|
Post by Boomzilla on Jan 8, 2013 14:24:47 GMT -5
Bookmark added - Thanks!
|
|
|
Post by audiohead on Jan 8, 2013 17:00:55 GMT -5
This tread sure gets ones thinking cap on.Very good read!
|
|
|
Post by Boomzilla on Jan 8, 2013 17:11:23 GMT -5
If ya aint learnin' anything, whats the fun?
:-)
|
|
xki
Emo VIPs
Gwack!
Posts: 1,756
|
Post by xki on Jan 8, 2013 19:29:09 GMT -5
Time-alignment IS phase... Agreed! I can't agree. At least at this point. I'm thinking that the distance setting will result in time alignment but since correction tools don't alter phase, phase could be incorrect. At least not that I know of. I could be wrong on room correction tools but, from the diagram below, I believe time and phase are not the same thing. Thoughts?
|
|
|
Post by Chuck Elliot on Jan 8, 2013 21:49:46 GMT -5
Frequency, phase and distance are all related as follows: WaveLength = Speed-of-Sound/Frequency As an example: A 1000Hz frequency has a wavelength of 13.512" at 70 degrees* A signal that starts off at 0 amplitude going positive would be: Phase Amplitude Distance ----------------------------- 0_______0_____0 90______Max___13.512*.25 180_____0_____13.512*.5 270_____Min___13.512*.75 360_____0_____13.512*1 * www.1728.org/freqwavf.htm
|
|
|
Post by Boomzilla on Jan 9, 2013 5:47:17 GMT -5
cfelliot got it right, as did DYohn in a previous post. Time is phase, so far as sound waves are concerned.
This is NOT opinion.
|
|
|
Post by Boomzilla on Jan 9, 2013 5:52:41 GMT -5
...I'm thinking that the distance setting will result in time alignment but since correction tools don't alter phase, phase could be incorrect. Hi xki! Your diagram is correct insofar as it goes. Your diagram is misleading, however, since the distance of physical offset is exactly equal to one wavelength. Despite the physical displacement, in your diagram the peaks & dips remain "in phase." This is due to the one wavelength offset and would occur ONLY at the specific frequency illustrated. In real life, the waves' lengths are not the same due to frequency changes. In other words, physically offset drivers can be "in phase" ONLY at one frequency. That frequency is the one that has a wave length identical to the physical offset. At all other frequencies, physical displacement DOES result in out-of-phase peaks & dips.
|
|
|
Post by Boomzilla on Jan 9, 2013 9:52:48 GMT -5
Note how the polling is skewed (17 to 5) toward the idea that time-alignment in speakers is "clearly audible." Now note how that perception clearly contradicts the study done by the Klipsch Corp. In the papers posted the study showed that delays of up to three feet between drivers could NOT be heard by 100% of their test subjects.
Here again, we seem to have a dogma that, although deeply ingrained in the listening public, is just plain wrong. How many other similar things do we think we "know" that are also wrong?
Interesting...
|
|
|
Post by Nemesis.ie on Jan 9, 2013 10:03:14 GMT -5
Indeed, however, I can attest that even having the UMC-1 distance setting off by 1 or 2 ft in the L or R results (with my system in my room) in the vocals (and other front effects sounds) being noticeably skewed left or right.
Is that not a result of the time alignment being incorrect? I.e. if the start of a sound appears to be coming from a particular place, that is one of the first cues the brain uses to determine position?
|
|
|
Post by Chuck Elliot on Jan 9, 2013 10:38:17 GMT -5
Note how the polling is skewed (17 to 5) toward the idea that time-alignment in speakers is "clearly audible." Now note how that perception clearly contradicts the study done by the Klipsch Corp. In the papers posted the study showed that delays of up to three feet between drivers could NOT be heard by 100% of their test subjects. Here again, we seem to have a dogma that, although deeply ingrained in the listening public, is just plain wrong. How many other similar things do we think we "know" that are also wrong? Interesting... Although I pretty much look at PWK as an audio god, being a huge Klipsch fan, I disagree none the less. I look at this more as a rationalization than a true test. Complete time alignment may not be the most important of loudspeaker attributes, but it is certainly some icing for the cake. You said in a previous post that you liked the Dahlquist DQ10. So did I. It had an open beautiful sound stage that few speakers ever had. It was time aligned. I didn’t run out and replace my Cornwalls because what the DQ10 didn’t have was the effortless “punch” to the sound that the Cornwalls have. So, how can you have the best of both worlds? The misalignment of drivers is most prominent at the crossover frequency between drivers. As one driver become the sole source there is no issue. Most older Klipsch heritage speakers use very low order crossover networks. My Cornwalls have only 1st order B1 networks. With a first order network, both drivers on each side of the crossover point interact for a very wide bandwidth. Increasing the slope of the network can make this bandwidth much narrower. Take a look at ALKENG’s extreme slope crossovers for the heritage line. See: www.alkeng.comThe second solution is much more expensive and involves DSP and active crossovers. With this solution both steeper crossover slopes and digital delay of the closer drivers is possible. See: www.minidsp.com
|
|
|
Post by jmilton on Jan 9, 2013 11:40:08 GMT -5
Time alignment issues are not just that speaker "A" is slightly closer than speaker "B", but the drivers themselves propagate sound at slightly different "speeds"... i.e. higher frequencies travel faster. This company specializes in "time alignment" issues and has gotten great reviews for fixing these anomalies. Read up on their FAQ page: www.deqx.com/product-deqxmate-overview.php
|
|
|
Post by Boomzilla on Jan 9, 2013 12:10:42 GMT -5
Time alignment issues are not just that speaker "A" is slightly closer than speaker "B", but the drivers themselves propagate sound at slightly different "speeds"... i.e. higher frequencies travel faster. The speed of sound in air is constant and does not vary with frequency. This is NOT an opinion. OK - After research at sites including NASA, I am (technically) wrong: The speed of sound in air depends on density and humidity. At extreme conditions, it may vary with frequency but only "slightly." At altitudes where stereo listening can occur, however, my original statement is completely correct: The speed of sound in air IS a constant and does NOT vary with frequency. Just to be exact...
|
|
|
Post by jmilton on Jan 9, 2013 14:49:12 GMT -5
Correct. But the tweeter is placed more forward than the woofer in a typical speaker design which causes the high frequency to reach your ear first. You can "correct" this issue buy mechanical (tilting) or electrical (phase shift ) means. The Deqx system does the latter.
|
|
|
Post by Boomzilla on Jan 9, 2013 16:36:41 GMT -5
Yep, that's what they do.
|
|
xki
Emo VIPs
Gwack!
Posts: 1,756
|
Post by xki on Jan 9, 2013 17:14:20 GMT -5
...I'm thinking that the distance setting will result in time alignment but since correction tools don't alter phase, phase could be incorrect. Hi xki! Your diagram is correct insofar as it goes. Your diagram is misleading, however, since the distance of physical offset is exactly equal to one wavelength. Despite the physical displacement, in your diagram the peaks & dips remain "in phase." This is due to the one wavelength offset and would occur ONLY at the specific frequency illustrated. In real life, the waves' lengths are not the same due to frequency changes. In other words, physically offset drivers can be "in phase" ONLY at one frequency. That frequency is the one that has a wave length identical to the physical offset. At all other frequencies, physical displacement DOES result in out-of-phase peaks & dips. Boom - In the picture you posted, the intent (as you know) was to align the tweeter to the mid horn and bass driver so the distance from the actual driver faces were 'more' equal. And as CF pointed out, the slopes on the Klipsch crossovers are around 12db/octave. So logically, having the 3 drivers distance aligned would keep sound in phase when 2 of the 3 drivers are reproducing the same source material regardless of frequency. Now, would I be able to hear a difference if I moved my tweeters back as in the picture? I have no idea whatsoever!
|
|
|
Post by Boomzilla on Jan 9, 2013 17:20:39 GMT -5
I still plan to try the "time aligned" method. I've gotten some small cardboard boxes. I'll yank the tweeters & install them on the backs of the boxes with a pocket knife & some twist ties. Moving the boxes so that the tweeters are in the same plane as the midrange driver will tell the tale.
I'll post photos!
|
|
|
Post by Boomzilla on Jan 10, 2013 18:13:45 GMT -5
OK - Here's the test setup: (Special thanks to the USPS for their assistance in this experiment). And the results are... . . . . . . . . . . . . . . . . . Maybe. Yeah, I know you were hoping for something definitive. The soundstage might have opened up a bit, but I couldn't say for sure if it was because the tweeter was back at the voice coil of the midrange, because I set up the tweeters vertically instead of horizontally, or because the tweeters were on the front of the mount, not behind 3/4 inches of plywood. If I turn the tweeters horizontally, it isn't a fair comparison because the reflections of the top of the speaker would cause horrible interference. If I had a 3/4" plywood board, I could cut a slot and mount the tweeter, but it would be a pain to do. Even if I wanted to remount the tweeters at the back of the box, the "spouse-acceptance-factor" drops astronomically. I could just take the entire top cap off the speakers and tell my better half that I was going for "steampunk style" speakers, but I don't know if she'd go for it... My conclusion is that since the difference isn't startling, I'll leave the tweeters in the box (for the time being). Apologies to those who'll say I have a tin ear for not being able to immediately tell the difference, but I'm suspicious that if you tried this test in a "blind" setting, that you might not be able to tell either. I think that the Klipsch technical papers may be right - differences of less than a few feet are not audible in typical listening environments. Does anyone have reference to a proper test with sufficient trials to do Chi-square confidence testing on? I know of none, but I've not researched the issue either. In any case, back to the stock configuration.
|
|
xki
Emo VIPs
Gwack!
Posts: 1,756
|
Post by xki on Jan 10, 2013 20:27:41 GMT -5
Seems you would get a bad reflection anyway. Maybe elevate with another box? Say 6-8" That way any reflection would go right over your head!
|
|
|
Post by Chuck Elliot on Jan 10, 2013 21:51:02 GMT -5
I agree with xki, you're getting horrible reflections off of the top. Got a piece of foam to cover the top?
|
|
|
Post by AudioHTIT on Jan 10, 2013 21:55:51 GMT -5
Huh? You changed the horn orientation to vertical? It's not a line source, that will give poor horizontal dispersion and more reflection. I'm sure it's a fun experiment but I think if you want a time aligned speaker buy one, Paul is groaning.
|
|