current system betters those in the mids and HF, the lower mids have suffer with floor bounce I suspect (I haven't measured but it sounds like there's a dip 'in room'.
The rubbish 3ways I had back then had the woofers close to the floor, lower mids and the sense of spaciousness in that region, was excellent. It was just everything above about 2kHz that sucked.
A very real advantage of arrays is mitigating floor bounce.
Corner loading can only help in a similar way, it would seem like it would move first reflection surfaces behind the listener, or make them later in time, delayed.
Wish I had full height corners I could try
I have a pitched ceiling :-(Id personally like to corner load an array in a short horn, with loading down to maybe 250Hz, taking advantage of corner loading in a slightly different way.
Last edited:
IMHO it is the same thing. At these frequencies anyway where the source direction is less precise, and where the floor and ceiling are 'odd' surfaces to a point source at listening height and the line source being the natural way to fit them without a vertical pressure gradient.
Last edited: