Dustbusting

Dustbusting is a somewhat laborious and challenging part of the Digital Intermediate process however it is a major part of the process that is hardly ever mentioned. Whenever I have reviewed other documentation, dustbusting gets just a cursory mention. It is however a major part of the workflow, which is extremely important and if neglected can detract from an otherwise beautiful final negative.

It is linked because no matter how clean, how careful and meticulous you are with the negative(s) somewhere there is going to be a small scratch, piece of dirt or worse, neg damage that needs to be cleaned, lovingly restored, or beautified. A lot is introduced through the processing stages, some within the camera itself but ultimately there will always be something lurking on the negative the human eye will register when viewing. Traditionally a lot of these small scratches and dirt would be removed during a wet gate print and closing some of the scratches through the transfer process. Also the process can conceal smaller dust particles – however because you are locking the image when you scan it, you are effectively locking in these scratches and dirt.

Before the final scanning process, it is necessary to clean the neg and hopefully remove as much of any lose dirt or material. Most post production houses worth their salt will have a sonic cleaner, this device passes the neg through a solvent which washes and instantly evaporates, while gently vibrating to loosen particles away from the negative. They may even employ a specialist to clean the negative for them, or they may not and ultimately rely on the eagle eyes of their in house dustbusting team. Or just simply pray to the dust fairies that it will not be too bad!!

Eventually every piece of negative, while lazily winding through the scanner, hopefully residing within a clean room, will acquire a tiny amount of dust or dirt floating around the atmosphere. Gently laying itself down on the negative and ultimately image processed into the scan. So even if it is cleansed, the slow scanning procedure allows dust to miraculously and infuriatingly to reappear.

You may start to realise it’s a never ending circle, like cleaning your house, you basically just move it around a little, or at least the stuff you remove is then replaced by some other dust. So ultimately the whole image has to be ‘digitally’ cleaned after scanning and this will hopefully remove the last of the offending material!

 

Techniques

Everyone will argue their way is the best however their dust removal is only as good as the eyes of the operator(s) in charge of the software or equipment, plus the communication between departments and said operators. There are dedicated hardware options, such as Revival from Da Vinci, software options such as MTI and then you can employ a myriad of paint packages. In my opinion they are all pretty good and all do the same job, however there are similarities where they all fall over.

One of the biggest problems I’ve seen is automated processing – software or dedicated hardware although programmed extremely well, in fairness cannot give the same objective view we can. Computers compare images both forward and backwards, possibly over a pre designated parameter (number of frames) and if there is an anomaly (piece of dirt) they use a copied section to effectively paste over the top. I hope you can see this can be dangerous, especially if you rely heavily on this technique. A computer cannot actually ascertain what is dirt and what is not. Unlike our eagle eyed vision and instinct, as a piece of dirt flashes by on screen it registers in our brain – well hopefully. It is an art and some people are unable to get the hang of it!

My first encounter with automated processing was a long time ago and in fairness it has progressed in leaps and bounds since. We were working on the restoration of an Italian film called L’avventura, at Cinecitta Studios in Rome. If you are familiar with Italian films, the leading ladies tend to wear a lot of intricate lace work and the processes used on this could not differentiate a hole in the lace from a piece of dirt and the lace came back after processing looking a pixelated mess. We spent a considerable amount of the restoration cutting and painting through areas by hand using frames forwards and backwards out of sequence – the final result was nothing short of spectacular but it was an extremely long slog and not for the feint hearted.

For new feature film work this is not going to cause as much hassle using automated process’ because for a start you are not working against years of neglect, contrast differences between frames and thirdly the frames either side are more than likely going to be perfect – so the dustbusting procedure can certainly be a lot less demanding.

 

Scheduling

Every post facility will have their own way of carrying out the clean up process. There is no right or wrong way of doing it, as long as the final images, which are printed back to negative are ultimately clean.

Method 1 would be to dustbust and clean every segment or individual edit after the initial scan before the conform. However this can pose problematic because you would need to keep on top of the material, especially if your conform performs across a network and you need to re-link to the media. For example if you clean a section and then delete the original scan leaving only the cleaned version, a hole would appear in the conform because it would no longer be linked via the database. However if you use this methodology, it can cause a bottle neck while waiting for material to be cleaned and made available to the timeline.

Method 2 is to clean after the conform - therefore as work progress’, cleanup and dustbusting takes place alongside the colour correction. As the edit is reviewed and checked the project will gradually become clean, it will seem more of a seamless project. Plus from a workflow point of view this is a far easier way of progressing – colour correction during the day with the client. Cleanup at night – on review the following day, the client will see a double whammy, the new colour and cleaner image. With this methodology, when the client signs off on the grade, they can ultimately sign off the show because what they will have seen will have been the final virtual filmout.

Method 3 is slightly more controversial as the images are dustbusted after the client has signed off on the colour correction. In my opinion you will need to have a very close trusting working relationship with your client for this to be acceptable. It is one thing to sign off on the colour correction but you never know how good the clean up will be after this stage. The client will probably not see the film again until it has hit negative and they are viewing the final print. If they are unhappy with the cleanliness of the master negative, you could be forced to output another, after another pass on the cleaning. Trust me that would not be a good position to be in and it will cost someone somewhere for the additional output stock.

I can see a need for this kid of workflow if your clients deadlines are inflexible and you are tight for time but they need to be made aware of the possible outcome if they push the deadline and ultimately adopt this workflow.

Of course these are all operator assisted processes. There are automated systems that can work as early as the scanning stage which can cut out a lot of this laborious work.

 

Infrared (IR)

Infrared (IR) enabled film scanners have been available in the consumer world for several years, only recently are they beginning to be optimised. These scanners use an IR illumination source and a sensor capable of measuring into the IR range to provide an additional channel (like a matte or alpha channel) along with the traditional red, green, and blue (RGB) channels. This IR channel is subsequently analyzed to detect minor defects in the scanned film image. These defects are then corrected by a robust filtering method.

The principal behind IR scanning is relatively simple: if the base and colour dyes in colour negative film absorb no IR light, then any deviation in an IR image is caused by defects (dust and scratches). Scratches and small dust tend to scatter incident light rather than block it completely; hence, some colour information still reaches the sensor in the film scanner. Image processing algorithms can analyze the residual colour information in "defective pixels" and make a correction based on the estimated scattering of light.

 

The image on the left is the scanned frame, the image on the right is the defect matte used for the dustbusting.

 

Infrared enabled dust busting and scratch removal have become a reality with the extension of Digital ICE technology by Kodak added to motion picture film scanners. The marriage of superior hardware design with robust signal processing algorithms allows for the automatic detection and correction of a wide variety of film artefacts. Moreover, the software is capable of indicating, through the defect matte, those defects that are severe enough to require manual intervention. This will significantly reduce the costs of dustbusting and scratch removal while improving overall image quality.