Sysex

AnthonyB

I had forgotten that ST4 was adding SYSEX support (something I had originally requested >12 months ago). I'll be in a position to do some testing today and will try the SYSEX stuff.

I have SYSEX commands in my DAW that I'm sending to a Behringer X32 mixer.
For ST3, I've setup Midiflow to do command remapping of some CCs to SYSEX.
I'll be able to try those raw SYSEX commands in ST4 and see what happens.

@peter I'm still a little concerned about the lack of a MIDI channel requirement for SYSEX commands. I know the assumption is that the SYSEX header contains the MIDI Channel and/or a bunch of other device specific info that will (should) make other devices ignore the message, but wouldn't it be cleaner to still specifiy a MIDI channel so the message isn't be sprayed everywhere to all devices and across all interfaces? Also, it would keep the lyric sheet command syntax consistent with existing MIDI commands (i.e. have @<Ch>} at the end of the command) which may help prevent people making mistakes?

peter

jfduval Okay, test done. The monitor receives the sysex commands correctly. It receives the three blocks contained in the command I am using. If I remove one or two blocks, it receives what is sent correctly. However, the device for which the sysex is intended (VoiceLive Play) does not respond. I did an additional test using OnSong on the same iPad that I use for ST4, which has the ability to send sysex, and it works fine; the VoiceLive Play receives the command correctly.

That's good to know. From my perspective I would say the problem is somewhere in the connection to the VoiceLive Play. Are you sure it works with iOS? We had a discussion here with the result that it does not work on iOS: https://support.stagetraxx.com/d/1625-connecting-a-ipad-to-voice-live-3-play-to-send-midi-commands/3

AnthonyB I'm still a little concerned about the lack of a MIDI channel requirement for SYSEX commands

SYSEX is not sprayed out over channels. MIDI has channel level messages like note, program change, control change that require a channel and global messages like start, stop and also sysex. And global message have no parameter to specify a channel.

AnthonyB Your syntax is wrong. This should work:

[midi@00:00.750: SXF0002032322F66782F322F7479706520444C59F7]

But I guess I should add support for adding whitespace between the bytes. The messages are way easier to read with whitespace.

jfduval

Hello Anthony, I also really want sysex integration; it's the ultimate feature that would make this application perfect!
I did another test on my end, where I put a typical MIDI command in the ST4 editor that is sent to OnSong. When OnSong receives the MIDI command from ST4, it then sends the sysex command to my VoiceLive Play. It works perfectly. So the problem seems to be related to sysex management in ST4.

AnthonyB

So, my initial testing did not go well.

The song I tested with has the following SYSEX commands in my DAW that load a specific type of effect in a specific FX slot and set the parameters on my X32 mixer:

I just grabbed 4 of the commands that select the effect type and slot and 3 params for testing.

I tested 4 different variations of the commands as I wanted to see what did and did not work (in my case, none worked). I know @peter said no MIDI Channel target, but I wanted to see if including would break stuff.
Below is what I had in my lyric sheet, and I removed the semi-colon to uncomment each group of commands to test.

;VARIATION 1 - SYSEX commands with spaces, include a MIDI Channel target
[midi@00:00.750: F0 00 20 32 32 2F 66 78 2F 32 2F 74 79 70 65 20 44 4C 59 F7@1]
[midi@00:00.776: F0 00 20 32 32 2F 66 78 2F 32 2F 70 61 72 2F 30 32 20 37 32 30 F7@1]
[midi@00:00.802: F0 00 20 32 32 2F 66 78 2F 32 2F 70 61 72 2F 31 30 20 35 30 F7@1]
[midi@00:00.828: F0 00 20 32 32 2F 66 78 2F 32 2F 70 61 72 2F 31 31 20 35 30 F7@1]
;
;VARIATION 2 - SYSEX commands with spaces, no MIDI Channel target
;[midi@00:00.750: F0 00 20 32 32 2F 66 78 2F 32 2F 74 79 70 65 20 44 4C 59 F7]
;[midi@00:00.776: F0 00 20 32 32 2F 66 78 2F 32 2F 70 61 72 2F 30 32 20 37 32 30 F7]
;[midi@00:00.802: F0 00 20 32 32 2F 66 78 2F 32 2F 70 61 72 2F 31 30 20 35 30 F7]
;[midi@00:00.828: F0 00 20 32 32 2F 66 78 2F 32 2F 70 61 72 2F 31 31 20 35 30 F7]
;
;VARIATION 3 - SYSEX commands without spaces, include a MIDI Channel target
;[midi@00:00.750: F0002032322F66782F322F7479706520444C59F7@1]
;[midi@00:00.776: F0002032322F66782F322F7061722F303220373230F7@1]
;[midi@00:00.802: F0002032322F66782F322F7061722F3130203530F7@1]
;[midi@00:00.828: F0002032322F66782F322F7061722F3131203530F7@1]
;
;VARIATION 4 - SYSEX commands without spaces, no MIDI Channel target
;[midi@00:00.750: F0002032322F66782F322F7479706520444C59F7]
;[midi@00:00.776: F0002032322F66782F322F7061722F303220373230F7]
;[midi@00:00.802: F0002032322F66782F322F7061722F3130203530F7]
;[midi@00:00.828: F0002032322F66782F322F7061722F3131203530F7]

Assuming Variation 4 was the 'right' way to do it based on earlier discussion, I fired up MIDI Monitor (https://apps.apple.com/us/app/midispy/id1444652196) and saw no SYSEX, only all the expected CC and PC that are also in the lyric sheet.

The only events I can't immediately account for are the 2 events listed as going to CH16 (not sure if this is relevant or not)
All other destaination MIDI Channels are accounted for:

MIDI Ch 2 - X32 Mixer effects bus sends/returns mutes
MIDI Ch 3 - guitar effects
MIDI Ch 5 - guitar effects
MIDI Ch 10 - drum trigger

Full list of MIDI events in my lyric sheet

; MIDI automation for Dunc's Hotone Ampero II Stage
[midi@00:00.234: CC28.127@3]
[midi@00:00.283: CC0.1@3]
[midi@00:00.307: PC69@3]
[midi@00:00.332: CC25.1@3]
[midi@00:00.356: CC56.127@3]
[midi@00:00.380: CC74.1@3]
[midi@00:00.405: CC75.0@3]
[midi@03:24.609: CC56.0@3]
;
; MIDI automation for Roland SPD-SX PRO
; Set to default Kit (Kit 200)
[midi@00:00.100: CC0.0@10,CC32.1@10,PC71@10]
;[midi@00:00.100: CC0.0@10]
;[midi@00:00.150: CC32.1@10]
;[midi@00:00.200: PC71@10]
;
; MIDI automation for Charlie's Helix LT
[midi@00:00.250: CC32.002@5,PC001@5,CC69.000@5]
[midi@00:00.432: CC94.127@5]
[midi@02:59.750: CC94.0@5]
;
; MIDI automation for Behringer X32
; Mute sends/returns for special effects on FX2, FX3, FX4
[midi@00:00.250: CC42.127@2]
[midi@00:00.276: CC61.127@2]
[midi@00:00.302: CC44.127@2]
[midi@00:00.328: CC62.127@2]
[midi@00:00.354: CC46.127@2]
[midi@00:00.380: CC63.127@2]
;
[midi@00:10.000: CC42.0@2]
[midi@00:10.125: CC61.0@2]
[midi@00:12.000: CC61.127@2]

The supprt email Log Files shows nothing of interest:

2025-10-16 17:19:14.927 🔵INFO  MidiHandler:321 [b] Connecting MIDI source MIDISpy (id 1916997628)
2025-10-16 17:19:37.435 🔵INFO  SongPlayer:197 [b] nextScheduleTime cleared (set to nil)
2025-10-16 17:19:37.436 🔵INFO  SongPlayer:1519 [b] Scheduling full playback from 0.000 to 180.250, adjustTime: false
2025-10-16 17:19:37.445 🔵INFO  SongPlayer:1258 [b] Re-scheduled audio segments at time 0.000
2025-10-16 17:19:37.447 🔵INFO  TrackPlayer:286 [b] Track 0: Starting playback at frame 0 on time 3794867949 volume 1.0 gain 0.0, lastrenderframe 3794867229, outputLatency 0.0010416667209938169
2025-10-16 17:19:37.448 🔵INFO  TrackPlayer:286 [b] Track 1: Starting playback at frame 0 on time 3794867949 volume 1.0 gain 0.0, lastrenderframe 3794867229, outputLatency 0.0010416667209938169
2025-10-16 17:19:37.449 🔵INFO  TrackPlayer:286 [b] Track 2: Starting playback at frame 0 on time 3794867949 volume 1.0 gain 0.0, lastrenderframe 3794867229, outputLatency 0.0010416667209938169
2025-10-16 17:19:37.451 🔵INFO  SongPlayer:1040 [b] Playback started in 15.64ms, delayAudioTime: 15.00ms
2025-10-16 17:19:51.206 🔵INFO  SongPlayer:1160 [b] Seeking to 13.734, isPlaying: false
2025-10-16 17:19:51.211 🔵INFO  SongPlayer:197 [b] nextScheduleTime cleared (set to nil)
2025-10-16 17:19:51.214 🔵INFO  SongPlayer:1519 [b] Scheduling full playback from 13.734 to 180.250, adjustTime: false
2025-10-16 17:19:51.318 🔵INFO  SongPlayer:1160 [b] Seeking to 0.000, isPlaying: false
2025-10-16 17:19:51.319 🔵INFO  SongPlayer:197 [b] nextScheduleTime cleared (set to nil)
2025-10-16 17:19:51.320 🔵INFO  SongPlayer:1519 [b] Scheduling full playback from 0.000 to 180.250, adjustTime: false
2025-10-16 17:19:51.322 🔵INFO  AppState:1087 [b] Rewound to beginning of song after manual stop

@peter what can I test, capture, send to you, to further assist troubleshooting assuming nothing above jumps out at you as a clear and obvious problem?

AnthonyB

peter Your syntax is wrong. This should work:

[midi@00:00.750: SXF0002032322F66782F322F7479706520444C59F7]

Bingo. Apologies if I missed some note somewhere saying 'Add 'SX' in front of the SYSEX command'!
OK, now that my functional test is working, I'll go and convert all my old MIDIFlow remappings into native SYSEX in ST4 and given it a thorough stress test.

But I guess I should add support for adding whitespace between the bytes. The messages are way easier to read with whitespace.

I don't mind either way. Some SYSEX commands will get really long and line-wrap in the lyric editor, so having a blank space between the bytes will only make that worse (i.e. potentially push things to additional lines).

jfduval

peter

Hello Peter, yes, I'm sure it works because I can do it with OnSong and Multitracker. It's the VoiceLive Play (blue) and not the VLP3. There's nothing in the user manual about this, but I assure you it works very well. As I mentioned in my other post, I can even do it with OnSong using a standard MIDI command (PC or CC) as a trigger from ST4.

peter

Can you compare the MIDI Monitor output of OnSong with the one from ST3? Is there a difference in the MIDI messages?

jfduval

I think we have a lead. The first photo is the sysex sent by ST4 with the three parts separated. The second photo is the one sent by OnSong, in a single block.

jfduval

Error, the second photo does not show the command sent by OnSong but the response from the VLP, I believe. I'll check...

jfduval

The message in the second photo is indeed from the VLP, when it apparently received the sysex from OnSong. I am currently unable to obtain (in the monitor) the sysex sent by OnSong. I'll try again later, I have to go to work.

peter

One other thing worth checking out could be to only send one sysex message at a time and see if that is processed correctly. Some devices need a bit of time to process sysex messages.

AnthonyB

peter One other thing worth checking out could be to only send one sysex message at a time and see if that is processed correctly. Some devices need a bit of time to process sysex messages.

Yeah, that is something I'm definitely focused on testing. Especially given SYSEX messages are a lot larger than other MIDI messages and so I might have to be more careful about how much time I leave between each message.

For example, on my X32 mixer, to load a specific effect into one of the effects block and then set all the parameters (e.g. maybe I use the same timed delay, but with different delay time and decay time, in different songs) is generally 12-13 SYSEX commands.

jfduval

Hello Peter, here is what I get from OnSong with the MIDI Monitor when I send the sysex sequence.

jfduval

...

jfduval

The same message sent by ST4.

Oh... There are some zeros missing in the one sent by ST4, and I just figured out why. It's the tool I used to remove the spaces. I'll try again by doing it manually.

AnthonyB

jfduval In case it is relevant, don't forget that you no longer need to remove the spaces between the bytes which removes the risk of 'bad' search & replaces. Peter made this change to allow the spaces a few builds ago.

jfduval

Thanks Anthony, I missed that information. I was about to give up... Unfortunately, it still isn't working, so maybe I'm doing something wrong. That's enough for me tonight. I'll take a fresh look at it this weekend! Thank you.

AnthonyB

I just did a bit of research on SYSEX for the TC Helicon VoiceLive stuff... OMG, what an absolute shitshow. I though having to deal with OSC to SYSEX conversion with my Behringer X32 mixer was bad! All this 'the commands have to be packed a certain way and certain size' mularchy with the VoiceLive stuff!!

Maybe someone else in this community using one of these devices that also wants to move to using native SYSEX commands from ST4 (might have previously used MIDI remapping with ST3) might want to join the party.

peter

So, am I correct that this issue just resolved itself? If yes, I would leave the thread open for further discussion but mark is as "closed".

jfduval

peter

Yes, of course. Thank you, Peter.

Thank you too, Anthony. I'll keep looking into it, and any help will of course be appreciated!