Sherk SCOTT

I’ve become confused with the options for normalization in the HOA patchers of spat5. There is an option for both FuMa and MaxN which has confused me.
In the help patch for spat5.hoa.converter~ it states “FuMa and MaxN are almost the same. The only difference is a sqrt(2) = 3dB gain on the Wchannel”. However, I am not certain that this is correct.
Is not Classic Gerzon – FuMa actually the following?
W: 1/sqrt(2); 3dB
X: 1: 0dB
Y: 1; 0dB
Z: 1; 0dB
Where as maxN is the following:
W: 1; 0dB
X: 1; 0dB
Y: 1; 0dB
Z: 1; 0dB
If this is the case, isn’t option listed as MaxN is actually maxN?
Thanks,
Scott

T.

Hi Scott,
Sorry for the late reply.
I’m not quite sure to understand your question/problem.
“Is not Classic Gerzon – FuMa actually the following?
W: 1/sqrt(2); 3dB
X: 1: 0dB
Y: 1; 0dB
Z: 1; 0dB
Where as maxN is the following:
W: 1; 0dB
X: 1; 0dB
Y: 1; 0dB
Z: 1; 0dB”
–> Yes, I agree with that.
“If this is the case, isn’t option listed as MaxN is actually maxN?”
–> What do you mean ? Do you discriminate between MaxN (upper case) and maxN (lower case) ??
I might be missing something, but the equations implemented in spat5.hoa.converter~ seem correct to me.
Best,
T.

Sherk SCOTT

Yes, T. It is my understanding that MaxN (upper case) is the same as classic Gerson / FuMa / MaxN
W : 1/sqrt(2) ; 3dB
X : 1 ; 0dB
Y : 1 ; 0dB
Z : 1 ; 0dB
whereas, maxN (lower case) is the following
W : 1 ; 0dB
X : 1 ; 0dB
Y : 1 ; 0dB
Z : 1 ; 0dB
Two different normalizations with almost identical names.
I’m confused…

T.

Hi Scott,
I am not aware of any such distinction between upper/lower case maxn (could you point me to relevant bibliography ?)
My best understanding of HOA normalization is detailed in the following conference paper:
“T. Carpentier. ‘Normalization schemes in Ambisonic: does it matter?’ In Proc of the 142nd AES Convention, Berlin, Germany, May 2017.”
and this corresponds to the spat5 implementation.
But anyway: it is recommended to avoid maxN (not matter the case). N3D or SN3D should now be considered ‘the right thing to do’.
T.
