Here's the same filter presented in Hypex Filter Designer. I selected AsymShelf as the filter type. It looks like the gain would need to be set to -6.8 dB or so to get the response at 10,000 to be zero dB.
I see! So it is basically a way of achieving a LR4 (or near) curve in the resulting total acoustic response, as opposed to just applying a LR4 filter electrically. Well, if the objective is a LR4 acoustical, then this is one way to get to it. As many people say - DSP makes things simpler to implement. A powerful tool.Sure. Say the tweeter has a natural mechanical low frequency roll off at 1350 Hz, it is effectively a 2nd order high pass filter with some arbitrary Q set by the driver properties. If you just cascade a 4th order LR filter at the 2000 Hz desired crossover frequency, you end up with a sort of 6th order filter response and resulting phase shift that won't play nicely with the low pass filter on the midrange or woofer. Starting with the Linkwitz transform (asymmetric 2nd order shelf filter) the 1500 Hz and arbitrary Q response can be canceled and replaced with the desired 2000 Hz, Q = .707 response desired to form half of the LR 4th order filter. I have tried to measure the Q and low frequency resonance using DATS or other impedance sweeps, but a simple near field acoustic measurement and just eyeballing the Q value works well.
I made a copy of the miniDSP spreadsheet in Google Sheets to calculate the coefficients of the Linkwitz transform. The example below shows how the natural response of a tweeter with a 1350 Hz Q = 1.2 ( Dark Blue ) can be shifted with the filter ( Light Blue ) to produce a 2000 Hz 0.7 Q high pass response ( Medium Blue? ). I'm a bit color blind.
I'm not sure but the Hypex software may have the 2nd order asymmetric shelf filter in the menu. The miniDSP requires typing in the filter coefficients, which isn't great. The link below the figure takes you to the spreadsheet.
View attachment 1449642
https://docs.google.com/spreadsheets/d/1SxDCrRJwczpLxShdq1TUdDULetdWTp43npEBMxRZWfw/edit?usp=sharing
X-overs are so much fun. Options are endless. You could add a couple of resistors and a cap to pad the tweeter about 5dB, and adjust the values, and layout to present the amplifier with a more resistive load. That may allow the amp to sound better, or it may have zero effect. It would protect the tweeter from any amplifier turn on thumps, and such. Might lower the noise floor slightly. Who knows. Fun to experiment sometimes.
If you look back at post #505, you will see that with a 2nd order passive filter on MY tweeter, I followed a LR4 target at 2k.
Filters don't have to be that close to a target, but it doesn't hurt either.
If you look back at post #505, you will see that with a 2nd order passive filter on MY tweeter, I followed a LR4 target at 2k.
Filters don't have to be that close to a target, but it doesn't hurt either.
Last edited:
Here is a question:
Looking at options for better measurements, I have googled for anechoic chambers and there is one very close and available for hire. Problem is the minimum they do is half day (4hrs) and they charge a lot. I would only need about an hour. They would allow sharing the four hours between a few people. Good to 125Hz.
Anyone want to share the time and cost?
Anyhow, that is long shot and the other options are outdoors or in a large room.
Looking at options for better measurements, I have googled for anechoic chambers and there is one very close and available for hire. Problem is the minimum they do is half day (4hrs) and they charge a lot. I would only need about an hour. They would allow sharing the four hours between a few people. Good to 125Hz.
Anyone want to share the time and cost?
Anyhow, that is long shot and the other options are outdoors or in a large room.
If you consider where your lowest cross is, where the mid is going to be down and out, then compare that to the lowest frequency you're confident of measuring then you'll know the band you're in need of. Now consider the frequency where driver nearfield measurements will be accurate up to, and consider the frequency up to where you'll tame the room for bass with your multi-sub setup, and you can assess the worth of using an anechoic chamber.
@AlmaAtaKZ: I suggest you stick to VituixCAD2 REW manual.
Build a turntable. Get speaker lifted to decent distance from nearest boundary. 1-1.5 meter is fine to get you started.
Build a turntable. Get speaker lifted to decent distance from nearest boundary. 1-1.5 meter is fine to get you started.
There's always some guesswork, and estimating involved. You don't need anything more than a medium size room, and effort. If the weather permits, outdoor measurements may make it easier, but you will likely end up with the same results as measuring indoors.
What is the longest distance that you can get the mic, woofers, and mids away from walls, , floor, and ceiling? (Not necessarily at the same time. )
I have 8' ceilings, so for me it's 4'. If I measured outdoors, and used some ladders, etc. I could probably get 8', but I would need a helper at that height to get a 50lb speaker up safely. I don't measure outside anymore, but I might, just to see how low I can gate, and compare with indoor measurements. I've done so before, but long ago. I'm not talking about multiple off axis. I just took on axis, and a few at around thirty, and forty five maybe.
I have 8' ceilings, so for me it's 4'. If I measured outdoors, and used some ladders, etc. I could probably get 8', but I would need a helper at that height to get a 50lb speaker up safely. I don't measure outside anymore, but I might, just to see how low I can gate, and compare with indoor measurements. I've done so before, but long ago. I'm not talking about multiple off axis. I just took on axis, and a few at around thirty, and forty five maybe.
Last edited:
Second attempt at indoor off axis and nearfileds, as well as the overall speaker with a starter crossover set. Files attached.
I have paid more attention to speaker and mic positions, a bit of strategic pillow and blanket placement etc. The near filed are 2cm mic distance. Both woofers playing, mid painting at the top woofer, no screening of the bottom woofer.
Fro the overall speaker, used LR4 at 250Hz and 3kHz, dropped tweeter -12dB and mid -10dB and applied LT 60/30Hz. No other EQ whatsoever. Speakers already sound pretty good.
I have paid more attention to speaker and mic positions, a bit of strategic pillow and blanket placement etc. The near filed are 2cm mic distance. Both woofers playing, mid painting at the top woofer, no screening of the bottom woofer.
Fro the overall speaker, used LR4 at 250Hz and 3kHz, dropped tweeter -12dB and mid -10dB and applied LT 60/30Hz. No other EQ whatsoever. Speakers already sound pretty good.
Attachments
-
baseline 2 tw.mdat2.9 MB · Views: 16
-
baseline 2 mid.mdat3.1 MB · Views: 12
-
baseline 2 w.mdat3.1 MB · Views: 17
-
baseline 2 nearf tw.mdat605.6 KB · Views: 11
-
baseline 2 nearf mid.mdat638 KB · Views: 10
-
baseline 2 nearf woofers top.mdat642.3 KB · Views: 12
-
WLT XoLR4@250 M-10bB XoLR4@3k TW-12dB.mdat1.2 MB · Views: 12
Yes, I have 3 presers that I can switch with button press on the Hypex. Plus I can upload DSP configurations via USB, but I have to connect to each speaker in turn.Can you do presets, and store several x-overs, that you can compare?
Last edited:
Great!Mid
Gate 5.3ms
Adjust vertical scale to around 50 dB range for meaningful graphs. (you used 210 dB in your graph!)
Looks like I do not need it.If you consider where your lowest cross is, where the mid is going to be down and out, then compare that to the lowest frequency you're confident of measuring then you'll know the band you're in need of. Now consider the frequency where driver nearfield measurements will be accurate up to, and consider the frequency up to where you'll tame the room for bass with your multi-sub setup, and you can assess the worth of using an anechoic chamber.
- Home
- Loudspeakers
- Multi-Way
- 3-way to active - Hypex FA253 - learning project