They really do kill the illusion. Every clue that makes us able to find the source (speaker) of the sound will work against it. With floor and ceiling reflections being mayor culprits with normal speakers.
Ra7's corner placement should go a long way to maximise all these factors.
In my setup I can't place damping behind the arrays on the front wall (without causing problems with my girl). DCR cures some of it, but not all. Though I have to say the S-curve mid-side processing helps to get more sense of depth.
I use an RME FireFace 800 which provides an ASIO driver or you can choose how many WDM devices it exposes (up to a max 14 which gives you access to the 28 output channels). The RME device is set with a buffer size of 64 samples. The JRiver output device screen has a 50ms buffer set + "use large hardware buffers" checked + "device only uses most significant 24 bits" (as the RME driver is really 24 bit in a 32 bit wrapper so this makes jriver dither correctly)
I use either jriver wdm or asio depending on what I'm doing and just pick some unused output channels for REW. For example using ASIO;
REW
- output: ASIO Channel 7
JRiver
- ASIO Line In: 2 channels, channel offset 6 (i.e. is reading the RME asio channels 7 & 8 as its input channels)
- output format set to "no change" (i.e. no resampling, no jrss) and no dsp blocks turned on
RME TotalMixFX
- output channel 10 configured to output input channel 1 & loopback on (i.e. is routing the "L" channel output from jriver to output channel 10 and then looping that back to input channel 10)
REW
- input: ASIO Channel 10 (i.e. reading that looped back input)
You can use this approach with an actual mic & measurement to provide a timing reference btw.
When using jriver wdm, it is basically the same setup except for which devices are chosen. In that case REW is set to Java mode and to output to the JRiver WDM device, I have the WDM device configured as a 5.1 device so, with no signal processing applied in jriver, I get output down channels 1-6 (as REW sends the signal to all output channels). I set output channel 10 in totalmix to pick up any 1 of those channels and turn loopback on. The REW input device is set as the corresponding RME WDM device.
In the ASIO case, the sample rate is chosen by REW and everything switches automatically. In the WDM case, I have to make sure I pick the sample rate the RME device is configured as otherwise the windows resampler kicks in.
which behringer device is it? IIRC some of them have crosstalk issues when using loopback.
Ahh ok that makes sense. I still find it odd that you're getting strange results with the loopback.
I'll have to do a dac/adc loopback with my interface run thru MC's loopback and see what I get. It's been quite a while but iirc it was fine. Either way if you have a method that works that's all that matters.
3ll3d00d thanks sharing details for setup, regarding what ra7 call pretty impressive IR wonder if is because analog I/O amps is bypassed so the loop is clean digital without a physical wire between analog I/O amps, something ala you have a switch in your "RME TotalMixFX" controlpanel that do the loopback.
What i mean is same as below that is M-Audio AP192 SPDIF I/O physical wired verse its analog I/O physical wired, clean digital is textbook IR as was its bandwidth DC to light-speed and calibrated analog I/O loop show bandpass.
What i mean is same as below that is M-Audio AP192 SPDIF I/O physical wired verse its analog I/O physical wired, clean digital is textbook IR as was its bandwidth DC to light-speed and calibrated analog I/O loop show bandpass.
Attachments
Last edited:
yeah I would have thought that is the case as going through a AD/DA loop will naturally limit the bandwidth of the signal
Impedance Measurements
Nothing much to see, except a high resonant frequency. Doesn't matter much coz the drivers are barely moving at 100 Hz. And they drop off like a cliff below 90 Hz. No other squiggles that might indicate cabinet vibrations or internal standing waves. And the impedances are well matched. Carpenter did an awesome job with the build. And Vifa did a good job with manufacturing tolerances. I am doing a great job listening to them right now
Happy new year, everybody!
Nothing much to see, except a high resonant frequency. Doesn't matter much coz the drivers are barely moving at 100 Hz. And they drop off like a cliff below 90 Hz. No other squiggles that might indicate cabinet vibrations or internal standing waves. And the impedances are well matched. Carpenter did an awesome job with the build. And Vifa did a good job with manufacturing tolerances. I am doing a great job listening to them right now
Happy new year, everybody!
Attachments
Wow, that looks more complicated than my setup. It probably isn't. I tried again with using the JRiver WDM driver as output in REW and getting input from the EMU 0204 line in using JAVA drivers in REW. Output format is set to do nothing for 44.1 kHz and no JRSS. Convolution was off but the parametric EQ is on to map channels. I still get a funky impulse. I'm going to stick with my convoluted ASIO setup when measuring with convolution. To get measurements for EQ though, I am going straight from the EMU card to the amp.
It's a cheapo UCA 222, btw.
Interesting, thanks BRYTT!
The loopback should include analog somewhere, right? You have to have the output go into the mic input. I'm not sure how you can get mic input "digitally."
Oh, don't get me started. Yes, there are tons of new things I am hearing. But like I said before, the most rewarding thing is understanding the meaning of the music. Something so simple as the song "Let it be" by the Beatles has sooo much great music crammed into it that it is just a privelege to hear what they have created. On that song, the piano is slightly down and to the side of the center and John Lennon's voice is in the center. Kind of gives the impression that he is sitting playing the piano. This I could hear before on previous systems, but now, it is stronger and easier to locate. But the biggest improvement is the ability to hear gaps between his piano notes, the strength of each note, and the harmonics. These things are revealed so easily and it really draws you in.
That song "Come Together" has complex stuff going on in the low frequencies. I never "got it." But now, it sounds so much in groove. It makes me want to move. The starts, the stops, the beats are in perfect sync. Just very rewarding.
I could go on and on. Just some other tracks I really enjoyed recently:
Carole King, Tapestry, track 4. Really enjoy the piano.
A La Carte, Boogeyin Swamprock (Mapleshade Records - A La Carte Brass & Percussion). On "Papa was a rolling stone," the percussion playing is pretty fast, but now each beat is clearly discernible and again, the strength and pauses are what makes it great!
(btw, if you haven't heard the album above, you should. Great music. Try the free samples on the website.)
It
Despite your unwillingness to speculate (or mine �� I do believe there is an important difference between planar/ribbon and array. BUT....I also believe it is entirely dependant on the specific design of the ribbon/planar.
From a purely speculative POV I note a number of things:
1/ If a floor to ceiling ribbon of comparable FR is operating perfectly under the control of the motor (pure theoretical case) then the physical displacement antinode is at the centre. This is of course before mass/motor control prevents this, and breakup occurs.
2/ Above the breakup corner, the ribbon radiation breaks into harmonic operation, like that of a simple string. Multiple antinodes distributed evenly across the length.
However, comparatively the ribbon mass is miniscule, so this mass limitation is also comparatively lesser than that of a cone radiator. I believe the more relevant factor for ribbons is the BL and getting flux high enough to maintain a (not pistonic but fundamental string type movement)
However
Isn't the goal for both ribbon and array, the same? Of course they are.
In a perfect line source consisting of a single driver, no break up should exist, and max displacement is at 1/2 line length and at all frequencies. The ribbon moves in unison at all frequencies, and the node of displacement exist at each end, where z=0.
In striving for a perfect line source, then it seems amplitude shading is an integral part of the goal
In the equivalent perfect line array of multiple point sources, then also no breakup should occur. Destructive interference still exists, on a fashion this emulates the harmonic string diaphragm type radiation antinodes and nodes of displacement.
I draw the conclusion that an perfect array will always display this combing, where as a planar/ribbon will not (or rather, the perfect ribbon will not)
However, its a little moot. Neither planar /ribbon or array of drivers will move all the diaphragm area in unison, at all frequencies of interest.
This is my biggest issue with arrays. To be as close to a single vertically large transducer, shading should be employed. (after all a perfect array is attempting to emulate a perfect ribbon).
It is also my biggest issue with convolution. I'm sure with work, a good dirac can be obtained in many cases. And by no means does that belittle the efforts some put in. But I question: Why bother? Why not shade instead. The end result is the same....and nearer to the theoretical ideal....it just seems like using a sledgehammer to crack the proverbial walnut.
However, I like the ingenuity, and this is exactly how iwould attempt such an array. (except id use shading...)
Despite your unwillingness to speculate (or mine ��
I do believe there is an important difference between planar/ribbon and array. BUT....I also believe it is entirely dependant on the specific design of the ribbon/planar.From a purely speculative POV I note a number of things:
1/ If a floor to ceiling ribbon of comparable FR is operating perfectly under the control of the motor (pure theoretical case) then the physical displacement antinode is at the centre. This is of course before mass/motor control prevents this, and breakup occurs.
2/ Above the breakup corner, the ribbon radiation breaks into harmonic operation, like that of a simple string. Multiple antinodes distributed evenly across the length.
However, comparatively the ribbon mass is miniscule, so this mass limitation is also comparatively lesser than that of a cone radiator. I believe the more relevant factor for ribbons is the BL and getting flux high enough to maintain a (not pistonic but fundamental string type movement)
However
Isn't the goal for both ribbon and array, the same? Of course they are.
In a perfect line source consisting of a single driver, no break up should exist, and max displacement is at 1/2 line length and at all frequencies. The ribbon moves in unison at all frequencies, and the node of displacement exist at each end, where z=0.
In striving for a perfect line source, then it seems amplitude shading is an integral part of the goal
In the equivalent perfect line array of multiple point sources, then also no breakup should occur. Destructive interference still exists, on a fashion this emulates the harmonic string diaphragm type radiation antinodes and nodes of displacement.
I draw the conclusion that an perfect array will always display this combing, where as a planar/ribbon will not (or rather, the perfect ribbon will not)
However, its a little moot. Neither planar /ribbon or array of drivers will move all the diaphragm area in unison, at all frequencies of interest.
This is my biggest issue with arrays. To be as close to a single vertically large transducer, shading should be employed. (after all a perfect array is attempting to emulate a perfect ribbon).
It is also my biggest issue with convolution. I'm sure with work, a good dirac can be obtained in many cases. And by no means does that belittle the efforts some put in. But I question: Why bother? Why not shade instead. The end result is the same....and nearer to the theoretical ideal....it just seems like using a sledgehammer to crack the proverbial walnut.
However, I like the ingenuity, and this is exactly how iwould attempt such an array. (except id use shading...)
Last edited:
I am not sure how the problem of breakup can be addressed by shading. The two things are completely different. Care to explain further?
Shading is not necessary for a floor-to-ceiling array. Shading is necessary for shorter arrays to come close to the performance of a floor-to-ceiling array.
I'm hoping to prove this via measurements and theory. Just haven't got there yet.
Shading is not necessary for a floor-to-ceiling array. Shading is necessary for shorter arrays to come close to the performance of a floor-to-ceiling array.
I'm hoping to prove this via measurements and theory. Just haven't got there yet.
Last edited:
Ra7,
I'm.hardly an expert ��
I was merely speculating that in a perfect line source consisting of a single radiating element, the diaphragm displacement is at minimum at each end and maximum at the centre.
How does one address that,in an array WITHOUT shading?
(btw I'm.one of those weirdos that believes all reflection is bad, whether it is 'used' to create the illusion of more perfect behaviour, or whether it is perceives as detracting from perfection)
I see a case for IB subs operating push pull to fire a wave 'thru' the room. I see how an perfect array approximates a line source. I just believe an imperfect line source of a single diaphragm as being nearer the perfect line source than an imperfect array.
My point (over laboured and lost in the cr@p):
An imperfect ribbon has breakup. But does it have combing? Or rather 'self combing'
...
I don't believe it does. (I could be wrong, but I've never seen evidence one way or the other)
Essentially,I believe that even using DSP and all.its trickery, a more perfect array will utilise shading, and (if I were to do this myself, n channels of DSP, where n= number of drivers in the array, for a true perfect Dirac)
I probably have a fundamental misunderstanding about how a line source can produce a perfect cylindrical radiation pattern, unless diaphragm displacement is equal alone the entire length (which I assume it cannot be in a true ribbon diaphragm due to termination
I'm.hardly an expert ��
I was merely speculating that in a perfect line source consisting of a single radiating element, the diaphragm displacement is at minimum at each end and maximum at the centre.
How does one address that,in an array WITHOUT shading?
(btw I'm.one of those weirdos that believes all reflection is bad, whether it is 'used' to create the illusion of more perfect behaviour, or whether it is perceives as detracting from perfection)
I see a case for IB subs operating push pull to fire a wave 'thru' the room. I see how an perfect array approximates a line source. I just believe an imperfect line source of a single diaphragm as being nearer the perfect line source than an imperfect array.
My point (over laboured and lost in the cr@p):
An imperfect ribbon has breakup. But does it have combing? Or rather 'self combing'
...
I don't believe it does. (I could be wrong, but I've never seen evidence one way or the other)
Essentially,I believe that even using DSP and all.its trickery, a more perfect array will utilise shading, and (if I were to do this myself, n channels of DSP, where n= number of drivers in the array, for a true perfect Dirac)
I probably have a fundamental misunderstanding about how a line source can produce a perfect cylindrical radiation pattern, unless diaphragm displacement is equal alone the entire length (which I assume it cannot be in a true ribbon diaphragm due to termination
Last edited:
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.
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.
Happy New Year and good health to all in 2016!
Hi guys,
Back in 2012 I tried shading with a 4, 8, and then a 16 driver array using DSP and I did not like the results.
Subjectively the system lost its ability to throw a detailed and natural sound-stage out into the room and the dynamics and max SPL suffered severely.
Objectively the CSD plots deteriorated due to all the stress being focused on the centre driver which displayed extended energy storage as it was exceeding Xmax whilst the outer drivers were hardly working.
I believe shading is a fundamentally flawed principal and it damages the sound of any array because it turns that array into a point source with low level assisted drivers.
The "issue" that the low level assist drivers are supposed to address is actually a non issue ie comb filtering is only audible close up with pink noise.....Playing music at the listening position its not audible.
Taking a step back, fundamentally all loudspeaker drivers are so flawed (bordering on unfit for purpose in some cases!) and the single biggest advantage of full length arrays (preferably mounted on wall or in wall) is that they spread the load evenly over a very large (the larger the better) Sd and therefore minimise the driver movement.
Hope this helps and all the best
Derek.
Hi guys,
Back in 2012 I tried shading with a 4, 8, and then a 16 driver array using DSP and I did not like the results.
Subjectively the system lost its ability to throw a detailed and natural sound-stage out into the room and the dynamics and max SPL suffered severely.
Objectively the CSD plots deteriorated due to all the stress being focused on the centre driver which displayed extended energy storage as it was exceeding Xmax whilst the outer drivers were hardly working.
I believe shading is a fundamentally flawed principal and it damages the sound of any array because it turns that array into a point source with low level assisted drivers.
The "issue" that the low level assist drivers are supposed to address is actually a non issue ie comb filtering is only audible close up with pink noise.....Playing music at the listening position its not audible.
Taking a step back, fundamentally all loudspeaker drivers are so flawed (bordering on unfit for purpose in some cases!) and the single biggest advantage of full length arrays (preferably mounted on wall or in wall) is that they spread the load evenly over a very large (the larger the better) Sd and therefore minimise the driver movement.
Hope this helps and all the best
Derek.
Attachments
I'm glad to know I'm not the only one
. I will try the recording you mentioned. Mapleshade, I remember that company name from my Car audio projects.One of the biggest hero's in the Car audio world, Earl Zausmer had a BMW equipped with B&W speakers and used Mapleshade recordings to demo.
Here's a link from Earl himself telling the story: Did anyone get to hear Earl Zausmer's BMW? - Page 4 - Car Audio | DiyMobileAudio.com | Car Stereo Forum
His car: Milbert, the Most Musical Amplifiers
Here's an other interesting read: Bessel arrays
Making the arrays a huge Bessel arrangement. I never tried it, I was having way to much fun with my more conventional wiring scheme and a bit of processing. It's interesting though. But as Derek said above for the shaded array, even here SPL capabilities are going down.
Making the arrays a huge Bessel arrangement. I never tried it, I was having way to much fun with my more conventional wiring scheme and a bit of processing. It's interesting though. But as Derek said above for the shaded array, even here SPL capabilities are going down.
Just listened to the Mapleshade example. Fun track! Very real in space and placement. Lots of little stuff going on but well defined. Must be a blast in pure lossless quality.
It reminded me a bit of Rodrigo y Gabriella and C.U.B.A. - Area 52. Though this particular album (Area 52) is way more busy. What can I say, I'm a sucker for guitar. Though this one has a complete Latin band included. Got to try it to grasp the scale.
Here's a promo clip: https://www.youtube.com/watch?v=LBBM-02joFU
The beauty for me is in the amount of separation you're getting, even in tracks this busy.
And the groove and timing that draws you into that music. Close your eyes and...
By the way, a fun short track for you to try, ImpulseRecord.com - Impulse Response Convolution Reverb Library Samples!
Try the first track, March of the Mantidae. It's an example of using convolution to add space to a recording. It's fun!
It reminded me a bit of Rodrigo y Gabriella and C.U.B.A. - Area 52. Though this particular album (Area 52) is way more busy. What can I say, I'm a sucker for guitar. Though this one has a complete Latin band included. Got to try it to grasp the scale.
Here's a promo clip: https://www.youtube.com/watch?v=LBBM-02joFU
The beauty for me is in the amount of separation you're getting, even in tracks this busy.
And the groove and timing that draws you into that music. Close your eyes and...
By the way, a fun short track for you to try, ImpulseRecord.com - Impulse Response Convolution Reverb Library Samples!
Try the first track, March of the Mantidae. It's an example of using convolution to add space to a recording. It's fun!
Last edited:
I believe both the 202 and 222 have crosstalk issues when used for loopback, some details in UCA202 owners, please read! - AVS Forum | Home Theater Discussions And Reviews
you need an analogue input for a true dual channel measurement btw but not if you just want to capture flight time accurately (e.g. for excess phase removal or for manipulating/comparing multiple measurements). REW just needs a recognisable copy (i.e. don't try to put a low passed signal into the timing input) of the measurement signal to arrive on its timing reference input channel, it doesn't care how it gets there.