Depth reading - DC vs Subsurface mis-match

[gobbledegeek]

Hi Folks

I imported the dives on Hollis DG03 on OSX and I find an inconsistency in the max depth shown on my DC screen vs Subsurface 4.9.1.

For example I am attaching a screenshot of a dive that my DC told me was to 39.8m - it was an unplanned deco dive due to down currents. The same dive imported into subsurface shows 38.5 in the RHS graph while  the LHS information pane shows a max-depth of 40m??!

Attaching the screenshot, the dive log ssrf and a snapshot of my DC dial screen....

What could be the issue here?

Thanks

Gobbledegeek

[Paul Buxton]

Quickly browsing the  ssrf file

There is an enrty

<depth max='39.795 m' mean='21.301 m' />

But the actual log of the dive itself is in 30 second intervals, and the deepest any of these are is at

<sample time='5:00 min' depth='38.519 m' temp='23.333 C' ndl='2:00 min' />

So I think the 40m max depth reported is the depthmax entry rounded up.

and the 38.5 in the RHS graph is deepest point in the sample data rounded up.

Paul

--
You received this message because you are subscribed to the Google Groups "Subsurface Divelog" group.
To unsubscribe from this group and stop receiving emails from it, send an email to subsurface-dive...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/subsurface-divelog/3edc7caf-5a7f-4068-ae2c-eb1bfd001a23%40googlegroups.com.

[gobbledegeek]

On Friday, August 30, 2019 at 3:44:46 PM UTC+5:30, Paul Buxton wrote:

Quickly browsing the  ssrf file

There is an enrty

<depth max='39.795 m' mean='21.301 m' />

But the actual log of the dive itself is in 30 second intervals, and the deepest any of these are is at

<sample time='5:00 min' depth='38.519 m' temp='23.333 C' ndl='2:00 min' />

So I think the 40m max depth reported is the depthmax entry rounded up.

and the 38.5 in the RHS graph is deepest point in the sample data rounded up.

Paul

Thanks. I missed the max-depth before the sample entries. Is this the DC doing the rounding/mean or Subsurface? I did set the sampling interval to 30s so looks like its the DC doing it -  and since I pulled myself  up from 39.8 before 30s when I realized I had been pushed down deeper by the current, I guess it is not logged in the running series of the sample data.. is that the right interpretation?

G 

[Paul Buxton]

Yes if reached your  max depth between sample points then it will not necesarily show up in the time series log, it would likely depend on how your computer logs the depth, it could be an average over the sampling period, snapshot at that instant, or something else. I do not have enough background knowledge to say.

So I think there are two levels of rounding going on.

Your dive computer is rounding from 39.75 to 39.8 when it is displaying the max depth.

And Subsurface is rounding from 39.75 to 40 when it is displaying the max depth. Internally the depth looks to be stored in mm, so the rounding is just to make it easier to read.

Hopfully someone with more knowledge of Subsurface will correct me if I am wrong! :-)

Paul

--

You received this message because you are subscribed to the Google Groups "Subsurface Divelog" group.
To unsubscribe from this group and stop receiving emails from it, send an email to subsurface-dive...@googlegroups.com.

To view this discussion on the web visit https://groups.google.com/d/msgid/subsurface-divelog/a1b6b988-2622-4d8e-b2e6-78e43f6a3c92%40googlegroups.com.

[Dirk Hohndel]

A few more details for this one:

- many dive computers keep a max depth reading that is the "real" maximum depth reached during the dive, even if it happens between samples (you should almost always expect this to be different from the deepest sample since as far as I know most dive computers report "average depth during the sample interval" as the sample depth

- the UI rounds values when displaying them. Given that most pressure sensors only have about a 1-3% accuracy, giving more than two significant digits for depth is kinda foolish. Yes, we store the depth in mm (as that makes calculations easy) and never round those stored values, but that doesn't mean that these values really have 5 valid digits.

/D

[Linus Torvalds]

On Fri, Aug 30, 2019 at 3:39 AM gobbledegeek <gobble...@gmail.com> wrote:
>
> Thanks. I missed the max-depth before the sample entries. Is this the DC doing the rounding/mean or Subsurface?

Most (but not all) dive computers give a separate max depth from the
maximum sample depth. As Paul noted, it may be that the maximum
happened between samples, and the dive computer itself may be sampling
depth every second, but only saving them as samples every 10-30
seconds.

(Whether the sample depths are averaged for that sample interval or
not, I don't know. Might be).

Some dive computers will also give us an average depth independently
of what we can calculate using the samples.

We will *fix* whatever maximum depth that the dive computer gives us
if we find a sample that is deeper (because then clearly the dive
computer was just wrong). But if the dive computer reports a max depth
deeper than the samples, we'll just assume that "yeah, there was
something going on between samplkes".

We won't really take it into account in any other way. It doesn't
affect our SAC rate calculations etc - those are purely based on
samples. But you can see a max depth that is deeper than any sample
you can find.

Linus