Digitising Photo Collections

This is an edited version of a presentation by James Walters of Vivid Recollection at the History and Genealogical Society of Canberra in April 2017. This presentation outlines the benefits of scanning, some technical aspects of the scanning process, and what can be done once this material is digitised.

Reasons to digitise photos
Why use a good scanner?
DPI or PPI?
What makes a quality digital image?
Scanning resolution guidelines
Scanning slides and negatives
Scanning software
Scan in TIFF format
Adobe Lightroom
Adobe Photoshop
Cloud storage
Giclée printing
Final thoughts

 

Reasons to digitise photos

There are four main reasons to digitise photographs and documents:

  1. Safeguard against loss by creating digital copies for storage on cloud servers;
  2. Share images with family and friends / wider audience;
  3. Digitally enhance and restore the material;
  4. Enable printing of high quality duplicates. These can be used as backup, or the originals can be stored and the duplicates put on display.

We’re all familiar with news footage of fires and floods in Australia, with disbelieving homeowners picking over the debris. When interviewed, victims rarely regret the loss of their plasma screen TVs, leather sofas or dining sets. These things can be replaced. The real treasures are so often the photos, and these are the things sorely missed.

Archival boxes and albums are good at keeping out light, dust, bugs and other slow-working agents of entropy, but are ineffectual when Mother Nature unleashes her fury. Thus, our first and primary reason for digitising photos: avoid the single point of failure!

Don’t let a rogue storm wipe everything out

Once we’ve created a set of digital copies we can make this material available to a wider audience online. We can convert the images to web resolution JPEGs and share on social media, and on platforms like Flickr. Adding metadata to images assists with cross-referencing, and facial recognition software (now embedded in most popular photo programs) can help us put together missing pieces of the puzzle.
Top

 

Why use a good scanner?

A scanner with high quality optics accurately captures detail, colour and tonal information in the original print or negative.

The ‘prosumer’ revolution is putting high-end technologies into the hands of home users and small organisations, and scanning is no exception. The output from an $800 scanner today might well be indiscernible from that of a $30,000 drum scanner a few decades ago. And they’re much easier and much quicker to use.

Our workhorse scanners are the Epson V700 and V800. We often notice a big difference in the quality of the images we scan and the digital files that are sent to us from customers with cheaper scanners. We can perform a few digital tricks to enhance these inferior scans, but the results are never as good.

Making claims about quality is all well and good, but I wanted to show some evidence for the variation between scanners. We used scanned the same photo with the V700 and a 10 year old all-in-one Canon MP600.

Note: We talk here about specific models but ‘V700’ may be more or less interchangeable with other ‘prosumer’ scanners on the market today, and MP600 with the average product in the $200-$300 range.

I was expecting to be able to point out the vastly better result from the V700. However the MP600 scan was surprisingly good:

Left: Canon MP600, right: Epson V700

The V700 better captured the sepia tone of the photograph; and there was less contrast between shadows and mid tones. There wasn’t a discernible difference in the sharpness of the image. There was slightly less detail in the highlights in the MP600 scan.

The differences were far more apparent in scanning a Kodachrome slide from the 1950s. We put the V700 head to head against a cheap slide scanner from Aldi.

This is what we got:

Aldi scanner slide scan comparison with Epson V700

Left: Aldi el-cheapo scanner, right: Epson V700

The scan on the left is very clearly inferior: the clouds in the background are gone, yellowy colour cast, and the highlights are badly blown out.
Top

The need for speed

One of the most compelling reasons to use a good scanner is speed. We set the clock on the preview scan of both the MP600 and V700.

The preview scan is a good metric for measuring speed, as this action has to be performed with every scan to take a snapshot of what’s on the platen.

The V700 comes in with a snappy 6.5 seconds for the preview – at least 3 times quicker than the agonisingly slow MP600. Multiply that over hundreds of impressions when scanning a whole collection and you see how the difference in time taken may be measured in days.

Tip: there’s a lively trade in second hand scanners on eBay and Gumtree and you can buy one here and sell when the scanning is done. Good scanners also hold their value nicely and you may almost recoup the costs on a new one.

Top

 

DPI = PPI

Dots per inch = pixels per inch

The term DPI, or dots per inch is a print industry term referring to the number of ink dots printed in the space of a square inch. The standard print resolution is 300dpi. This is the resolution at which at normal viewing distance the eye resolves the dots into continuous tone, i.e., can no longer discern the dots that compose the image.

Billboards are printed at a much lower resolution as they’re usually viewed from much further away.

PPI is the correct term for resolution in the digital space, as images are made of pixels, not dots. Even so, DPI is still the commonly used term.
Top

 

What makes a quality digital image?

From a technical standpoint there are seven determinants in image quality:

  • The camera used to take the photo (Hasselblad or Brownie?)
  • The skill of the photographer (choice of favourable lighting, sharp focus, settings etc.)
  • Darkroom technique – correct focus, exposure and developing (in the case of prints)
  • The condition of the photo / slide / negative
  • The optical quality of the scanner
  • The scanning resolution
  • Lossless / lossy image formats (e.g. TIFF or JPEG – we’ll return to this)

It’s a common misconception that scanning at 800 or 1200dpi will produce a crisper digital image. But as we can see above, resolution is only one factor among many. In a lot of cases, this inflation of pixels only results in unnecessarily huge files.

It’s said that a photographic print contains about 300dpi worth of information. To test this claim I carried out some scans on a 3×4” photo to see what differences were discernible at various resolutions.

All images were scanned in TIFF format and enlarged 200% to help with close examination. The 72dpi and 150dpi versions are clearly very pixelated. The scratches on the higher resolution images are more apparent, but there’s not much difference between the 300dpi and 1200dpi image in terms of the clarity of the subject. The glaring difference is in the file size: 3 MB at 300dpi and 48 MB at 1200dpi!

Comparison of image scanned at various resolutions

Click to enlarge


Top

Now to complicate things a bit

We’ve demonstrated the validity of the claim there’s not much more than 300 dpi’s worth of information in photographic prints. There’s a lot of data in a 300dpi scan and this is perfectly adequate for scanning most scanning jobs.

However there are situations in which it helps to scan at a higher resolution.

The first one is for the purposes of photo restoration. We usually scan images for restoration at 600dpi because this helps when we zoom in to the ‘molecular level’ to repair an image.

The second, and most common reason to scan at higher resolutions, is to get enough pixels to print an enlargement. If you scan a 4×6” photo at 300dpi and double the size to 8×12” the resolution drops to 150dpi. Many (but not all) print services will only accept files at 300dpi at the print size.

There are four ways to calculate the resolution needed for enlargement:

  1. Let the software do the calculation. In the Epson scan program, for example, you can specify the target resolution and size and the program will automatically up-size the file. This is the easiest way to do it.
  2. Go to www.scantips.com/calc, put in your target print size and resolution and it will calculate it for you. Then enter that resolution into the settings in your scanning program.
  3. Use Photoshop or Lightroom to upscale the image to the desired size. The resampling algorithms in these programs do a good job in interpolating the pixels. However it’s still better to start with a higher resolution scan.
  4. Some basic maths: Short side of output size ÷ short side of original image x 300 dpi. Thus, to get an 8” x 12” print from a 4” x 6” original, you would need to scan at 600dpi. (8÷4 x 300).

Top

 

First impressions count:
Scanning slides and negatives

If you have a choice between scanning transparencies (slides and negatives) and scanning prints, always choose the transparencies. This is because the transparency is the product of the initial exposure, and as such contains the maximum possible information.

Owing to their small size and information density, transparencies must be scanned at a much higher resolution. A 300dpi scan of a 35mm slide will yield a 35mm print – a tad small.

Once again, you can go to www.scantips.com/calc to calculate the resolution for each scan, depending on intended print size.

As a rough guide: For enlargements at:4×6” (100x150mm) scan at 1500dpi8x12” (around A4 size) scan at 2500dpi

TIP: use cotton gloves when scanning negatives to avoid fingerprints.

 

Removing dust from slides and negatives

All transparencies have small scratches and dust that appear exaggerated when scanning for big enlargement factors. There are a few ways to minimise dust and it’s best to try to remove as much as possible before scanning.

  1. One way to physically remove dust is to use a Hurricane Blower. These are available quite cheaply at stores like Ted’s.
  2. Gently wipe with a clean microfibre cloth, preferably a new one to ensure there’s not any embedded grit.

Compressed air cans are not recommended as the propellant can condense and blow droplets on to the film, which may cause damage.

There are three ways to remove dust with technology:

  1. Digital ICE technology. ICE stands for Image Correction and Enhancement. The scanner shoots up an infrared ray and determines what is a surface artefact (i.e., not wanted) and then interpolates pixels to blot out those areas. It works well but is excruciatingly slow – it takes two or three times longer per slide. Bear in mind that slide scanning is already a slow process. Most higher end consumer scanners come with this function.
  2. Software can reduce or remove dust. You can use the dust and scratches filter in Photoshop to clean up images. Note that the software does this by bleeding pixels together, so some loss of information results. You can also use a spot healing brush in Photoshop or Lightroom. Typically it’s the sky in images that shows up dust and this can be a fairly quick fix.
  3. Wet mounting transparencies is another way of reducing the appearance of dust, and also flattening the film. Wet mounting kits are available for scanners including the Epson V700 / 800.

Top

 

Scanning software

All scanners ship with software and these programs can also be downloaded free online.

For scanning reflective material (photos and documents) the bundled software is usually fine. The Epson Scan software that we’re familiar with has an automatic mode which crops and rotates scans automatically. This is a big time saver when many photos have to be scanned.

Another program is VueScan. The virtue of this program is that it works with just about any scanner and is thus useful when manufacturers no longer release software updates for new operating systems. We used this for the MP600 tests for this very reason. The downside of VueScan is its clunky interface. See www.hamrick.com

For serious scanning of transparencies (slides and negs) SilverFast is recommended. This program enables high-end colour and tonal adjustment and can also scan the image in two passes – one that records maximum detail in shadows, and one that records maximum detail in highlights – and combines them into one file. See www.silverfast.com
Top

 

File Formats

Scan in TIFF format.
TIFF is a lossless file format, which means that no data is lost to compression.
JPEG is a ‘lossy’ format, which means that every time the image is saved, more data is lost. There’s no discernible difference between TIFF and a new JPEG at maximum quality, however, editing and saving in JPEG format causes progressive deterioration.

Use TIFF as your master format and output your edited files as JPEGs.
Top

 

Adobe Lightroom

I’m only a fairly recent convert to Lightroom, having for a long time regarded it as ‘Photoshop Light’. Now I can’t recommend it highly enough. Here’s why it’s so useful:

Single click colour correction

Speedy image editing – Rapid editing of large numbers of images. Automatic tone and colour controls brighten dull images and correct colour casts with a click of a button. Sometimes it gets it slightly wrong but it’s easy to adjust with sliders. Auto rotate and quick cropping speeds up this otherwise time-consuming and boring task. Spots are easy to remove.

Add location and metadata – Metadata is the digital equivalent of written notations on the back of photos. Metadata can easily be added to single or multiple photos to assist with cataloguing and search.Many programs enable users to add metadata to images but this data is lost to other programs and operating systems. Lightroom metadata is universally readable.Images can be dragged onto a map to geotag the location.

Facial recognition – The software can identify the same person across an entire catalogue of images. If you enter the name of that person in one of the images it applies that data to all instances of that person. This may be very useful for family researchers and may have wider applications.

Non-destructive editing – Lightroom enables users to work ‘non-destructively’ with the master files. These are left untouched by the program. Instead a separate catalogue file is created which contains all the edits. These are applied to the images when exported. So you can keep the raw scans untouched and output the finished images as JPEGs.

It’s easy to learn and use, and it’s quite affordableLightroom is AU$12 per month on the Creative Cloud plan, or buy outright for around $200.

Lightroom screenshot

Top

Adobe Photoshop

When the editing task is beyond the powers of Lightroom it’s time to boot up Photoshop. For much of the restoration work we do this program is our stock-in-trade. The process is part technique and trickery and part artistry. Badly damaged faces need careful airbrushing and this is where our background in drawing and portraiture helps.

Here are some examples of work we’ve done using Photoshop. See more examples here.


Top

 

Cloud storage

Once you’ve scanned, edited and added metadata to your images it’s time to upload your JPEGs to the cloud. We use a paid Dropbox account to store all our images – $10/month for one terrabyte of data. There are other services which provide more than enough data for free.

  • Google Drive – 20GB
  • Microsoft OneDrive – 30GB

Amazon Drive provides unlimited storage for $100 per year.

We recommend cloud storage over physical media such as CDs and external hard drives, as these devices tend to break down or get lost, accidentally wiped etc. Data saved to local computers is not always migrated to new computers when upgraded.

Cloud storage is not fail-safe as the host companies may go out of business – but it’s a safe bet that Google, Microsoft and Amazon will be around for a long time to come.
Top

 

Giclée printing

There’s an emerging consensus among archivists that digital storage is not enough to safely backup collections. Printed backups are also recommended for the most important material.

In the last 20 years advances in inkjet printing technologies have finally brought permanence to colour photography, as this technology uses pigment based inks rather than the dye based chemical processes of standard lab photos. Image permanence is further improved with the use of archival acid-free paper.

These pigment ink inkjet prints are often referred to as giclée prints. This derives from a French word meaning spurt, or jet. It’s a marketing term, but the distinction is a necessary one as the quality and permanence is superior to standard inkjet technology.

The independent US researcher Henry Wilhelm, who has worked for decades with institutions such as the Smithsonian as a consultant on image permanence, has conducted accelerated exposure tests on these materials and found them to last 200+ years without noticeable fading or change in colour balance if in dark storage or albums. See www.wilhelm-research.com
Giclée printing is more expensive than standard lab prints and we recommend reserving this method for the best and most important material.

We passed around two photos of Bart Cummings, one printed at Officeworks and the other on a giclée printer. The difference was stark: the Officeworks print was flimsy, had a blueish cast and blown out highlights. The giclée print closely resembled a classic silver gelatin print.

Comparison between cheap Officeworks print and giclée print – note blue tint and blown and highlights on the cheaper print

Top

Final thoughts

  • Appoint a family archivist – keep abreast of changes to cloud services. Make sure everyone has saved the files to their own cloud account or at least their own local computer.
  • Put all the usernames and passwords for your research into your will.
  • Create printed duplicates of the most important photos and documents. Go for quality pigment ink prints on archival paper, and store away from light in acid-free containers – or frame using conservation materials.
  • Keep originals in archival storage boxes, out of basements and flood-prone places, and readily accessible in case of fire evacuation.

Top

2017-07-17T14:14:32+00:00

Leave A Comment