Ensuring time alignment takes comparison over a band of frequencies, not just looking at the null, so that's not very revealing. Considering that you also often want a certain phase shift and the null method looks even less certain, perhaps more of a quick and simple method.
Take a look at the freq response in the link i gave Allen, it reveals the issue quite clearly and cannot be equalized as suggested in the graph ( and as my own try to eq failed too... on same serie of driver but different diameter).
Experienced the same with my 3 ways which use 15" to 4" at 750hz ( stock), as expected destructive interference can't be eqed....
Post #18 Allen.
On my Tannoy it was easy to determine it was a Z axys issue ( it's more or less common issue with these drivers, really coincident on Z axys being the exception rather than the norm with coax).
On my threeway it is related to directivity mismatch but it didn't stopped me to identify if it was related to Z offset too/phase issue from initial passive xover ( using either 'J.Bagby method' - involving taking measurement of all individual drivers, then by group of two and with three of them and then load the individual response into lspcad and play with Z ( depth) parameter until overall shape of simulation match measurement/ or wavelet method as once suggested by Mark100 in an article he linked in the past- both are ok but i prefer wavelet as it's quicker and give info about time delay and phase( if applicable) in same measurement).
On my Tannoy it was easy to determine it was a Z axys issue ( it's more or less common issue with these drivers, really coincident on Z axys being the exception rather than the norm with coax).
On my threeway it is related to directivity mismatch but it didn't stopped me to identify if it was related to Z offset too/phase issue from initial passive xover ( using either 'J.Bagby method' - involving taking measurement of all individual drivers, then by group of two and with three of them and then load the individual response into lspcad and play with Z ( depth) parameter until overall shape of simulation match measurement/ or wavelet method as once suggested by Mark100 in an article he linked in the past- both are ok but i prefer wavelet as it's quicker and give info about time delay and phase( if applicable) in same measurement).
There are a whole bunch of other issues like diffraction or directivity mismatches that can cause such phase jumps that it will never be equal off-axis.
But that all depends how a system is constructed.
But just saying that it's easy to fix a hard null (perfect phase behavior) over a wide range of axis, doesn't hold much practical value.
Besides the fact that it's debatable that it's even mandatory to create a good sounding system to begin with.
But that all depends how a system is constructed.
But just saying that it's easy to fix a hard null (perfect phase behavior) over a wide range of axis, doesn't hold much practical value.
Besides the fact that it's debatable that it's even mandatory to create a good sounding system to begin with.
Ok, here's what's on the bench right now for a little demo...
Working on seeing if i want to use a permanently imbedded house curve. Indoor non-gated, just fooling around really....
5-way synergy, syn11.
Here's on-axis. Blue = all polarities are correct
Notches at 110Hz and 375Hz, were caused by swapping low driver polarity.
375Hz and 900Hz caused by mid swap.
900Hz and 3000Hz by HF swap.
Done swapping one section at at time, and saving each trace.
Xover points of course...110, 375, 900, and 3000Hz.
All notches plain and clear as expected with on-axis tuning.
Here's same thing but now at 35 deg off-axis.
Horn is a 75 deg conical, so almost outside of its pattern.
Looks like nulls can be seen pretty easily, huh
Without their frequency budging a single Hz....
try it yourself using an eq set to full cut around 3K
Planar speakers have a natural null to either side. The null is very apparent - there is no bass!
At higher frequencies, the null becomes more localized and less apparent.
Ed
At higher frequencies, the null becomes more localized and less apparent.
Ed
Very good example Edgr. Remind me of a video demonstrating ASR's Directiva cardio prototype ( with headphones it's obvious when there is trotation around the loudspeakers. It's a good example of what to listen when looking for room's critical frequency):
I don't see anything relevant - https://www.audiosciencereview.com/...y-system-600-speaker-review.11919/post-345503
offtopic:
Traditionally (very) bad and way to quick of an advice as always to "just EQ the crap out of things".
(I am referring to the screenshot)
Bizarre that some people just keep doing that without knowing and EXPLAINING the ins and outs of a system as well as the (lack of) resolution of the graph and/or measurements. 😡👎
Can you determine whether this is a timing or phase issue by looking at the plot? Can you determine the amount of delay by looking at the plot?
@b_force: Yes and if you read the whole thing you'll see he never listened to the loudspeakers...
Ok i get he have much more interesting thing to do and ASR is focused on measurements but it demonstrate a lack of inquiring mind imo.
Not that it is something mandatory but this is a character trait i encountered with the best maintenance engineer i met and which is commonly seen iwith members here.
I like to understand things, even more when something gone wrong with 2 designs using same drivers ( System800a doesn't have the dip) and when it's easy to find schematic of both designs... but he doesn't have time to spare, it's an 'old' design,... so wrong conclusion about the issue = wrong recommendation...
@AllenB : no of course not with this plot only, but if i see this behavior i know what to look for in the schematic of the version which doesn't have the issue and/or do what i described earlier to identify if it comes from this effect or not ( wavelet is good to identify both (delay/phase) in one shot).
Iow if a dip like that appear with a coax i know who the culprit can be.
Ok i get he have much more interesting thing to do and ASR is focused on measurements but it demonstrate a lack of inquiring mind imo.
Not that it is something mandatory but this is a character trait i encountered with the best maintenance engineer i met and which is commonly seen iwith members here.
I like to understand things, even more when something gone wrong with 2 designs using same drivers ( System800a doesn't have the dip) and when it's easy to find schematic of both designs... but he doesn't have time to spare, it's an 'old' design,... so wrong conclusion about the issue = wrong recommendation...
@AllenB : no of course not with this plot only, but if i see this behavior i know what to look for in the schematic of the version which doesn't have the issue and/or do what i described earlier to identify if it comes from this effect or not ( wavelet is good to identify both (delay/phase) in one shot).
Iow if a dip like that appear with a coax i know who the culprit can be.
Last edited:
No, minimum and excess phase components can combine making this method near unusable by visual means alone.
Last edited:
To put it very straight and simple, I don't.
In fact, I REALLY don't.
You can't claim to be all objective and "scientific" , without putting in ALL the effort of making sure you don't make mistakes.
Or at least trying to get better at it.
Take a extremely good and positive example with some other guy with a similar measuring setup.
If not, simply don't post it or take on the job. That's just your ethical obligation imo.
Why? Because how someone portrays himself with such a name and measuring setup. This will put up expectations with people.
Or worse, put companies in a bad position.
Saying, advising, to "just" EQ boost 5-7 dB is absolutely nuts.
Say that for something you don't know or understand the root cause from, is just absolutely totally insane.
I have seen him "advising" something similar for an issue, that was obviously destructive interference. Go figure.........
An inquiring mind is one of the very first things you learn at most science universities for this reason.
As a researcher you have the simple obligation to see if your measurements make sense and to second guess them if needed.
If there is anything out of the ordinary, you just try to find an explanation or you just postpone the test entirely.
Understanding, in this case also listening, is an absolutely vital part.
Yes i share your view but... i remember members of Pass's subforum being bothered by same kind of shortcuts.
Nelson Pass answered it's all a matter to have an audience and entertaining it, and HE prefered the audience he have here ( we can be more funny! 🙂 ) to the one ASR have.
I think it resume things well. To bad there is some really good members at their forum with very interesting projects ( Directiva i linked the video i like a lot).
I suppose same kind of things could be targeted on us in here, about different points maybe but...
Nelson Pass answered it's all a matter to have an audience and entertaining it, and HE prefered the audience he have here ( we can be more funny! 🙂 ) to the one ASR have.
I think it resume things well. To bad there is some really good members at their forum with very interesting projects ( Directiva i linked the video i like a lot).
I suppose same kind of things could be targeted on us in here, about different points maybe but...