Hello,
Just want to say this is an excellent thread! Really enjoyed reading it. Good hands on information.
I should also drag my speakers once to a bigger room to compare measurements results, but having a 3-way active system, I am a bit to lazy...
Keep up the good work! Looking forwards to seeing some polars (I am also starting a waveguide project in the near future, so this will be interesting to see your measurement results).
Greetz,
Wim
Just want to say this is an excellent thread! Really enjoyed reading it. Good hands on information.
I should also drag my speakers once to a bigger room to compare measurements results, but having a 3-way active system, I am a bit to lazy...
Keep up the good work! Looking forwards to seeing some polars (I am also starting a waveguide project in the near future, so this will be interesting to see your measurement results).
Greetz,
Wim
I don't know if this will interest you, but I documented a fair amount of my thoughts and procedures for a small 2-way I made last year. There's a page on the close-mic to 4-pi splicing with some comparison of methods. Not complete, but may add some food for thought.
Chameleon 2-way
There's a bit of a mix of closed box vs. passive-radiator boxes, but above 100Hz the responses are the same anyway.
Dave
Last edited:
From the site:
"ARTA Software is classified as "shareware". To use it you have to pay for the user license key. If you have no license key, you can freely use this software only in a demo mode. The demo mode of programs is fully functional except loading and saving of files.
There are two types of licenses:
1) Personal license enables the single user personal use of the ARTA Software. The price for the personal license is 79 euros."
It seems that you cannot save or load data files in the demo version, but the price is reasonable at 79 euros. It has an impressive feature list and the author has published papers on acoustical measurements:There are two types of licenses:
1) Personal license enables the single user personal use of the ARTA Software. The price for the personal license is 79 euros."
ARTA Home 03012010
Back again!
Travel and work has kept me occupied lately, so progress haven't quite been as hoped for.
Anyway..
As I now feel that I have a basis for working on baffle step correction, it is time to look at crossing over to the tweeter.
So as previously recommended, I made a plot with on- and off-axis response.
Comparing it with the original SEAS plot, it looks pretty much as expected.
Measurements are taken on axis, 30 and 60 degrees off.
Now, I was also instructed to do a similar measurement for the tweeter, so that will be next.
And Dave, thanks for the link, I'll certainly read up on that! 🙂
Travel and work has kept me occupied lately, so progress haven't quite been as hoped for.
Anyway..
As I now feel that I have a basis for working on baffle step correction, it is time to look at crossing over to the tweeter.
So as previously recommended, I made a plot with on- and off-axis response.
Comparing it with the original SEAS plot, it looks pretty much as expected.
Measurements are taken on axis, 30 and 60 degrees off.
Now, I was also instructed to do a similar measurement for the tweeter, so that will be next.
And Dave, thanks for the link, I'll certainly read up on that! 🙂
Attachments
If the measurments are made in-situ then "baffle step" is not required. If not, then they should be.
I use 7.5 degree increments. 30 degrees is far too coarse. basically its only a "glimpse".
These measurements were made during the "Workshop session" so, Baffle step will be required. Now that I know the basic slope and the magnitude baseline, I should have a good chance of making a correction circuit and adjusting the required compensation in situ by ungated measurements.
Now, for measuring off angle in finer increments; why not? After all, I took the trouble to build the mic! 🙂
I also compared gated "farfield" results from my living-room with those made during the workshop session, and the "home-measurements" compare for the higher frequencies. Thus I can make a series of finer off angle increment measurements knowing they are good for the crossover region.
The only "problem" I have is that when I rotate the loudspeaker on the turntable i have, the axis of rotation does not coincide with the physical centre of the loudspeaker, i.e i get an increasing offset as the rotation increase.
But I guess that if I do this consistently, both when measuring the woofer and the tweeter, it should still be usable.
Another challenge will be to get multiple curves (more than 3) on the same plot in HolmImpulse, but perhaps there is a function for this?
Now, for measuring off angle in finer increments; why not? After all, I took the trouble to build the mic! 🙂
I also compared gated "farfield" results from my living-room with those made during the workshop session, and the "home-measurements" compare for the higher frequencies. Thus I can make a series of finer off angle increment measurements knowing they are good for the crossover region.
The only "problem" I have is that when I rotate the loudspeaker on the turntable i have, the axis of rotation does not coincide with the physical centre of the loudspeaker, i.e i get an increasing offset as the rotation increase.
But I guess that if I do this consistently, both when measuring the woofer and the tweeter, it should still be usable.
Another challenge will be to get multiple curves (more than 3) on the same plot in HolmImpulse, but perhaps there is a function for this?
To clarify, if you did not measure the speakers in some type of infinite baffle setup (if you measured them in the middle of a large room, for example), then you have already measured the baffle step. From your measurements, it appears you have measured the baffle step. So you only have to correct the response you measured when you design your crossover design - you don't have to make any additional allowance for baffle step.
Why is it that people do not measure the drivers performance in-situ? This seems to me the only reasonable way to do it. Am I missing something?
I do. I don't see the value of a drivers performance when it is measured in something other than the box it will be in.
You don't seem to understand my comment. "In-situ" means, placed in the location where they will be used. If they are measured like that then no "baffle step correction" is used.
I was specific about the driver because baffle step involves the actual baffle width. So no, I don't understand you. 😕
If you measure the drivers "in-situ", i.e. in the actual enclosure in which they will be used, then the "baffle step" is already included in the data set and nothing more needs to be done. Only when the crossover is done based on data taken in an infinite baffle (as usually supplied by the manufacturer) does the "baffle step compensation" need to be used. But, as we both seem to agree, this is not the ideal. The drivers should be measured for use in the crossover design as they will actually be used in practice.
So when you said "I do.", you meant that you "do" measure them "in-situ"? The statement was confusing because you seemed to disagree with me, but maybe you were just agreeing and I misinterpreted.
Last edited:
Not that you don't know this Earl, but driver manufacturers measure in infinite baffles because it's an easy to define setup, it's repeatable (when set up right), it's easy to correlate it to modeling which makes it easy to tweak designs, it's easy to quickly mount drivers, and it's easily adaptable to different size drivers. The only reason I see for diyers to measure in anything other than the final enclosure or something like the final enclosure is if you do a ton of speaker designing and you want to evaluate bunch of drivers' high frequency responses without building a box. But at that point you're blurring the line between small time manufacturer and hobbyist imo... And in that case I'd still then put the selected driver in the box and remeasure it to do the real design.
I measure the driver in the ~finished speaker and in the room. I used to take them outdoors but found this to be extremely inconvenient and it didn't really tell me anything more. Add to that problems with wind and ambient noise, etc. You know.
Of course John. If you don't know where your driver is going to be used, then the IB is the logical choice. If you know precisely how the driver is going to be mounted then it makes no sense not to mount it that way.
But lets be very clear here, no part of the room should be in the data used to develop the crossover. That would be a big mistake.
Yes, true. I don't use passive crossovers anymore though - all active with a digital crossover.
- Status
- Not open for further replies.