[ardour-dev] latest (MAMMOTH) CVS commit

[Paul Davis]

>> About 10 days, we noticed that the panning model inside Ardour was
>> completely wrong. Not just the mathematics of the pan law, but the
>> actual software architecture of panning. Fixing this has required a
>> massive set of changes to Ardour's basic data flow and a lot of subtle
>> changes to a lot of code. 
>
>Wow, I don't know how I missed this discussion on the list.  If it's

We don't use ardour-dev for much discussion any more. The important
stuff tends to happen in real-time over on #ardour at irc.freenode.net.

>not too much trouble, can you give me a one-liner about what was
>wrong?  I'd hate to be making the same mistake in my own code.

you need 1 pan control per data stream. if you have 2 data streams
(either 2 inputs to a route, or a mono->stereo plugin that has 2
outputs), you need independent pan control for each of them. without
this, you cannot do accurate placement in the stereo field, and you
cannot alter the "stereo width" of the acoustic image.

everything else mostly follows from that. when ardour was mono-only,
as in its original design, the pan model was fine. when we added
support for N-input tracks, it broke.

--p