[Elphel-support] Best way to save consecutive jpg's to disk?

[Brian Burke]

The "several lines of mandatory vertical blanking" - how does that affect
the final composite image?  Does it result in gaps in the image, or rather a
fraction of time between sensor frame captures (meaning some small fraction
of time passed between two adjacent 2 line frames)?

Also - what do you mean by "it is very easy to get camera wait for very
long"?  Do you mean the camera will complete a composite frame, and then get
"stuck", not realizing it has to start another?  If so, would having
"narrower" composite frames (fewer sensor frames per composite frame) help?

Thanks for the quick reply.



On Thu, Mar 3, 2011 at 11:55 PM, Andrey Filippov <
support-list at support.elphel.com> wrote:

> Brian,
>
> In the camera you can specify how many sensor frames will be combined into
> one frame for processing - and sensor frame can be as narrow as 2 lines
> (with several lines of mandatory vertical blanking, unfortunately. There is
> no high-level software, and it is very easy to get camera wait for very long
> - driver is interrupt-based and interrupts are generated once per composite
> frame. There are two tips to recover from the camera getting stuck with
> that:
>
> 1) with parsedit.php change default "3" to "0" frames ahead, in that case
> the new parameters will be written to the sensor immediately, not to the
> sequencer that will send them after several frame interrupts.
> 2) "killall -9 lighttpd" - that will restart web server and php (there are
> only 3 instances currently enabled in lighttpd.conf) if all php instances
> are waiting for something
>
> As for recording - camogm in *.mov mode is most efficient. It is rather
> easy to split mov into JPEG sequences - just using start of frame/end of
> frame tags, as in the php script Oleg recently posted.
>
> Andrey
>
>
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://support.elphel.com/pipermail/support-list_support.elphel.com/attachments/20110304/8123221c/attachment-0002.html>