Page 1 of 1
Posted: 31 Dec 2016, 02:22
by woodslanding
I have a moderately complex wkp (simple compared to my last one) and I keep getting in situations that create a constant string of these messages to the console.
Can anyone tell me what they mean, where they come from, how to stop them?
So far, I've been rebuilding my wkp from patches, and then they go away. But once they are in a wkp, they never seem to disappear. They seem to keep any midi from showing up in patches, but that's all I've figured out so far.
I am doing dynamic loading of patches containing VSTs. Maybe this is triggering it....
Thanks for any clues.
-eric
Posted: 31 Dec 2016, 16:59
by woodslanding
Okay, this definitely is triggered by loading a VST. I have it in 'process idle'. Is there any more low priority way to load it?? It is peaking the CPU, causing audio dropouts, and then it seems that HH can never recover....
Here's my script:
Code: Select all
/////////////////////////////////////////////
// Load subpatch with VST
////////////////////////////////////////////
CONST folder = 'INSTRUMENTSPATCHES';
// parameters declaration
var filenameIN,prevNameIN, filenameOUT : Tparameter;
var running : boolean;
var filename : string;
// destroy
procedure Destroy;
begin
end;
// initialisation : create parameters
procedure init;
begin
filenameIN := CreateParam('file name',ptTextfield);
SetIsOutPut(filenameIN,false);
//prevNameIN := CreateParam('previous file',ptTextfield);
//SetIsOutPut(prevNameIN,false);
//filenameOUT := CreateParam('file name',ptTextfield);
//SetIsOutPut(prevNameIN,false);
running := false;
//SetFastCallback(filenameIN,false);
end;
procedure Callback(n:integer);
begin
if (n = filenameIN) and(getValue(filenameIN) > 0)// and (getValue(filenameIN) <> getValue(prevNameIN))
then begin
strace('MOON____LOADING PATCH: ' + getStringValue(filenameIN));
filename := getApplicationPath() + 'config'
+ folder + ''
+ trim(GetStringValue(filenameIN)) + '.pat';
//test if it exists!
running := true;
end;
end;
// no process bloc
procedure processidle;
begin
(*
CREATE_MODULE george 6 300 300 mysubpatch.pat
CREATE_LINK 1 0 george 0
CREATE_LINK 2 1 george gain
CREATE_LINK george 2 3 0
*)
if running then begin
SendInternalMsg2('SET_TARGET_PATCH','SENDER_PATCH');
SendInternalMsg2('DELETE_MODULE','VST');
SendInternalMsg6('CREATE_MODULE','VST','6','300','50',filename);
//SendInternalMsg5('CREATE_LINK','1','0','VST','0'); //midi
SendInternalMsg5('CREATE_LINK','audioINproc','2','VST','0'); //audio in 1
SendInternalMsg5('CREATE_LINK','audioINproc','3','VST','1'); //audio in 2
SendInternalMsg5('CREATE_LINK','midiINproc','1','VST','2'); //midi in
SendInternalMsg5('CREATE_LINK','VST','3','audioOUTproc','0'); //audio out 1
SendInternalMsg5('CREATE_LINK','VST','4','audioOUTproc','1'); //audio out 2
strace('MOON____MODULE REPLACED');
running := false;
end;
end;
and here is a sample subpatch:
http://www.sensomusic.org/forums/upload ... synth1.pat
thanks,
-eric
Posted: 31 Dec 2016, 18:47
by 23fx23
when dealing with iml should often mosly set inputs fast callback false
procedure SetFastCallBack (Param: TParameter; val: boolean); // set to true to use a fast callback insteed the
// Windows message processing
tho i don't know if this apply to processidle, never used that procedure, usually used callback with a on init pulse, i have to check this
Posted: 01 Jan 2017, 01:45
by woodslanding
good tip though, I will try that!
Okay, I am getting an access error now. Probably got it before, but lost in the stream of cycle restart messages.
first:
[5:27:12 PM] Error : 'NAN' is not a valid floating point value //lots of these
[5:27:12 PM] Error : E033 audio dropout //lots of these
[5:27:22 PM] Error : Access violation at address 0000000000DF9713 in module 'Usine.exe'. Write of address 000000011BE5B408
[5:27:22 PM] Load User Module : PluginWrapper.usr-win64
[5:27:22 PM] Create user module : D:_USINEUsine Hollyhock Pro Win64ConfigPlug-InsVSTMassive.plugin
Here's the log file:
http://www.sensomusic.com/forums/upload ... gUsine.txt
Posted: 01 Jan 2017, 23:13
by woodslanding
Okay comes from the HSL object having a color output that is NOT A NUMBER.
Any chance of a patch on that before HH3?
If not, any ideas for a workaround? I think I have an old HSL script around somewhere, I'll try to dig it up.
edit: found a nearly identical patch that doesn't do this.... so it's working.
Posted: 04 Jan 2017, 00:58
by woodslanding
well, it's back again, and no sense of how to trace it....
Posted: 04 Jan 2017, 03:07
by shawnb
Have you learned about 'repair IDs'?
It's magic...
Try it at the patch level..,
Posted: 04 Jan 2017, 03:46
by woodslanding
not working here, sadly. I feel like it must be something wierd in a script, but have no idea how to track it down....
Posted: 04 Jan 2017, 05:26
by woodslanding
Well, I got my rack down to a couple dozen simple modules, most of them not even connected to each other, and it was still blazing away.
Rebuilt the rack from component patches and.... it's gone. Still would love to know what's going on.
Posted: 11 Jan 2017, 05:17
by woodslanding
DOH!
Well, restart cycle refers to the cycle in the sync module. If you send it a value of 1, it continuously resets! I just put a pass if changed module in, and it's fine.
Maybe this helps someone else someday.