[Ardour-Dev] Some issues with 2.8.10

[Paul Davis]

On Fri, Aug 13, 2010 at 11:51 AM,  <fons at kokkinizita.net> wrote:

> I just transferred the session that showed the 'solo problem'
> to another machine that is running 2.8.7 instead of 2.8.10.
> And - surprise surprise - here the busses *are* muted by a
> solo on another strip.

Please forward the .ardour file (no audio required), and which
tracks/busses to solo and in what order. I find this rather unlikely,
but certainly possible.The basic (broken) rule (for 2.X) is that
soloing tracks has no effect on busses, and soloing busses has no
effect on tracks.

> It's this sort of thing that makes me feel uneasy. For how
> long does Ardour have solo ? One would assume that this is by
> now a stable part of the code, and fairly independent of any
> new features. Yet it isn't - apparently improvements to other
> unrelated parts, or adding new ones, can still break it easily.

very little of the signal processing pathway is "independent". the
interaction of gain, declicking, solo, mute, etc. is really incredibly
complex. this is one of the reasons why there are some deep
architectural changes in 3.X, but even so, the addition of true
PFL/AFL and a monitor path doesn't really reduce the complexity or
interdependency. i spent about 2 months working to get the solo design
in 3.X "right" with input from several console owners and users, and
even now, because of Ardour's "route anything anywhere" topology,
there are still a few (obscure) corner cases where its behaviour will
be "surprising" (multi-layer bussing, primarily).

>> we have worked fairly hard to hide MIDI features when they are not relevant.
>
> I believe you when you say you're working hard :-). But I wouldn't
> be surprised that in order to exclude all risks related to midi I'll
> have to disable it over and over again, and try not to forget it.

you don't enable or disable MIDI in 3.X. you either have MIDI tracks
or you don't. if you don't, there are just a few extra items in the
GUI to confuse you.