For some years there have been few options when it came to film scanning - but recently that has changed dramatically, and with the change has come the ability to select film scanners based on real-world business models, rather than simply selecting a scanner to get the film into the data world.

This may sound like a rather strange thing to say, but with 'next generation' film scanners - from traditional scanners such as Imagica, NorthLight, ArriScan and diTTo, to data telecines such as Spirit, DSX and dataMill - comes a virtual guarantee that they are all capable of extracting the best possible quality from the underlying film image. What that quality is, is a separate question, and one looked into within the section on Thoughts on Resolution, as well as within the DI Guide white paper available from the download menu above.

However, it's worth stating that the required resolution could be anything from SD, through HD, to 2K and 4K, depending on the project's requirements.

What is so nice about this state of affairs is that film scanning systems can be assessed based on the business model of the facility in question; and the expansion of Digital Intermediate into the wider world of post-production is impacting the choice of scanner, or more accurately film transfer engine, greatly.

One of the key factors now affecting film transfer engine choice is the need to perform virtually all post-production work via a Digital Intermediate workflow, regardless of the final product requirement. This generally means having two immediate options - telecine based film transfer engines, or more traditional film scanners.

The above images show Cintel's dataMill film transfer engine, and diTTo film scanner.

Film Scanning Business Model
The decision as to what system is best suited to your needs depends on the projects and client base you expect to be working with, as well as the actual workflow to be used to complete the work. This could be a data centric network based workflow, or a 'video' based workflow utilising the benefits of Digital Intermediate.

For example, the traditional telecine approach to film transfer to 'video' entails the use of a real-time colour corrector attached to the telecine, making colour decisions about the transferred video at the point of image extraction form the original film source. Such decisions are almost impossible to alter later in the post-production chain due to limited dynamic range of the transferred video. And as the transfer and grade usually takes place from the original OCN film, it is very difficult to 'see' one shot's grade in context within the final edit. And to overcome such limitations very complex and costly telecine grading systems have been developed from the likes of Pandora and da Vinci.

Is it not better to simply 'clone' all the image data from the film master, perform the edit, and then grade the material within context of the final edit?

This is the Digital Intermediate approach, and the benefits for all post-production businesses - film (2K/4K); HD video, and SD video - are obvious.

Therefore, we are seeing a new approach to film transfer engines; one that regardless of the end result requirement (theatrical film, HD video, or SD video) encompasses a self calibrated film transfer, extracting all the dynamic range from the film master, in what ever resolution is required for the project in hand - no telecine control system with colour grading capability attached to the film transfer engine.

Such an approach means that the film transfer engine in question can self-calibrate to an industry standard - usually Cineon/dpx 10bit Log - at the required resolution for the project in question, and wrap the image data within either a data file, HD or SD video. More information on this approach to post-production can be found within the Digital Film: Scene-to-Screen white paper from the download menu above; especially the 'HD - An Alternative Workflow for DI' section, as well as within the HD as a DI Workflow discussion.

From the available film transfer engines, it is Cintel's dataMill that best meets these requirements.
The beauty of the dataMill is its low cost base, combined with its ability to self calibrate to Cineon/dpx 10bit Log file formats, and wrap the resultant images in and file format and resolution, from 4K, through 2K and HD, to SD. This ,in the eyes of Light Illusion, makes dataMill the most flexible film transfer engine available to date.

To aid with such an approach to film transfer, Cintel have also introduced the imageMill, a real-time grain and image management system that can lessen the unwanted film artefacts associated with film scanning; notably grain reduction without resolution loss.

For operations who require less in the way of real-time film scanning, a lower cost, non-real-time film scanner can be a cheaper alternative, outputting the scanned image data as 'files' to be saved via a computer network onto disc storage.

Such film scanners do not possess the ability to output the scanned image data wrapped in a 'video' wrapper, outputting data files in dpx/Cineon 10bit Log form as discreet data frames.

Again, one of the more interesting film scanner is Cintel's new diTTo film scanner. A desktop film transfer engine capable of multiple frames per second for 2K and 4K output, at a quality point equal to any more costly alternative film scanner.

The following two diagrams assist in understanding Steve's approach to film scanning for DI workflow:

• HD 10bit LOG Real-Time Workflow
• Data 10bit LOG DPX Network Workflow

For more information on the Cintel range of Film Transfer Engines check out Cintel's website.