pd-cvs-request@iem.at wrote:
Message: 6 Date: Wed, 11 Jul 2007 17:34:40 +0000 From: Martin Peach mrpeach@users.sourceforge.net Subject: [PD-cvs] externals/mrpeach/osc packOSC-help.pd, 1.6, 1.7 packOSC.c, 1.6, 1.7 unpackOSC.c, 1.3, 1.4 routeOSC-help.pd, 1.2, 1.3 To: pd-cvs@iem.at Message-ID: mailman.9.1184234402.17743.pd-cvs@iem.at
Update of /cvsroot/pure-data/externals/mrpeach/osc In directory sc8-pr-cvs1.sourceforge.net:/tmp/cvs-serv5352
Modified Files: packOSC-help.pd packOSC.c unpackOSC.c routeOSC-help.pd Log Message: unpackOSC now simply outputs (through the right outlet) a delay in milliseconds when it receives a timetag, as the timetag itself is not useable inside pd. The delay is zero when the timetag is in the past. Help patches have been updated to match. Also corrected a mistake in packOSC.c timetag calculation.
cool feature, thanks martin.
however, i have 2 remarks:
the simple one: is it possible to set the "current" timetag within the [packOSC] and the [unpackOSC]? usually 2 osc-hosts should be synchronized via ntp, but who knows... otoh, this might be rather low priority...
the more complicated one:
"0" for past events is (imho) not a very good choice. how should we distinguish between bundles that have just arrived "in time" (where the natural delay is "0"), bundles that are tagged "immediately", and bundles that arrived too late.
one solution might be to use negative delays: delaying objects, such as [delay] and [pipe] just ignore negative values (so they behave the same as when fed with "0"), but the user has the option to determine whether the message arrived to late and can act accordingly (e.g. discard messages that should have happened yesterday)
i haven't looked at the code, but what happens when a bundle is to be processed "immediately": a delay of "0" is _not_ the same as no delay, as can be seen in:
[bng] | [trigger bang bang bang] | | [print c] [print a] | [del 0] | [print b]
which will result in "c, a, b"
so if i am to honour the timetags (via [del] or whatever), my "immediate" messages will not be handled immediately (it sounds worse than it is: due to the nature of network traffic, we could always assume that the "immediate" bundle arrived a fraction of a second later...; nevertheless i think it might be very good if i could distinguish between the 3 "types" of timetags)
mfg.asdr IOhannes
IOhannes m zmoelnig wrote:
pd-cvs-request@iem.at wrote:
Message: 6 Date: Wed, 11 Jul 2007 17:34:40 +0000 From: Martin Peach mrpeach@users.sourceforge.net Subject: [PD-cvs] externals/mrpeach/osc packOSC-help.pd, 1.6, 1.7 packOSC.c, 1.6, 1.7 unpackOSC.c, 1.3, 1.4 routeOSC-help.pd, 1.2, 1.3 To: pd-cvs@iem.at Message-ID: mailman.9.1184234402.17743.pd-cvs@iem.at
Update of /cvsroot/pure-data/externals/mrpeach/osc In directory sc8-pr-cvs1.sourceforge.net:/tmp/cvs-serv5352
Modified Files: packOSC-help.pd packOSC.c unpackOSC.c routeOSC-help.pd Log Message: unpackOSC now simply outputs (through the right outlet) a delay in milliseconds when it receives a timetag, as the timetag itself is not useable inside pd. The delay is zero when the timetag is in the past. Help patches have been updated to match. Also corrected a mistake in packOSC.c timetag calculation.
cool feature, thanks martin.
however, i have 2 remarks:
the simple one: is it possible to set the "current" timetag within the [packOSC] and the [unpackOSC]? usually 2 osc-hosts should be synchronized via ntp, but who knows... otoh, this might be rather low priority...
You can set the offset in packOSC so you can send a packet from the future as it were ;) I think that setting the actual time should be done outside of pd.
the more complicated one:
"0" for past events is (imho) not a very good choice. how should we distinguish between bundles that have just arrived "in time" (where the natural delay is "0"), bundles that are tagged "immediately", and bundles that arrived too late.
one solution might be to use negative delays: delaying objects, such as [delay] and [pipe] just ignore negative values (so they behave the same as when fed with "0"), but the user has the option to determine whether the message arrived to late and can act accordingly (e.g. discard messages that should have happened yesterday)
OK, I'll put in negative delays.
i haven't looked at the code, but what happens when a bundle is to be processed "immediately": a delay of "0" is _not_ the same as no delay, as can be seen in:
[bng] | [trigger bang bang bang] | | [print c] [print a] | [del 0] | [print b]
which will result in "c, a, b"
so if i am to honour the timetags (via [del] or whatever), my "immediate" messages will not be handled immediately (it sounds worse than it is: due to the nature of network traffic, we could always assume that the "immediate" bundle arrived a fraction of a second later...; nevertheless i think it might be very good if i could distinguish between the 3 "types" of timetags)
That's a difficult problem. What's the difference between zero and zero? I mean how does one tag no delay as being different from a delay of zero without adding another outlet? There's a problem with unpackOSC at the moment in that if no time tag is received there is no output on the second outlet, so you have to externally zero the delay line after every message.
Martin
mfg.asdr IOhannes
PD-dev mailing list PD-dev@iem.at http://lists.puredata.info/listinfo/pd-dev
Martin Peach wrote:
one solution might be to use negative delays: delaying objects, such as [delay] and [pipe] just ignore negative values (so they behave the same as when fed with "0"), but the user has the option to determine whether the message arrived to late and can act accordingly (e.g. discard messages that should have happened yesterday)
OK, I'll put in negative delays.
great. and i forgot to mention, that you can always connect the 2nd outlet to a [max 0] so you get the original behaviour.
That's a difficult problem. What's the difference between zero and zero? I mean how does one tag no delay as being different from a delay of zero without adding another outlet? There's a problem with unpackOSC at the moment in that if no time tag is received there is no output on the second outlet, so you have to externally zero the delay line after every message.
that is _good_ news, as it allows one to detect the absence of a timetag. (no message is a message too...)
one could also use "bang" as a more explicit indicator, but then one would _have_ to acknowledge the type too. (it's not such a good idea to send the bangs to the right inlet of [del]; but then this would enforce people to really take care of the timetags (if used at all), which is not necessarily a bad thing)
so i am happy with anything that allows me to distinguish between "now" and "immediately".
another question: in osc, timetags are per-bundle (not per message). is the scheduling information sent to the outlet for each message or only once for each bundle?
fgmas.dör IOhannes
-
IOhannes m zmoelnig -
Martin Peach