Open andliszmmu opened 4 months ago
Can confirm with reproducible example below (colormodel="cmyk" results in legend values being darker than the raster), but... is this not because the RGB hues are beyond the CMYK range?
pal8 <- c("#33A02C", "#B2DF8A", "#FDBF6F", "#1F78B4", "#999999", "#E31A1C", "#E6E6E6", "#A6CEE3")
example_map <- tm_shape(land, ylim = c(-88,88)) +
tm_raster("cover_cls",
values = pal8,
col.legend = tm_legend("Global Land Cover")) +
tm_layout(legend.only=TRUE)
tmap_save(example_map, filename = "example_map_cmyk.pdf", device = pdf(), colormodel="cmyk", width = 205, height = 260, units='mm')
tmap_save(example_map, filename = "example_map_rgb.pdf", device = pdf(), colormodel="rgb", width = 205, height = 260, units='mm')
Thank you, marine-ecologist!
but... is this not because the RGB hues are beyond the CMYK range?
Generally, no. RGB and CMYK looks absolutely different because one is adopted for eye/display and the second for print. However the range of colours are the same. Nevertheless, I checked coincidence of the legend and map in wide spectrum of colours, as well as in different functions. The problem with tm_raster repeats with any colours, including "red", "green" etc. Besides the problem appears with tm_raster() only. For example, tm_fill() works properly.
Also confirm. This answer may be useful here: https://stackoverflow.com/a/68920794
The strange thing is that the raster image is not affected. Perhaps this cannot be easily converted to CMYK so it remains in RGB (?). Polygons, lines, dots, etc also get other colors:
pal8 <- c("#33A02C", "#B2DF8A", "#FDBF6F", "#1F78B4", "#999999", "#E31A1C", "#E6E6E6", "#A6CEE3")
example_map <- tm_shape(land, ylim = c(-88,88)) +
tm_raster("cover_cls",
col.scale = tm_scale_categorical(values = pal8),
col.legend = tm_legend("Global Land Cover"))
tmap_save(example_map, filename = "example_map_cmyk.pdf", colormodel="cmyk", width = 205, height = 260, units='mm')
tmap_save(example_map, filename = "example_map_rgb.pdf", colormodel="rgb", width = 205, height = 260, units='mm')
example_map2 <- tm_shape(World) +
tm_polygons("HPI",
fill.scale = tm_scale_intervals(n = 8, values = pal8))
tmap_save(example_map2, filename = "example_map2_cmyk.pdf", colormodel="cmyk", width = 205, height = 260, units='mm')
tmap_save(example_map2, filename = "example_map2_rgb.pdf", colormodel="rgb", width = 205, height = 260, units='mm')
``
@marine-ecologist note the difference with our code: tm_scale is needed to pass on `col8`. The second difference is that `device = pdf()` is not needed. Somehow, with `device = pdf()`, the device is not closed correctly (maybe a bug in `tmap_save`) or at least something to be improved....
I am working on an Atlas with a number of raster maps. I want complete correspondence between the colours in the legend and on the map, naturally. I need CMYK pdf for print. When I produce a pdf, colours in the legend and the map are very different. The problem is absent if I save the map in RGB. Shorten code: model.pal <- c("#4DAF4A","#BCE395","#FFFF33","#FD9243","#DF171C") ... model <- raster(paste0("./tif/",species.name,".tif")) ... tm <- ... tm_shape(model)+ tm_raster(style='fixed', breaks = c(0.445,0.521,0.635,0.729,0.85,1.2), palette=model.pal, alpha = 1, n=5, stretch.palette = FALSE, title = species.name, legend.show = TRUE, legend.reverse = TRUE, legend.z=0)+ ... tmapsave(tm, filename = (paste0(dsn,file.name,'', file.suff,'.pdf')), device = pdf(), colormodel="cmyk", width = 205, height = 260, units='mm')
I guess that it is not a problem of pdf device, since both legend and the map come there from tmap.