Is it broken on an M as well?

If you post a user/library file we can load it on our M8X and take a 'look see'.

Is this a Montage performance? The SoundCloud description appears to indicate it is a MODX performance but you posted in the Montage forum which is why we ask the question.

Hard to say without examining the perf. But if this is a Modx perf that breaks on a Montage some places to look:

1. functional areas that are different between Montage and Modx.

2. the Tx/Rx screens for all parts where the CC switches are located.

3. dump the montage version and compare the dump to the dump from soundcloud. Any differences?

As you may know from doing dumps certain SysEx commands can appear to freeze things: sending a bulk header but never sending a footer - the system just waits forever for the rest of the data. So unlikely, but there could be some embedded SysEx in the dump that causes the same effect.

Try sync-save-system init-load and see if the perf and montage work properly from a factory reset and then local load of the saved perf that was downloaded.

Suggesting the above test since you didn't indicate if the 'freeze' problem only occurs immediately after the download/sync of the perf or not. The above test should indicate whether it is the performance itself that is causing the issue.

Given the various reported SoundMondo issues our hypotheses would be that any issue is due to the soundmondo/sync/download/conversion process rather than the stored content of the perf.

... just being lazy and waiting for the SysEx mirror site to index this perf.  Last indexing done on 12/27 and doesn't have this yet.

Here is a copy of the JSON data from SoundMondo for that performance, which includes the sysex. I've also processed the data into .syx files for Montage and MODX, so you can play them directly to the synth using something like Sysex Librarian.

https://drive.google.com/drive/folders/10D8VbwQ183BSPGOyxEWKOJRLltK5lRzQ?usp=sharing

The metadata suggests that it was created with a MODX.

Extremely confusingly, the problem I (and others) were having appear to be gone for me right now for reasons I cannot explain.

Regardless, this performance did cause major problems, so I'd be interested to see if it causes similar issues for you guys.

M8X won't load that Montage sysex file. If you post a user/library file  we will try loading and looking at it.

Extremely confusingly, the problem I (and others) were having appear to be gone for me right now for reasons I cannot explain.

What does that mean? Did you repeat the EXACT test of a new sync/download and it is 'gone'? Or did you just load a saved version and it is 'gone'? If the internet link (webmidi) is still connected but had failed it could be interfering with other operations.

But loading the saved download after a power down, power up, cycle is different than the environment exists after a failed sync.

Did you rule out what we mentioned?

Given the various reported SoundMondo issues our hypotheses would be that any issue is due to the soundmondo/sync/download/conversion process rather than the stored content of the perf.

Not enough info to know just what steps you followed for each test you did. Did the initial problem occur AFTER:

1. sync/download of perf

2. save perf

3. disconnect cable used for download

4. power down and then back up

5. load perf

6. use PC to change to another perf

7. instrument frozen

As noted before there have been past reports of SoundMondo/WebMidi issues around syncing and downloading. So it wouldn't surprise us if the sync/download itself never completed properly.

There have also been reports of problems with model info in sysex files. The Montage sysex file you posted won't load on a M8X. MIDI-OX shows it being sent but there is never any 'BULK DATA' being received message on the M. But we haven't seen any official source that says an M can load a sysex file of an earlier model.

Does a sync/download on a Modx manifest the problem?

What 'others' were having the problem and how does their process compare to yours? Same instrument model?

In the past of I was doing experiments of "sort of" invalid stuff, I could notice normal things breaking.  Maybe it's a combination of other experimental actions combined with the Performance that is the real culprit and lack of those other things has ghosted the problem.

Not sure,  but I can look at it on my Montage.

A little background:

KDx has a tool that loops through user/library performances in order to extract their names for use with the Live Set editor (the Live Set sysex contains just mdb/lsb/program addresses, rather than names).

A MODX user reported to me that his sync resulted in hundreds of performances in the KDx db with the same name.

This is indicative of KDx being unable to instruct the synth to load a new performance via MIDI program change - which causes the same name to be read for every address visited after any performance that prevents program changes.

After checking the obvious things (program change being disabled in Part 1 > Part Settings > Receive Sw), I asked the user to push the offending performance to SoundMondo so I could investigate it.

After I synced the performance to my Montage, I too was unable to instruct the synth to change program via MIDI by any means, even after manually switching to a different performance.

Eventually I had to reboot the Montage in order for program change functionality to be restored.

When I came back to further investigate the behaviour today, I can no longer reproduce the problem - program change works fine now, but yesterday I had exactly the same problem experienced by the original user, on repeated attempts.

FWIW, Montage M / MODX M absolutely can not load sysex from Montage / MODX - the parameter addresses are completely different, and even the address length is different (4 bytes for M, 3 for classic)

We have the same fundamental question: did the problem occur immediately after the sync/download of the perf from soundmondo?

After I synced the performance to my Montage, I too was unable to instruct the synth to change program via MIDI by any means, even after manually switching to a different performance.

So has the problem EVER occured after:

1. download the perf from soundmondo

2. save it

3. power down

4. power up

5. load the saved perf

We are still wanting to know if you have ruled out the synd/download issues reported in the past. The soundmondo download could be leaving the instrument in an infinite wait loop waiting for the final 'bulk footer'.

You can reproduce that problem by sending the instrument a 'bulk header' - you won't be able to do anything else until the instrument receives the 'bulk footer' saying the transmission is complete.

Once you initiate a 'bulk load' to the instrument it essentially switches 'modes' to code that handles the reception of the bulk data. Until that code gets its 'done' signal (bulk footer) it will wait forever. There  is no 'timeout' as far as we know.

That is a much different problem than a standalone perf in the user area that, when loaded, locks everything up.

SoundMondo uses sysex so anything that prevents a 'bulk footer' being received correctly can cause the issue. That is why we are asking if the problem has ever occurred with that performance when soundmondo isn't involved.

Nor does Soundmondo translate between M and non M cousins.

Posted by: Anonymous

↑

That is why we are asking if the problem has ever occurred with that performance when soundmondo isn't involved.

SoundMondo only became involved when I directed the user to push the problematic performance to SoundMondo so I could inspect it.