Digitising Anne-Mie Van Kerckhoven's slides
In 2014, Anne-Mie Van Kerckhoven wanted to use slides of her installations and artistic images for a book – and the slides had to be digitised for this. PACKED (now meemoo) helped her to find a suitable digitisation company and check the quality of their digitisation process. You can read about the various steps taken and difficulties encountered in this case study.
Inhoud
Status
- Make an inventory of the slides: autumn 2014
- Draw up a request for quotation and find suppliers: spring 2015
- Test phase: autumn 2015
- Digitisation and finishing: 2016
Problem definition
In July 2013, expertise centre BAM (now Flanders Arts Institute) contacted PACKED with a question about digitising visual artist Anne-Mie Van Kerckhoven's slides. She wanted around a thousand slides digitised for a new book, including her own images of installations and other pictures that she'd presented in various ways. The artist had already digitised some of the slides herself with her own scanner, but they weren't good enough quality in terms of sharpness to be usable for the book. She therefore asked PACKED for technical advice and to recommend digitisation companies to request a quote from.
Method
- Preliminary discussion
- Make inventory of slides
- Determine quality requirements and draw up a request for quotation
- Select digitisation company and award contract
- Test phase
- Digitisation phase
- Finishing
Preliminary discussion
We talked about the objectives of the digitisation assignment in a preliminary discussion. The scans would mainly be used in a publication, but Anne-Mie Van Kerckhoven also wanted to be able to print out the digitised slides in A3 format. Furthermore, it came to light that some of the sides were dirty and needed cleaning, and some were set behind glass which we expected would need removing to achieve the best results.
Make inventory of slides
The slides that needed digitising were in various sizes, so Anne-Mie Van Kerckhoven started by creating an overview of the different requirements. The size classification was created to help determine the cost of digitisation, which can vary depending on the various sizes of slides to be scanned. Together with her assistant, Anne-Mie Van Kerckhoven gathered all the slides together, put them in order, and described them in an Excel spreadsheet. Each slide was given an identification number and title, and they also identified the ones that had become dirty over time.
The inventory showed that 1,111 slides needed digitising in total, in the following breakdown of sizes:
- 24 mm x 36 mm: 828 slides;
- 42 mm x 55 mm: 11 slides;
- 55 mm x 55 mm: 92 slides;
- 65 mm x 55 mm: 15 slides;
- 70 mm x 55 mm: 4 slides;
- 75 mm x 55 mm: 155 slides;
- 57 mm x 85 mm: 4 slides;
- 90 mm x 115 mm: 2 slides.
Determine quality requirements and draw up a request for quotation
PACKED used this classification to draw up a brief request for proposals that included the following elements:
- Objective: what do you want to do and achieve with the digital files? In the case of Anne-Mie Van Kerckhoven, she wanted to use the slides to print out images in minimum A3 format.
- Then we described the assignment. This included the size of the slides collection and how it needed to be processed along with technical requirements for the master files.
- Scope: a summary of the number of slides and their various sizes.
- Processing: the condition of the slides and any processing requirements for the digitisation company to be able to create good scans. In Anne-Mie Van Kerckhoven's case, this included cleaning slides with sawdust on them. We also added how we wanted the digital images to be edited. We requested for dust and small scratches to be filtered out, and for the edges around the images to be cropped[1].
- Technical requirements for the master file, such as image resolution, colour space, bit depth, compression, file format, filename, bit integrity and metadata. We followed the guideline for digitising photographic collections (link in Dutch) for this, and requested AdobeRGB 1998 as the colour space.
- The proposal also included the tools that PACKED would use to check these requirements.
- Finally, we also stated the practical agreements in the request, to establish what is expected from the digitisation company, e.g. to first create test images and how the files needed to be delivered. Anne-Mie Van Kerckhoven's proposal requested that test images should be processed first and to deliver the files on an external hard drive.
When digitising slides, the image capture resolution is different for each format depending on its size and target print size (i.e. A3 or A2). For this quote, the image resolution was calculated by multiplying the base resolution of 300 ppi (a resolution required for high-quality printed matter) by the enlargement factor. This is the ratio between the size of the slide (longest side) and the longest side of an A3 and A2 page respectively.[2] (link in Dutch)
Select digitisation company and award contract
We sent the request to seven specialist digitisation companies: Acmis, Sercu, Vanden Broele, Picturae, GMS, d/arch and Digidat. After comparing the quotes received, in consultation with Anne-Mie Van Kerckhoven, in May 2015 we awarded the contract to GMS, a Dutch digitisation company that had also handled large-scale newspaper digitisation projects run by the Flemish Institute for Archives (VIAA) (now: meemoo) and k.ERF heritage body (link in Dutch).
The first task for GMS was to digitise a test set of ten slides.
Test phase
We asked the supplier to do a test run with a small number of files to ensure they could fulfil the request, and made a set of ten slides available for this:
- six small image slides (24 mm x 36 mm);
- two slides of 55 mm x 55 mm;
- two slides of 55 mm x 75 mm.
Five of the six small image slides were framed (four in plastic, one in cardboard), and one of them was behind glass. The slides with 55 mm x 55 mm format were also framed and behind glass.
We asked GMS to digitise the slides to a resolution that allowed high-quality A2 printing with 300 dpi. The proposal also requested that the supplier should use the names given on the slide frame as the filenames, but they had so much text on them that it was then decided to create a separate filename for each slide. Anne-Mie Van Kerckhoven and her assistant added these filenames to the inventory.
Each digitisation was followed by a quality control in which we checked that GMS had digitised the slides in accordance with the set requirements. Specifically, we tested the following aspects:[3] (link in Dutch)
- Virus check
- Completeness: all slides digitised? All required files sent?
- Authenticity: are the files received identical to the ones sent?
- Naming: do all files have the correct names?
- File format: is the file format correct and valid?
- Requirements: do the files satisfy the quality requirements?
- Metadata: do all files have the required metadata?
In order for PACKED to carry out the quality control in an identical way to GMS, we requested the digital files together with the colour chart and software used along with the quality control results.
They delivered the test files at the end of October 2015. Despite the agreements made, however, the colour chart, measurement results and information about the software used by GMS for the quality control were missing, making it impossible for us to perform a proper quality control. Following extensive consultation with GMS, the test was performed again with a larger selection of images, which resulted in a small additional cost and delayed the digitisation process somewhat.
In order to ensure that we could perform a thorough quality control this time, we asked GMS to include the following details in the test:
- the adjustment slide used for each slide format;
- the measurement values obtained in this process;
- the reference values that GMS used for this;
- the test captures themselves;
- the embedded metadata for each capture:
- inventory number
- copyright statement (e.g. Copyright: AMVK)
- manufacturing number for the recording device
- model of recording device
- make of recording device
- time file was captured
- software used to edit file
- time file was edited
- aperture
- shutter speed
- ISO value
- special camera or RAW conversion settings (if applicable)
- colour profile
- colour space
- sampling rate (in pixels per inch)
- length and width of scan in pixels
- bit depth
- file format
- a spreadsheet with at least:
- inventory number per file
- filename per file
- MD5 checksum per file
GMS performed the test and supplied the details requested. The quality control was largely positive, but there were still a few problems. The scans weren't all equally sharpened, for example, and there was still a lot of dust visible. GMS assured us that they had cleaned the slides as much as possible. They had opted to blow the slides clean with compressed air because it would cost too much time and money to remove them all from their glass coverings. But if there was then still too much dust left in the slides, GMS would open them for cleaning.
Digitisation phase
Once the test results had been accepted and we completed the quality control for the second test in March 2016, we commissioned GMS to digitise the full set of slides. The files had to satisfy the following specifications:
- File format: Uncompressed Baseline IBM TIFF v6.0;
- Sampling rate: 300 ppi;
- Bit depth: 24 bits;
- Colour profile: AdobeRGB;
- Length and width in pixels: this had to be as close to 7017 x 4962 as possible in order to allow printing in A2 format with 300 ppi;
- Filenames: based on the names stated in the inventory, with a hyphen (-) and underscore (_) to replace dots (.);
- Embedded metadata as requested in the last test.
Together with the files, GMS had to deliver an Excel spreadsheet including the old filename, the new filename and an MD5 checksum for each digital file.
The delivery of these files followed in April 2016. When checking the files, however, it soon became apparent that a few things had gone wrong:
- The TIFF files were not in the right dimensions. Moreover, it appeared that the files had been compressed. The software that GMS had cropped the files with used LZW compression on the images, a type of lossless compression, even though the contract stated that uncompressed files were required.
- Some of the embedded metadata values were missing, including the value for the colour space used, which was probably due to an error in writing the metadata.
- The scans from the initial test were missing and not included in the set.
- Some scans had an incorrect name, and it wasn't possible to change the names manually because the delivered files were read-only.
- Not all targets and test results were sent, so it wasn't possible to perform the quality control.
- Some captures were much smaller than agreed (A3 and A4 instead of A2).
GMS acknowledged these errors. We asked them to create the TIFF files again and to send the targets and missing scans, and re-write the metadata in the files. These issues delayed the project further, and the quality control ultimately ran from April to October 2016.
Finishing
We completed the quality control at the start of October 2016 so the files were ready to send to Anne-Mie Van Kerckhoven, who received them in December 2016. There was still a final problem, however: Anne-Mie Van Kerckhoven could not edit the files or change the filenames because the hard drive was not NTFS formatted. The is a Windows formatting, whereas Anne-Mie Van Kerckhoven uses a Mac. We therefore took the hard drive to PACKED and converted it to a suitable formatting.
Results
1,111 of Anne-Mie Van Kerckhoven's slides were digitised into uncompressed TIFF files with a size and bit depth that allowed high-quality printing in A2 format, ensuring the artist could use them in her book. The digitisation process was not without obstacles, with the digitisation company delivering files that were not in accordance with the set requirements in each phase of the project. The project therefore taught us that the quality control on the work delivered by the supplier is extremely important and needs to take place regularly throughout the process.
Authors: Nastasia Vanderperren (meemoo) and Rony Vissers (meemoo)
- ↑ Removing the excess part of a scan.
- ↑ For more info about the ideal image resolution, see https://kennisbank.meemoo.be/inzichten-en-praktijk/wat-is-de-ideale-opnameresolutie-bij-digitaliseren.
- ↑ See also: https://www.slideshare.net/VVBAD/kwaliteitscontrole-bij-digitalisering-van-kranten-en-tekstaffiches