[PD] bendin bug (?)

[Alexandre Torres Porres]

2016-09-12 15:31 GMT-03:00 Miller Puckette <msp at ucsd.edu>:

> In fact I thought there already was one in Cyclone.
>

The one in cyclone is [xbendin] and [xbendout] and they both go from 0 to
1383. I find this pair of objects quite obsolete and I don't know why they
were put in cyclone, as they don't do anything different than Pd's
bendin/bendout - they are consistent between themselves though.

Max also has bendin/benout, but the 'x' variance seems to be newer and with
a different range, cause the old - but still existing - bendin/bendout
objects in Max are only 7 bits (from 0 to 127). BTW, Miller, can you
confirm pitch bend messages were only 7 bits back in the day? And do you
know when this changed to 14 bits?

Funny enough, max has inconsistencies with other objects.

Quite recently, in the 7.2.0 version (Max is now at 7.2.4), the objects
[midiformat] and [midiparse] were finally updated to include 14 bits pitch
bend messages, via a '@hires' attribute. The ones in cyclone are still only
7 bits (0-127).

We (Me and Derek) just updated these objects in our currently unreleased
cyclone update project to include 14 bit pitch bend messages. In
conjunction with Pd's [midiin] / [midiout], they can replace [bendin] /
[bendout] related objects, but it is just too clumsy for that. And the
silly thing is that they chose for them the more intuitive range from -8192
to 8191, making it inconsistent to xbendin/xbendout. We adopted this more
intuitive range anyway.

I was then checking bendin/bendout in Pd to see if they were also from
-8192 to 8191 and then found this inconsistency.

So, in short: yeah; xbendin/xbendout in cyclone can be used for a
consistent range between MIDI in/output. But then I still think these
objects are quite obsolete, cause bendin/out are basic elements present in
vanilla. And I it is also weird to me to have them as a way to replace
existing objects in vanilla.

> Yet another idea would be to make a new object name (pitchbendin/out)

Cool. Whatever the solution is, I'm hoping for a vanilla fix so both pitch
bend in/out messages are consistent in vanilla. I'm ok with any solution,
but I'm just hoping that both make use of the more intuitive -8192 to 8191
range.

cheers
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.puredata.info/pipermail/pd-list/attachments/20160912/8675cd6b/attachment.html>