Another oil painting question:

At my job I am instructed to split up shots of a painting and then merge them together in Photoshop after to achieve the highest resolution possible.

I'm doing it and the results are adequate but why does it feel so wrong?

Am I really getting better images through this method rather than just shooting a full shot of the painting?

Is the color balance affected by this method?

  • "split up the shots" it is not clear what that means. Explain what that is and how you are doing it. Do your employers have any experience or real knowledge of photography at all? i am starting to question there directives to you. Do they know what resolution is? If you are using a decent newer camera then it has more then adequate resolution for online images for the purpose of sales. What, if anything, about your photos led them to ask you to improve them? – Alaska Man Jul 10 at 18:58
  • Shooting art and in particular oil painting is challenging. You need a solid tripod (as you know). You need to have the art and the film or sensor plane lined up just right (true and parallel, each corner, or top bottom sides, of the piece equidistant to the sensor) Preferable on a black wall with and the light strobes from both sides, (equal as well), should have polarizing gel filters, a polarizing filter on the lens in order to eliminate the reflections from the many many different angles of the shiny oil paint. All other lights in the room should be off when exposing. – Alaska Man Jul 10 at 19:17
  • @AlaskaMan Maybe split up the subject would be a better way to put it. Fill the frame with the left 1/3 of the painting, next shot is the middle 1/3 and another shot of the right 1/3 of the painting. The camera is a Canon EOS 5D Mk III. Lens is not well suited for the task but the Art Director is stingy so it'll be a while before I can convince them of any lens changes. They really haven't asked me to improve the images, I'm taking that initiative myself. – Eric Jul 10 at 19:53
  • @AlaskaMan As for my boss' experience, I'm not sure. I know he's got a Master's is Graphic Design but that's hardly the same thing. And yes I question everything this guy directs me to do but he's stubborn and loud. I'm hear to elarn because this photographing of oil paintings is very interesting to me and they haven't set a very high bar in terms of quality so I though I could pad my resume by helping these guys out (as much as I can with my limited experience in any art job). – Eric Jul 10 at 19:58
  • 2
    How does the position of the camera change between the different views? 1/ The position is fixed and the different views are obtained by a rotation. Around which point is this rotation made? 2/ The camera is moved laterally and/or vertically between the different views in order to keep the painting and the camera sensor parallel to each other. – hpchavaz Jul 11 at 10:43

Shooting a reproduction shot as tiles is absolutely normal, and standard operating procedure. It's just not done with the sort of makeshift and low-budget set-up you've been describing to us over a number of questions very often, at least not on a regular/professional/production basis. (And it was a regular thing before digital photography as well.)

Ideally, you would want a setup very similar to a rostrum camera. Either the camera or the painting you're shooting can be the moving element - normally, the thing to be copied is on a table that can be moved in X and Y directions if it's relatively small, and for larger or more fragile things the camera is moved (using a gantry on tracks for a horizontal set-up, or a vertically-adjustable stand on tracks for a vertical set-up). Using a tripod, things get a bit trickier, especially if you're not free to build a jury-rigged track-and-dolly system. Basically, you're being told to do the right thing, but not being allowed to use any tools that would allow you to do the right thing. Which is a pity, since the cost of making an adequate shooting setup - the materials cost, at least - would be about the same as the utterly inadequate tripod solution (which is almost never the right choice for working in a single location/studio). No, you wouldn't be able to take it outside or on location, but you'd be able to process a lot more items a lot more quickly.

| improve this answer | |
  • There are also very large flatbed scanners... – xenoid Jul 13 at 15:34
  • @xenoid There are limits to them. Paintings and hangings can be... rather big. Also, they're rather a lot more expensive than a camera, a stand and a dolly & track. – user93022 Jul 13 at 15:38

Consider that for a good result you have to:

  1. have some overlap between the shots
  2. fix the lens vignetting so that the overlap pixels have the same exposure
  3. fix the lens distortion so that the pixels really overlap
  4. interpolate pixels between the two images in the seams

So, you may have more pixels, but on the seams the image quality can be lower than what you would get with a single shot.

| improve this answer | |
  • For 2 and 3 should I go for a manual approach or does computation do an accurate job? I try to use the vignetting tool but it seems like I have zero control over it and feels more like when I used to abuse filters back when I was a kid. Same for lens distortion and I'm definitely fixing distortion manually. Am I hamstringing myself by ignoring the tools I have at my disposal for fixing those two aspects? – Eric Jul 13 at 14:27
  • Tools like Hugin do this automatically, but you get even better results if the image is corrected first using actual data obtained by calibration (lhis can be done by these tools, too, as part of a pre-processing step). – xenoid Jul 13 at 15:24
  • Never heard of Hugin but I'll check it out in the next round of photos. The grea thing about the company I work for is that the artists I work with can pump out so much art for me to practice reproducing them. – Eric Jul 13 at 20:56

You can and some cameras even do it automatically now. With a full single shot of the subject, the resolution you get is at most the maximum of the camera. It is usually slightly lower due to optical imperfection and even many cameras have an anti-alias filter to reduce sharpness slightly in order to avoid moire artifacts.

When you take multiple photos of either the whole image or part then you are gathering more data. These need to be assembled into a higher resolution output which is the principal for panoramic photography. What you get though is not the sum of resolution due to many reasons but you certainly should get more than a single shot.

Images have to overlap so that they can be matched and the common areas blended together. This blending processes the image so that details appear uniform which reduces a little sharpness compared to each shot on its own. Still, just take 4 photos will give you more than double the resolution. For best results, everything has to match between shots, exposure, focus, white-balance and any parameters like tone curve. Modern panorama software can compensate for some differences but this is still processing pixels which results in reduced image quality.

Similar results can also be achieved by taken multiple times the whole subject but with the camera slightly offset between frames. This has the added benefit of capturing more color information since the sensor can be moved to ensure that each color on the Bayer filter is used for all points in the image. This is called Super Resolution and requires processing in order to merge images together. Many mirrorless cameras currently do this.

| improve this answer | |

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.