[OSM-talk] Continuous audio in JOSM on tracks without waypoints
david at frankieandshadow.com
Wed Feb 20 18:16:42 GMT 2008
To augment continuous audio synchronized with waypoints that I mailed
about earlier this week, I have now also added the facility to work with
continuous audio recordings on tracks where you don't have or don't want
to use explicit GPS waypoints. This will be in tomorrow's JOSM build.
After loading your GPX track, right click on the GPX layer and choose
the new "Make Sampled Audio Layer" option. This will ask for your WAV
file; then it will create a new layer combining the audio track with the
GPX trackpoints to produce a set of audio markers laid out along the
track. These will be at least 15 seconds and 75 metres apart, or
whatever values you choose for these in Advanced Preferences settings
for "marker.audiosampleminsecs" and "marker.audiosampleminmetres"
You're then in a similar position to applying audio to explicit
waypoints as per my previous changes: you can synchronise to a marker
near the beginning of the track, play by reference to the visual
position on the map, jump forward and back in the commentary, pause and
resume and so on.
The sampled markers are named according to the time offset from the
beginning of the sound track (e.g. "1:37", "1:09:07"). To facilitate
this, I've reversed the default for whether to show text for button
markers (audio, image and web), but you can turn these off as before,
transiently from the right button layer menu, or permanently by setting
marker.buttonlabels to false in Advanced Preferences.
(*) the defaults are chosen so that they are about the same for a
cyclist travelling at 5 metres per second (about 11mph or 18km/h), so
you get a useful but not overwhelming number of samples, but if you stop
or slow down, you don't suddenly get a concentration of points close
together. Of course if you stop and then record intermittently, you'll
still find it hard to locate the bit of commentary you want - the whole
idea is it is related to landmark junctions or loops you make in the
road or whatever. If you're in a car you might want to set the sample
time a bit shorter, say 7.5 in an urban environment, though the distance
would probably be the same unless you actually want a higher or lower
density, and maybe 60 to 90 seconds apart if walking at a typical 1 m/s.
More information about the talk