Just wondering
a) if I've been a dumbass and overlooked it?
b) if we could have one?
I dissected the multiband gater and discovered the use of 2 filters in series with identical frequencies and with Q of 0.94 in order to slice off a frequency range.
What would be nice is a dedicated frequency splitter module, that is perfectly set up so that if the 2 output signals are recombined they would be identical to the input.
Dedicated crossover module (frequency splitter)?
-
- KVRist
- Topic Starter
- 335 posts since 20 Jul, 2010
Just to be absolutely clear what I mean, this does the job (but is 32 bit only)
http://www.rs-met.com/freebies.html
I can't get the filters in MUX to split the frequencies properly. When recombined and inverted the outputs of a crossover *should* sum to zero when added to the input. I hope that is clear
http://www.rs-met.com/freebies.html
I can't get the filters in MUX to split the frequencies properly. When recombined and inverted the outputs of a crossover *should* sum to zero when added to the input. I hope that is clear
-
- KVRian
- 855 posts since 3 Mar, 2009
I uploaded one patch in Mutools Dropbox that does split the input in 3 bands:
https://www.dropbox.com/s/4nj6h4rachoir ... 0Split.Mux
Recombined the frequency response is better than +- 0.2dB. As the filters are not linear phase there is an overall phase shift of 720 deg.
Filter slope is 24dB/octave.
It's currently setup as a "Band Kill". Replace the amplifiers with whatever you need or create 3 outputs for the MUX.
I do have a patch that does a flat frequency response with zero phase shift, but it's a 6dB slope filter only.
[edit]
I checked the rs-met crossover and it has exactly the same behaviour as mine.
For your own experiments, the trick is q=0.707
https://www.dropbox.com/s/4nj6h4rachoir ... 0Split.Mux
Recombined the frequency response is better than +- 0.2dB. As the filters are not linear phase there is an overall phase shift of 720 deg.
Filter slope is 24dB/octave.
It's currently setup as a "Band Kill". Replace the amplifiers with whatever you need or create 3 outputs for the MUX.
I do have a patch that does a flat frequency response with zero phase shift, but it's a 6dB slope filter only.
[edit]
I checked the rs-met crossover and it has exactly the same behaviour as mine.
For your own experiments, the trick is q=0.707
-
- KVRist
- Topic Starter
- 335 posts since 20 Jul, 2010
Huge thanks for uploading that example too.the trick is q=0.707
I think the dev should consider a module for this though : 1 in 2 out with a single knob for the split frequency (possibly another for db slope selection). It would make things neater and easier.
-
- KVRian
- 855 posts since 3 Mar, 2009
[edit:] Here is the 2 way crossover:
1 stereo input
4 stereo output: 1) Low 24dB; 2) High 24dB; 3) Low 12dB; 4) High 12dB
Now you can mix and match.
Enjoy
Link:https://www.dropbox.com/s/wsimqjksbqudz ... X-Over.Mux
1 stereo input
4 stereo output: 1) Low 24dB; 2) High 24dB; 3) Low 12dB; 4) High 12dB
Now you can mix and match.
Enjoy
Link:https://www.dropbox.com/s/wsimqjksbqudz ... X-Over.Mux