Thanks dwk123, your note gives some perspective to my thoughts. The Optimal Source Distribution approach is quite nice. As a smooth transition from traditional stereo to ambiophonic stereo, it incorparates the best of both worlds and i wonder how such a device will sound - and what kind of new tradeoffs it introduces.
Hi syntheticwave, wave field synthesis is a nice idea but not really connected to the diy world, since it is just too complex.
Hi syntheticwave, wave field synthesis is a nice idea but not really connected to the diy world, since it is just too complex.
WOW worth some experimentation!!!
dwk123
Thanks EVER so much for mentioning this.
At first read OSD appears barely practical, but a quick look at the little public material at the ISVR
site this morning was enough to get ready to give it a go - little to lose. Fortunately my speakers are modular with B,M,T in individual boxes, so some improvisation of stands let me set up a OSD array with B far apart, M at +/-30 degrees and T at +/-5 degrees. I generated stereo-convolution impulses to provide the correct filtering (300Hz and 2kHz crossovers) and crosstalk cancellation as per the general guideline given (though not, it seems, using the ISVR algorithm - there is something I don't yet understand - in effect I used modified RACE type algorithms, combined appropriately over the 3 frequency bands).
I'm astonished: even this "poor" approximation to the ideal works great on several recordings.
It only works excellently, as expected, in a rather narrow left-right zone (couple of cm each way). The filters are almost flat so the radiated power has, as predicted, almost the same tone balance as before. This is totally different from every other crosstalk canceller I've tried: they never avoided unwanted colouration due to enhanced room reflections.
The image is as solid as I've experienced, and robust against small head rotations. The front-back zone is smaller than it should be: something to inverstigate; also my aim was off and the listening distance is 40cm smaller than it should have been.
The tweeters need mm-precise positioning to set the listening axis, the rest is fairly uncritical to set up (a couple of cm precision works and is easy).
The most amusing thing to hear was a high-rhat solidly imaged to come right out of the 12" bass drive.
A bit more thinking, calculating and planning to do to see how to optimise this setup and find the bugs (the main one of which at the moment is that the tweeters block my fire). Meantime it is more than listenable.
More later.
Ken
dwk123 said:
dwk123
Thanks EVER so much for mentioning this.
At first read OSD appears barely practical, but a quick look at the little public material at the ISVR
site this morning was enough to get ready to give it a go - little to lose. Fortunately my speakers are modular with B,M,T in individual boxes, so some improvisation of stands let me set up a OSD array with B far apart, M at +/-30 degrees and T at +/-5 degrees. I generated stereo-convolution impulses to provide the correct filtering (300Hz and 2kHz crossovers) and crosstalk cancellation as per the general guideline given (though not, it seems, using the ISVR algorithm - there is something I don't yet understand - in effect I used modified RACE type algorithms, combined appropriately over the 3 frequency bands).
I'm astonished: even this "poor" approximation to the ideal works great on several recordings.
It only works excellently, as expected, in a rather narrow left-right zone (couple of cm each way). The filters are almost flat so the radiated power has, as predicted, almost the same tone balance as before. This is totally different from every other crosstalk canceller I've tried: they never avoided unwanted colouration due to enhanced room reflections.
The image is as solid as I've experienced, and robust against small head rotations. The front-back zone is smaller than it should be: something to inverstigate; also my aim was off and the listening distance is 40cm smaller than it should have been.
The tweeters need mm-precise positioning to set the listening axis, the rest is fairly uncritical to set up (a couple of cm precision works and is easy).
The most amusing thing to hear was a high-rhat solidly imaged to come right out of the 12" bass drive.
A bit more thinking, calculating and planning to do to see how to optimise this setup and find the bugs (the main one of which at the moment is that the tweeters block my fire). Meantime it is more than listenable.
More later.
Ken
I´m looking to trying opsodis with my system. It now consists of crosstalk cancelled dipole stereo crossed at 90hz and wide-spaced (sub)woofers below that frequency.
200hz seems to be the lower operational limit for a crosstalk cancelled dipole.
The idea is to cross them higher than 90hz. What x-over frequency and order would you guys suggest? Would 200hz be best? I´ll be using passive line level crossovers from Marchand so I want to be sure before ordering.
200hz seems to be the lower operational limit for a crosstalk cancelled dipole.
The idea is to cross them higher than 90hz. What x-over frequency and order would you guys suggest? Would 200hz be best? I´ll be using passive line level crossovers from Marchand so I want to be sure before ordering.
lower cutoff and extended opsodis
At the moment my lowest pair of speakers is only about 57 degrees apart (span), with mono driven subs taking over from them below about 80Hz. My hilbert transformer was made to work down to 200Hz (1 dB amplitude error) and together with the 4th order "crossover" in my implementation of the osd (I dislike "opsodis") algorithim, gives an effective cutoff a bit above that.
After studying the paper, I felt that might be a bit low so changed the crossover filter to 400Hz. So I have mono up to ~80Hz, then stereo up to around 400Hz and cancellation above that. This is probably tonaly better (subtly) and no worse imaging as far as I can hear.
It depends on the angle of the speakers (and the size of your head). I think working down to an effective "n" of 0.3 is about as low as it is worth going. See figure 18 of Takeuchi and Nelson.
Perhaps of interest: I see Takeuchi gave a talk at Acoustics 2008 "Extension of Optimal Source Distribution principle". I've only read the abstract, but it implies that separating out sum (presumably for playback through a single central speaker) and out of phase signal (for opsodis) is an improvement. I'll see if there is a published version available. I'm not in a position to try it as I don't have a spare speaker that would be good enough, and I don't immediately have the right software setup either (at least I can't see how to do it with what I have). So something for the future, perhaps.
I believe you will be very satisfied with opsodis (with at least 3 bands) compared to crosstalk cancellation on a pair of speakers. Mainly because the room reflections will have the right tonal balance.
Ken
At the moment my lowest pair of speakers is only about 57 degrees apart (span), with mono driven subs taking over from them below about 80Hz. My hilbert transformer was made to work down to 200Hz (1 dB amplitude error) and together with the 4th order "crossover" in my implementation of the osd (I dislike "opsodis") algorithim, gives an effective cutoff a bit above that.
After studying the paper, I felt that might be a bit low so changed the crossover filter to 400Hz. So I have mono up to ~80Hz, then stereo up to around 400Hz and cancellation above that. This is probably tonaly better (subtly) and no worse imaging as far as I can hear.
It depends on the angle of the speakers (and the size of your head). I think working down to an effective "n" of 0.3 is about as low as it is worth going. See figure 18 of Takeuchi and Nelson.
Perhaps of interest: I see Takeuchi gave a talk at Acoustics 2008 "Extension of Optimal Source Distribution principle". I've only read the abstract, but it implies that separating out sum (presumably for playback through a single central speaker) and out of phase signal (for opsodis) is an improvement. I'll see if there is a published version available. I'm not in a position to try it as I don't have a spare speaker that would be good enough, and I don't immediately have the right software setup either (at least I can't see how to do it with what I have). So something for the future, perhaps.
I believe you will be very satisfied with opsodis (with at least 3 bands) compared to crosstalk cancellation on a pair of speakers. Mainly because the room reflections will have the right tonal balance.
Ken
Thanks for your response, Ken. I guess I´ll try a 4th order at 300hz.
m50sniper:
The barrier is a real drag, more so if you want to watch movies. I have to remove the top of the barrier for that.Plus I had to rig up 6 relay switches to derive the center channel to the dipole and left and right to an additional stereo pair in cinema mode. For me,however, it´s worth the hassle. I can´t enjoy music from standard stereo any more.
m50sniper:
The barrier is a real drag, more so if you want to watch movies. I have to remove the top of the barrier for that.Plus I had to rig up 6 relay switches to derive the center channel to the dipole and left and right to an additional stereo pair in cinema mode. For me,however, it´s worth the hassle. I can´t enjoy music from standard stereo any more.
If you want to use a TV for Ambiophonics, you basically can't use a sound barrier. You have to use RACE software and a Computer. Although I successfully have build small 5" LCD Screens inside the sound barrier but that is no fun. According to Ralph Glasgal, a Sound Barrier has to be at least 3m high or alternatively reach up to the ceiling to work properly. It also should be made of 6cm MDF and some Rockwool to be able to hold frequencies successfully down to 90HZ.
Concerning crosstalk: I've often read on forums that some crosstalk is needed, because that is how it is in real life, but I am sorry to say that that is nonsense. A stereophonic system produces uncontrollable crosstalk and the one sound is represented by two speakers set widely apart. The sound stage is also limited to the incidental angle of the speakers. In Ambiophonics there is no crosstalk and with two closely spaced speakers the direction of the sound can be carefully controlled. The speakers are very close to each other, so it just seemes that one ear hears the sound louder than the other, which is the best way to produce sound as natural as possible. In stereophonics crosstalk happens twice instead of once.
Regards,
Wolfgang
Concerning crosstalk: I've often read on forums that some crosstalk is needed, because that is how it is in real life, but I am sorry to say that that is nonsense. A stereophonic system produces uncontrollable crosstalk and the one sound is represented by two speakers set widely apart. The sound stage is also limited to the incidental angle of the speakers. In Ambiophonics there is no crosstalk and with two closely spaced speakers the direction of the sound can be carefully controlled. The speakers are very close to each other, so it just seemes that one ear hears the sound louder than the other, which is the best way to produce sound as natural as possible. In stereophonics crosstalk happens twice instead of once.
Regards,
Wolfgang
Great News Everyone!
I've found a way to use a TV (if not a beamer) in Ambiophonics with a physical sound barrier. Ralph Glasgal was very enthusiastic of the concept and it should definately work. Call it a tube sound barrier.
The positive changes to the plain sound barrier are that, first of all, barely any sound escapes into the room and you need no room treatment at all except for the bass (Bass traps only). You have space for the legs and the barrier should be easy to shift in the room. If you would make the plain sound barrier properly (6cm MDF) you would even save a lot of money (depends how large the speakers are. You should be as close to them as possible, but large speakers with lots of drivers require a longer distance from the listener. Better stick to small 2-way ones) And of course, you can use a TV.
The negative changes are that it will be a little more complex to build (compared to a simple plain) and that you cannot move back and forth as much; you cannot move your head so well.
If you have a wive which doesn't really like your hobby, this is great news for her. Far less sound escapes out into the room because of the tube barrier and it disturbes noone.
I am building my tube barrier ambio system in about 4 months and will go through each step exactly.
Please post if you have questions! I will answer them as soon as possible.
Stay tuned!
Kindest Regards,
Wolfgang
I've found a way to use a TV (if not a beamer) in Ambiophonics with a physical sound barrier. Ralph Glasgal was very enthusiastic of the concept and it should definately work. Call it a tube sound barrier.
The positive changes to the plain sound barrier are that, first of all, barely any sound escapes into the room and you need no room treatment at all except for the bass (Bass traps only). You have space for the legs and the barrier should be easy to shift in the room. If you would make the plain sound barrier properly (6cm MDF) you would even save a lot of money (depends how large the speakers are. You should be as close to them as possible, but large speakers with lots of drivers require a longer distance from the listener. Better stick to small 2-way ones) And of course, you can use a TV.
The negative changes are that it will be a little more complex to build (compared to a simple plain) and that you cannot move back and forth as much; you cannot move your head so well.
If you have a wive which doesn't really like your hobby, this is great news for her. Far less sound escapes out into the room because of the tube barrier and it disturbes noone.
I am building my tube barrier ambio system in about 4 months and will go through each step exactly.
Please post if you have questions! I will answer them as soon as possible.
Stay tuned!
Kindest Regards,
Wolfgang
Attachments
Wolfgang,
That is a very, very interesting idea indeed. One I must have a stab at as soon as I can. Please keep us posted on your effort.
The need for the inner side of the cavities to be free of absorption seems intriguing to me since common sense would suggest otherwise. Could you explain the reasoning behind this?
Also I´m thinking a non absorbent cavity may add a honky signature to the sound as horns usually do ?
That is a very, very interesting idea indeed. One I must have a stab at as soon as I can. Please keep us posted on your effort.
The need for the inner side of the cavities to be free of absorption seems intriguing to me since common sense would suggest otherwise. Could you explain the reasoning behind this?
Also I´m thinking a non absorbent cavity may add a honky signature to the sound as horns usually do ?
Dear Poldus!
I really respect your intelligence in not rejecting new ideas of how to be using Ambio. Most people are far more narrow minded! ->Happy!
The reflection on the inside of the tube has a reflection angle of less than 5° so there is barely any distortion. If you use absorption material on the reflection point certain frequencies get absorbed more than others. The remaining sound is strongly distorted and with a reflection angle of 5° the sound doesn't decrease in volume. What you end up is a very strong distortion in the frequency response magnitude.
This is a very common mistake people do when implementing a sound barrier in Ambio. Make sure to remove the fibergless or rockwool from the wood on your plain sound barrier.
YES! I was worried about honky sound too when the sound gets reflected by the wood. But Ralph Glasgal explained a 5° reflection angle doesn't distort the sound enough to make a difference. (Make sure to use MDF!)
A quote from one of Ralph's e-mail:
"I still don't really believe that any reflections within the tube will be audible. But if the internal sound proofing will keep sound from escaping the tube then this is fine if it doesn't cost too much. A reflection angle of 5 degrees vertical or horizontal will not be audible in my opinion. If you actually eliminate every reflection in the tube you will lose loudness and this will strain your vacuum tube amplifiers and thus increase distortion. Also the frequency response of the sound will be changed because different frequencies will be absorbed at different rates no matter how you build the thing."
When you have a normal room with or without room treatment, the sound gets reflected by the walls and ceiling. You usually are only worrying about the reverberation times.
I have heard, Aluminium is used as reflection panels (for example the acoustic lense thechnology from Bang & Olufsen) so it might be possible to cover the Sound barrier with a layer of Aluminium. I am not sure about it, it's just an idea that I got when writing this post; I will get back to you on this in a couple of days.
I am awfully sorry not to be able to build this thing quickly. I am lacking money, and am studying hard for university. I just don't have the time for it.
You might be the very first person in the entire world to try this method which might open new possibilities in Ambiophonics.
KIndest Regards,
Wolfgang
I really respect your intelligence in not rejecting new ideas of how to be using Ambio. Most people are far more narrow minded! ->Happy!
The reflection on the inside of the tube has a reflection angle of less than 5° so there is barely any distortion. If you use absorption material on the reflection point certain frequencies get absorbed more than others. The remaining sound is strongly distorted and with a reflection angle of 5° the sound doesn't decrease in volume. What you end up is a very strong distortion in the frequency response magnitude.
This is a very common mistake people do when implementing a sound barrier in Ambio. Make sure to remove the fibergless or rockwool from the wood on your plain sound barrier.
YES! I was worried about honky sound too when the sound gets reflected by the wood. But Ralph Glasgal explained a 5° reflection angle doesn't distort the sound enough to make a difference. (Make sure to use MDF!)
A quote from one of Ralph's e-mail:
"I still don't really believe that any reflections within the tube will be audible. But if the internal sound proofing will keep sound from escaping the tube then this is fine if it doesn't cost too much. A reflection angle of 5 degrees vertical or horizontal will not be audible in my opinion. If you actually eliminate every reflection in the tube you will lose loudness and this will strain your vacuum tube amplifiers and thus increase distortion. Also the frequency response of the sound will be changed because different frequencies will be absorbed at different rates no matter how you build the thing."
When you have a normal room with or without room treatment, the sound gets reflected by the walls and ceiling. You usually are only worrying about the reverberation times.
I have heard, Aluminium is used as reflection panels (for example the acoustic lense thechnology from Bang & Olufsen) so it might be possible to cover the Sound barrier with a layer of Aluminium. I am not sure about it, it's just an idea that I got when writing this post; I will get back to you on this in a couple of days.
I am awfully sorry not to be able to build this thing quickly. I am lacking money, and am studying hard for university. I just don't have the time for it.
You might be the very first person in the entire world to try this method which might open new possibilities in Ambiophonics.
KIndest Regards,
Wolfgang
According to Ralph Glasgal, there isn't much difference between wood sound reflection and Aluminium sound reflection. A thin layer of Aluminium on wood might reflect the high frequencies better but distort the lower ones. The barrier would need to be made of pure 25mm Aluminium. I do not know about the costs of Aluminium nor where you can get it, but Ralph Glasgal believes there will be, if only minute, improvement in treble.
This is not guaranteed and I have not yet decidet what to build exactly. Feel free to experiment. Fact is, the sound cannot get worse when using Aluminium.
This is not guaranteed and I have not yet decidet what to build exactly. Feel free to experiment. Fact is, the sound cannot get worse when using Aluminium.
I have jury-rigged a tube dipole and guess what? It works!! It gets rid of deleterious room interactions and offers a clear, detailed sound. On the flip side it does require a lot of amplifier power for the same volume.
It will take me a while to work out a good practical design for my speakers. (I need a two meters long barrier).
It will take me a while to work out a good practical design for my speakers. (I need a two meters long barrier).
Hi Poldus!
Juhee! Someone finally put some meaning into my life! I'm very happy it works, although I have no idea what you mean by "jury-rigged a tube dipole"... How loud are your speakers per Watt? Any idea what would be optimal?
Do you have the e-mail to M50SNIPER, because he seemed to have the same problems like you.
Kindest Regards,
Wolfgang
Juhee! Someone finally put some meaning into my life! I'm very happy it works, although I have no idea what you mean by "jury-rigged a tube dipole"... How loud are your speakers per Watt? Any idea what would be optimal?
Do you have the e-mail to M50SNIPER, because he seemed to have the same problems like you.
Kindest Regards,
Wolfgang
Wolfgang,
I just put some acoustic-absorbent material above and to the sides of my speakers and barrier. My crosstalk barrier is itself very absorbent, hence the severe db loss and need for more amp power.
Following your indications I´ll next substitute plain mdf for the current barrier and hopefully restore some of the lost db.
I do not have M50SNIPER´s email. He is welcome to share his thoughts and experience here in the forum.
Cheers
I just put some acoustic-absorbent material above and to the sides of my speakers and barrier. My crosstalk barrier is itself very absorbent, hence the severe db loss and need for more amp power.
Following your indications I´ll next substitute plain mdf for the current barrier and hopefully restore some of the lost db.
I do not have M50SNIPER´s email. He is welcome to share his thoughts and experience here in the forum.
Cheers
Dear Poldus,
Make sure the tube barrier is tall enough for the speakers you will use. You might get a nervous breakdown when you realise your speakers don't fit in it. The speakers have to be very loud, because as you said the volume decreases in the tube barrier. I did not decide on my speaker project, but it's probably going to be some kind of PA two-way speaker.
In a normal room 100db is a nightmare. The sound rings in you ears and you get very aggresive. In a treated room it should be possible to play music at 100db and more without the sound quality decreasing. Using the Tube barrier, there is no limit. The sound gets as loud and as good as the speakers allow. Therefore I would recommend a pair of small sealed PA speakers. Hi-End PA speakers are very expensive, so better save some money first.
PS: I know this is a little exaggerated, but I think it would help to choose the right speaker for you and me; Can you measure the difference in Volume of the speakers using the tube barrier technique (With the MDF crosstalk canceller) versus common stereo?
Kindest Regards,
Wolfgang
Make sure the tube barrier is tall enough for the speakers you will use. You might get a nervous breakdown when you realise your speakers don't fit in it. The speakers have to be very loud, because as you said the volume decreases in the tube barrier. I did not decide on my speaker project, but it's probably going to be some kind of PA two-way speaker.
In a normal room 100db is a nightmare. The sound rings in you ears and you get very aggresive. In a treated room it should be possible to play music at 100db and more without the sound quality decreasing. Using the Tube barrier, there is no limit. The sound gets as loud and as good as the speakers allow. Therefore I would recommend a pair of small sealed PA speakers. Hi-End PA speakers are very expensive, so better save some money first.
PS: I know this is a little exaggerated, but I think it would help to choose the right speaker for you and me; Can you measure the difference in Volume of the speakers using the tube barrier technique (With the MDF crosstalk canceller) versus common stereo?
Kindest Regards,
Wolfgang
Hi poldus!
It seems like it's better to use a slant of 30° not 45° for the top wood plate. Otherwise the remaining bits of sound get reflected into the room horizontally, reflected by the wall and more of the distorted sound will enter your ear. Better use a slant of 30°.
And here is my concept of the tube barrier when using tall speakers. Use Rockwool and Wood in the same manner as in the smaller one. This one might be less practical.
It seems like it's better to use a slant of 30° not 45° for the top wood plate. Otherwise the remaining bits of sound get reflected into the room horizontally, reflected by the wall and more of the distorted sound will enter your ear. Better use a slant of 30°.
And here is my concept of the tube barrier when using tall speakers. Use Rockwool and Wood in the same manner as in the smaller one. This one might be less practical.
Attachments
Hi Wolfgang,
After a hectic couple of days testing different iterations based on your design I have returned to my previous plain barrier. Two main reasons:
I found that a lot of absorbent material is needed for effective inter-channel interference, which along with the refflective inner sides of the tubes wreaks havoc with the tonal balance at the listener´s end.
This can be overcome by voicing the speakers filter in a way that counteracts the inbalance. Personally that´s a task I am not ready to tackle right now (spent a few years tweaking my filter to adapt to the absorbent plain barrier).
Also, an effective tube barrier turns out to be pretty bulky and at least as high as the listener´s ears, so it´s finally not as inconspicuous either.
I have to say that the idea warrants further testing, however. It may not be long before I give it another shot.
After a hectic couple of days testing different iterations based on your design I have returned to my previous plain barrier. Two main reasons:
I found that a lot of absorbent material is needed for effective inter-channel interference, which along with the refflective inner sides of the tubes wreaks havoc with the tonal balance at the listener´s end.
This can be overcome by voicing the speakers filter in a way that counteracts the inbalance. Personally that´s a task I am not ready to tackle right now (spent a few years tweaking my filter to adapt to the absorbent plain barrier).
Also, an effective tube barrier turns out to be pretty bulky and at least as high as the listener´s ears, so it´s finally not as inconspicuous either.
I have to say that the idea warrants further testing, however. It may not be long before I give it another shot.