Thanks for the reply. Could you name some of these aspects ?
I am here to learn sincerely.
Speakers are my present main concern. I think that the speakers choice is the most important one because it sets the limit of the performance obtainable from a chain.
If i am not wrong in Duntech they regard highly the SW response of a speaker.
I guess it is one of their most important tests.
I understand also that other manufactures do not think it being so important, and often their speakers are not able to reproduce a decent SW.
The fact that a SW does not exist in nature is not important to me.
Also a loudspeaker does not exist in nature and still we use them to reproduce sound, maybe wrongly ... we should find a natural loudspeaker maybe.
For instance speakers are tested in anechoic chambers, an environment so different from the real listening environment.
And still manufacturers perform this test almost always i think.
What is important to me is if this test is telling something important or not
Honestly speaking of transducers (i.e. woofers, tweeters, microphones, cartdriges) i would prefer one able to give back a decent SW.
The same for crossovers.
But i stop here. I understand that SW is an annoying issue.
Thanks and regards,
gino
Last edited:
JCX,
Read through the first section. This is completely against what I have read before in many places. It could be new more correct information and as you have suggested, we have been reprinting incorrect information for years. Not being a member if IEEE, I of course can't read the second abstract.
So, this suggests we can hear phase relationships below 3K and can use these for localization rather than the previously accepted time difference and the even more significant frequency contour provided by our outer ear. I did not quite get that from the chapter, but did get that the hair bundles will fire locked to the phase. It only hints that we process this for localization. Other chapters may demonstrate this.
I go back to a test my Father told me about when he was at Wright Pat. In an anechoic chamber, they could move a pure tone in the plane from in front of you, over your head, to behind you, and you were not able to tell the sound had moved. But if it was noise, it was easily tracked. They then proved that even moving it right to left, you could not track the pure sound. I do not know what frequency was used. This test suggests we are most sensitive to the equalization from the pinna than phase or time.
Poole suggests that out localization is reinforced primary by environment near reflections. Excessive close reflections will "muddy" our ability to localize. Thinking about this, could this be swamping our ability to use time or phase? Don't know, but I know what happens when you deal with cabinet edge diffraction correctly and then put a set of 20 sq ft, 3 inch thick compressed FG absorbers behind the speakers reducing the short reflections. Almost magic , and it does not matter how well the speaker is time or phase coherent. (none of mine are, they are aligned for smooth crossover transition in the frequency domain)
So, I will consider this answer still under dispute as opposed to settled. I would need a lot more evidence to override what I have demonstrated in my movie room. I see not reason to believe one text book is correct and every one I have read before incorrect. Maybe so, maybe not. I would like to see a controlled test using in-ear sources, in a chamber so as to remove all possibility of environment giving us the clues. Vary phase only and see what happens. Pure tones, mixed tones and broadband. It would be interesting how "clicks" were processed. I suspect this has been done. Anyone have any reference?
Read through the first section. This is completely against what I have read before in many places. It could be new more correct information and as you have suggested, we have been reprinting incorrect information for years. Not being a member if IEEE, I of course can't read the second abstract.
So, this suggests we can hear phase relationships below 3K and can use these for localization rather than the previously accepted time difference and the even more significant frequency contour provided by our outer ear. I did not quite get that from the chapter, but did get that the hair bundles will fire locked to the phase. It only hints that we process this for localization. Other chapters may demonstrate this.
I go back to a test my Father told me about when he was at Wright Pat. In an anechoic chamber, they could move a pure tone in the plane from in front of you, over your head, to behind you, and you were not able to tell the sound had moved. But if it was noise, it was easily tracked. They then proved that even moving it right to left, you could not track the pure sound. I do not know what frequency was used. This test suggests we are most sensitive to the equalization from the pinna than phase or time.
Poole suggests that out localization is reinforced primary by environment near reflections. Excessive close reflections will "muddy" our ability to localize. Thinking about this, could this be swamping our ability to use time or phase? Don't know, but I know what happens when you deal with cabinet edge diffraction correctly and then put a set of 20 sq ft, 3 inch thick compressed FG absorbers behind the speakers reducing the short reflections. Almost magic , and it does not matter how well the speaker is time or phase coherent. (none of mine are, they are aligned for smooth crossover transition in the frequency domain)
So, I will consider this answer still under dispute as opposed to settled. I would need a lot more evidence to override what I have demonstrated in my movie room. I see not reason to believe one text book is correct and every one I have read before incorrect. Maybe so, maybe not. I would like to see a controlled test using in-ear sources, in a chamber so as to remove all possibility of environment giving us the clues. Vary phase only and see what happens. Pure tones, mixed tones and broadband. It would be interesting how "clicks" were processed. I suspect this has been done. Anyone have any reference?
you could be the 2nd person to download and listen then: http://www.diyaudio.com/forums/everything-else/54596-audibility-absolute-phase-5.html#post3841138
I got 10/10 with foobar2000, ABX plugin and Senn HD600 headphones: http://www.diyaudio.com/forums/everything-else/54596-audibility-absolute-phase-6.html#post3868300
my recent posts (back to march this year) on head-fi that collect my reading, explainations: created by jcx
that absolute polarity/phase sensitivity isn't very important with music and loudspeakers in home listening is likely - but it is a basic capability of human hearing for some signals, lower frequency range
I got 10/10 with foobar2000, ABX plugin and Senn HD600 headphones: http://www.diyaudio.com/forums/everything-else/54596-audibility-absolute-phase-6.html#post3868300
my recent posts (back to march this year) on head-fi that collect my reading, explainations: created by jcx
that absolute polarity/phase sensitivity isn't very important with music and loudspeakers in home listening is likely - but it is a basic capability of human hearing for some signals, lower frequency range
Last edited:
Of COURSE a two tone harmonic will change in timber as their relative phase changes. But this is pure amplitude addition.
Now generate a stereo signal with a very pure single frequency tone. With cans, sitting on a ladder out in the middle of a field to reduce any environment bleed through, the sound will be in the middle of your head. Then have one side shift. Does either location or timber change? That would tell us if we process phase.
Now generate a stereo signal with a very pure single frequency tone. With cans, sitting on a ladder out in the middle of a field to reduce any environment bleed through, the sound will be in the middle of your head. Then have one side shift. Does either location or timber change? That would tell us if we process phase.
a continuous sinewave tone may not have enough "features", but an infamous result is that we can tell the difference in arrival times between our two ears of some transient signals down to a few microseconds - 10 us is very robust, reports in the literature include lower # - certainly 30 us is a agreed textbook number for ITD descrimination
https://www.google.com/#q=ITD+microsecond
inter channel time delay is used in mixing, although the literature suggests the competing effects are very complex, for the simple audio mastering view read page 3: Moulton Laboratories :: Principles of Multitrack Mixing: The Phantom Image
https://www.google.com/#q=ITD+microsecond
inter channel time delay is used in mixing, although the literature suggests the competing effects are very complex, for the simple audio mastering view read page 3: Moulton Laboratories :: Principles of Multitrack Mixing: The Phantom Image
Last edited:
A very simple experiment to check the sensitivity to arrival times can be done by sending the same signal to the L and R speakers and then delay one of the two channel.
I have been told that up to few us the signals are perceived as identical in time and level
Then increasing the delay the delayed channel sounds lower in level
Increasing the delay even more you get an echo effect.
I do not remember the exact point at which the differences start by the way.
I think that above 10us some sort of echo can be perceived.
But going back to the topic i have another basic question, that is always the same
Let's say that i have a group of woofers
All of them give back in certain range a good sine wave
If we test them with a SW some are better than others
on this basis i would say that the SW test is more discriminating compared to the sine wave test
In general i would prefer a driver able to pass both signals with accuracy, and i stress accuracy
I like accuracy in sound reproduction
Thanks to all for the very interesting advice and explanations
Kind regards, gino
I have been told that up to few us the signals are perceived as identical in time and level
Then increasing the delay the delayed channel sounds lower in level
Increasing the delay even more you get an echo effect.
I do not remember the exact point at which the differences start by the way.
I think that above 10us some sort of echo can be perceived.
But going back to the topic i have another basic question, that is always the same
Let's say that i have a group of woofers
All of them give back in certain range a good sine wave
If we test them with a SW some are better than others
on this basis i would say that the SW test is more discriminating compared to the sine wave test
In general i would prefer a driver able to pass both signals with accuracy, and i stress accuracy
I like accuracy in sound reproduction
Thanks to all for the very interesting advice and explanations
Kind regards, gino
Last edited:
Square wave woofer test? Better read up on signal theory my friend. The results of a test may be different, but what is to say better? This has all to do with frequency response.
I believe where this thread is really talking is about mid to tweeter integration, if time/phase delays outside simple frequency addition is relevant' and if a square wave is a valid way to look at it. Personally, I don't think so.
I believe where this thread is really talking is about mid to tweeter integration, if time/phase delays outside simple frequency addition is relevant' and if a square wave is a valid way to look at it. Personally, I don't think so.
jcx,
If a pure tone does not have enough features, that that is almost proof it is not simple phase differences that give us the clues. As soon as you switch to noise, then the frequency additions and the mechanical filtering of the pinna and environment provide out location clues. Yes, we do have some discrimination on time with pulses. I did not think that was in question, only phase.
If a pure tone does not have enough features, that that is almost proof it is not simple phase differences that give us the clues. As soon as you switch to noise, then the frequency additions and the mechanical filtering of the pinna and environment provide out location clues. Yes, we do have some discrimination on time with pulses. I did not think that was in question, only phase.
Woofers are notoriously bad for producing higher harmonics, in fact much of the time the audible "bass" is mainly distortion - it's quite easy to distinguish the quality of the driver by feeding in a signal of a very low frequency sine wave, where the amplitude is ramped up from very low levels to a maximum, and then back down again. Set it on repeat, and listen as the harmonics start to kick in badly, and fade out, at certain amplitude levels.
To make it easier, if necessary, I mix in a sine wave at a frequency a few Hz different from the suspected harmonic at roughly the right level; hearing the slightly different tones beat - warble in amplitude - confirms what I've got ...
Last edited:
Thanks for the explanation
If i understand well better a distortion test for drivers selection.
Maybe we can end saying that the SW is impossible to reproduce for a driver. Anyway this is not so important because in nature there is no such thing as a square wave
And this is more than enough for me.
Now my interest is moving to crossover design.
I have seen schematics from well regarded brands very different from a textbook design.
I guess as someone has said already the xover can make or break the sound more than the drivers themselves.
Thanks again and best regards, gino
Hello and agreed.
Still if i had a nice set of instrumental tests for drivers selection and xover testing i would appreciate it.
I like instrumental tests mich more than listening tests.
Moreover i cannot believe that some fine tuning is done by ear ... i just cannot believe this as instead someone keep on saying (i think it is a marketing issue).
I think that the best designers have developed original test protocols to test their prototypes.
Of course they do not disclose this info can understand.
Regards, gino
sorry but this is wrong - I thought earlier you were just referring to 180 degree front/back ambiguity
small phase differences in pure, level matched sines are heard with pure tones in headphones
I got 10/10 1st try, no "training" in foobar2000 abx plugin for these +/- 15 degree vs 0/0 400 Hz sines in R/L channels in HD600 SPL < 80 dB (69 dB in fact with my RS sound meter)
of course in headphones you don't get a clear "bearing angle" sensation - more of a position along the "line between your ears" that stereo feels like to me in headphones
the +/-15 degrees @400 Hz handily works out to +/- 104 us delay
Last edited:
Hi and sorry but i have thought a lot during the weekend about your advice.
I do not fully agree, differently from what i have said previously.
I believe that superior drivers have indeed the "potential" for superior sound.
Then it is up to designer not to waste this potential.
Now the question is how to spot a high potential driver.
Drivers selection procedure is key to get a very high quality sound.
This is all very trivial i know, but still i believe that some sort of instrumental screening should be performed.
Again, i am not sayin the the SWR is the test ... but it could say something
IMD measurements can also be very important of course.
But at least for solid state devices i see in the datasheets always a SWR test
So i assume that it is telling something about the quality of the signal transfer
And, repeating myself, if i have also a passive xover distorting SWs and an active one not distorting them i would wonder if active it is not really the right way to go.
Thanks again and kind regards. gino
Last edited:
The most severe test for a driver is to operate over the whole audio bandwidth.
Most speakers cannot achieve adequate performance over the whole bandwidth and I suspect no driver can give better than good performance over the whole bandwidth.
I.e. there are no excellent full range speaker drivers.
But if you filter the signal to suit the driver, I suspect that all drivers have at least some good capability over a part of the frequency range.
The trick in design is getting a balance of good and/or excellent performance over a reasonable range from that audio bandwidth.
Size of the driver cone is a very good indicator of which part of the whole range they might be good at.
I further suspect that all drivers have at least good performance over at least one octave.
But I don't want a 10 way active speaker.
A 4 way is probably my upper limit and that requires every one of those 4 drivers to have good to excellent performance over at least 2 octaves. 15", 6"to 8", 3" to 4" and 3/4"
Now you just have to integrate those suggestions to make for good impulse response and see if anything close to good phase behaviour is achievable.
I wish I knew how to do that !
Most speakers cannot achieve adequate performance over the whole bandwidth and I suspect no driver can give better than good performance over the whole bandwidth.
I.e. there are no excellent full range speaker drivers.
But if you filter the signal to suit the driver, I suspect that all drivers have at least some good capability over a part of the frequency range.
The trick in design is getting a balance of good and/or excellent performance over a reasonable range from that audio bandwidth.
Size of the driver cone is a very good indicator of which part of the whole range they might be good at.
I further suspect that all drivers have at least good performance over at least one octave.
But I don't want a 10 way active speaker.
A 4 way is probably my upper limit and that requires every one of those 4 drivers to have good to excellent performance over at least 2 octaves. 15", 6"to 8", 3" to 4" and 3/4"
Now you just have to integrate those suggestions to make for good impulse response and see if anything close to good phase behaviour is achievable.
I wish I knew how to do that !
Last edited:
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.