[Fwd: [Subsurface-trac] #486: Manually entered profile not transferred correctly to divelogs.de]

[Jef Driesen]

On 2014-04-07 11:59, Rainer Mohr wrote:
> Am 07.04.14 11:42, schrieb Jef Driesen:
>> I'm not sure if you are aware of this, but there are dive computers 
>> out there that don't have a fixed sample rate. Some (older) Oceanics 
>> support a depth based sample rate. They record a new sample whenever a 
>> certain depth threshold is exceeded. The timestamps of such samples 
>> are not regularly spaced.
>> 
> True, the smarttrak data format also has a "repeat x times" if the
> depth does not change IIRC
> Fact is though, that until today, every DC data format that I have an
> import for does give me regular samples inside their own format, even
> if the computer does not specifically record one (except for the
> Smarttrak format, where i create x dummy samples for the "repeat x
> times").
> 
> if there is need for irregular sample intervals, I will be happy to 
> adapt

Technically, the Uwated sample interval is still regular. Omitting 
samples where nothing has changed is basically some simple form of data 
compression technique to reduce the amount of storage. After 
decompressing the data again, you end up with evenly spaced samples 
again. That's what libdivecomputer does internally for the Uwatecs.

For the Oceanics the situation is a bit different. It's a real event 
based mechanism, without a base sample rate (like the 4s used by the 
Uwatecs), at least not that I know about. To make things worse, is that 
the resolution of the recorded timestamp is only 1 minute. The result is 
that you can end up with zero or more values per minute. Currently 
libdivecomputer tries to spread those evenly within that minute. Far 
from perfect, but it does result in something reasonable. Multiple 
values with the same timestamps is probably even more confusing for 
applications. But the consequence is that the sample interval is 
variable.

I can provide some data if you want to have a look.

Jef