[PD] MIDI input problems in PD

Pierre-Olivier Boulant po.boulant at free.fr
Sun Feb 26 20:15:32 CET 2012


Hi,

I have the same problem. It affects several computers running XP, Vista 
or Win7 and several Midi controllers; either connected to a RME 
multiface or through USB (Korg NanoPad / Berinhger BCR2000 or BCF2000 / 
Akai LPD8). Mostly using CC, hardly Midi-notes. And most of the time I 
go through Midi-Ox and a virtual midi port (midi yoke) to merge all 
controllers before sending to Pd. Midi-Ox checks for looping and should 
be able to prevent it and my patches try to limit looping and Midi data 
flow too.

It does look like there is some overflow of the midi I/O buffers.

I tend to limit the flow of data outgoing with [speedlim] and have a 
separate sessions of Pd for the control/GUI and audio with the 
audio/midi properties set differently between the two sessions.

For me it happens after big rushes of data: if I instantiate all the 
parameters of a BCR2000 at once (over 100 parameters) at the same time 
as updating the GUI in Pd it's not that hard to overload the midi 
buffers and create this problem and it doesn't take two hours to crash 
then... :)
And if you send a lot of stuff from the controllers too then it can lead 
to the same issue.
I'm not sure it's only the flow of midi data, but it could be the Midi 
data plus a lot of processing in Pd, especially with GUI objects moving 
in response to the Midi input.

Re-instantiating the Midi-settings seems to flush the buffers, but 
sometimes it takes a reboot to make the problem go away once everything 
is stuck.

Cheers
Pierre-Olivier




On 26/02/2012 08:28, Ingo wrote:
> I had the same problem with Windows XP and a RME Hammerfall DSP card using
> only the normal [notein], [ctlin], [toutchin] and [pgmin] for a sampling
> synth. I don't think it's the midi interface. RME has some of the best
> drivers - very stable.
>
> Sometimes after a certain amount of time it started playing notes by itself
> and didn't seem to stop anymore. It sounded like notes or melodies that I
> had been playing before. Nothing random.
>
> Could it be possible that midi data is written into a buffer that does not
> get emptied anymore? Then something might trigger reading out that buffer?
>
> I couldn't recreate why it happened. Most of the time it didn't do it but
> every once in a while it did.
>
> I had no problem ever running Nuendo for days on that same machine
> programming midi. The same Pd patch was doing fine on a Linux computer after
> switching to Linux.
>
> Ingo
>
>
>
>> -----Ursprüngliche Nachricht-----
>> Von: pd-list-bounces at iem.at [mailto:pd-list-bounces at iem.at] Im Auftrag von
>> Miller Puckette
>> Gesendet: Sonntag, 26. Februar 2012 00:04
>> An: Villa Anna
>> Cc: pd-list
>> Betreff: Re: [PD] MIDI input problems in PD
>>
>> I haven't seen this problem, but I have to admit I don't think I ever
>> ran MIDI into a Windows machine for more than 2 hours at a time (back
>> when I acually used MIDI Windows itself wasn't stable enough to run for
>> that long at a time :)
>>
>> It's hard to know whether this is the MOTU driver misbehaving in windows 7
>> or something with Pd itself - my only suggestion would be to try either
>> (1) switching interfaces to something more modern than 828MKII or (2)
>> try using some other software to see if the problem is specific to PD.
>>
>> cheers
>> Miller
>>
>> On Sat, Feb 25, 2012 at 11:03:32AM +0100, Villa Anna wrote:
>>> Dear list,
>>>
>>> I have problems with Midi input in PD.
>>> I use Windows 7 and make use of a Motu 828MKII soundcard.
>>> I receive MIDI via a polytouchin object (make use of FSRs and a coridium
>>> armmite to translate pressure on the FSRs to Midi messages).
>>> All goes fine in the beginning, but sometimes after a while (f.e. 2
>> hours)
>>> my Midi input starts to be random (FSRs trigger that are not supposed to
>> be
>>> triggering). If I restart PD everything is back to normal. Any idea what
>>> the cause of this problem might be?
>>> It's an installation project so it is difficult to restart everything
>> from
>>> time to time.
>>>
>>> Thanks in advance for your help!
>>>
>>> Laura




More information about the Pd-list mailing list