Re: [LAU] send a midi CC signal to my midi controller from bash

On Tue, 13 Oct 2015 14:05:35 +0300, Athanasios Silis wrote:
>On Tue, Oct 13, 2015 at 2:05 AM, Ralf Mardorf wrote:
>> Control Change bits of the three bytes are
>>
>> Byte 1, nnnn is for channel 0 to 15
>>
>> 1011 nnnn
>>
>> Byte 2, ccc cccc number of controller
>>
>> 0ccc cccc
>>
>> Byte 3, vvv vvvv value
>>
>> 0vvv vvvv
>>
>> It's not too hard to write a function using amidi that allows you to
>> use it like that
>>
>> my_function interface channel controller value
>>
>> my_function 0 10 7 127
>>
>> It's comfortable to have a function that alows
>> values from -63 to 0 to 64 and

Oops, I made a mistake. Sometimes MIDI does use two's complement, IOW
from -64 to 0 to 63 but IIRC this never is used for control change
values.

>> values from 0 to 127.
>>
>> Unfortunately I don't have time to search my scripts for such a
>> function, but IIRC it wasn't hard to write it.
>thank you Ralph, i'll look into it.
>
>I took a midi dump (through arecordmidi) from my BCR2000 controller and
>realised while reading the midi file that 2 byte values are stored
>locally in a midi file lsb first. This is opposite to what I see here
>http://www.somascape.org/midi/tech/mfile.html#midi . for example the
>midi *MThd* : Header chunk, was written in the midi file as 54 4D 64
>68 . and chunklen as 00 00 06 00 . So this is one thing that I should
>keep in mind when I send hex numbers.
>
>also based on your analysis, can I just forget about all those headers
>when I send a midi cc command? can I just send the midi even directly?
>
>yes writing a function is not too hard = i do work in low lelvel
>programming after all, biut I dunno if by mistake I send some cmd that
>can fuck up the controller :)

I don't have knowledge about MIDI files, for simply sending MIDI data
headers are only needed for SysEx. I would use variables, e.g.

mx_header="F0 10 06"

F0 Beginn SysEx transmission
10 Oberheim ID
06 Matrix-6/6R/1000 Device ID

then follows the data
and after that a checksum could be required,
but sometimes it's not necessary

eox="F7"

F7 End of SysEx

Now to your MIDI data, 54 is binary 0011 0110, hence not a MIDI
command/not a status byte.
For MIDI commands you don't need SysEx headers, but the first byte must
be a status byte.

FWIW ignore running status. When using amidi add all bytes to the send
option, amidi likely will transform to running status. Running status
allows to send data without a leading status byte, but again, you don't
need to know how this works.

For command bytes, aka status bytes bit 7 is set: 1xxx xxxx
For data bytes bit 7 is cleared: 0xxx xxxx

http://www.midi.org/techspecs/midimessages.php

A control change example:

1011nnnn Status byte, were n is for the MIDI channel 0 - 15.
         1011 is for control change, IOW hex Bn were n is the MIDI
         channel hex 0 - F, IOW dec 0 - 15.

So the control change status byte must be

B0, B1, B2, B3 ... B9, BA, BB, BC ... BF regarding the used MIDI channel

After the status byte follow 2 data bytes.
The data bytes are explained by your device's user manual.

Most likely the used control changes for your device are for sending LSB
and MSB and they are vendor/device specific.

To send e.g. volume control to an averaged synth on MIDI channel 1, max
volume send

B1 07 7F

B1 Control Change, MIDI channel 1
07 First data byte, 07 is for volume control
7F Second data byte, the max value, dec 127 and binary 0111 1111

LSB and MSB is needed if more than 128 values are needed
                  (value 0 and 1 up to 127 = 128 values)
then a pair needs to be send

B1 xx yy
B1 xx yy

instead of a single

B1 xx yy

Regards,
Ralf
_______________________________________________
Linux-audio-user mailing list
Linux-audio-user@email-addr-hidden
http://lists.linuxaudio.org/listinfo/linux-audio-user
Received on Tue Oct 13 16:15:02 2015