Hi,
A general question; If If more than one set of speakers are used in any setup, how much of a mess would phase mismatches cause.
To expand: Pop a second set of speakers on top of your first: (e.g. oddly with a set of Quad 22L I can add some improvements with a speaker responding to certain bands).
Many of you would probably think this a terrible thing to do: Firstly many 3 ways using second order crossovers would put the mid out of phase with woofer/tweeter, so an additional midrange would need to be out of phase with the main speaker to be in-phase with the main speakers direction of travel at that frequency.
Even if this were overcome, the phase shift as a crossover nears cut off frequency would surely interfere with the response of the other speakers, so unless the same circuit were used in each (i.e. another set of the same speakers), then a complicated combination of wave forms would be heard by the listener.
Whilst playing with additional speakers would clearly be detrimental to accuracy, it may add to reality of sound stage and listening enjoyment, which I personally find does not go hand in hand with speaker accuracy (e.g. a set of reference studio montiors are very accurate, and great to get a mix right, but pretty dull to sit and enjoy music on!)
It strikes me that by using rear speakers in cinema type setups, or by using centres, each with their own crossover circuitry, people have complicated phase relationships all the time. I've not noticed this mentioned to any extent (please point me the correct direction if I've missed it), but would others say the modification to the accurate reproduction of sound by more than one set of crossovers/speakers is always a bad thing?
...and do rear speakers (I don't use them) not cause terrible phase problems both because of crossover characteristics AND because at certain frequencies the sound they produce will be out of phase with the front speakers?
...and doesn't the room cause all sorts of phase issues anyhow. Pretending an accurate reproduction of sound from speakers results in the listener hearing what the mixer intended is surely optimistic?
I'm almost sceptical that some reality /depth might be added by such anamolies resulting from phase shifts near crossover, by mismatch between speakers, or by subtle differences in the distance bewteen the various drivers and ears. In short I can't explain why a £900 set of speakers can be improved with certain frequency/phase combinations of additional drivers from old wharfdale/leak units at certain bandwidths.
Thanks for your thoughts. I'm a kind of noob/experienced person: fairly familiar with the basic physics, electronics & music listening, but new at speaker building. As a result, I may have asked something with some very complicated answers, or something very stupid, in which case I also look forward to hearing why!
Thanks, Robin
A general question; If If more than one set of speakers are used in any setup, how much of a mess would phase mismatches cause.
To expand: Pop a second set of speakers on top of your first: (e.g. oddly with a set of Quad 22L I can add some improvements with a speaker responding to certain bands).
Many of you would probably think this a terrible thing to do: Firstly many 3 ways using second order crossovers would put the mid out of phase with woofer/tweeter, so an additional midrange would need to be out of phase with the main speaker to be in-phase with the main speakers direction of travel at that frequency.
Even if this were overcome, the phase shift as a crossover nears cut off frequency would surely interfere with the response of the other speakers, so unless the same circuit were used in each (i.e. another set of the same speakers), then a complicated combination of wave forms would be heard by the listener.
Whilst playing with additional speakers would clearly be detrimental to accuracy, it may add to reality of sound stage and listening enjoyment, which I personally find does not go hand in hand with speaker accuracy (e.g. a set of reference studio montiors are very accurate, and great to get a mix right, but pretty dull to sit and enjoy music on!)
It strikes me that by using rear speakers in cinema type setups, or by using centres, each with their own crossover circuitry, people have complicated phase relationships all the time. I've not noticed this mentioned to any extent (please point me the correct direction if I've missed it), but would others say the modification to the accurate reproduction of sound by more than one set of crossovers/speakers is always a bad thing?
...and do rear speakers (I don't use them) not cause terrible phase problems both because of crossover characteristics AND because at certain frequencies the sound they produce will be out of phase with the front speakers?
...and doesn't the room cause all sorts of phase issues anyhow. Pretending an accurate reproduction of sound from speakers results in the listener hearing what the mixer intended is surely optimistic?
I'm almost sceptical that some reality /depth might be added by such anamolies resulting from phase shifts near crossover, by mismatch between speakers, or by subtle differences in the distance bewteen the various drivers and ears. In short I can't explain why a £900 set of speakers can be improved with certain frequency/phase combinations of additional drivers from old wharfdale/leak units at certain bandwidths.
Thanks for your thoughts. I'm a kind of noob/experienced person: fairly familiar with the basic physics, electronics & music listening, but new at speaker building. As a result, I may have asked something with some very complicated answers, or something very stupid, in which case I also look forward to hearing why!
Thanks, Robin
Hi,
Rear speakers should incorporate a delay such that they do not
contribute to the direct sound heard from the main pair, but they
do help in recreating the acoustics of the original venue, given
appropriate multi-channel source material, there is the problem....
🙂/sreten.
Rear speakers should incorporate a delay such that they do not
contribute to the direct sound heard from the main pair, but they
do help in recreating the acoustics of the original venue, given
appropriate multi-channel source material, there is the problem....
🙂/sreten.
Thanks for reply.
Hmm, interesting. Delay makes sense, but raises even more questions (sorry!): eg I guess if front speakers are 4m further away than rear speakers, then with sound travelling at 344m/s, a delay of 12ms is about right.
But. Delay or no delay, wouldn't the sound meeting at the end user be out of phase at higher frequencies where tiny distances would show up errors in delay?
Put another way:
what if we imagine the rear speakers were sat on top of the main speakers (more similar to the setup I'm playing with), but for arguments sake these were 10cm further back than the main speakers. A wave with a wavelength of 20cm (1.715KHz) would be completely out of phase at the point it passes the main speaker, causing a dip in frequency response between the speakers? Strengely, like misalignment between two front speakers, the expected dip isn't very obvious (presumably as it has a narrow Q).
Another way this cancelling happens using only two speakers would be if there is a wall behind a listener, there will always be cancelling or enforcing (depending on wavelength) of some sound waves from the rear wall?
I have a feeling there's a light not turned on somewhere, and I've missed something fundamental, but until then, back on the original question, i'm still very interested in comments regarding the issues of wave phase.
Thanks, Robin
Hmm, interesting. Delay makes sense, but raises even more questions (sorry!): eg I guess if front speakers are 4m further away than rear speakers, then with sound travelling at 344m/s, a delay of 12ms is about right.
But. Delay or no delay, wouldn't the sound meeting at the end user be out of phase at higher frequencies where tiny distances would show up errors in delay?
Put another way:
what if we imagine the rear speakers were sat on top of the main speakers (more similar to the setup I'm playing with), but for arguments sake these were 10cm further back than the main speakers. A wave with a wavelength of 20cm (1.715KHz) would be completely out of phase at the point it passes the main speaker, causing a dip in frequency response between the speakers? Strengely, like misalignment between two front speakers, the expected dip isn't very obvious (presumably as it has a narrow Q).
Another way this cancelling happens using only two speakers would be if there is a wall behind a listener, there will always be cancelling or enforcing (depending on wavelength) of some sound waves from the rear wall?
I have a feeling there's a light not turned on somewhere, and I've missed something fundamental, but until then, back on the original question, i'm still very interested in comments regarding the issues of wave phase.
Thanks, Robin
About two months ago, I´ve done some work on my frontloaded horns. In its response, there are similar dips caused by interference of direct and reflected wave from the mouth of the horn. Same as dips caused by wave interferences in hearing room. They are very narrow and deep. But when you set up some subjective listening tests you´ll reveal that they are almost inaudible. One possible reason is its narrow shape, and the second is that this dips are very common as linkwitz said.
Same dips are also very common in every room, and the dips are caused by reflections of floor and roof. As linkwitz says in his issues in loudspeaker design, we get used to it, through the years.
Same dips are also very common in every room, and the dips are caused by reflections of floor and roof. As linkwitz says in his issues in loudspeaker design, we get used to it, through the years.
Further to this, I should be honest about what I'm doing. I made an accidental discovery that sitting another speaker on top of my Quad 22ls (oddly enough an old Leak sandwich 200) made certain improvements (and equally had negative effects) on the sound. With a few frequency recordings, much of the improved 'reality' of the vocals came from a speaker wired to a band pass filter (narrow midrange) which amongst other effects, decreased some frequencies (it was out of phase with the Quads).
Now, the Quads are interesting in that they're a sort of phase linear speaker: in that bass and mid-bass are both second order low pass filters (the base having much larger inductor, and resistor in series with the capacitor). The tweeter is a third order high pass, so all are wired in phase.
Imitating the same change in cumulative frequency response from adding such a speaker with a graphic equaliser does not have the same effect as actually adding the speaker. I suppose it is a lot more complicated an effect than layers of parametric equalisation can imitate. I'm inquisitive as to why now. I wonder if it is to do with hearing: I know from my limited knowledge that ears, to some extent, perform Fourier analysis in their processing of sound, and that the process of sensing sound is incredibly complicated. I just wonder whether inducing some out of phase signals on a sound allows perception of some strange three-dimensional element. I'm aware that a stereo image can give the impression of sound stage, and tiny subtleties in element of hifi can effect this soundstage. I just wonder if some inaccuracies resulting from driver, crossover or even speaker positioning, room reflections etc are partly responsible for that sound stage, and that a 'real' and absorbing sound is not necessarily the result of a highly accurate transformation of a recorded waveform into air movement?
Coffee,
Thanks - I appreciate that response: it made me realise those who have crossover theories/circuitry named after them have probably written some valuable papers and I should hunt them out. Very much in line with what I was thinking. I'm glad i'm not too off the wall (no pun intended) in thinking these phase issues exist in narrow bands that occuer everyday don't have a great impression on the sound.
The fact all this hasn't been discussed greatly anywhere makes me very suspicious that despite friends agreeing on an improvement in sound, it is probably just that I have a unique living room with such silly acoustics that such circuitry improves the sound. If on the other hand, anyone has any pointers to resources which mention of out of phase speakers improving sound I'd be very grateful.
And thanks for reading such a rant if you're really down here :s
Now, the Quads are interesting in that they're a sort of phase linear speaker: in that bass and mid-bass are both second order low pass filters (the base having much larger inductor, and resistor in series with the capacitor). The tweeter is a third order high pass, so all are wired in phase.
Imitating the same change in cumulative frequency response from adding such a speaker with a graphic equaliser does not have the same effect as actually adding the speaker. I suppose it is a lot more complicated an effect than layers of parametric equalisation can imitate. I'm inquisitive as to why now. I wonder if it is to do with hearing: I know from my limited knowledge that ears, to some extent, perform Fourier analysis in their processing of sound, and that the process of sensing sound is incredibly complicated. I just wonder whether inducing some out of phase signals on a sound allows perception of some strange three-dimensional element. I'm aware that a stereo image can give the impression of sound stage, and tiny subtleties in element of hifi can effect this soundstage. I just wonder if some inaccuracies resulting from driver, crossover or even speaker positioning, room reflections etc are partly responsible for that sound stage, and that a 'real' and absorbing sound is not necessarily the result of a highly accurate transformation of a recorded waveform into air movement?
Coffee,
Thanks - I appreciate that response: it made me realise those who have crossover theories/circuitry named after them have probably written some valuable papers and I should hunt them out. Very much in line with what I was thinking. I'm glad i'm not too off the wall (no pun intended) in thinking these phase issues exist in narrow bands that occuer everyday don't have a great impression on the sound.
The fact all this hasn't been discussed greatly anywhere makes me very suspicious that despite friends agreeing on an improvement in sound, it is probably just that I have a unique living room with such silly acoustics that such circuitry improves the sound. If on the other hand, anyone has any pointers to resources which mention of out of phase speakers improving sound I'd be very grateful.
And thanks for reading such a rant if you're really down here :s
charliemouse said:
Hi,
High frequencies are not sine waves, they are musical events that
start and stop. It is this the ear/brain is attuned to, there is no
confusion caused by the random cancellation of sound waves at
the ear, the ear/brain ignores these for what is coherent in the
signal.
The ear/brain auto correlates (similar to radar) and associates
the delayed sound with the initial direct sound to give the
impression of the acoustics / effects in the recording.
What is poor is if the ear/brain cannot seperate the initial
and reflected/ effected information, which happens if rear
delay is too low.
🙂/sreten.
Sreten,
Thanks; makes sense high requencies not being sine waves. I was thinking in those terms as the most simple form of high frequency sound to conceptualise is a longitudinal sine wave. I suppose that is as far as such a conept should go, because as you say a naturally occuring high frequency sound such as from a musical instrument is an 'event' (complicated wave form). That could be fabricated from multiple combined sine waves (fourier), but it is presented as one complex waveform.
Interesting how the brain/ear puts things together that are received out of phase.
Generally,
I should add that I'm not interfering with the sound of the quads as much it might be imagined: filters for additional speakers are of a narrow band and those speakers are being driven quietly. Also, as this goes on at high frequencies, phase differences I guess have a more comlicated effect? e.g. at low frequencies, woofers out of phase kill the sound as they cancel each other, but at high frequencies, the effect of wiring tweeters out of phase is much more complicated and a bit less devastating on the overall sound.
I'll keep playing and not ask too many questions - I presume interactions of the physics / biology of sound production and perception is rather too comlicated to be of practial use in speaker building :s
Very useful forum by the way, I've learned a lot in other areas from reading through threads on here. Thanks all.
Thanks; makes sense high requencies not being sine waves. I was thinking in those terms as the most simple form of high frequency sound to conceptualise is a longitudinal sine wave. I suppose that is as far as such a conept should go, because as you say a naturally occuring high frequency sound such as from a musical instrument is an 'event' (complicated wave form). That could be fabricated from multiple combined sine waves (fourier), but it is presented as one complex waveform.
Interesting how the brain/ear puts things together that are received out of phase.
Generally,
I should add that I'm not interfering with the sound of the quads as much it might be imagined: filters for additional speakers are of a narrow band and those speakers are being driven quietly. Also, as this goes on at high frequencies, phase differences I guess have a more comlicated effect? e.g. at low frequencies, woofers out of phase kill the sound as they cancel each other, but at high frequencies, the effect of wiring tweeters out of phase is much more complicated and a bit less devastating on the overall sound.
I'll keep playing and not ask too many questions - I presume interactions of the physics / biology of sound production and perception is rather too comlicated to be of practial use in speaker building :s
Very useful forum by the way, I've learned a lot in other areas from reading through threads on here. Thanks all.
🙂 I've resolved my confusion about the positive effect of an out of phase speaker with my Quad 22Ls.
The excellent 'speaker designer' spreadsheet by Jeff Bagby allowed me to draw the crossover from Quads, and the crossover used for my additional speaker(s).
It was simply a strange series of coincidences that resulted in this out of phase speaker improving the sound:
I could be wrong on any of these points:
1: The quad is phase linear apparently (though i'm not entirely sure about that). I do know it is not really a 3 way, the two lower units are pretty much in phase with each other, using parallel but different low pass filters. Tweeter wired in phase with woofers, but at crossover it is 180 out of phase with woofer. I'm not entirely sure how this is termed 'phase linear', but still; That's irrelevant to the explanation.
2: At Crossover there is theoretical dip in the response of the speakers. About 2K above cross over there is a peak. The dip isn't very clear on my measurements, but the peak is.
3: The out of phase speaker added has a good response at the Quad's crossover, and 'speaker designer' says the quad is 180 out of phase with source at the point. This is therefore *in phease* with the extra speaker.
4: By crossover + 2K, the Quad exhibits a bit of a peak in response. the output of the little extra speaker is at its strongest at this point, but is now 130 out of phase with Quad, so acts to cancel some of the output peak of the quad.
Could be load of cobblers, but if it isn't it might be a pointer for someone equally new to speaker design and severly confused as to why a driver seems to do the opposite of what was thought. I needed to consider not just its polarity with respect to other speakers wiring, but it's phase reponse with respect to other speakers phase at the frequencies at which it operates.
In conclusion, my rambling questions earlier in the thread regarding adding out of phase speakers were, as I'd suspected, ungrounded and irrelevant to the experimentation I was doing. useful though: some things I now won't overlook when I start trying to design speakers 😱
The excellent 'speaker designer' spreadsheet by Jeff Bagby allowed me to draw the crossover from Quads, and the crossover used for my additional speaker(s).
It was simply a strange series of coincidences that resulted in this out of phase speaker improving the sound:
I could be wrong on any of these points:
1: The quad is phase linear apparently (though i'm not entirely sure about that). I do know it is not really a 3 way, the two lower units are pretty much in phase with each other, using parallel but different low pass filters. Tweeter wired in phase with woofers, but at crossover it is 180 out of phase with woofer. I'm not entirely sure how this is termed 'phase linear', but still; That's irrelevant to the explanation.
2: At Crossover there is theoretical dip in the response of the speakers. About 2K above cross over there is a peak. The dip isn't very clear on my measurements, but the peak is.
3: The out of phase speaker added has a good response at the Quad's crossover, and 'speaker designer' says the quad is 180 out of phase with source at the point. This is therefore *in phease* with the extra speaker.
4: By crossover + 2K, the Quad exhibits a bit of a peak in response. the output of the little extra speaker is at its strongest at this point, but is now 130 out of phase with Quad, so acts to cancel some of the output peak of the quad.
Could be load of cobblers, but if it isn't it might be a pointer for someone equally new to speaker design and severly confused as to why a driver seems to do the opposite of what was thought. I needed to consider not just its polarity with respect to other speakers wiring, but it's phase reponse with respect to other speakers phase at the frequencies at which it operates.
In conclusion, my rambling questions earlier in the thread regarding adding out of phase speakers were, as I'd suspected, ungrounded and irrelevant to the experimentation I was doing. useful though: some things I now won't overlook when I start trying to design speakers 😱
- Status
- Not open for further replies.