Hi Demian,
I work as an Application Specialist for a film recorder manufacturer based in Norway, called Cinevation. The technology we use is quite different to that of an Arrilaser, but the basic recording process is very similar, so hopefully I can be of help.
I think your idea of recording to camera stock would be ideal; you would retain the best recording quality possible, with a very natural film grain structure. As you mention, there are significant time and cost implications for a recording facility that needs to set up this support from scratch, which is why they may be hesitant to go down this path. As Dirk mentions, CRT recorders are often better suited (and more commonly calibrated) for camera speed stocks, so this may be one avenue to explore. Keep in mind that if you go down this path, the colour of the final image will look slightly different to what you have already seen, so I'd recommend you run a short test to check things out before committing to the full job.
From experience, pushing the stock or using a bleach bypass process can be highly unpredictable, and I would steer clear of these methods if possible. Additionally, bleach-bypassed negative does not have a good shelf-life, and over a few short years you will see a dramatic shift in the colour and contrast of the image, so only explore the bleach-bypass idea if you are happy that the negative will not need to be archived or re-used in years to come.
Adding digital grain may be a good alternative. Just remember to try it, and run a quick test recording to check the results on film. If the recording facility are using a sharpening filter in the recorder, this may emphasise the digital grain in a negative manner, and you should check this out in advance.
One last suggestion, although it is a more expensive option... You may try recording the film as an Interpositive, rather than a negative, then produce a standard photochemical Dupe Negative from this IP. It is a more costly option, but could provide the slight hint of grain and jitter that you are looking for. Again, I'd run tests, and make sure you talk to your lab about the implications on both cost and workflow.
Hope this is helpful.
Best,
Peter
