"Scanning" with an SLR camera (NSFW images)
14 years ago
Sometimes you can't carry a scanner around with you. Maybe your scanner is too large. Maybe your scanner requires wall power. Maybe your scanner was one of those awesome slim glass-top ones but it started developing stuttering and streaking and you can't buy a new one anymore because all anybody makes nowadays are those huge horrible generic box scanners GOD DAMMIT YOU BEAUTIFUL BASTARD WHY DID YOU DIE ON ME ;_;
Sometimes you can't carry a scanner around with you. Those of you who saw me at FC may have seen me taking pictures of sketches with my camera, a Nikon D3000 (bottom-of-the-line SLR, I got it at the Costco). I've had a few people ask me how to get pictures of sketches taken with a camera to come out not all blurry and washed out. This is how I do it:
First, camera settings. Flash on, small aperture, short exposure. You want to use your flash so you won't need a tripod, and room light is usually not to be trusted. You want your aperture small so distance variation won't cause blur. You want your exposure a little on the short side so no part of the image gets completely washed out.
Next, arrangement. If your flash isn't external or indirect, you'll want to take your picture at a 20-30 degree angle, because both ink and pencil can reflect light harshly from head-on. You'll also want to make sure that nothing is casting a strong shadow on the page; Even with the flash on, these can show.
If all goes well, your picture should look something like this example photo of a sketch
argonvile put in my book:
http://nanimoose.furryhome.com/stuf.....LR-Demo-01.jpg
This picture makes a good example for this technique because there are large swaths of very light gray coloring throughout the picture that I want to preserve. If you try doing plain levels on it now, some of the light grays will wash out while other parts of the image are still too dark.
We load the image into Photoshop or some similar image editing program. We select the canvas and do free transform, nudging the corners (ctrl-click) until the page is completely rectangular (move it to the edges of the canvas to check for straightness), and also stretching the sides to get the correct aspect (it helps to hold up the original and compare). My example's transformation box looks like this:
http://nanimoose.furryhome.com/stuf.....LR-Demo-02.jpg
If properly transformed, you should be able to do a rectangular clip at the edges of the page and not get more than a few pixels worth of page edges, like so:
http://nanimoose.furryhome.com/stuf.....LR-Demo-03.jpg
The basic theory of what to do from here on out is to make a layer that represents the colors we'd see on a blank white page under identical lighting conditions. The difference between the fakey blank paper layer and the real photo layer should be just the image on the page.
To accomplish this, I first make a new layer and set its Fill to 0%, so it can be edited but not seen (you could also just put it behind the photo layer). Then I go in with a picker and pick the color at a place where the blank white page can be seen. I switch over to an airbrush tool and put down a large, soft patch of that color at the same place on the hidden layer.
http://nanimoose.furryhome.com/stuf.....LR-Demo-04.jpg
I repeat this sample-and-airbrush process on other white areas around the page until I've covered all the real estate of the image to some degree. After this, my "blank page" layer looks something like this (transparency changed to red for effect):
http://nanimoose.furryhome.com/stuf.....LR-Demo-05.jpg
I set up a macro to do "duplicate layer" followed by "merge down". I run this macro until no more transparency exists in the layer:
http://nanimoose.furryhome.com/stuf.....LR-Demo-06.jpg
So that completes our fakey blank paper layer. Keen eyes may be able to see that it's ever so slightly darker around the edges, especially in the upper left. That difference is subtle, but important! So now, if we take this paper layer, making sure it's fully opaque and above the photo layer, and set its blending mode to "Difference", we get this:
http://nanimoose.furryhome.com/stuf.....LR-Demo-07.jpg
Yeah, that's inverted. Easily fixed; Merge the layers together and then invert the whole thing again. Then we can use levels to do finishing adjustments. I tweak levels to trim just a hair off the bright end to account for noise, turning the paper pure white; and also trim off from the dark end enough that ink shows up as pure black. Like so:
http://nanimoose.furryhome.com/stuf.....LR-Demo-08.jpg
So there you go! An off-kilter, under-exposed photo from an easily portable camera is transformed into a vivid scan-like image that can be presented online without shame (content aside). Our final product looks like this:
http://nanimoose.furryhome.com/stuf.....OnATuesday.jpg
Sometimes you can't carry a scanner around with you. Those of you who saw me at FC may have seen me taking pictures of sketches with my camera, a Nikon D3000 (bottom-of-the-line SLR, I got it at the Costco). I've had a few people ask me how to get pictures of sketches taken with a camera to come out not all blurry and washed out. This is how I do it:
First, camera settings. Flash on, small aperture, short exposure. You want to use your flash so you won't need a tripod, and room light is usually not to be trusted. You want your aperture small so distance variation won't cause blur. You want your exposure a little on the short side so no part of the image gets completely washed out.
Next, arrangement. If your flash isn't external or indirect, you'll want to take your picture at a 20-30 degree angle, because both ink and pencil can reflect light harshly from head-on. You'll also want to make sure that nothing is casting a strong shadow on the page; Even with the flash on, these can show.
If all goes well, your picture should look something like this example photo of a sketch

http://nanimoose.furryhome.com/stuf.....LR-Demo-01.jpg
This picture makes a good example for this technique because there are large swaths of very light gray coloring throughout the picture that I want to preserve. If you try doing plain levels on it now, some of the light grays will wash out while other parts of the image are still too dark.
We load the image into Photoshop or some similar image editing program. We select the canvas and do free transform, nudging the corners (ctrl-click) until the page is completely rectangular (move it to the edges of the canvas to check for straightness), and also stretching the sides to get the correct aspect (it helps to hold up the original and compare). My example's transformation box looks like this:
http://nanimoose.furryhome.com/stuf.....LR-Demo-02.jpg
If properly transformed, you should be able to do a rectangular clip at the edges of the page and not get more than a few pixels worth of page edges, like so:
http://nanimoose.furryhome.com/stuf.....LR-Demo-03.jpg
The basic theory of what to do from here on out is to make a layer that represents the colors we'd see on a blank white page under identical lighting conditions. The difference between the fakey blank paper layer and the real photo layer should be just the image on the page.
To accomplish this, I first make a new layer and set its Fill to 0%, so it can be edited but not seen (you could also just put it behind the photo layer). Then I go in with a picker and pick the color at a place where the blank white page can be seen. I switch over to an airbrush tool and put down a large, soft patch of that color at the same place on the hidden layer.
http://nanimoose.furryhome.com/stuf.....LR-Demo-04.jpg
I repeat this sample-and-airbrush process on other white areas around the page until I've covered all the real estate of the image to some degree. After this, my "blank page" layer looks something like this (transparency changed to red for effect):
http://nanimoose.furryhome.com/stuf.....LR-Demo-05.jpg
I set up a macro to do "duplicate layer" followed by "merge down". I run this macro until no more transparency exists in the layer:
http://nanimoose.furryhome.com/stuf.....LR-Demo-06.jpg
So that completes our fakey blank paper layer. Keen eyes may be able to see that it's ever so slightly darker around the edges, especially in the upper left. That difference is subtle, but important! So now, if we take this paper layer, making sure it's fully opaque and above the photo layer, and set its blending mode to "Difference", we get this:
http://nanimoose.furryhome.com/stuf.....LR-Demo-07.jpg
Yeah, that's inverted. Easily fixed; Merge the layers together and then invert the whole thing again. Then we can use levels to do finishing adjustments. I tweak levels to trim just a hair off the bright end to account for noise, turning the paper pure white; and also trim off from the dark end enough that ink shows up as pure black. Like so:
http://nanimoose.furryhome.com/stuf.....LR-Demo-08.jpg
So there you go! An off-kilter, under-exposed photo from an easily portable camera is transformed into a vivid scan-like image that can be presented online without shame (content aside). Our final product looks like this:
http://nanimoose.furryhome.com/stuf.....OnATuesday.jpg
You could have easily done the same "tutorial" in either as a plain adult tagged upload or if it had to be a journal entry, then with SFW content.