The No.1 Website for Pro Audio
High end nearfield test
Old 22nd September 2019
  #3451
Lives for gear
 
dotl's Avatar
 

Quote:
Originally Posted by Karloff70 View Post
Better for what?
For precise monitoring the high frequency range, what else?!
Old 23rd September 2019
  #3452
Gear Guru
 
Karloff70's Avatar
 

Quote:
Originally Posted by dotl View Post
For precise monitoring the high frequency range, what else?!
Well, 'better bass' isn't very clear. The PSI will go loudest without breaking up by far, the one15 have a limit you will find and the 906 break up early. That is one angle on reality. Then there is the amount of actual bass you get into the room, the PSI might win there too, and if you're talking amount of clean bass the one15 is second. And then there is how useful the information in the bass is for making decisions. And there, for me the amphion slays both the other two. So, there.
Old 23rd September 2019
  #3453
Lives for gear
 
dotl's Avatar
 

Quote:
Originally Posted by Karloff70 View Post
Well, 'better bass' isn't very clear. The PSI will go loudest without breaking up by far, the one15 have a limit you will find and the 906 break up early. That is one angle on reality. Then there is the amount of actual bass you get into the room, the PSI might win there too, and if you're talking amount of clean bass the one15 is second. And then there is how useful the information in the bass is for making decisions. And there, for me the amphion slays both the other two. So, there.
Thank you!

Yes, i guess usefulness is the most important thing here. I need a relatively small monitor for sound design and im torn between these 3: One15, PSIa17m and RL906. I don’t need them to go louder than 85 dBs anyway, so regarding how loud they can go all 3 are good enough. For bass, im satisfied with the range of upto 45-50 hz.
What i need is a precise signal representation in near field monitoring for the money at moderate listening levels. Lowest bass i can check on my headphones.
Old 23rd September 2019
  #3454
Gear Guru
 
Karloff70's Avatar
 

Quote:
Originally Posted by dotl View Post
Thank you!

Yes, i guess usefulness is the most important thing here. I need a relatively small monitor for sound design and im torn between these 3: One15, PSIa17m and RL906. I don’t need them to go louder than 85 dBs anyway, so regarding how loud they can go all 3 are good enough. For bass, im satisfied with the range of upto 45-50 hz.
What i need is a precise signal representation in near field monitoring for the money at moderate listening levels. Lowest bass i can check on my headphones.
No, what you need is a dealer who will let you take them to try at yours and do actual work on them. Then you'll know which one suits you. The rest is guessing.
Old 23rd September 2019
  #3455
Lives for gear
 
dotl's Avatar
 

Quote:
Originally Posted by Karloff70 View Post
No, what you need is a dealer who will let you take them to try at yours and do actual work on them. Then you'll know which one suits you. The rest is guessing.
I agree! But, until then...
Gearslutz - From dusk till dawn 137

Really, im not asking for objectivity.
Old 23rd September 2019
  #3456
Gear Addict
 
syncussion's Avatar
 

Quote:
Originally Posted by dotl View Post
I agree! But, until then...
Gearslutz - From dusk till dawn 137

Really, im not asking for objectivity.
Never having heard the PSI A17M, it doesn't look like a great design. Has a 7" polypropylene woofer taken up to 3kHz. Won't be nearly as linear as for instance the aluminium 5" woofer taken to 1.6kHz of the Amphion.
Also, not an expert on waveguides, but that doesn't look like much of a waveguide on the A17M and I've often seen bad measuring results of a tweeter mounted from the back with a chamfer like that. (and if fr corrected on-axis the off-axis will suffer badly)
Also a 7" will be quite directional by the time it reaches 3kHz resulting in bad off-axis. If you're not in an anechoic room this matters, often a great deal.
Should have much more bass than the Amphion 5" though, but for bass quality cone stiffness and the quality of the motor and suspension matters and you don't know this untill you measure distortion etc at higher SPL.
Old 23rd September 2019
  #3457
Gear Guru
 
Karloff70's Avatar
 

Quote:
Originally Posted by syncussion View Post
Never having heard the PSI A17M, it doesn't look like a great design. Has a 7" polypropylene woofer taken up to 3kHz. Won't be nearly as linear as for instance the aluminium 5" woofer taken to 1.6kHz of the Amphion.
Also, not an expert on waveguides, but that doesn't look like much of a waveguide on the A17M and I've often seen bad measuring results of a tweeter mounted from the back with a chamfer like that. (and if fr corrected on-axis the off-axis will suffer badly)
Also a 7" will be quite directional by the time it reaches 3kHz resulting in bad off-axis. If you're not in an anechoic room this matters, often a great deal.
Should have much more bass than the Amphion 5" though, but for bass quality cone stiffness and the quality of the motor and suspension matters and you don't know this untill you measure distortion etc at higher SPL.
Won't get much distortion from PSI's.....they have a weird driver compensation thing where the bass driver actually never looks like it moves at all. I didn't get on with that, need my cones moving.
Old 24th September 2019
  #3458
Lives for gear
 
dotl's Avatar
 

Yes, test results for the a17m looks good!

http://downloads.psiaudio.com/studio...S_techdata.pdf
Old 11th December 2019
  #3459
Here for the gear
 

Quote:
Originally Posted by dotl View Post
Thank you!

Yes, i guess usefulness is the most important thing here. I need a relatively small monitor for sound design and im torn between these 3: One15, PSIa17m and RL906. I don’t need them to go louder than 85 dBs anyway, so regarding how loud they can go all 3 are good enough. For bass, im satisfied with the range of upto 45-50 hz.
What i need is a precise signal representation in near field monitoring for the money at moderate listening levels. Lowest bass i can check on my headphones.
Hey man, I am in the same boat as you were and looking at Ultra Near Field Monitors. I am currently deciding between One 15 and RL906. One 15 would be better because I just bought Medium IsoAcoustics stands for my Kef LS50 and I can use the same for them, not sure about RL906.

What did you end up buying?
Old 3 weeks ago
  #3460
Gear Addict
 

@ syncussion

You were closer than you think to great monitors with your experiment with Ikea bowls.
A couple things first. Metal drivers might sound too harsh at close range. Also, the Markaudio drivers I tried make some buzzing noise well before they reach x-max.
But mainly, you shouldn't have equalized them flat (I made the same mistake years ago, by the way. I was surprised at how much a truly flat response sounds lifeless).

There is some consensus on the 'ideal' speaker response (@ the listening position) to be tilted downward of a certain degree.
I would steal those speakers back (great choice making them not ported, by the way) and equalize them differently.
For a starting point of what the EQ should be like, I suggest 'Accurate Sound Reproduction Using DSP' by Mitch Barnett.
Old 3 weeks ago
  #3461
Gear Addict
 
syncussion's Avatar
 

Quote:
Originally Posted by sax512 View Post
@ syncussion

You were closer than you think to great monitors with your experiment with Ikea bowls.
A couple things first. Metal drivers might sound too harsh at close range. Also, the Markaudio drivers I tried make some buzzing noise well before they reach x-max.
But mainly, you shouldn't have equalized them flat (I made the same mistake years ago, by the way. I was surprised at how much a truly flat response sounds lifeless).

There is some consensus on the 'ideal' speaker response (@ the listening position) to be tilted downward of a certain degree.
I would steal those speakers back (great choice making them not ported, by the way) and equalize them differently.
For a starting point of what the EQ should be like, I suggest 'Accurate Sound Reproduction Using DSP' by Mitch Barnett.
Oh those Ikea bowl full range driver speakers were many years years ago as an experiment for speakers really close to the ears like headphones.
I've since become much wiser Several problems with that idea, still not correct HRTF when too close, and as far as the brain is concerned the far-field starts at >1m as it detects the level differences between left and right ear, etc.
And single full range drivers are all horrible due to (managed) cone breakup, directivity issues etc etc but for that project I had no choice.

But yes, a large roundover or round overall shape is very good against edge diffraction / a smooth baffle step. But it should be done with a dedicated mid driver and tweeter with close center to center spacing etc.
Metal drivers (or ceramic) are usually better than paper etc drivers regarding frequency response due to cone edge resonance in mid drivers and early breakup in tweeters. Though Purifi-audio now have a paper mid driver that is as well behaved as a metal one in the passband due mainly to an innovative surround.

As for a flat response sounding lifeless. I have no idea where you got that idea but it's completely not true.
Old 3 weeks ago
  #3462
Gear Addict
 

Quote:
Originally Posted by syncussion View Post
I've since become much wiser Several problems with that idea, still not correct HRTF when too close, and as far as the brain is concerned the far-field starts at >1m as it detects the level differences between left and right ear, etc.
The HRTF is ALWAYS correct. It is a parameter associated (mainly) to the head-source direction, not the source distance. If you're referring to the fact that at close source range the HRTF is also dependent on distance and not only the direction, it's true. However, when the source is fixed (speaker), the source can be equalized so to take its position out of the equation.
Buy that's beside the point.
One of the many advantages of a setup like that is to provide left and right ears with a better, less cross talk tainted version of the stereo signal. Why would you care at what distance the brain starts to detect level differences between left and right (which happens at any distance, by the way, and has more or less weight on the localization recognition of the source depending on its frequency content)?
The far field to consider in this scenario, on the other hand, is associated to the speaker, and it's the distance at which the signal path difference between the same frequencies, reproduced from 2 drivers in the cross over region, becomes 'small enough'. It depends on the geometry of the speaker, and it could be much more than 1m, even for 2 ways. That distance is very small when using a single driver, which is why you can put a speaker like that very close to the head (20 cm might be pushing it, though), and decrease the contribution of room reflections to the total sound (the other, less proper, definition of far field is the distance at which the power of reflected sound is equal to the power of direct sound).



Quote:
Originally Posted by syncussion View Post
And single full range drivers are all horrible due to (managed) cone breakup, directivity issues etc etc but for that project I had no choice.
That's a myth. I guess there are horrible single drivers out there, just like there are horrible mids, woofers and tweeters. But not all single drivers are horrible. Some don't exhibit phase inversion (cone break-up) at all and have very smooth frequency response. Not flat, but not bumpy either which is all you need to be able to use DSP to equalize them to the target response you want.
Directivity is high, but it doesn't matter in a setup like that, since you're trying to engage the room as little as possible. Limited directivity is actually a plus, in this case.

Quote:
Originally Posted by syncussion View Post
Metal drivers (or ceramic) are usually better than paper etc drivers regarding frequency response due to cone edge resonance in mid drivers and early breakup in tweeters. Though Purifi-audio now have a paper mid driver that is as well behaved as a metal one in the passband due mainly to an innovative surround.
Look at Zaph audio and you'll see that, measurements at hand, some of the best behaved wide band drivers are not metal.
I guess it depends on a lot more factors than simply the material, but metal cones have a specific resonance frequency that sometimes, if not properly designed, can fall too close to the pass band and create potential issues.
With wide band metal drivers, as you correctly say, if the resonance is not managed correctly, you could do more damage than good. Fiber cones have distributed dampening throughout the surface, and while there are badly designed ones out there, there are also a few very good ones.

Quote:
Originally Posted by syncussion View Post
As for a flat response sounding lifeless. I have no idea where you got that idea but it's completely not true.
It's completely true. But mind you, I specified @ the listening position. It also depends on how you measure the response. If you're interested in how good a pair of speakers like the ones you made a couple years ago can truly sound, I can't recommend that book on DSP enough. It also covers a lot of the studies on speaker measurements done previously. In a nutshell, it says that how the speakers measure at the listening position matters more than their anechoic response. Speakers that measure flat anechoically sound good or bad in a room depending on their directivity. But as long as the response is tilted downward in the highs (by the right amount) at the listening position, it doesn't matter how they measure anechoically.
Old 3 weeks ago
  #3463
Gear Addict
 
syncussion's Avatar
 

Quote:
Originally Posted by sax512 View Post
The HRTF is ALWAYS correct. It is a parameter associated (mainly) to the head-source direction, not the source distance. If you're referring to the fact that at close source range the HRTF is also dependent on distance and not only the direction, it's true. However, when the source is fixed (speaker), the source can be equalized so to take its position out of the equation.
Buy that's beside the point.
One of the many advantages of a setup like that is to provide left and right ears with a better, less cross talk tainted version of the stereo signal. Why would you care at what distance the brain starts to detect level differences between left and right (which happens at any distance, by the way, and has more or less weight on the localization recognition of the source depending on its frequency content)?
The far field to consider in this scenario, on the other hand, is associated to the speaker, and it's the distance at which the signal path difference between the same frequencies, reproduced from 2 drivers in the cross over region, becomes 'small enough'. It depends on the geometry of the speaker, and it could be much more than 1m, even for 2 ways. That distance is very small when using a single driver, which is why you can put a speaker like that very close to the head (20 cm might be pushing it, though), and decrease the contribution of room reflections to the total sound (the other, less proper, definition of far field is the distance at which the power of reflected sound is equal to the power of direct sound).





That's a myth. I guess there are horrible single drivers out there, just like there are horrible mids, woofers and tweeters. But not all single drivers are horrible. Some don't exhibit phase inversion (cone break-up) at all and have very smooth frequency response. Not flat, but not bumpy either which is all you need to be able to use DSP to equalize them to the target response you want.
Directivity is high, but it doesn't matter in a setup like that, since you're trying to engage the room as little as possible. Limited directivity is actually a plus, in this case.



Look at Zaph audio and you'll see that, measurements at hand, some of the best behaved wide band drivers are not metal.
I guess it depends on a lot more factors than simply the material, but metal cones have a specific resonance frequency that sometimes, if not properly designed, can fall too close to the pass band and create potential issues.
With wide band metal drivers, as you correctly say, if the resonance is not managed correctly, you could do more damage than good. Fiber cones have distributed dampening throughout the surface, and while there are badly designed ones out there, there are also a few very good ones.



It's completely true. But mind you, I specified @ the listening position. It also depends on how you measure the response. If you're interested in how good a pair of speakers like the ones you made a couple years ago can truly sound, I can't recommend that book on DSP enough. It also covers a lot of the studies on speaker measurements done previously. In a nutshell, it says that how the speakers measure at the listening position matters more than their anechoic response. Speakers that measure flat anechoically sound good or bad in a room depending on their directivity. But as long as the response is tilted downward in the highs (by the right amount) at the listening position, it doesn't matter how they measure anechoically.
No, there's a normal far field HRTF unique to each person and that HRTF changes noticeable for small distances. The angle to various parts of the face vs ear changes including level differences as the ear is closer than the rest of the face and level differences between the ear as I said before (and for HRTF both ears count). This is well known to start at distances <1m which is called the nearfield as far as HRTF is concerned. And is very noticeable for the distances I had planned my idea back then <0.4m). So there's a change in fr in the HRTF and there's a change in amplitude between the ears which our ears naturally uses to estimate sources that are close to us. Crossfeed is natural and only sources very close to us produce this level difference between our ears.
I've tested my idea in a near anechoic room and the speakers sounded bad in a half headphone way at 20-40cm and started to sound more natural at >60-80cm if I recall correctly.

As for metal vs paper, I'm talking about dedicated mid drivers and tweeters. Yes of course a metal mid driver shouldn't be taken too high and one should watch out for their stronger cone breakup. I think for instance a 6-7" driver shouldn't be crossed higher than 2k, preferably lower whether it's paper or metal. Also due to directivity match with the tweeter and not wanting any directivity issues either in the 2-5kHz region that our ear is so sensitive.

As for speaker flatness. I know what you're talking about and research done on this by Toole and at Harman etc. But I think that is flawed to the core. It also doesn't apply to studio use and is dependent not just on listener and use and room but also very much on the recording. Besides this I know another reason why flat fr can get one into trouble and I happen to developing a commercial product to remedy this but can't share about this now Will do so later.
But in the basis, with qulity loudspeakers, flat is natural and can be completely transparent. But will argue that with proof at a different time
Old 3 weeks ago
  #3464
Gear Addict
 

Quote:
Originally Posted by syncussion View Post
No, there's a normal far field HRTF unique to each person and that HRTF changes noticeable for small distances. The angle to various parts of the face vs ear changes including level differences as the ear is closer than the rest of the face and level differences between the ear as I said before (and for HRTF both ears count). This is well known to start at distances <1m which is called the nearfield as far as HRTF is concerned. And is very noticeable for the distances I had planned my idea back then <0.4m). So there's a change in fr in the HRTF and there's a change in amplitude between the ears which our ears naturally uses to estimate sources that are close to us.
I had a hunch that is what you were referring to, and I acknowledged this fact. I agree that for 20 cm, the HRTFs could vary enough that small displacement of the head around the sweet spot can cause too much variation. Also, at 20 cm, one could argue that even a single driver is not exactly in far field, as it is still a distributed source, not an ideal infinitesimal point.

My speakers look a lot like yours (only bigger cabinet and drivers), and they are to be listened to between 0.3 and 1 m from the center of the sweet spot. Definitely not any closer than 0.3, and at more than 1 m they may leave you wanting for more SPL, if you like it loud (I don't).
At that distance, the source proximity effect is not an issue, as relatively reasonable displacements of the head around the sweet spot will cause a much smaller change than the one caused by 2+ way speakers due to their combined reproduction of the same frequencies in the cross-over range. As long as you make sure your ears stay in the on axis projection cone of the respective driver (around +/- 10-15 deg, depending on the driver design and size), you get a surprisingly wide sweet spot.
Under these conditions, the equalization of the speakers can be done the same way that any other 'normal' mid-far field speakers would be equalized, and that's the underlying assumption I made when I said that the HRTF is not dependent on distance. Sorry I probably didn't make it clear enough.

Quote:
Originally Posted by syncussion View Post
Crossfeed is natural and only sources very close to us produce this level difference between our ears.
Crosstalk might be natural and people are used to it, but there is a very marked increase in sound transparency and intelligibility when it is reduced, which happens as a side bonus when the speakers are close to the head and the distance between one speaker and the opposite ear is not anymore about the same as with the same side ear, so that the square law adds to the head shadow effect and further decreases crosstalk as a consequence. This is something that should be wanted. It helps in making the speakers disappear (sonically speaking) and makes it easier for the brain to extract the binaural cues that are present even in traditional stereo recordings, expanding the sound stage further left and right than the speaker location, and closer and farther than their physical distance in front of you, depending on what's actually in the recording (which is what we are all ultimately interested in extracting).

Quote:
Originally Posted by syncussion View Post
I've tested my idea in a near anechoic room and the speakers sounded bad in a half headphone way at 20-40cm and started to sound more natural at >60-80cm if I recall correctly.
That's because at that distance the actual frequency response is tilted downward in the highs, as a consequence of what the room adds to the frequency balance.

Quote:
Originally Posted by syncussion View Post
As for metal vs paper, I'm talking about dedicated mid drivers and tweeters. Yes of course a metal mid driver shouldn't be taken too high and one should watch out for their stronger cone breakup. I think for instance a 6-7" driver shouldn't be crossed higher than 2k, preferably lower whether it's paper or metal. Also due to directivity match with the tweeter and not wanting any directivity issues either in the 2-5kHz region that our ear is so sensitive.
If you're concerned (as you should be!) with the 2-5 kHz region, that's the main reason why you should go with a single driver.
If you forget about directivity for an instant, and even discounting the benefits of taking down room reflections by 20-25 dB (at least) as a consequence of placing the sound source closer to your head, there is no substitute for the benefit of lack of crossover. Even with dedicated midrange (3+ ways speakers), there are problems related to the time alignment of the drivers that are hard to fix even in an individual point in space and impossible to fix when you move around that point (imagine the point being the center of your head, and the movement being the actual location of your ears, a few cm left and right).
But if you want to fill the room with sound, for group listening, then yes, you should be concerned with directivity of the speakers (on top of driver time alignment). In that case, single drivers are not a good choice. They simply can't produce enough bass for more than one listener.
Although some speakers, like Eclipse TD, seem to have filled this gap, by using ported cabinets.
Since I am interested in transducing the signal in the recording into a pressure wave as close to that signal as possible, ported cabinets are not an option for me.

Quote:
Originally Posted by syncussion View Post
As for speaker flatness. I know what you're talking about and research done on this by Toole and at Harman etc. But I think that is flawed to the core. It also doesn't apply to studio use and is dependent not just on listener and use and room but also very much on the recording. Besides this I know another reason why flat fr can get one into trouble and I happen to developing a commercial product to remedy this but can't share about this now Will do so later.
But in the basis, with qulity loudspeakers, flat is natural and can be completely transparent. But will argue that with proof at a different time
Every listening situation is dependent on the recording. That's to be expected. No speaker design can fix that for all recordings.
The book I suggested actually departs from the findings of Toole (I agree that it is somewhat flawed), shifting the attention from how the speaker measures in an anechoic chamber to how the speakers directivity affect the perceived tonal balance by the time the room they are put in adds to the sound that gets to the listener.
What it basically does is it tries to find a target curve (@ the listening position) that is good enough so that it works for the average recording (tonally speaking), where recordings that are bright don't sound too thin, and the ones that are bass heavy don't sound too dark. A happy medium, that is useful for the mixing/mastering engineer. Knowing full well that if the recording is unbalanced there's nothing that can fix it other than specific added EQ for that specific recording, the information in there is very helpful, and I think this book is a fundamental reading, especially if you intend to make a product that aims at addressing these matters.
It is waaaay beyond the research done by Toole. It barely even builds up from it.
Old 3 weeks ago
  #3465
Gear Addict
 
syncussion's Avatar
 

Quote:
Originally Posted by sax512 View Post
I had a hunch that is what you were referring to, and I acknowledged this fact. I agree that for 20 cm, the HRTFs could vary enough that small displacement of the head around the sweet spot can cause too much variation. Also, at 20 cm, one could argue that even a single driver is not exactly in far field, as it is still a distributed source, not an ideal infinitesimal point.

My speakers look a lot like yours (only bigger cabinet and drivers), and they are to be listened to between 0.3 and 1 m from the center of the sweet spot. Definitely not any closer than 0.3, and at more than 1 m they may leave you wanting for more SPL, if you like it loud (I don't).
At that distance, the source proximity effect is not an issue, as relatively reasonable displacements of the head around the sweet spot will cause a much smaller change than the one caused by 2+ way speakers due to their combined reproduction of the same frequencies in the cross-over range. As long as you make sure your ears stay in the on axis projection cone of the respective driver (around +/- 10-15 deg, depending on the driver design and size), you get a surprisingly wide sweet spot.
Under these conditions, the equalization of the speakers can be done the same way that any other 'normal' mid-far field speakers would be equalized, and that's the underlying assumption I made when I said that the HRTF is not dependent on distance. Sorry I probably didn't make it clear enough.



Crosstalk might be natural and people are used to it, but there is a very marked increase in sound transparency and intelligibility when it is reduced, which happens as a side bonus when the speakers are close to the head and the distance between one speaker and the opposite ear is not anymore about the same as with the same side ear, so that the square law adds to the head shadow effect and further decreases crosstalk as a consequence. This is something that should be wanted. It helps in making the speakers disappear (sonically speaking) and makes it easier for the brain to extract the binaural cues that are present even in traditional stereo recordings, expanding the sound stage further left and right than the speaker location, and closer and farther than their physical distance in front of you, depending on what's actually in the recording (which is what we are all ultimately interested in extracting).



That's because at that distance the actual frequency response is tilted downward in the highs, as a consequence of what the room adds to the frequency balance.



If you're concerned (as you should be!) with the 2-5 kHz region, that's the main reason why you should go with a single driver.
If you forget about directivity for an instant, and even discounting the benefits of taking down room reflections by 20-25 dB (at least) as a consequence of placing the sound source closer to your head, there is no substitute for the benefit of lack of crossover. Even with dedicated midrange (3+ ways speakers), there are problems related to the time alignment of the drivers that are hard to fix even in an individual point in space and impossible to fix when you move around that point (imagine the point being the center of your head, and the movement being the actual location of your ears, a few cm left and right).
But if you want to fill the room with sound, for group listening, then yes, you should be concerned with directivity of the speakers (on top of driver time alignment). In that case, single drivers are not a good choice. They simply can't produce enough bass for more than one listener.
Although some speakers, like Eclipse TD, seem to have filled this gap, by using ported cabinets.
Since I am interested in transducing the signal in the recording into a pressure wave as close to that signal as possible, ported cabinets are not an option for me.



Every listening situation is dependent on the recording. That's to be expected. No speaker design can fix that for all recordings.
The book I suggested actually departs from the findings of Toole (I agree that it is somewhat flawed), shifting the attention from how the speaker measures in an anechoic chamber to how the speakers directivity affect the perceived tonal balance by the time the room they are put in adds to the sound that gets to the listener.
What it basically does is it tries to find a target curve that is good enough so that it works for the average recording (tonally speaking), where recordings that are bright don't sound too thin, and the ones that are bass heavy don't sound too dark. A happy medium, that is useful for the mixing/mastering engineer. Knowing full well that if the recording is unbalanced there's nothing that can fix it other than specific added EQ for that specific recording, the information in there is very helpful, and I think this book is a fundamental reading, especially if you intend to make a product that aims at addressing these matters.
It is waaaay beyond the research done by Toole. It barely even builds up from it.
The test I did with the single full range was done in near anechoic conditions (only the floor not anechoic) and measurements were equally flat in fr close and further away (as far as a single driver gets flat).

As for single drivers, yes they do not produce enough SPL in the bass. That is one of their problems. Though there are 10" full ranges but what you gain in the bass you lose in the highs. In any case, a full range small or large will never give you the treble quality of a dedicated tweeter. Full ranges all have cone breakup in the treble. They are like headphones in that regard, terrible treble quality that cannot be fixed by EQ.
Also the most audible form of distortion from drivers is IMD and a single full range driver always has a lot of IMD.

And there is nothing wrong with a crossover. A properly done crossover is proven to be 100% undetectable on-axis. It is only in the vertical off-axis that they will give problems with most designs but if you wish to get rid of this then use a tweeter that can cross very low combined with a small mid so you get close CTC spacing so that the cancellation axis is more than 90 degrees from the on-axis (so it doesn't materialize). You then get coaxial without the problems. When done properly you will also get a smooth off-axis between mid and tweeter overall without a waveguide as the mid won't drop off yet off-axis before crossing to the tweeter. This type of design is in everyway vastly superior to any single driver, also in non treated rooms.
But the best room is no room. And that is what studio's often try to achieve in a large part. I'd much rather have a speaker with bad off-axis and cheap drivers etc in a great room that is very well treated than any top speaker in a bad untreated room.

Balancing all, the best sound comes from at least 3-way speakers in a great very well treated room.
Even better if you can mount the speakers in-wall. And this is exactly what you see the more experienced studio designers do, 3 or 4-ways in-wall in great well treated rooms.
Listening close to a single driver speaker doesn't reach that level at all. Though it is much cheaper
Old 3 weeks ago
  #3466
Gear Addict
 

Quote:
Originally Posted by syncussion View Post
The test I did with the single full range was done in near anechoic conditions (only the floor not anechoic) and measurements were equally flat in fr close and further away (as far as a single driver gets flat).

As for single drivers, yes they do not produce enough SPL in the bass. That is one of their problems. Though there are 10" full ranges but what you gain in the bass you lose in the highs. In any case, a full range small or large will never give you the treble quality of a dedicated tweeter. Full ranges all have cone breakup in the treble. They are like headphones in that regard, terrible treble quality that cannot be fixed by EQ.
Also the most audible form of distortion from drivers is IMD and a single full range driver always has a lot of IMD.

And there is nothing wrong with a crossover. A properly done crossover is proven to be 100% undetectable on-axis. It is only in the vertical off-axis that they will give problems with most designs but if you wish to get rid of this then use a tweeter that can cross very low combined with a small mid so you get close CTC spacing so that the cancellation axis is more than 90 degrees from the on-axis (so it doesn't materialize). You then get coaxial without the problems. When done properly you will also get a smooth off-axis between mid and tweeter overall without a waveguide as the mid won't drop off yet off-axis before crossing to the tweeter. This type of design is in everyway vastly superior to any single driver, also in non treated rooms.
But the best room is no room. And that is what studio's often try to achieve in a large part. I'd much rather have a speaker with bad off-axis and cheap drivers etc in a great room that is very well treated than any top speaker in a bad untreated room.

Balancing all, the best sound comes from at least 3-way speakers in a great very well treated room.
Even better if you can mount the speakers in-wall. And this is exactly what you see the more experienced studio designers do, 3 or 4-ways in-wall in great well treated rooms.
Listening close to a single driver speaker doesn't reach that level at all. Though it is much cheaper
Hmmm. There is a lot to argue about what you just said.
Inaudibility of cross over is straight up wishful thinking, for one.
I'm sorry to see that you started on the right path and then joined the dark side
Oh, well. To each their own. I just thought I might help you get great sound, since you already built the DIY speakers, which is the hard part. But since you're not into it... Let's just agree to disagree.
Old 3 weeks ago
  #3467
Gear Addict
 
syncussion's Avatar
 

Quote:
Originally Posted by sax512 View Post
Hmmm. There is a lot to argue about what you just said.
Inaudibility of cross over is straight up wishful thinking, for one.
I'm sorry to see that you started on the right path and then joined the dark side
Oh, well. To each their own. I just thought I might help you get great sound, since you already built the DIY speakers, which is the hard part. But since you're not into it... Let's just agree to disagree.
You can look up many studies regarding audibility of group delay.
A fourth order Linkwitz-Riley crossover done right is not audible under even lab conditions with specially constructed test signals, let alone music.
You can test this yourself. Take some music, run it through an equalizer which has 2 12dB/oct low pass filters Q1 -3dB at for instance 1500Hz so the combined filters give -6dB at 1500Hz. Now process the same music in parallel with 2 highpass filters set to the same. Now recombine them and you'll have what is a Linkwitz-Riley 24dB/oct crossover giving a group delay of one period. You can now listen to this with your single full range driver speaker or headphones etc. And test the output to check if you've done it right. When done correctly it is inaudible, completely inaudible even with for instance square waves or impulse test signals etc.
Now you can also delay for instance the highpassed signal before recombining and simulate the vertical off-axis of the crossover this way. This will give you the "dark side" of the crossover. You can also simulate mid driver dropoff / beaming off-axis etc.
With the right design you can avoid these things. Close CTC spacing and low crossover freq (pick the right tweeter capable of this).
Old 3 weeks ago
  #3468
Gear Addict
 

Quote:
Originally Posted by syncussion View Post
You can look up many studies regarding audibility of group delay.
A fourth order Linkwitz-Riley crossover done right is not audible under even lab conditions with specially constructed test signals, let alone music.
You can test this yourself. Take some music, run it through an equalizer which has 2 12dB/oct low pass filters Q1 -3dB at for instance 1500Hz so the combined filters give -6dB at 1500Hz. Now process the same music in parallel with 2 highpass filters set to the same. Now recombine them and you'll have what is a Linkwitz-Riley 24dB/oct crossover giving a group delay of one period. You can now listen to this with your single full range driver speaker or headphones etc. And test the output to check if you've done it right. When done correctly it is inaudible, completely inaudible even with for instance square waves or impulse test signals etc.
Now you can also delay for instance the highpassed signal before recombining and simulate the vertical off-axis of the crossover this way. This will give you the "dark side" of the crossover. You can also simulate mid driver dropoff / beaming off-axis etc.
With the right design you can avoid these things. Close CTC spacing and low crossover freq (pick the right tweeter capable of this).
You're totally disregarding transient response and time alignment. I'll leave it at that for you to investigate this.. if you want, it's all in that book. $10 on kindle. It's a good investment, in my opinion.
Old 3 weeks ago
  #3469
Gear Addict
 
syncussion's Avatar
 

Quote:
Originally Posted by sax512 View Post
You're totally disregarding transient response and time alignment. I'll leave it at that for you to investigate this.. if you want, it's all in that book. $10 on kindle. It's a good investment, in my opinion.
Maybe you're not familiar with group delay? This is what a minimal phase crossover gives. This concerns transient response / perfect time allignment. As i said, it is inaudible.
Old 3 weeks ago
  #3470
Gear Addict
 

Quote:
Originally Posted by syncussion View Post
Maybe you're not familiar with group delay? This is what a minimal phase crossover gives. This concerns transient response / perfect time allignment. As i said, it is inaudible.
As I said, you're wrong about that.
I'm quite familiar with group delay, as opposed to phase delay, of which group delay is the first derivative.
I'm sorry, I'm talking about more advanced stuff than minimal phase.
It's 2020. Phase linearizing DSPs have been around for decades.
Let me know when you catch up
Old 3 weeks ago
  #3471
Gear Addict
 
syncussion's Avatar
 

Quote:
Originally Posted by sax512 View Post
As I said, you're wrong about that.
I'm quite familiar with group delay, as opposed to phase delay, of which group delay is the first derivative.
I'm sorry, I'm talking about more advanced stuff than minimal phase.
It's 2020. Phase linearizing DSPs have been around for decades.
Let me know when you catch up
You're really misinformed about this aspect.
Again, it has been proven many times that up till a certain point (which is not reached by fourth order filters) that the group delay by minimal phase crossovers is inaudible.
You can test this yourself as I described above.
You can also construct the same crossover simulation in linear phase. In linear phase this fourth order Linkwitz-Riley crossover will be 100% identical to the signal before the crossover. Bit perfect reconstruction. The minimal phase crossover is identical with the exception that it gives a group delay. As I said, it has been tested over and over and the conclusion is clear, this group delay is inaudible under any and all circumstances including transients etc.
The difference between the linear phase and minimal phase crossover lies in the off-axis. Here the linear phase crossover will do pre-ringing and the minimal phase crossover will have normal post-ringing. The linear phase crossover pre-ringing is actually slightly more audible because the ringing is masked less (look up auditory masking).
These are not things which I'm making up myself but well researched and published subjects. And again, follow my instructions in the post above and you can reproduce a simple test of this. To argue against this is like arguing against the fact the earth is round..
Crossover have gotten a bad name because of limitations of passive crossovers (which can only be made right if the resistance and inductance of drivers is flat and stable both of which they are not) and where the tweeter and mid are not time aligned and the crossover is happening too high with respect to CTC distance and mid driver beaming etc.
In 2020 these are not issues anymore with either active crossovers or better still DSP crossovers, combined with a good speaker design.
Old 3 weeks ago
  #3472
Gear Addict
 

Quote:
Originally Posted by syncussion View Post
You're really misinformed about this aspect.
Again, it has been proven many times that up till a certain point (which is not reached by fourth order filters) that the group delay by minimal phase crossovers is inaudible.
You can test this yourself as I described above.
You can also construct the same crossover simulation in linear phase. In linear phase this fourth order Linkwitz-Riley crossover will be 100% identical to the signal before the crossover. Bit perfect reconstruction. The minimal phase crossover is identical with the exception that it gives a group delay. As I said, it has been tested over and over and the conclusion is clear, this group delay is inaudible under any and all circumstances including transients etc.
The difference between the linear phase and minimal phase crossover lies in the off-axis. Here the linear phase crossover will do pre-ringing and the minimal phase crossover will have normal post-ringing. The linear phase crossover pre-ringing is actually slightly more audible because the ringing is masked less (look up auditory masking).
These are not things which I'm making up myself but well researched and published subjects. And again, follow my instructions in the post above and you can reproduce a simple test of this. To argue against this is like arguing against the fact the earth is round..
Crossover have gotten a bad name because of limitations of passive crossovers (which can only be made right if the resistance and inductance of drivers is flat and stable both of which they are not) and where the tweeter and mid are not time aligned and the crossover is happening too high with respect to CTC distance and mid driver beaming etc.
In 2020 these are not issues anymore with either active crossovers or better still DSP crossovers, combined with a good speaker design.
Look man. I just wanted to help you out.
Inaudibility of cross-over is a BS concept. To really test if cross overs are inaudible, you would have to test all types of cross overs, with all types of cabinets, in all types of rooms. Are YOU going to do it? Because nobody has ever done it, and your little test doesn't prove anything. Especially since you can't do it in real time and ABX.
Also, I'm talking about very near field listening, where the room reflected sound is a lower order error and doesn't mask cross over artifacts as it does in normal listening situations.
Cross over is definitely audible in that situation, and to be honest, it is in normal listening situations as well. Sorry to burst your bubble.

Also, just the fact that you mention pre-ringing, tells me that I made a mis-judgement when I thought you were knowingly onto something with your Ikea test experiment. You were, but you were also fumbling in the dark.
Go learn something about electronics then come back talk to me..
So much for showing interest in somebody's experiments
Old 3 weeks ago
  #3473
Gear Addict
 
syncussion's Avatar
 

Quote:
Originally Posted by sax512 View Post
Look man. I just wanted to help you out.
Inaudibility of cross-over is a BS concept. To really test if cross overs are inaudible, you would have to test all types of cross overs, with all types of cabinets, in all types of rooms. Are YOU going to do it? Because nobody has ever done it, and your little test doesn't prove anything. Especially since you can't do it in real time and ABX.
Also, I'm talking about very near field listening, where the room reflected sound is a lower order error and doesn't mask cross over artifacts as it does in normal listening situations.
Cross over is definitely audible in that situation, and to be honest, it is in normal listening situations as well. Sorry to burst your bubble.

Also, just the fact that you mention pre-ringing, tells me that I made a mis-judgement when I thought you were knowingly onto something with your Ikea test experiment. You were, but you were also fumbling in the dark.
Go learn something about electronics then come back talk to me..
So much for showing interest in somebody's experiments
There are no artifacts on-axis for a correctly done Linkwitz-Riley crossover. Again you can test this yourself with headphones or in an anechoic chamber if you wish. This is basic digital signal processing and you can find it in any textbook.
The artifacts happen off-axis and one can design these to not matter in the way I described above. And no, this is not masked by room reflections as yuo suggest as in an anechoic room the only axis that matters is the on-axis of course..
As for testing it real time and ABX yes I have done exactly that extensively with simulations.
As for testing the off-axis with actual speakers, one can see the effects in fr polar plots. One can find the measurements of this of many speakers.
Old 3 weeks ago
  #3474
Gear Addict
 

Quote:
Originally Posted by syncussion View Post
There are no artifacts on-axis for a correctly done Linkwitz-Riley crossover. Again you can test this yourself with headphones or in an anechoic chamber if you wish. This is basic digital signal processing and you can find it in any textbook.
It's not the digital signal processing part that I have a problem with. It's the assumption that it bears more resemblance to a real speaker's behavior than it actually does in reality.
By simulating a cross over and playing it back through single drivers or headphones, as you suggest, you take away a lot of what is physically causing the problem, which is the fact that the sound is not emanating from the same location in space, with 2+ way speakers. That's not something you can get past by with 'close enough' CTC placement. This stuff needs to be addressed with scientific methodology, not 'close-enoughness'.

Quote:
Originally Posted by syncussion View Post
The artifacts happen off-axis and one can design these to not matter in the way I described above. And no, this is not masked by room reflections as yuo suggest as in an anechoic room the only axis that matters is the on-axis of course..
The artifacts happen everywhere else other than the point at which you place the measurement microphone, with 2+ way design. It's simply physics. No escaping that, unfortunately.

Quote:
Originally Posted by syncussion View Post
As for testing it real time and ABX yes I have done exactly that extensively with simulations.
As for testing the off-axis with actual speakers, one can see the effects in fr polar plots. One can find the measurements of this of many speakers.
Simulations are NOT the real thing. To test the real thing, you would need instantaneous swap of speaker, room and crossover design. Obviously impossible, which is why people can claim that it's not audible. To test otherwise is impossible, if one wants to do that with a scientifically valid process.
It's a technique typical of snake-oil salesmen. Claim something whose opposite can't be easily demonstrated.

Lucky for us, it can be measured, and the artifacts of a REAL speaker playing crossed over parts of the spectrum from 2+ sources in space are way above the audibility threshold, when you move slightly off from the measurement mic location. Less than the distance between your ears, to be specific.
Old 3 weeks ago
  #3475
Gear Addict
 
syncussion's Avatar
 

Quote:
Originally Posted by sax512 View Post
It's not the digital signal processing part that I have a problem with. It's the assumption that it bears more resemblance to a real speaker's behavior than it actually does in reality.
By simulating a cross over and playing it back through single drivers or headphones, as you suggest, you take away a lot of what is physically causing the problem, which is the fact that the sound is not emanating from the same location in space, with 2+ way speakers. That's not something you can get past by with 'close enough' CTC placement. This stuff needs to be addressed with scientific methodology, not 'close-enoughness'.



The artifacts happen everywhere else other than the point at which you place the measurement microphone, with 2+ way design. It's simply physics. No escaping that, unfortunately.



Simulations are NOT the real thing. To test the real thing, you would need instantaneous swap of speaker, room and crossover design. Obviously impossible, which is why people can claim that it's not audible. To test otherwise is impossible, if one wants to do that with a scientifically valid process.
It's a technique typical of snake-oil salesmen. Claim something whose opposite can't be easily demonstrated.

Lucky for us, it can be measured, and the artifacts of a REAL speaker playing crossed over parts of the spectrum from 2+ sources in space are way above the audibility threshold, when you move slightly off from the measurement mic location. Less than the distance between your ears, to be specific.
Ah, so you're assuming our ear can differentiate for instance half wavelength CTC distance at for instance 1500Hz at for instance >1meter. Which is what I'm talking about. And vertically at that. It can't. You can look these things up, just wiki "directional hearing" as a start. Scientific methodology is exactly what you should use here and what I've been arguing these posts.
You focus on a problem which can be mitigated to the point that it doesn't matter anymore in the big picture while you seem to be blind to the elephant in the room - the major errors of single full range drivers that cannot be fixed. In the grand scheme of things you put the balance completely wrong. As I said, vertical off-axis can be a problem for a well implemented crossover (and often is) but with the right design (close CTC distance and low crossover point) this is fixed to the point that it doesn't matter anymore. And yes that is a scientific sound argument.
Old 3 weeks ago
  #3476
Gear Addict
 
syncussion's Avatar
 

Quote:
Originally Posted by sax512 View Post
Lucky for us, it can be measured, and the artifacts of a REAL speaker playing crossed over parts of the spectrum from 2+ sources in space are way above the audibility threshold, when you move slightly off from the measurement mic location. Less than the distance between your ears, to be specific.
And what you write here is simply not true. First of all if you move horizontally the distance between the ears with the measurement mic you will not get any difference in crossover other than off-axis mid dropoff depending on the mid drivers off-axis behaviour and the crossover freq. This is why I say cross low as you won't have mid driver beaming yet. Alternatively you can use a waveguide/horn to match the directivity of the tweeter to the directivity of the mid driver at the crossover freq.
As for vertically moving the mic, this relates to change of angle to the speaker and it depends on the crossover freq and CTC spacing how much this is affected.
For what I wrote, half wavelength of the crossover freq as CTC distance you get the cancellation axis 90 degrees and -3dB at 45degrees. It will be inaudible over at least +-30 degrees and the total room reflection effect will be inaudible as well in this region. That is taking the problem away in a realistic manner.
All the problems of a single driver are not going away and are far larger to start with than even a suboptimal crossover design.
Old 3 weeks ago
  #3477
Gear Addict
 

At 1.5 kHz the wavelength is about 22 cm. You need half of that distance to cause complete cancellation. That's 11 cm.
If you move horizontally, there is still a little difference as the distance changes, so it is as if you move back and forth. As a matter of fact, let's just consider the case of moving back and forth, as moving laterally would cause also a difference in amplitude sensed by the left and right ears.
Depending on where your sweet spot is, you're changing the vertex of a triangle whose base is the distance between the drivers (fixed), and the height is your head's distance from the speaker.
So if you move enough so that the difference between the two remaining sides of that triangle is 11 cm, you have gone from complete reinforcement to complete cancellation.
Now, it takes quite a movement back or forth for that difference to be 11 cm, I'll give you that. However...
You double the frequency and you're still smack in the middle of the most sensitive range of your hearing. Now it takes only 5 cm to go from reinforcement to cancellation.
True, if you have a higher order cross over, you're already 24 dB lower on the output from the woofer to that frequency, and the sound coming from the tweeter overcomes it. However..
24 dB is about 20 dB higher in power than a good single driver IMD, so if you can't hear the cross over, you definitely can't hear IMD (and in fact, you can't).
But more than that, not all speaker designs imply the use of perfectly time aligned drivers with 4th order cross over.
You're describing a very specific, rarely implemented design of 2 way speaker.
This design requires high number of taps FIR filters to make sure that no cancellation happens throughout the whole cross over range. It's not a simple 4th order cross over. It's a filter with a precisely controlled phase response that depends on the physical attributes of the speaker and the electro-mechanical attributes of the drivers. This means higher latency (not a problem in my book, but just to point it out).
I'm sure in that case the cross over is less audible than the typical 2 way design you can buy (even very expensive ones, by the way). But you still have to listen at a longer distance, which implies higher room reflections contributions to the total sound (or a considerably larger room to get that contribution back down).
Also, consider that a small change in height of your head location is pretty common. You just rest on your listening chair slightly differently and you have your 5 cm right there, without even realizing it.

On the other hand, with a full range, you have embedded time alignment, ability to listen at close range (room is not a prevalent factor anymore, and that's HUGE), and, personal opinion here, a way cooler look
It's a win win, in my book. Provided that you intend to use the speakers for individual listening purposes only, of course.
Old 3 weeks ago
  #3478
Gear Addict
 
syncussion's Avatar
 

Quote:
Originally Posted by sax512 View Post
At 1.5 kHz the wavelength is about 22 cm. You need half of that distance to cause complete cancellation. That's 11 cm.
If you move horizontally, there is still a little difference as the distance changes, so it is as if you move back and forth. As a matter of fact, let's just consider the case of moving back and forth, as moving laterally would cause also a difference in amplitude sensed by the left and right ears.
Depending on where your sweet spot is, you're changing the vertex of a triangle whose base is the distance between the drivers (fixed), and the height is your head's distance from the speaker.
So if you move enough so that the difference between the two remaining sides of that triangle is 11 cm, you have gone from complete reinforcement to complete cancellation.
Now, it takes quite a movement back or forth for that difference to be 11 cm, I'll give you that. However...
You double the frequency and you're still smack in the middle of the most sensitive range of your hearing. Now it takes only 5 cm to go from reinforcement to cancellation.
True, if you have a higher order cross over, you're already 24 dB lower on the output from the woofer to that frequency, and the sound coming from the tweeter overcomes it. However..
24 dB is about 20 dB higher in power than a good single driver IMD, so if you can't hear the cross over, you definitely can't hear IMD (and in fact, you can't).
But more than that, not all speaker designs imply the use of perfectly time aligned drivers with 4th order cross over.
You're describing a very specific, rarely implemented design of 2 way speaker.
This design requires high number of taps FIR filters to make sure that no cancellation happens throughout the whole cross over range. It's not a simple 4th order cross over. It's a filter with a precisely controlled phase response that depends on the physical attributes of the speaker and the electro-mechanical attributes of the drivers. This means higher latency (not a problem in my book, but just to point it out).
I'm sure in that case the cross over is less audible than the typical 2 way design you can buy (even very expensive ones, by the way). But you still have to listen at a longer distance, which implies higher room reflections contributions to the total sound (or a considerably larger room to get that contribution back down).
Also, consider that a small change in height of your head location is pretty common. You just rest on your listening chair slightly differently and you have your 5 cm right there, without even realizing it.

On the other hand, with a full range, you have embedded time alignment, ability to listen at close range (room is not a prevalent factor anymore, and that's HUGE), and, personal opinion here, a way cooler look
It's a win win, in my book. Provided that you intend to use the speakers for individual listening purposes only, of course.
Perhaps your time spent with single drivers has made you inexperienced with crossovers
There's no difference when moving back and forth. The difference is when moving down and up vertically that is when we move from the on-axis line to the vertical off-axis.
As I stated, in the above example the cancellation axis is at 90 degrees vertical relative to the on-axis. So in a proper speaker design this is parallel to the baffle. (btw I personally aim to have the cancellation axis at >90 degrees so there is no cancellation axis at all as it would be virtually "inside" the speaker). (also there isn't a true full notch at the cancellation axis like there would be if we inverted the phase of the mid or tweeter but this is not important to this discussion)
The -3dB vertical axis is at 45 degrees up and down. Yes, this means that at 3kHz it cancels at 45 degrees, but as you said the mid is 24dB down already so in reality we see a normal -3dB bell dip at 1500Hz which has a certain slope and almost no effect anymore at 3kHz. (in reality here is also where the mid driver off-axis dropoff / beaming plays a role, another reason to cross low as I said).
With a 4th order crossover this 3dB bell dip is already quite narrow so it's hard to detect it (really talking minute differences here in reality with such a small narrow dip). If the speaker is otherwise perfectly flat on-axis, the narrow -3dB dip at 45 degrees vertical off-axis makes the speaker +- 1.5dB. If you make the crossover steper it's already not really detectable with actual music at this angle. But with a 24dB I would say that under perfect conditions otherwise and in an anechoic room it would be undetectable at a 30 degree angle up or down (so a 60degree angle in total in which it is undetectable). In a non anechoic room this isn't even worth talking about anymore / totally undetectable and less important than other things which affect off-axis.

And no I'm not describing a 2-way. There are more 3-ways which have close CTC spacing as a 3-way allow the use of a smaller mid driver.
Also most studio speakers have active crossovers and many (other than some of the cheapest) have time alligned /corrected drivers (though most with too high crossover points and lacking tweeters in my opinion.)
Reason I'm not talking about woofer to mid driver crossover is because the wavelengths are long and almost always these will be within a quarter wavelength at crossover freq CTC spacing, it's hard to mess up a design there.

As for the merits of taking out the room in nearfield listening. Nearfield listening is still best in a good well treated room and I wouldn't recommend listening nearfield within 1m with any type of speaker. Also not with Ikea bowl single full range driver speakers.
I would definately recommend a truly good active 3-way at more than 1m listening distance in a well treated room with measurements and some final room correction in the bass with EQ. Proven method to work best and with good reason. Improvements to this are in the form of better drivers, better 3-way design (including huge roundover against edge diffraction), DSP crossovers, Hypex NCore or Purifi class-D amps and massive room treatment. That's how you get the best audio quality currently possible. Not by going to full range drivers etc.
Old 3 weeks ago
  #3479
Gear Addict
 

Quote:
Originally Posted by syncussion View Post
Perhaps your time spent with single drivers has made you inexperienced with crossovers
There's no difference when moving back and forth. The difference is when moving down and up vertically that is when we move from the on-axis line to the vertical off-axis.
Nah, it's basic trigonometry. You move straight forward and you get closer to both tweeter and woofer, but not by the same amount, depending on where you start (which usually they suggest being in the tweeter on-axis).
The only case where what you say is true is when the the triangle is isosceles, which means you started from the axis that cuts right in between tweeter and woofer. And even then, same distance difference doesn't mean anything, as for different frequencies that adds up to different phase rotation due to the different acoustic origin of the drivers for different frequencies in the cross over range. That's one of the main problems with cross overs.
You can't be that imprecise about this stuff.

Quote:
Originally Posted by syncussion View Post
As I stated, in the above example the cancellation axis is at 90 degrees vertical relative to the on-axis. So in a proper speaker design this is parallel to the baffle. (btw I personally aim to have the cancellation axis at >90 degrees so there is no cancellation axis at all as it would be virtually "inside" the speaker). (also there isn't a true full notch at the cancellation axis like there would be if we inverted the phase of the mid or tweeter but this is not important to this discussion)
The -3dB vertical axis is at 45 degrees up and down. Yes, this means that at 3kHz it cancels at 45 degrees, but as you said the mid is 24dB down already so in reality we see a normal -3dB bell dip at 1500Hz which has a certain slope and almost no effect anymore at 3kHz. (in reality here is also where the mid driver off-axis dropoff / beaming plays a role, another reason to cross low as I said).
With a 4th order crossover this 3dB bell dip is already quite narrow so it's hard to detect it (really talking minute differences here in reality with such a small narrow dip). If the speaker is otherwise perfectly flat on-axis, the narrow -3dB dip at 45 degrees vertical off-axis makes the speaker +- 1.5dB. If you make the crossover steper it's already not really detectable with actual music at this angle. But with a 24dB I would say that under perfect conditions otherwise and in an anechoic room it would be undetectable at a 30 degree angle up or down (so a 60degree angle in total in which it is undetectable). In a non anechoic room this isn't even worth talking about anymore / totally undetectable and less important than other things which affect off-axis.
What you're saying here has some degree of truth here and there, but you're really too approximate in your analysis of what goes on with cross overs. No speaker measures even remotely close to what you describe. For one thing, cancellation is not relative to axes but points in space. Again, this derives from an excessively crude model of the system.

Quote:
Originally Posted by syncussion View Post
And no I'm not describing a 2-way. There are more 3-ways which have close CTC spacing as a 3-way allow the use of a smaller mid driver.
Also most studio speakers have active crossovers and many (other than some of the cheapest) have time alligned /corrected drivers (though most with too high crossover points and lacking tweeters in my opinion.)
Reason I'm not talking about woofer to mid driver crossover is because the wavelengths are long and almost always these will be within a quarter wavelength at crossover freq CTC spacing, it's hard to mess up a design there.
Please name specific monitors with such characteristics and relative measurements. The only thing that comes remotely close to what you're saying, to my knowledge, are the Kii Three.

Quote:
Originally Posted by syncussion View Post
As for the merits of taking out the room in nearfield listening. Nearfield listening is still best in a good well treated room
No argument here.

Quote:
Originally Posted by syncussion View Post
and I wouldn't recommend listening nearfield within 1m with any type of speaker.
It depends on the design and driver used. It can be done. I have done it. Anybody can do it. You were extremely close to get very good sound from your Ikea speakers.

Quote:
Originally Posted by syncussion View Post
Also not with Ikea bowl single full range driver speakers.
Close, but it can be done better. Admittedly, Ikea bowls are not the best speaker cabinets one can make. You're definitely right about that.

Quote:
Originally Posted by syncussion View Post
I would definately recommend a truly good active 3-way at more than 1m listening distance in a well treated room with measurements and some final room correction in the bass with EQ. Proven method to work best and with good reason. Improvements to this are in the form of better drivers, better 3-way design (including huge roundover against edge diffraction), DSP crossovers, Hypex NCore or Purifi class-D amps and massive room treatment. That's how you get the best audio quality currently possible. Not by going to full range drivers etc.
Best is subjective. Single drivers will give you the most accurate transducer from electric signal to pressure wave. It's not even debatable. But if you don't like accuracy, you should definitely look into something else, which you have done. I just think if you took the time to investigate how to EQ your Ikea speakers the right way, you might be at a very different place, right now.
But to each their own..
Old 1 week ago
  #3480
Gear Maniac
Hey Audiovision
Have u heard the ocean way HR4 orPro2a?
Topic:
Post Reply

Welcome to the Gearslutz Pro Audio Community!

Registration benefits include:
  • The ability to reply to and create new discussions
  • Access to members-only giveaways & competitions
  • Interact with VIP industry experts in our guest Q&As
  • Access to members-only sub forum discussions
  • Access to members-only Chat Room
  • Get INSTANT ACCESS to the world's best private pro audio Classifieds for only USD $20/year
  • Promote your eBay auctions and Reverb.com listings for free
  • Remove this message!
You need an account to post a reply. Create a username and password below and an account will be created and your post entered.


 
 
Slide to join now Processing…
Thread Tools
Search this Thread
Search this Thread:

Advanced Search
Forum Jump
Forum Jump