I'm a programmer doing a client project, and I'm trying to write an app that will do this (the photograph of the woman) to user-uploaded images. I think it's some combination of Sepia and some other filters, but I'm not sure. Anyone have any ideas?
-
To clarify: you're asking for the names of the effects that have been applied to that photo - including the colouring, texture, high-contrast, etc? (So that you can then go and find out how to make them work in automated image processors like imagemagick?)– user56reinstatemonica8Commented Mar 14, 2013 at 11:23
-
Yes, thanks. I've tried sepia, sketch, and a few others in imagemagick, and it is part of the way there but doesn't really look right.– mlstudentCommented Mar 14, 2013 at 11:24
-
Have you been through Imagemagick's options page? imagemagick.org/script/command-line-options.php– user9447Commented Mar 14, 2013 at 15:54
-
@graphicsman (s)he's not asking for programming help, as this isn't a programming Q&A site. They're asking us about what needs to happen, generally, to make a photo look like that (then, they will go and figure out how to program a similar process in whatever they choose to use).– user56reinstatemonica8Commented Mar 14, 2013 at 18:47
3 Answers
I'm not sure what the effects would be in ImageMagick, but I'd start by trying to create the following steps:
- Increase contrast quite a lot
- Slightly blur
- Add a transparent texture for the paper
- Overlay coffee rings and misc stains (random placement if possible)
- Overlay a lighter texture to give the mottled effect on the darker areas
- Sepia tint/filter
-
1+1 I'd suggest maybe reducing saturation before increasing contrast but other than that looks good to me Commented Mar 14, 2013 at 12:52
-
Interesting idea - what effect would lowering the contrast first produce? Commented Mar 14, 2013 at 14:32
-
If you increase the contrast without first reducing the colour saturation, then depending on how it's done you might find that areas with relatively lighter hues like greens end up too light and areas with relatively darker hues like reds end up too dark (or it might make no difference, or have the opposite effect, it depends on how the contrast is done: it's something to play with and see what comes out better) Commented Mar 14, 2013 at 14:51
-
The real trick here is the source photo: it has been vignetted and selected for suitability.
That said, you can provide the user with a way to create an alpha mask in order to isolate the portion they want to feature. The effect allows for finger-painting level accuracy.
In Photoshop, you can get the source ready by doing something like the following. (layers, in bottom to top order):
- Base Photo
- very light grey layer (such as 245,245,245) with a layer mask to isolate the portion you want
- black & white adjustment layer with colors tweaked brighter (suit to taste)
- color fill adjustment layer with a dark brown/purple light similar to the darkest colors in sample. Set the layer mode to "color"
- brightness contrast adjustment layer with contrast ramped near 100 and brightness tweaked brighter
- levels adjustment layer with darkest level nearer to mid-point and mid point near 80%
- photo filter adjustment layer with a color set to the "paper color" in sample, do not preserve luminosity, density 100%
The overlay and scale of the coffee mug stains and watercolor spreading are just icing.
I am sure you can reduce or combine steps. The mask is the only user interaction needed, but some course control over contrast etc might be required to really polish it..
- Print photo
- Place coffee cup on it a few times - create stains
- Scan photo
DONE! ;-)
-
I believe the OP is looking for a way to script the process per
I'm a programmer doing a client project, and I'm trying to write an app that will do this
– user9447Commented Mar 14, 2013 at 17:37 -
2Also, add steps to leave the photo out in the sun for 7-8 years so that it's faded but not too faded, near a source of humidity so it's slightly wrinkled but not too wrinkled... Sounds like a labour-intensive process to do to every photo every app user uploads ;-) would need a very unusual business model to be profitable... Commented Mar 14, 2013 at 18:51