wiki:Help/AudioMapping/SeparateClipsByTime

Version 4 (modified by skyper, 3 years ago) ( diff )

removed unneeded empty lines and white spaces; use trac listings

Audio Mapping: Audio clips using audio file time stamps

This method of audio mapping uses the modified time stamps of a set of separate WAV files recorded while surveying to make a set of audio markers at the corresponding times on the GPX track.

Calibration is critical because the clock in the voice recorder which applies the time stamps is being matched to the times recorded for points along the track. However, you cannot synchronize in JOSM with this method, so set the recorder's clock carefully when you start.

Also, try to use an setting on your audio recorder that makes WAV files in a format that JOSM can understand directly - if you have to transform the files, it is hard to keep the time stamps (though the Unix command touch is your friend if you do have to do this).

Note that the modified time stamp is taken to refer to the time at the end of each recording, but the audio markers are placed at the beginning of the recording. Make sure you get the time zones of all the equipment set so it is all consistent. If your times are off by an hour, chances are it is a time zone or daylight savings / summer time problem.

Procedure:

Before you start

  1. Calibrate your voice recorder. Though a recorder may not be very accurate, it is unlikely to vary much, so you will probably only need to do this once.

While surveying

  1. Obtain a fix with your GPS.
  2. Set the clock in your recorder to exactly match the time shown on the GPS display.
  3. Do your surveying. Whenever you want to identify a location, start the recorder, dictate notes and stop the recorder again.

On the computer

  1. Extract your tracks from the GPS as a GPX file, and your sound track from the recorder as WAV files, taking care to preserve the modified time stamps on the files.
    • JOSM (actually, Java's built-in audio facilities) doesn't recognise every variety of WAV file encodings. If you need to (JOSM will tell you), convert your recording to a suitable format using e.g. Audacity. 8,000 16-bit samples per second is a reasonable format. You will need to apply the timestamps of the old files onto the new ones if you have to do this; the Unix command touch can do this easily if you use consistent naming. On Windows, you can get touch as part of cygwin.
  2. Open your GPX file. This will create a GPX layer showing the track.
  3. Make sure the Audio Preference "Modified times (time stamps) of audio files" option is checked.
  4. Import Audio using the context menu (right mouse button) for the GPX layer. This then asks for your WAV files. Select all the relevant WAV files (Ctrl+Shift + click to add and remove files in the selection). This should a produce Marker Layer containing one audio marker for each file labelled with the file name (excluding the extension .wav).
    • if you don't see labels with the Markers, check that Show/Hide Text/Icons on the context menu for the Marker Layer is not off.
  5. Make the map using your commentary. Click on the markers to play the corresponding audio. You can move the play head within any one recording's range and use the jump forward, fast forward and slow forward audio controls, but you can't drag the play head to positions between recordings.
    • If you start finding play back isn't happening where you expect:
      1. you may not have calibrated correctly
      2. you may have been speaking too soon or too late while mapping: you can insert a lead-in time to compensate for this using 'Lead-in time' in Audio Preferences.

Back to Audio Mapping
Back to Audio Menu
Back to Main Menu
Back to Main Help

Note: See TracWiki for help on using the wiki.