Ask an Acoustic Engineer (me)

[Jdunk54nl]

Attachment 18351

This is my example. my tweeter and midrange 120 degree shift and sum at 2450hz with 0db gain. on what frequency should i attempt to use first order all pass filter? thank you.

Can you post your actual data. A picture of that would be worth 1000 words. I understand what you are saying about your summing (you get 0 summing) but without seeing the data, anyone that gives advice on using all pass would just be guessing.

[Kb86]

[Jdunk54nl]

Can you link your rew file? A online storage link works best like google drive or dropbox or onedrive link to the file.
I am confused as to why you are getting summing at 400hz to like 1000hz. That should just overlap with the mid line exactly.

Also if you could tell me your desired crossover for those drivers.

[Kb86]

I will find a way late. im kinda dumb on computer thing. thank you.

[dumdum]

If your speakers sum together flat, then what problem are you trying to solve?

Not flat… zero summation, none… he is only ever guessing that he has 120 degrees out

put simply no he needs a phase graph to accurately see how the best way to tackle it is…

Also I’d question how he has set the timing?

[Jdunk54nl]

Not flat… zero summation, none… he is only ever guessing that he has 120 degrees out

put simply no he needs a phase graph to accurately see how the best way to tackle it is…

Also I’d question how he has set the timing?

Yup, and for those that don't have access to real time phase, this is where the JL target curve stuff is SUPER powerful. You can take your speakers measured response, use that to create an identical target, then view the phase and make adjustments to your system based on those targets in the jl tun software.....which all of that is free!

You can even mimic reflections and what they will impact with the tun software by having a similar acoustic response just different levels.

[Jdunk54nl]

Like this example here: This is a "perfect" tweeter response, with a "perfect" reflection that is -6db and delayed in time by 1.04ms (or about a 1ft)

Yellow trace is your sum, red is your original, and white is the reflection




now if I throw another "reflection" into here that is now -12db and delayed by 2.08ms.. Now green is the second reflection

[Jdunk54nl]

And even more fun when you do a midrange driver and you get a response like so out of the driver.....

Attachment 18356


Notice anything fun (read semi common for people) around the 450hz mark in this midbass driver
Attachment 18357

[Jdunk54nl]

Just very crudely making your measured results, I can get a fairly similar graph if I delay the midrange a little bit more than it should be. Not saying this is your issue, but it could be. The closer you could get to having your exact information into Jl tun, the better the model would be and you could then use that to make changes to your measured system.....very powerful stuff right here people!

If one didn't have access to smaart or open sound meter, you could use your usb microphone and measure via tun, make a target curve so you can then get phase information, and line up everything that way. Obviously all this becomes easier and quicker if you have smaart or open sound meter (I don't expect many on here to have a jl max like I do now...for reasons)

[Kb86]

i tried adjust timing, not better sum. Could be reflection at that frequency.
i tried phase slider on helix dsp, around 20 degree adjust on tweeter, sound much more smooth at high and definition at drum/kick drum. but it screwed my imaging.
i move crossover point out of that problematic zone. Move to 2800hz acoustic xover. Happy result. - thanks @dumdum for advice.

im still asking/messing with 2450hz xover zone for a sake of learning how to do first order all pass filter using only RTA and ear.

Nick demo that technique but it was easy when there is big cancellation. just use 2nd apf at that exact frequency dip.
for around zero summation, how to determine what frequency should i attempt to use 1st order apf?