[Qgis-developer] Feedback for contrast enhancement ... enhancements ?

[Mathieu Pellerin]

Please don't change the default behavior for 3-band rasters to min/max;
while it does make in many scenarios for 1-band rasters (DEM, precipitation
value, vegetation index, etc.), this is a terrible default value for most -
if not all - satellite imagery products (i.e. Sentinel-2, LANDSAT-{5,7,8},
ASTER, etc.).

IMO, we need to have two separate default option: one for 3-band raster
min/max (which doesn't change), and one for 1-band raster (which is changed
to min/max).



On Fri, Sep 2, 2016 at 7:00 PM, Nyall Dawson <nyall.dawson at gmail.com> wrote:

> On 2 September 2016 at 21:50, G. Allegri <giohappy at gmail.com> wrote:
> > +1
> >
> > I would also make min/max the default option, because I've has many users
> > (mostly newcomers) were struggling too find out why some pixels were not
> > displaying as expected. It was because the statistics weren't including
> > some, few, important pixels.
>
> Yes please! There's so many users on stackexchange who get confused
> with this same issue.
>
> Nyall
>
> > Anyway I know that, generally, cutting the tails provide a better visual
> > result.
> >
> > giovanni
> >
> >
> > Il 02/set/2016 12:39, "Even Rouault" <even.rouault at spatialys.com> ha
> > scritto:
> >>
> >> Hi,
> >>
> >> There has been interest expressed to improve the usability of the
> contrast
> >> enhancement function. Basically there are workflows where you want the
> >> min/max
> >> values used by contrast enhancement to be updated each time you pan or
> >> zoom
> >> the canvas. From what I've seen, here's what currently exists:
> >> - a long method (6 clicks): through the Style tab of the layer
> properties
> >> - a faster method (2 clicks): through the "Strech using current extent"
> of
> >> the
> >> contextual menu of the layer panel.
> >>
> >> When you have several rasters loaded, even the 2-click method is
> >> inconvenient,
> >> so there's a need for a 0-click solution (once the layers have been
> >> configured)
> >>
> >> Another shortcoming in the current implementation is that the settings
> of
> >> the
> >> "Load min/max values" foldable group are not persistant, which requires
> >> re-
> >> setting them if you're not happy with the default Cumlative count cut
> >> method.
> >> And if you use the "Strech using current extent" menu item, the genuine
> >> min/max values are used (not the 2-98% cumulative count cut)
> >>
> >> So I was thinking to something along the following lines:
> >> - make the min/max settings persistant
> >> - remove the "Load" button and "Clip extent to canvas" checkbox, and
> >> replace
> >> them by 3 radio buttons to determine the scope of statistics : "Whole
> >> raster",
> >> "Current canvas" and "Updated canvas". See the attached
> >> proposal_min_max.png.
> >> The function of the Load button would be replaced by the general Apply /
> >> OK
> >> buttons.
> >> - when the user manually enters the min/max values, the 2 groups of
> radio
> >> buttons (method to compute min/max and scope of statistics) would be
> >> unchecked
> >> so that is is clear that they don't come from computed statistics. I
> also
> >> think that when you select "Current canvas", once the Apply/OK buttons
> >> have
> >> been pushed, the checked state of Current canvas shouldn't be saved.
> This
> >> way
> >> when you display the layer properties you have a good idea of where the
> >> current min/max values come from.
> >> - make the "Strech using current extent" method honour the way min/max
> are
> >> computed in the min/max settings instead of using systematically the
> >> min/max.
> >> So when "Updated canvas" would be selected, it would have no (extra)
> >> effect as
> >> expected.
> >>
> >> Any opinions ?
> >>
> >> Even
> >>
> >> --
> >> Spatialys - Geospatial professional services
> >> http://www.spatialys.com
> >>
> >> _______________________________________________
> >> Qgis-developer mailing list
> >> Qgis-developer at lists.osgeo.org
> >> List info: http://lists.osgeo.org/mailman/listinfo/qgis-developer
> >> Unsubscribe: http://lists.osgeo.org/mailman/listinfo/qgis-developer
> >
> >
> > _______________________________________________
> > Qgis-developer mailing list
> > Qgis-developer at lists.osgeo.org
> > List info: http://lists.osgeo.org/mailman/listinfo/qgis-developer
> > Unsubscribe: http://lists.osgeo.org/mailman/listinfo/qgis-developer
> _______________________________________________
> Qgis-developer mailing list
> Qgis-developer at lists.osgeo.org
> List info: http://lists.osgeo.org/mailman/listinfo/qgis-developer
> Unsubscribe: http://lists.osgeo.org/mailman/listinfo/qgis-developer
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.osgeo.org/pipermail/qgis-developer/attachments/20160903/a9de3adb/attachment-0001.html>