Yes, but.
When I exported the phase from his measurements, It looked super weird. So I recommended we just ignore it altogether (I didn't export it) for the first pass at this, and get him learning how to use everything. So what I'm working with doesn't have the measured phase response, just a simulation.
Basically my impulse was to lay the blame at the indoors far field measurements (might have been me of course). I figure once he gets some measurements either outside or a blended nearfield the phase data might be useful again.
When I exported the phase from his measurements, It looked super weird. So I recommended we just ignore it altogether (I didn't export it) for the first pass at this, and get him learning how to use everything. So what I'm working with doesn't have the measured phase response, just a simulation.
Basically my impulse was to lay the blame at the indoors far field measurements (might have been me of course). I figure once he gets some measurements either outside or a blended nearfield the phase data might be useful again.
Woofer from REW
Midrange
I'm not the best at phase analysis, but it does look like room nodes and floor/ceiling bounce type phase anomalies around the frequencies we're talking about.
Midrange
I'm not the best at phase analysis, but it does look like room nodes and floor/ceiling bounce type phase anomalies around the frequencies we're talking about.
I think I just realized what you're saying, shifting the various crossover freqs to align the phase simulation better.
That would certainly not be a bad thing.
This last one is getting more into small filters. In my limited experience, small filters based on indoor measurements don't work.
My initial design on my own speakers was done indoors, my first measurements outside on that initial filter design were rough. Lots of garbage in there.
My initial design on my own speakers was done indoors, my first measurements outside on that initial filter design were rough. Lots of garbage in there.
I don't find nulls useful to me when I have phase available.
As for precision, a few degrees here and there won't hurt anything. If you need to make some kind of unrelated demonstration and don't have anything reliably prepared, a little smoothing might not go astray 😉
I'd agree with this concern. In agreeing with what Jim said, unless you have full polars there will be room for doubt. There will also be doubt about the needed crossover alignment.
As for precision, a few degrees here and there won't hurt anything. If you need to make some kind of unrelated demonstration and don't have anything reliably prepared, a little smoothing might not go astray 😉
Just for fun, show us this with 2nd order electrical filters. I'd do it myself, but I'm not set up for active Sims. Actually, my software probably can do it, I just never used the feature.
Last edited:
Quick and dirty, second order slopes, set crossover points and moves the mid and tweeter delays to get the phase in the ballpark
I think this is sensible at this stage. I will definitely do nearfield soon.
I will then try to do new and better far field either outdoor, in a very large room or even find a better option. Already started working on these.
So, for the 'good far flied', do I just do a full 0 to 180 degrees sweeps at 10 deg intervals, horizontally? Do I need to do anything else? Is this what is meant by 'the full polar'? Can you help me to scope this 'good far filed' measurement?
Last edited:
Do you have that measurement in REW, or was that posted? If in REW, let's see it gated.
Yes, those pictures are REW. The midrange is gated to 100hz, the woofer is un-gated.
The point of the REW measurement is the phase is borked, and I don't know how to gate the export for use in VCAD.
The point of the REW measurement is the phase is borked, and I don't know how to gate the export for use in VCAD.
How much does the mid, and tweeter clean up gating at around 250hz?
When I measure. The shape of the response changes a lot when gated. Even on a smooth tweeter response.
When I measure. The shape of the response changes a lot when gated. Even on a smooth tweeter response.
I know there's too many cooks in the kitchen, including me, but off axis is important after you get on axis flat. Measuring off-axis will help you decide on the tweeter / mid x-over point, but a quick look at about 45 or 60 degrees will show most of what you need. Easily done indoors as well. Horizontal off axis on mid and woofers will be fine at 400hz. It's the tweeter, and mid that take some consideration.
Last edited:
I'm traveling now, so no access for a bit. He measured at 2m indoors if I remember correctly. If you gate the mid response correctly, 250hz is off/at the bottom,.
This is all a first pass, and he needs better measurements to produce better results.
Last edited:
Better measurements will improve accuracy. I'm just itching to play along. As better measurements are posted, I will update my sim.
I downloaded his measurements, so I've seen the gated. . My PCD software can sim active filters. So far, I'm having trouble making an frd from the measurements. Seems easy, and I've done this before maybe a month ago, but I must be missing a detail. I export, open, and see it's text file. I save as something like midresponse.frd, and select all files, but it does not load. Any suggestions? I almost always have issues doing this, but eventually get files that work.
I downloaded his measurements, so I've seen the gated. . My PCD software can sim active filters. So far, I'm having trouble making an frd from the measurements. Seems easy, and I've done this before maybe a month ago, but I must be missing a detail. I export, open, and see it's text file. I save as something like midresponse.frd, and select all files, but it does not load. Any suggestions? I almost always have issues doing this, but eventually get files that work.
When you measure the tweeter, I suggest that you should sweep starting at 200hz or lower. If you want, you can use a cap for protection. In fact, the cap could be a permanent part of the speaker, and protect the tweeter from all sorts of potential future problems.
Last edited:
Yes. I see it says export as text. Then I open the text file, and do a save as, and add the .frd. I think PCD is picky about the files.