Re: Time synchronization [was: GPS geotagging: first steps to implementation]

[Jan Girlich <vollkorn cryptobitch de>]

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

Am 06.09.2010 12:29, schrieb Matija Cizmek:
> On Tue, Aug 31, 2010 at 11:16 AM, Jan Girlich
> <vollkorn cryptobitch de> wrote:
>> Am Dienstag, den 31.08.2010, 07:52 +0200 schrieb Mike Gemünde:
>>> Am Dienstag, den 31.08.2010, 01:17 +0200 schrieb Jan Girlich:
>>>> This is just a preparatory step for syncing GPS data and
>>>> pictures, for which accurate EXIF times are crucial.
>>>
>>> Just let the user adjust the time. If he or she locates one
>>> photo properly, you exactly know the difference of UTC and Exif
>>> time. So, I don't see a problem there. From my experience, Exif
>>> time is normally far away from be set properly. I often forget
>>> to do the clock change and also my camera clock is anyway some
>>> minutes out of sync.
>>
>> I am sorry, but from this statement I must assume you did not
>> understand my idea at all.
>>
>> You just repeated the problem I'm trying to solve with my idea:
>> the camera clocks are way off. I proposed a semi-automatic way to
>> correct the EXIF timestamps properly.
>>
>> Positioning one photo correctly on a map is not a very good
>> solution because people usually spend more time than one second
>> at one spot to take a picture. So by placing one picture
>> correctly you only determine a time frame of several minutes,
>> which is not accurate enough and causes the user to position
>> several pictures until he pinpointed the time properly for the
>> whole import roll.
>>
>> My way for time synchronization has two advantages:
>>
>> 1) More accurate (and easier to use, I think).
>>
>> 2) Can be implemented right now, all the necessary libs are
>> available (No need for a map widget).
>>
>> Cheers Jan
>>
>
>
> Hi, sorry for late contribution to this discussion, but your
> solution seems really complicated. What I do on my trips is that
> when my GPS acquires signal lock I take a picture of it's screen
> with visible time displayed. Later, I can easily calculate the
> time difference by comparing file time with photograph content.
That's exactly the idea. But instead of taking a picture of your GPS,
you take a picture of a 2D barcode encoding the time (this barcode
changes every second) and the time difference is calculated
automatically. So it's much easier than your usual way since it
automates half the process, I'd say. Also it works for people who have
a GPS tracker without any display like mine.

What the user would do:
1. Start f-spot and hit a button added by the extension. A 2D barcode
changing every second will  show up.
2. The user takes a photo of his screen with the barcode.
3. The user imports the photos of his camera. The user is done here.
4. The f-spot extension automatically reads the barcode, calculates
the time difference and sets the pictures properly.

It's as simple as that.

Cheers
Jan
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.10 (GNU/Linux)
Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org/
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=39ie
-----END PGP SIGNATURE-----