Skip to content

Latest commit

 

History

History
38 lines (37 loc) · 2.18 KB

notes.md

File metadata and controls

38 lines (37 loc) · 2.18 KB
Raw

Reverse engineering process for PMIDIPD30:

  • Couldn't find editor software on Pyle website.
  • Noticed that the bank change sysex message matches Korg Nanokontrol.
  • Found the Nanokontrol MIDI implementation chart, tried other sysex messages, without luck.
  • Contacted Pyle Support, got Chinese factory email address after two weeks.
  • Emailed factory, no response so far.
  • Factory email address was @worlde.com.cn, remembered that identical hardware is also sold as Worlde EasyControl.9.
  • Found pointer to editor software on Worlde website. Worlde website was down at the time, but Wayback Machine had saved a copy.
  • Mac version of software got stuck when trying to upload a new config to device.
  • Looked at saved configs, no sign of sysex content.
  • Recorded a memory dump of editor software, no sign of sysex content.
  • Tried Windows version via wine. Saw MIDI messages go by, but update still failed.
  • Had a hunch that the software doesn't like USB hubs. Tried Mac version again, with the device plugged in directly, and now the update succeeded.
  • Sort of. The device was updated, but the config was broken (transport buttons didn't work, etc.).
  • Turns out that updates aren't done with sysex but with a strangely indirect protocol piggybacking on MIDI note messages.
  • Looking at the messages and the nature of the misconfiguration, noticed that some of the data seemed misaligned by one byte.
  • Added a filler byte in what seemed like the right place, sent the corrected config to the device (via Web MIDI API), and it worked. Whew...
  • Once I had a working config and a way to upload it, the rest of the format was straightforward to reverse engineer (see pmidipd30.js).

Device issues:

  • Encoder counts 0, 2, 3, ... going up, but ..., 3, 2, 1, 0 going down.
  • The encoder channel is the channel of whichever controller was used just before it, i.e., the encoder is pretty much useless when using multiple MIDI channels.
  • By default, faders 7 and 8 map to CC 9 and 10, just like crossfader and encoder, so they're useless unless reconfigured (that was the main motivation for figuring out the configuration protocol).