Re: zp2sos error: cplxpair: could not pair all complex numbers

[Nicholas Jankowski]

On Thu, Jan 5, 2017 at 12:45 PM, Mike Miller <address@hidden> wrote:
> On Thu, Jan 05, 2017 at 12:30:40 -0500, Nicholas Jankowski wrote:
>> > If you can help identify what Matlab does in this situation, and what
>> > should be fixed and where, then maybe something can be done about it.
>>
>> Running the same script in Matlab 2016b produces no errors and the
>> following output.
>> Note Octave output from cheby2 is a row vector, Matlab is a column. I
>> don't know if this matters.
>
> Yes, that is https://savannah.gnu.org/bugs/?49318
>
>> Matlab output follows. Octave 4.2.0,
>> signal pkg v.1.3.2 ouput shown below that.
> […]
>>   0.999999629730721 + 0.000860545419704i
>>   0.999999629730721 - 0.000860545419704i
> […]
>>   0.999999629730722 - 0.000860545419713i
>>   0.999999629730722 + 0.000860545419685i
>
> Thanks for running that. So yeah, looks like the bilinear function could
> use some improvement here. If this last conjugate pair were correct,
> there would be no error in cplxpair.
>
> --
> mike


Yes, Matlab's cplxpair errors the same way on Octave's z values as
well, so octave's cplxpair seems okay.

nickj