I am sorry in poor English.
I am a programmer for the editor “VOX Tone Room” using JUCE framework.
I noticed that checking Android 8.0 Oreo on Nexus 5X / 6P,
but Android 8.0 Oreo’s MIDI port seems to stall with a few Kbytes sysEx receive.
It is strange that once you close and open that MIDI port it will return to normal.
JUCE’s MIDI receiver class did not specify the buffer size in the constructor.
Just to be sure, specifying the default 2GByte or setting the smaller value to MaxMessageSize was the same result.
Please advise me who should buzz on this JUCE forum or ask Google.
Can you let us know exactly which SysEx you are trying to write? We regularly test NOISE (which is built entirely with JUCE) on a Nexus 5X on Android Oreo. ROLI Blocks require hundreds of sysex messages to communicate with each other and we haven’t seen this particular issue yet.
Thank you fabian.
A parameter change of about 10 bytes is sent from VOX amp to Android App, but several tens of consecutive occurrences occur in large quantities only in case of FX type change. Only then will you stall.
Next week, I will try experimenting equipment using the latest MidiTest so please wait for about a week.
here is a sysEx data.
- VOX Adio sysex implementation document
- Nexus5X console log
- diff and APK package of MidiTest which added console output of sysEx made for comparison with ToneRoom.
but, MidiTest did not stop.
It will not be possible to judge by this alone, so I will investigate a little more.
ToneRoom has been made for several years and it has been variously adapted, so now it is JUCE 4.3.1 base, picking up Android related differences up to 5.0.1 from GitHub and patching it by hand. Since the project is still Android_Ant, I have to go to Gradle and be able to debug
I tried replacing these three OSs,It seems not to be improved.
Android Oreo 8.0.0 build OPR6.170623.023
Android Developer preview 8.1.0 build OPP6.171019.012
Android Oreo 8.1.0 build OPM1.171019.011