mizu

**mizu**

Hello,

I'm trying to install flext externals for Pd under WinXP compilation with MinGW. When trying to build I've got this error message:

mingw32-make[1]: *** No rule to make target
`pd-mingw\release-single/flbase.opp', needed by
`pd-mingw\release-single\libflext-pd_s.a'. Stop.
mingw32-make: *** [build-release-single] Error 2

Does anyone know what about it is? Do I need a flbase.opp file? I can't find it

Thank you

**mizu**

Hi all,

I meet problems writing realtime values in a text file. Here are two ways I used [textfile] (you can try these patches in the attachment)

This way works:

This way doesn't work (it only saves the first values):

I think this comes from the fact that in the first way, datas are generated instantaneously, where the second way's datas are generated in any time (here depending on the metronome period). And this is precisely what I aim to do.

Can anyone propose a method to acheive that? You would really really help me!
Thanks

http://www.pdpatchrepo.info/hurleur/using-textfile.pd

**mizu**

Hello,

I need to save a pitch and velocity list into a new midi file. Does anyone know which way I can do that?

The inlet below receives the list

[inlet]
|

|
[? write to midi file ?]

Thanks

**mizu**

Well I'm trying to use [xeq] to play midi on a [dac~] but I can't find any help file telling what must come to the inlet.

Anyone?

**mizu**

Maelstorm, thank you for the symbol trick, it's smarter than my pipe artifacts

Nevertheless the [sel] object is useless for my application: the number of lines in my text file isn't fixed but depends on the running.

Actually the values generated correspond to midi pitch and velocity values transcribed from a wave file (or a realtime audio input), so I can't predict how many of them will be created.

**mizu**

ow! what's the trick?

Indeed the output values are wrong! I can't even see the difference between both ways. Are you a magician?

**mizu**

Ok, that's done. Simply put an infinite delay on the [pipe] and add a [flush( message to the left inlet, so that the flush is sent when the data generation is stopped.

In the pocket

**mizu**

I can't find an object that simply passes a message through when it receives a bang! It doesn't seem so complicated, who's got the solution??

**mizu**

Hi,

well, it's not an error: I used that to output two values per line in my text file (here these values are the same, for the example, in the attachment, the second is multiplied by 2, anyway it doesn't matter for now). These values must be outputted at the same time, so no trigger is needed here.

Anyway, yestarday evening I found a solution that satisfies me. For who is interested, here is what I did:

In fact, adding values in the [textfile] is done before writing the [textfile] contents into the real text file, using a trigger as below:

The problem is that [trigger] sends signals sequentially but with no delay. So the write message is sent after 0 second, when the only first data has been generated by the metronome. So we need a delay before the trigger sends the write message. This can be achieved using the [pipe] object between the trigger and the message:

[t a b]
/ |
[pipe s 2000] generating data as above
|
[write $1(
|
[textfile]

Now, the datas are saved after 2000 ms on the text file. The datas generated after these 2 seconds are lost. This becomes a pseudo-realtime saving because the data flow is recorded in realtime but only saved at a precise moment.

I join the Pd file corrected.

Next step is to replace the [pipe] object by something that sends the message when the metronome is stopped. If anyone manages this, I'm interested.

Anyway, thank you hardoff for worrying about our issues, you do help!

Mizu

http://www.pdpatchrepo.info/hurleur/using-textfile-corrected.pd

**mizu**

Oh ok, I'm going to check that immediately. Tks