Max. Tiefe fehlerhaft?

Linus Torvalds torvalds at linux-foundation.org
Sat Feb 23 13:16:31 PST 2013


On Sat, Feb 23, 2013 at 12:58 PM, Arno Jansen <mail at arno-jansen.de> wrote:
>
> So the OSTC samples every 10sec (default setting in custom function 20 of the OSTC) but he also samples perhaps every second the actual deepness and saves the deepest point separately.

Note that there are other possible effects, and there is some reason
to believe that what you are seeing is possibly due to unit issues.

In particular many dive computers save their samples in a different
format from the "overview data". The reason for that can range from
just encoding issues (the sample data is often some delta-compressed
thing in order to save space) to actual unit and computational issues
(the sample data may be in pressure units rather than actual depth
units) to just pure confusion (the dive computer fundamentally senses
pressure and then converts it to depth for the UI, and the conversion
is done with a certain precision that is separate from the precision
that things are *shown* in).

For example, the overview of the dive data may well be saved in the
same prevision that it is shown to the user. Which may include various
rounding etc. The fact that the maximum depth is shown in tenth of a
meter, but the samples you have clearly have more digits of precision
(but not necessarily accuracy!) implies that there is likely rounding
and/or unit conversion going on.

For example, some dive computers internally save all the samples in
feet (and not necessarily because they are american either - it's
simply because saving samples in whole meters is obviously not
sufficient precision, while feet may be a measure that is
"sufficiently close" to the precision of the sensor taking things like
unknown ambient pressure into account etc). When we download the dive,
libdivecomputer will turn that into standard format, and will give us
depth in meters - and it has a lot of digits of precision for that
reason, even though it is *accurate* only to roughly 0.3m.

But then for the "overview", the dive computer may well save things in
some metric unit (like decimeters), just because it uses a different
data format.

So your "18.5m" max depth can be a mixture of all these issues. Higher
internal sampling rates, different units, and different rounding for
the "overview" view as opposed to the samples.

                 Linus


More information about the subsurface mailing list