How to inject gps data into DV videocam via firewire ?

Hi,

I can control my DV videocam from Ulead video studio run on PC via
firewire
like play, stop, rewind, record and the like.
I would like to use firewire to have my remote gps device to inject
gps data in video frames in record mode
to have them geotagged.
I don't expect each videocam showing such feature
but i would be nice to build and use such gps videocam.

Any idea how to send anything from PC to videocam using firewire
link ?

Darius

On Wed, 16 Jan 2008 17:30:27 -0800 (PST), Darius


Why not ask in one of these groups:

alt.video.dvd.authoring
rec.video

?

Yes, those commands are defined and pretty much universally implemented
by the firmware in video cameras.


It seems like there are a number of problems you'll have to solve for
this to work:
- Is there reserved space in the DV video data stream for camera
geographic location (I'd guess not).
- If not, is there any room in the data stream to add location (e.g.
data bits that are recorded and transmitted, but currently undefined)
- How will you capture this data on playback to use it?
- How are you going to write new firmware for the video cam, or modify
the existing firmware, to support recording position data?

The third one may not be too hard; there is probably open-source software
for reading DV data streams. You might have to modify a Windows driver,
which will be a bit more difficult.

But the last item is the killer. Without access to the original source
code for the video camera firmware, plus some information about the
hardware, adding an extra feature to the firmware is beyond the
abilities of most people. Have you done anything like this before?


There's probably a simple interface for this. If you can solve the
other problems above, this is likely to be trivial.

        Dave

On Jan 17, 9:24 pm, da...@cs.ubc.ca (Dave Martindale) wrote:

Thanks Dave,

First I need to test real gps-DV videocam to see how it works.

Darius

Do you know of any such device that exists now?

Also, do you have to do this is real time? If not, you could take
somewhat the same approach as is often used for geotagging still camera
images: Carry a GPS that records a track log. Set the time of day on
your video camera from the GPS, so there's at most a few seconds of
difference. Shoot as usual, then transfer video to computer later.
During the transfer, or as a second pass later, figure out the time of
each frame (via time code in DV data), where the GPS receiver was at
that time (interpolating between track log records), and attach that
location to the frame.

There will still be some time discrepancy between GPS receiver time and
video time code, but if you shoot a few seconds of video of the time
display on your GPS receiver at the beginning and end of the day, you
can probably calculate the video clock error and its drift rate pretty
accurately. Remember that many consumer GPS receivers *also* have some
error in the displayed time.

        Dave