Post ringing

[UliBru]

Ok, thanks. Please send me the pulse responses.

[PhilActiveAudio]

Here are the Left and Right pulses.

[UliBru]

Hello Phil,

do you use a Umik-1 microphone?
The zoomed pulse response looks like

Time.png (49.13 KiB) Viewed 2905 times

This is not normal. It is not caused by a brickwall filter.
Please do a recording of the tweeter in nearfield (about 40 cm at tweeter axis), you can simply run e.g. a sweep from 5 kHz to 24 kHz even without XO. Select the proper recording channel and the low frequency according to the tweeter specs. A sweep time of 10 sec is enough. In addition you can also check the Dirac option in the recorder if you use a Umik-1 or USB mic.
Please send me then the file Logsweep48_rec.wav to ulrich.brueggemann@audiovero.de

[PhilActiveAudio]

Hi Uli,

No, I use a Earthworks DM23 mic and a RME Fireface UCX card for the mic input and the 6 outputs.

These measurements are 3 ways with a basic XO setup (NT-1 550 5500) with no delay correction neither voice optimisation.
In my setup, the low voice is physically in first position, then the tweeter and finally the mid voice which is the most distant driver.
To protect drivers, I had installed Mundorf capacitors before the mid and tweeter.

I will measure the tweeter at 40cm this evening.
I already send you a tweeter on axis measurement at 100cm, 2.5 to 30 kHz (96 kHz).

[PhilActiveAudio]

I've made some new measurements of the tweeter this evening, on-axis 40 cm and found several things.

1) The physical position of the tweeter was probably not good: the tweeter was originally located on the back side of the Arai mid driver horn and was probably generating reverb on the pulse. Once relocated on the front of the Arai horn, reverb was less important.
2) The measurement of the range 2,5 kHz to 30 kHz shows oscillations that are highly corrected when measuring starting from 5 kHz. Is this can have an impact of the original problem, knowing that the measured filter was set at 5500 Hz between mid and tweeter? I probably need to set a new filter at 6,5 kHz or 7 kHz.

Tweeter in back of the horn, on-axis 40 cm, 2,5 kHz to 30 kHz

Tweeter avant.jpg (360.86 KiB) Viewed 2894 times

Tweeter in front of the horn, on-axis 40cm, 5 kHz to 30 kHz

Tweeter après.jpg (363.89 KiB) Viewed 2894 times

[UliBru]

This looks much better. How about a tweeter measurement up to 24 kHz with 48 kHz samplerate?

[PhilActiveAudio]

I will perform new measurements this WE with the new positioning.
The last measure at 48 kHz samplerate was before I found the problem and move to 96 khz to see if if was equal and then proceed with optimisation.

[PhilActiveAudio]

Here are the new 48 kHz measurements on axis at 40 cm. The tweeter is now located in front of the Arai horn so no possible early reflections except the room:
> Red curve is 5 kHz to 24 kHz.
> Green curve is 7 kHz to 24 kHz.

Pulse 48kHz.JPG (341.39 KiB) Viewed 2838 times

The pulse still looks weird to me. I was not able to remove the oscillations as I managed to to with my early post but I don't know how (except that the recording was done at 96 kHz from 7 khz to 30 kHz).

[PhilActiveAudio]

I did some more measurements and found that there are oscillations when I measure the tweeter above 20 kHz at 48 kHz and above 40 kHz at 96 kHz.
I did a RME Fireface UCX II loopback measurement and found that I have the same oscillations but with limited effects.
Can these oscillations be related to the card?
I don't know and would appreciate some feedback from other users.

However, I have good results when measuring at maximum 20 kHz at 48 kHz.
Here is the tweeter pulse at 3,5m distance (listening position).
Good enough to build a first "real" filter (UB jPol11 2nd order) with corrected delays for each driver.
Sound very good!

Left tweeter pulse at 3.50m.JPG (71.08 KiB) Viewed 2771 times