Analogue-sequencer

Folks,

I've uploaded v3.1.006 beta 34 to Yahoo files.
This build covers all the technical changes I'm planning on making 
before the next release build - there are just UI things to add, 
assuming there aren't any problems in this one.

Feedback is always welcome with P3 betas - even if you install it 
and have no trouble at all, I'd like to know. Otherwise I can't be 
sure whether it all works perfectly, or nobody tried it.

The most recent change is new external MIDI sync handling code, 
which should allow P3 to sync to dodgy PC based sequencers that send 
out clock bytes with the regularity of radioactive decay.
It now queues up clock bytes, and handles them at maximum tempo if 
they arrive in clumps.
I don't have any dodgy PC sequencers in my studio, so it would be 
nice if afflicted persons could confirm this resolves the issue 
(Miguel, this could save you the MIDI-Ox workaround).

There are some major internal changes to the way P3 handles poly 
patterns.
For mono patterns things will be exactly the same, but if you are 
using aux notes, a couple of things are different.

First, multiple auxes set to the same note are thinned out, so there 
is only a single note message for each note number.
This eliminates unneccessary use of MIDI bandwidth, and makes life 
easier for poorly written synth OSs that struggle with multiples of 
the same note...
There's now a limit of 8 notes that can be played by each track - 
you can achieve this by pairing two tracks together, with aux A on 
one track redirecting all auxes to the second track and the other 
three auxes doing aux notes.
Adding those three to the four aux notes plus one main note on the 
second track makes eight.
I have also changed the behaviour of the 'aux note abs Xd' and 'aux 
note rel Xd' events which are renamed 'aux note abs xF' and 'aux 
note rel xF'. xF = no FTS
Force-to-scale does not apply to these aux notes, but the current 
part transpose and playlist transpose settings now do.
This seems a lot more logical to me, but if you find it stops you 
doing something useful, let me know.
Another change is the rel aux notes are now calculated based on the 
FTS'd value of the primary note, not the stored value.
This makes more sense if you are using 'aux note rel xF' to do 
fixed 'chord memory' type effects against a root note pattern that 
is being forced to scale.

One other change in this build is not at all obvious unless you 
check P3s MIDI output on an oscilloscope.
After recent discussion about MIDI delay, I decided to make sure 
that there was the absolute minimum of CPU latency in P3s MIDI 
output, so the notes come out as fast as the MIDI bus can take them.
I found that with heavy use of CPU intensive aux events and all 
tracks active, CPU latency was reaching 3ms in some cases.

A re-ordering of the way each clock pulse is handled has eliminated 
this, by moving all aux event processing from the start of each 
clock pulse to the end of the previous one.
Personally I'm not sure I hear any difference even with quite a busy 
group of tracks triggering percussive sounds, but if you think this 
build is any rhythmically tighter, let me know.

Best regards,
Colin Fraser
Sequentix Music Systems Ltd
http://www.sequentix.com