As the title suggestions I’m looking to implement FIR (Convolution) crossovers using Roon convolution.
My signal path will be:
Roon Server > RPI4 USB > Motu MK5 Lite > 2x NAD c298 amps.
This setup is tested and working for procedural crossover upto 8 channels.
For My JBL m2 setup… Stereo source, 4 outputs……
44100 2 4 33 #44.1khz, stereo in, 4 channels out, up to 4 channel mapping
0 0 #Delay on inputs)
0 0 0 0 #Delay on outputs)
Drive:\high44.wav #Location of 44.1 wav)
0
0.0 #Channel 1 input left
0.0 #Channel 1 output left
Drive:\high44.wav #Location of 44.1 wav)
0
1.0 #Channel 2 input right
2.0 #Channel 3 output right
Drive:\low44.wav
0
0.0 #Channel 1 input left
1.0 #Channel 2 output left
Drive:\low44.wav
0
1.0 #Channel 2 input right
3.0 #Channel 4 output right
Does this look correct? I assume I just create .cfg for each sampling rate used if not resampling to a fixed sampling rate? I.e. I.e. 48000, 96000, 192000
I believe that I was told previously that not including higher resolution files will add to the processing load and if you have higher res files you should use them.
In my case I was trying to apply convolution to DSD and my sonicTransporter i7 (from several years ago) did not have the power to pull it off.
I assume these .cfg files are simply renamed text files.
One thing that does confuse me is the output channel mask… and how it would apply to my M2 2x4 solution.
The convolver website example suggests 33 to be used for a 2x4 setup, it seems to be an addition of the speaker position hex bits, 1,2,10,20… but I really don’t understand how its generated… or what the correct number would be for my situation, in the context to how Roon assigns channel mapping…
Looking at a basic procedural filter in Roon (Yes not Convolution), channel mapping follows the following:
So in this circumstance, 1+3 left, 2+4 is right.