Did this scan come out alright?

Does it appear balanced nicely, not un-straight or titled in any way? Also, would you recommend any tweaks with regard to colors and such?

White balance adjusted in Levels, Color Adjust, Crop and Sharpen.

@alan_olander, your rendition looks good to me. Blacks are black and whites are white. You can't put in detail (the gloves) where none exists, all you can do is grey it out (which is self-defeating). Flat lighting with an on-camera flash is not very flattering, but got the job done at the time. I took the lazy way out and used Photoshop Levels to normalize each color channel separately. That establishes a good base point from which you can do more subtle modifications. The results are a little warmer than above, which is not necessarily bad. The first thing to go in old negatives is blue, and too much correction can leave face tones too red.

In "the day", I wrapped the strobe head with Saran Wrap to remove most of the UV, which has a disproportionate effect on skin tones. I'm guessing this photo is from the 60's or early 70's.

Basically, the "white" border should be white.

A simple click of the white level dropper on the lightest part of the border should do the trick. Followed by setting the grey balance from the gentleman's lapels.

I would assume that a facsimile of the unfaded and unyellowed original would be preferred to the aged version.

Personally I think Ray's restoration is darned good from a small JPEG. The nearest lady's face is a bit too blue and overexposed, but that's an artefact of the direct flash. It could be improved with some layer work and rubbing through with a light eraser brush.

Here's my interpretation. Simply adjusted WB, in DxO Optics Pro, using the dropper on the shadow part of the lady on our right (the border of the print is not the correct spot). I saves as a TIFF, ran it through piccure+, which cleaned up the OOF condition nicely. No recropping.

KTgbbQI_DxO-TIFF_picc by David Stephens, on Flickr

Working with a full rez file off the scanner, I think that the sharpening could be even better.

Basically, the "white" border should be white.

True, but only to a first approximation. The orange color is the leuco base, and is lighter wherever there is greater density in the negative, because developing the image consumes part of the base. Making the border white usually causes the image to be on the magenta side. I find normalization by channel more reliable than using the sampling droppers, unless you happen to have a grey card in the field of view.

Color balancing is very subjective, so practically anything you do can be either acceptable or unbalanced, depending on the subject and the personal taste of the viewer. Even with the finest calibration, you can never balance more than two primary colors, and often only one.

Francis sent me a scan and asked me if I would work on it for him...this is what I came up with.

True, but only to a first approximation. The orange color is the leuco base, and is lighter wherever there is greater density in the negative, because developing the image consumes part of the base. Making the border white usually causes the image to be on the magenta side. I find normalization by channel more reliable than using the sampling droppers, unless you happen to have a grey card in the field of view.

 

Color balancing is very subjective, so practically anything you do can be either acceptable or unbalanced, depending on the subject and the personal taste of the viewer. Even with the finest calibration, you can never balance more than two primary colors, and often only one.

The "white" border has no photo information. It's just a border. You need something within the image to set white balance of your subject.

"The "white" border has no photo information. It's just a border."

- Exactly!

And that's why it gives you a white reference. There's no exposed dye density, only age yellowing. It's the perfect baseline for what should be white in the image.

"The orange color is the leuco base, and is lighter wherever there is greater density in the negative, because developing the image consumes part of the base."

- That's mumbo-jumbo nonsense. The border has been through the development process too. Nothing is "consumed" unless the print is a Cibachrome. In the neg/pos process dye is generated where the emulsion layers are exposed. The border has had no exposure and therefore should be perfectly white.

This is obviously a scan from a print, and not a film scan. You can easily see that.

There will be handling marks, but these can generally be avoided when picking a point for white reference.

"The "white" border has no photo information. It's just a border."

- Exactly!

And that's why it gives you a white reference. There's no exposed dye density, only age yellowing. It's the perfect baseline for what should be white in the image.

.

For one, in this example, it's not white, it's faded orange. Second, you don't want white-white, you want medium grey, in the image itself. The border of the print paper was not impacted by the white balance in the image, so you don't use that.

The border of the print paper was not impacted by the white balance in the image, so you don't use that

Really? The border color is orange due to fading. What color do you suppose it was when new? Do you support what happened to the border occurred throughout the print? Sampling the border is only the first step of restoration, but I don't recall good color, the kind I get from an inkjet, ever occurred in machine prints of this era.

It doesn't seem to matter if you set color balance from a grey card or white, as long it's neutral and the white, in particular, contained no UV brighteners. That's true for calibrating scanners and cameras too. I would not trust anything "grey" in the image to be neutral unless someone put it there as a reference. If you shoot white (or any neutral reference) with a scanner or camera, auto exposure renders it as something like 12% grey. Try it sometime! Using the white sampling tool (eyedropper) sets the white point, and won't wash out detail unless by some miracle the image is brighter than the print base.

Conversely the black sampler seldom works as intended. Nothing is truly black in the real world, certainly not a tuxedo jacket If you use the jacket, for example, all detail of the fabric is lost. You need an artificial version of black, such as smoked ( with acetylene) steel or an industrial "black body", which is hollow, the size of a shoebox, with a small port for your optics. It doesn't even need to be blackened on the inside, but capable of withstanding heat up to about 3000 deg F. (It's used to measure the temperature of a furnace, using an optical pyrometer. A matte black color patch works well enough for photography.

Does it appear balanced nicely, not un-straight or titled in any way? Also, would you recommend any tweaks with regard to colors and such?

I prefer the crispness of Ray House's and others' post productions. (comment is addressing posts #4 and #5). Unless there is a particular purpose to maintain a 'vintage look', I think that most people want their 'old photos' to 'look better', Keeping this approach, if an old image is important enough to make and effort in detailed Post Production, I would also address the Flash Shadows

The original is tilted. It also appears to have a slight barrel distortion.

Working from the original file posted, above is a 15 minute rough indicative, the original for comparison on the left.

I straighten the image (2 degrees); colour corrected; removed barrel distortion; sharpened; cropped to remove LH wall structure; dodged and burned to enhance various features; removed flash shadows.

More detailed (i.e."better") results could be had with more attention, more time and if one were working from the original Print or Negative.

It would be interesting to know what exactly was scanned and if it were a negative, what was the size of the negative.

WW

Really? The border color is orange due to fading. What color do you suppose it was when new? Do you support what happened to the border occurred throughout the print? Sampling the border is only the first step of restoration, but I don't recall good color, the kind I get from an inkjet, ever occurred in machine prints of this era.

....

Ed, I looked around for your example and I can't find it. Is there a link, buried in your verbiage, that I missed?

If the orange of the border is "white" and the shirt and blouse are "white", then why aren't they the same?

Ed, I looked around for your example and I can't find it. Is there a link, buried in your verbiage, that I missed?

I see it now. Sorry for the miss.

So you had to post your criticism twice. So what's so great about the "yellow" vintage look, and the contrast and whites weren't any worse than the original posted image which is all Ray had to work with. Let's see your rendition.

I hit a key twice.

Which is better, negative or print? The negative, if you have it. Most old photos are long separated from their negatives, if indeed the negatives were kept. The film has a dynamic range of capture of 10-12 stops, which it compresses into a dynamic range of 6-8 stops on film. A print has about 3.5 stops dynamic range, or less if old and faded or poorly exposed.

The image I adjusted using Photoshop Levels is the 8th post in this thread. Below is adjusted using the white dropper in the upper border, which is evenly faded but not stained.