Can has unislope?

[o-smirnov]

How difficult would it be to revive the "unislope" functionality in the delay_and_offset? Maybe a common_delay_and_offset?

[JSKenyon]

I actually wanted to add this as a scalar mode to all the appropriate solvers i.e. on all terms which aren't 2x2 by definition. Will try to come up with something - in principle it is just a tiny tweak prior to the computation of the update. That said, I will not support mixing scalar and non-scalar behaviour for a single term i.e. if one specifies a scalar delay_and_offset, you will get a single delay, and a single offset. More complex behaviour can be accomplished by mixing terms i.e. a non-scalar phase plus a scalar delay would result in a phase estimate per correlation and a delay estimate over both. Is that adequate?

[o-smirnov]

More complex behaviour can be accomplished by mixing terms i.e. a non-scalar phase plus a scalar delay

Yep, nice and clean and easy to express.

[JSKenyon]

An experimental version of this functionality has been added in #358. It could use some testing in the wild.