Perfect in my application would be the ideal acoustic target to hit the crossover point and slope to integrate with other drivers with a passband chosen to optimise dispersion and minimise non-linear behaviour. that's why I put perfect in speech marks. Sorry I thought that was obvious.
Ben, I don't think your square wave example holds water, for a number of reasons.
I agree that square waves are a good way to see phase, but honestly, they aren't much better than a good phase trace from REW...
And lots of folks are getting good phase traces nowadays.
When I use a real scope, square waves stay pretty consistent looking when moving the mic around small amounts like you did.
But neither of my full range e'stats (Acoustat X and CLS) make square waves that look all that good. Both crap out long before 500Hz.
Are you certain yours are as good at it as you think?
Neither of my stats have any form of processing on them...which has a great deal to do with proper phase delivery. Maybe I should try to tune them...
Because good square waves at 1000Hz can be done, as shown below.
And to say good phase info can't reach our ears just because there are reflections etc, simply doesn't make sense imo.
First screen shot below is Acoustat at 490Hz. Just showing it's already crapped out by then.
Next two are a DIY 3-way using a coax CD, at 640 and 1k.
I agree that square waves are a good way to see phase, but honestly, they aren't much better than a good phase trace from REW...
And lots of folks are getting good phase traces nowadays.
When I use a real scope, square waves stay pretty consistent looking when moving the mic around small amounts like you did.
But neither of my full range e'stats (Acoustat X and CLS) make square waves that look all that good. Both crap out long before 500Hz.
Are you certain yours are as good at it as you think?
Neither of my stats have any form of processing on them...which has a great deal to do with proper phase delivery. Maybe I should try to tune them...
Because good square waves at 1000Hz can be done, as shown below.
And to say good phase info can't reach our ears just because there are reflections etc, simply doesn't make sense imo.
First screen shot below is Acoustat at 490Hz. Just showing it's already crapped out by then.
Next two are a DIY 3-way using a coax CD, at 640 and 1k.
Attachments
Absolutely. A good example is the same soft dome in a 2 or 3 way speaker. I would cross this over differently; shift up the crossover point or steepen the slope in a 3 way for the tweeter for 2 reasons:
1. The midrange is likely smaller than a 2 way so dispersion is improved
2. The tweeter could become the excursion moreso than thermal weak link as larger woofers invariably need less sweep and /or are more sensitive to handle more power
Ben re. your post number 57.
Read my post again Ben, it asks for impulse response, frequency response and phase response. You have not give any of these, you appear to completely sidestep my questions by just putting up a smoke screen of responses showing a single tone (rich in odd order harmonics) that does not convey any phase data. and two amplitude responses in the time domain.
Neither picture shows an axis labelled phase in degrees. I'm not about to spend time trying to extract data from nothing. None of what you have provided in the way of data correlates with the questions I've asked.
Please don't take the opportunity to reply just to enjoy getting inebriated by the sound of your own verbosity, it just clutters up the thread and doesn't help the TS.
Peter
Read my post again Ben, it asks for impulse response, frequency response and phase response. You have not give any of these, you appear to completely sidestep my questions by just putting up a smoke screen of responses showing a single tone (rich in odd order harmonics) that does not convey any phase data. and two amplitude responses in the time domain.
Neither picture shows an axis labelled phase in degrees. I'm not about to spend time trying to extract data from nothing. None of what you have provided in the way of data correlates with the questions I've asked.
Please don't take the opportunity to reply just to enjoy getting inebriated by the sound of your own verbosity, it just clutters up the thread and doesn't help the TS.
Peter
Last edited:
Ben, I don't think your square wave example holds water, for a number of reasons.
I agree that square waves are a good way to see phase, but honestly, they aren't much better than a good phase trace from REW...
And lots of folks are getting good phase traces nowadays.
When I use a real scope, square waves stay pretty consistent looking when moving the mic around small amounts like you did.
But neither of my full range e'stats (Acoustat X and CLS) make square waves that look all that good. Both crap out long before 500Hz.
Are you certain yours are as good at it as you think?
Neither of my stats have any form of processing on them...which has a great deal to do with proper phase delivery. Maybe I should try to tune them...
Because good square waves at 1000Hz can be done, as shown below.
And to say good phase info can't reach our ears just because there are reflections etc, simply doesn't make sense imo.
First screen shot below is Acoustat at 490Hz. Just showing it's already crapped out by then.
Next two are a DIY 3-way using a coax CD, at 640 and 1k.
Nice square waves (although you seem to have something ringing at 2400Hz). Quads make great ones too. I sometimes get square square waves if I want to take the time to find the right spot. But as we all know, the sound is the same. Which itself, of course, further demonstrates that phase is not a crucial cue.
The question I was addressing was Peter's: do the phases go out of whack if you move your mic a small distance. And if they do scramble, even on axis, what does that say about the heart-felt if evidence-shy yearning on this forum for phase coherence.
I don't see anything in your otherwise meaningful post that criticizes my demonstration of scrambling.
B.
Last edited:
Ben, the point you always seem to miss is that there are things you can do to optimize the speaker *before* taking into account room interactions, which will greatly reduce the number of negative effects you will get from those interactions.
Just correcting in room for things that could have been avoided in the first place if engineered properly, is not my idea of proper speaker design.
IMO it is always better to treat the cause rather than the symptoms.
Tony.
Just correcting in room for things that could have been avoided in the first place if engineered properly, is not my idea of proper speaker design.
IMO it is always better to treat the cause rather than the symptoms.
Tony.
Quit nit-picking.
My two images show EXACTLY what you requested and they demonstrate phase scrambling clear as day, whether every harmonic is precisely labeled or not.
Scrambling is scrambling and can't be unscrambled. Phase coherence is a minor issue for hearing and psychoacoustics even if it is of any significance to any physics textbook.
Anybody have one of those "test records" with an announcer speaking with reversed polarity to one speaker? Anybody run any of their drivers with reversed polarity for a a couple of days before noticing it? Or might be doing it now? Anybody spent an afternoon trying to figure out which polarity they prefer for their sub-woofer?
B.
Not an entire afternoon - but long enough to know which setting complimented the woofers in the overlap region. I only have the one sub.
Are you saying relative phase between drivers - no matter what the value (between 0 and 180 degrees) is meaningless and does not affect the sound?
I am not sure what is meant by "phase coherence", but here is my take on phase:
- Absolute phase is irrelevant
- phase linearity with frequency can be an important aspect of perception, but it is not a dominate one, as is frequency response.
- phase linearity, i.e. phase(f) means that there is no variable group delay in the signal (only a uniform time delay across frequency is possible if the phase is linear,) i.e. all components of a transient are time aligned. If this is "phase coherence" then you are incorrect that it is unimportant. If "phase coherence" means something different to you then you must define it.
- Large changes in phase with frequency as measured by the impulse response mean that various portions of a signal arrive at different times, regardless of the phase aberrations caused in steady state signals by reflections etc.
- the ear hears only the transient parts of a signal at higher frequencies, the window length becoming shorter as the frequency goes up such that we hear steady state signals at LFs, but only the transient part at HFs.
- Hence the phase as measured by a windowed impulse response is audible, albeit, not a dominant effect.
This is true if one can measure anechoic farfield SPL that low. However, most DIY folks cannot easily do that. In these situations simulation is better than no data at all.
Yes. It depends on the recording, so if that's screwed up, phase wise, then not such an issue, otherwise it's beneficial without a doubt.
Not sure what your point is, especially due to the selective quoting which makes it look like I said something very different to what I did...
You may not be aware of Ben's premise, ie that the only thing that matters is designing the crossover in the actual room, because the room has much more effect than anything else. He sees designing for good performance using quasi anechoic measurements as a complete waste of time.
Tony.
O thank you, this is why there are now 'stereo' designed by you that sounds better than a full orchestra or a live jazz band.
We can go to the Metz and listen to a great stereo, DSP, no need for singers anymore or real instruments, digital replaced everything.
You guys live in a digital fantasy.
If a loudspeaker cannot be made to sound right with passive components it will never sound right with any modification of the signal sent to it.
Hi Earl, it is something that I've picked up after discussions across multiple threads. Ben does not distinguish between the crossover, and "corrections" that may be made for the particular room, that the speakers are playing in.
So whilst my approach is to design the crossover using quasi anechoic measurements, and to worry about specific room issues as a separate process, Ben lumps it all in together.
This may help to understand his viewpoint.
Tony.
So whilst my approach is to design the crossover using quasi anechoic measurements, and to worry about specific room issues as a separate process, Ben lumps it all in together.
This may help to understand his viewpoint.
Tony.
It's not that "I" lump sound together. It is that your ears lump direct, reflected, absorbed, and otherwise scrambled-phase sound reaching your search all together.... mostly.
For one example, there is no benefit in finally achieving a flat anechoic response because once back in the real world, the speaker polar response will interact with the specifics of your room, drywall, windows, carpet, and furnishings to render it almost a waste of time.
But as my "....mostly" implies, your ears do make certain distinctions like the Haas Effect. The Haas Effect, as Tony may or may not know, further interacts with the distance to reflecting walls nearer or further*. Again, anechoic FR tells you zero about that.
But some people find it comforting to embrace the illusion of personal control through three-decimal-place analysis step-by-step. This may help to understand Tony's viewpoint.
The question for stereo sound quality (which mostly means spatial localization in the horizontal plane) is to have available to the brain the cues supporting identification of sound objects and their angle and distance in the plane. System development should work toward that end.
B.
* I think I once posted two FR curves for my music room with the door to a closet which holds my bulky winter motorcycle riding gear left open versus closed; the FR measures better and the SQ is better with the door slid open. Anybody tested their closet doors or too busy with getting the anechoic response perfect?
Last edited:
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.