Digitizing Old Photos

[rpcrowe]

I have just located many hundreds of old family photos which I want to digitize.

My end product will be to distribute the digitized photos to various family members and to include these photos to illustrate my family tree.

I assume that the quickest way to digitize the photos will be with the use of a scanner but, that copying the photos with a digital camera and good lens might give the best results. IS THAT TRUE?

If the better results with the camera would be worthwhile, what would be the best way to keep the photos flat during copying?

This is an example - shot hand held using my Sony A6400 and Tamron 28-75mm lens. I would, of course do a setup with a stand or a tripod when copying the photos.



I wonder if this setup might be worthwhile?

https://smile.amazon.com/Photography...cx_mr_hp_atf_m

[Geoff F]

I digitise hundreds of images for my local history society. Mostly using a good quality scanner (Epson Perfection V800 Photo Scanner) for up to A4 size. This is the easiest method but it is expensive if you are purchasing a good quality scanner and aren't doing a lot of copying. For bigger items or anything within a photo frame etc which cannot be opened, I shoot with my Canon 7D Mk 11 and a suitable lens.

With camera copying, I always use a tripod and usually place the image vertically on a chair, or similar item, and as near to being vertical as possible without any risk of falling over. Images without frames are 'stuck' onto a piece of plywood, or similar object, using Blu Tack. (What is that substance called in other parts of the world?). That achieves a similar result to the expensive bit of kit in your link.

The camera must be set squarely on to the image. Check that horizontal and vertical edges are straight, not tapering. Getting a suitable lighting method and angles can be tricky. The light needs to be equal across the image and not creating a colour cast.

Sometimes, daylight is best but careful use of fluorescent tube lighting can also work when the image is carefully placed. But check your white balance. Glass covered images are particularly difficult to light correctly and many angles with a variety of light sources, may need to be tried before good results are achieved. More than one light source may be required.

My tripod allows for shooting directly above a subject but this can cause lighting and shadow problems. Two light sources, from opposite sides can work.

Photo negatives really require a bit of special kit but suitable options usually come with good quality scanners. As a last resort, I have photographed positive photo slides by cutting out a suitably shaped hole in a piece of card to hold the slide and placing a good light source behind the slide. Also requires a tripod mounted camera and a bit of experimentation.

Whatever method is employed will require great care and experimentation will be necessary before you find something which works for you.

[Donald]

I scan old photos with an Epson V500 scanner. The results are perfectly adequate. The trick is getting mastery of the Epson Scan, the software that comes with the unit.

[DanK]

I'm no expert, but I've done some, and by chance, that's what I'm working on this week.

We've done it both ways. Based on my meager experience, the main time a camera has an advantage is if the old photo has a surface that's both shiny and highly textured. We've had some where the close lamp in the scanner picked up that pattern.

Often, though, a scanner works just fine, and it's easy. I use a Canon CanoScan 9000 II, which I bought years ago and which is probably inferior to newer ones. For photos, it has a maximum resoltion of 4800 x 4800 (23 Mpx), which is adequate for most purposes and seems to exceed the detail in the old photos. It's smart enough to recognize the edges of the photograph and not waste sensor area on the background.

The key for me, which I only realized this afternoon, is that I guessed wrong about the default scanner settings. I was getting substantial JPEG compression artifacts. I've now set the software to store in TIF format if I specify that the document is a photo, and that solved that problem. So I'm sticking with the scanner for most things. I do have one print from roughly 1890 that is far too big for my scanner, so I'll have to do that with the camera or take it to a lab.

What is difficult is that if the photos aren't in great shape, processing is a whole new world for me, very different from the editing I'm used to doing. For example, many of the photos look washed out, and a gentle use of a multiply blend mode, globally or locally, often works better than simply raising the black point. And of course, you have to be ready to do a lot of tedious healing and cloning.

[Ken MT]

Blu Tack. (What is that substance called in other parts of the world?
Blu Tack in New Zealand

[xpatUSA]

The Scanner on my all-in-one Canon MG 8120 resolves 4800x9600 film thingies and comes with a 6-count 35mm film holder which rests over the scanner platen, so up to six captures a go. It works well enough on my color negatives and they're all I have after a good few moves ...

[DanK]

Richard,

I blew this up in the lightbox and dropped it into Photoshop to look at it more closely. I thought you had reflections from the surface, which is a big issue in copying photos, but when I blew it up, it looked more like severe JPEG artifacts:



This isn't what old photos look like when magnified; I think this is an artifact of the digital capture and storage. My unsolicited advice is that if you use a camera, shoot raw, and if you use a scanner, save in TIF format. I don't know why this is so, but some of the old photos I've copied seem particularly prone to JPEG artifacts like these. I wonder if it might stem from the fact that many old photos have areas with very limited tonal variation, but that's just an ill-informed guess.

Dan

[xpatUSA]

I thought you had reflections from the surface, which is a big issue in copying photos, but when I blew it up, it looked more like severe JPEG artifacts:



This isn't what old photos look like when magnified; I think this is an artifact of the digital capture and storage.

Interesting - because the posted image is not Chroma sub-sampled i.e. it is Y'CbCr 4:4:4 (best) according to Harvey's ExifTool, and the Quality is about 90% according to XnView. At those values, I would expect JPEG artifacts to be barely noticeable.

Looks to me like digitized grain.

[xpatUSA]

Blu Tack. (What is that substance called in other parts of the world?
Blu Tack in New Zealand

Over here we have many, many kinds including the original Bostik stuff:

https://www.walmart.com/c/kp/sticky-tack-for-walls

I use UHU White Tack ...

[DanK]

Interesting - because the posted image is not Chroma sub-sampled i.e. it is Y'CbCr 4:4:4 (best) according to Harvey's ExifTool, and the Quality is about 90% according to XnView. At those values, I would expect JPEG artifacts to be barely noticeable.

Looks to me like digitized grain.

This is very important to those of us doing this stuff, but I don't entirely understand what you're saying.

By "digitalized grain" do you mean posterized grain?

Richard shot this with a 24 Mpx camera, so there should have been at last 30,000 px on that face. There are far fewer squares in the blown up image of the face. What's causing that huge reduction in data, and why are the blocks so big relative to pixel size?

I don't know the answer, given the low level compression, but it has to be something in the processing or storing of the data, no?

[xpatUSA]

This is very important to those of us doing this stuff, but I don't entirely understand what you're saying.

Basically, I'm saying that the noise in the OP does not look like "severe JPEG artifacts". Here is a comparison between the OP and it converted to a much lower quality sub-sampled JPEG from the same original:



Each crop above is up-sized 800% with no smoothing (i.e. Nearest Neighbor). Therefore each group of 64 pixels on your screen represents one unchanged original image pixel.

By "digitalized grain" do you mean posterized grain?

Yes, I guess so - I meant to write "digitized grain".

Richard shot this with a 24 Mpx camera, so there should have been at last 30,000 px on that face. There are far fewer squares in the blown up image of the face. What's causing that huge reduction in data, and why are the blocks so big relative to pixel size?

My prior crop was up-sized 2000% with no smoothing, i e. no interpolation. Therefore, each original pixel was represented by 400 of your screen pixels all of which will have the same value as the original one.

In this post, the crop was up-sized at 800% to more clearly show any JPEG 8x8 DCT blocks. Therefore, each crop contains the same number of pixels but each pixel is represented on your screen by 64 screen pixels with no interpolation.

I don't know the answer, given the low level compression, but it has to be something in the processing or storing of the data, no?

I disagree.

I don't even know what that "something" could be ... anybody else?

My claim is that Richard's camera captured the surface of the photo well enough for his stated purpose.

Please note above that, in particular, the main effect of a low-Q fully chroma-sub-sampled JPEG is a smoothing of detail, not addition of artifacts - other than the DCT 8x8 blocking.

[DanK]

Richard,

here is a simple setup for photographing photos that you don't want to scan. This photo is far bigger than my scanner, so I had no choice. It required no equipment I didn't already have. I used a cheap (maybe $10) remote release. The black cloth is from my background, but normally, it wouldn't matter; I used it because of the curved cuts in the corners.



BTW, the photo itself is a mystery. There is no one living who knows for sure, but I believe it was taken around 1885 in Lithuania. I have no idea what the process was. It's on canvas, but the emulsion is thick enough that it is completely flat, with no trace of the canvas texture. The surface is chalky flat, with no reflection at all. In places, it looks almost like a drawing or painting, but the surface shows that it can't be. I may see if a local museum curator will take a look.

The boxes scattered around are ones I cut open this week to look for more old photos. I've found some, but none anywhere as old as this one.

Dan