This web page describes most of the efforts I've made so far, trying to come up with a libusb-based method of retrieving voice recordings from an Olympus VN480PC Digital Voice Recorder

.wav file as output by Olympus' software Suspected binary version of voice data copied from the USB Sniffer (during the largest bulk transfer, seemingly implying that this is the payload the other parts of the conversation are there to facilitate) and converted from hexdecimal to binary USB Sniffer Output File (I've sed'd in the meanings of the endpoints to help simplify the analysis process) My summary of what I believe happened during the transmission
retrieve-hello-delete-all.usbsnoop.log
hi.wav retrieve-hi-no-delete.binary retrieve-hi-no-delete.usbsnoop.log retrieve-hi-no-delete.summary
one.wav, two.wav retrieve-one-two-no-delete.usbsnoop.log retrieve-one-two-no-delete.summary
retrieve-whistle-no-delete.usbsnoop.log retrieve-whistle-no-delete.summary

  • An analysis of the lengths of the resulting .wav files (as output by Olympus' windows-only software) and what appears to be the part of the USB Sniffer output that pertains to the voice data most directly.
  • My best guess at this time, as to the general algorithm used to transfer the data (still in an unknown codec?) from the voice recorder to a windows machine
  • The result from when I symlinked the relevant-looking data from the USB sniff to a bunch of file extensions, and fed them to sox. Nothing useful turned up. :(
  • The python code I used to generate a bunch of .wav containers wrapping the relevant-looking data in the USB sniff, to see if it's a common codec used in .wav's. And digging through those .wav's for something useful. Nothing useful turned up. :(

    Hits: 3305
    Timestamp: 2024-04-18 04:20:44 PDT

    Back to Dan's tech tidbits

    You can e-mail the author with questions or comments: