Demystifying tempo related sysex

[question]

Hello

I was working on a sort of machine learning algorithm that involves MIDI data and a part of it estimates incoming tempo based on audio pulses or midi note durations. When I decided to add a tap tempo feature, my H9 became handy. I can get the tap tempo through CLK.OUT by counting the number of 248s per second, but this causes a lot of continuous data traffic, which is pretty redundant after the measurement is finished. Also there’s this known H9 control communication loss problem when CLK.OUT is set to ON. So I looked at the SYSEX solution. Obviously, along with its DUMP facilities, H9 sends some sysex info upon the tempo change (actually upon any change).

For example, upon an Dump -> Active operation within the DUMP menu I get

    [14] 4 5 3

    4 7fe0 7fe0 0 0 1f42 1545 3c62 3b60 4110 6420 7fe0

    0 0 0 0 0 0 0 0 7fe0 1f0e 0 0 0 3c93 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0

    0 22e5 1 0 9 5 4 0

    3 2 0 0 4 3 17 16.7165 50.88 65.3603 65000 65000

    C_cc06

    H910 -P5 (-7)

which is explained clear enough in Sysex description  pdf

 

But when I exit the DUMP menu and start playing with the tempo settings I get lines like

   240, 28, 112, 1, 96, 7, 0, 92, 2, 0, 0, 1, 49, 247

   240, 28, 112, 1, 96, 7, 0, 92, 2, 0, 0, 0, 47, 247

in hex format,

 

or like

   f0 1c 70 1 60 7 0 5c 2 0 0 1 31 f7 f0 1c 70 1 60 7 0 5c 2 0 0 0 2f f7

when I converted them to ASCII.

Which I have no idea about its meaning : )

I wonder if anyone could enlighten me about the part of this message that contains the current tempo value

Also, to be able to get sysex messages from the H9 during normal operation, ie. when tweaking the knob etc, it looks like one have to run H9 control first then close it. Otherwise H9 doesn’t send any sysex after a power cycle. Maybe I have to send an info request sysex from within the software I use to develop my patch (that’s Pure Data, to be precise). And I also wonder what this message should be : )

Cheers

Q

 

[nickrose]

Alas, we are not going to be much help.

The f0 1c 70 1 60 messages are specific to H9control and thus are not documented.

For reasons which will now be obvious, they are only sent when H9control is connected to the H9.

Duplicating the H9control connection procedure would be difficult, and involves some secret sauce.

The H9 does not send a specific message when tempo changes, as it was thought that this would be handled by MIDIclock. The activity that you see is the H9 notifying H9control of parameter changes, etc.

I'm not familiar with "there’s this known H9 control communication loss problem when CLK.OUT is set to ON" – not known to me, alas.

I fear that MIDIclock will be your best bet. You can turn it off with a sysex message if you need it no more.

 

 

[nickrose]

As mentioned, we believe that the bug you mention is actually a Windows issue, rather than an H9 proplem, in that Windows does not handle the mix of MIDIclock and MIDI messages.

There should not be any problem if the H9 is not sending messages to a Windows PC.

But, I'll nudge our H9control team and see if they know anything.

 

[question]

Hi,

On the contrary it was rather helpful, thanks for the clarification.

nickrose wrote:

The H9 does not send a specific message when tempo changes, as it was thought that this would be handled by MIDIclock. The activity that you see is the H9 notifying H9control of parameter changes, etc.

 

Actually, that’s also what I started to gather after an eye crossingly lengthy observation on the incoming sysex 

: )

nickrose wrote:

You can turn it off with a sysex message if you need it no more.

Good tip. This would eliminate the necessity of running H9 control beforehand which was making things impractical. I’ll try.

nickrose wrote:

I’m not familiar with “there’s this known H9 control communication loss problem when CLK.OUT is set to ON” – not known to me, alas.

 

It’s mentioned in this post:

https://www.eventideaudio.com/community/forum/stompboxes/h9-control-bug-when-clock-out-standard-h9

Best

Q

[Author]

Posts