Digital Production, Part 4 – The Shoot
After what seems like an eternity, we finally get to the actual shoot. How does everything discussed so far affect the shoot? What benefits does the digital process bring to the mix? Read on to find out…
Our shoot was scheduled for 3 1/2 days. The final shooting script was around 12 pages long, so we aimed to shoot 4 pages per day, leaving us with 1/2 a day to do pick-ups and to do some adlibs/improvised stuff. In actual fact, I figured the shoot would over-run, and so the 1/2 day would end up being used to finish the previous days’ shooting.
Contrary to earlier assumptions, we did not shoot on film. This was partly out of budgetary concerns, but also because we wanted a truly digital experience. We were toying with the notion of shooting using Grass Valley’s Viper FilmStream camera, but it would have been too expensive to get the supporting equipment and acrhive media, not to mention making the whole pipeline more expensive. So instead we revisited the idea of shooting HD, or more specifically, HDV.
We’d initially thrown out the idea of shooting HDV in our previous tests, mostly because the post workflow would be really painful. However, the choice we were faced with was quality vs. ease of use, which is a no-brainer for us. So it became a question of making it work. The solution lies in adding a “pre-post” stage to the workflow, which will be covered in its entirety in the next instalment of this series. In any case, on to the shoot.
We decided that although we could get a medium-sized crew (though not a large one), we would actually eschew that option in favour of a skeleton crew. This may seem counter-intuitive, but the benefits of a small crew are clear: we could work really, really quickly. We used found lighting for the whole thing (the location was selected specifically with this in mind, and the whole story is set in one place) and we were lucky in that the weather was very good throughout the shoot (bearing in mind we filmed in not-so-sunny London). All creative lighting was modelled using reflectors (and will be augmented with digital grading later).
Speaking of which, I feel I should point out that the idea was to avoid the “we’ll fix it in post” adage as much as possible. The idea of this whole process is not to offload production process onto the post-production process, but rather to try and streamline the whole thing. There are no visual effects called for in the script, and we felt there was no need to rely on anything of that nature. I should also point out that with the exception of pre-visualisation techniques discussed previously, much of the lead-up to the shoot was exactly as it would be for a non-digital production, with rehearsals with the actors, script breakdowns, and so on.
In terms of the camera, we used a Sony DVCAM with a variety of lenses, and a Sony HDR-Z1U, with just the built-in lens. On both we cranked the aperture wide open (to around f/1.6) to give us the best possible depth of field, and used neutral density filters when necessary (which was most of the time). Audio was recorded directly onto the camera tapes, which certainly made life much easier later on.
Here’s the interestig part of all of this: we didn’t use a slate/clapper throughout the whole shoot. Instead, we used a laptop with our production database loaded and called out timecodes before each take. This accomplished several things. First of all, we could see easily how many takes we’d got for each shot, which shots were still to do, and we didn’t waste time (or tape) shooting a slate in-camera. And all this without a script supervisor. As a sweetner, we were able to log notes for each take directly into the database, which was to help a great deal during post-production.
One of the problems we didn’t really anticipate was that the DVCAM shot at 29.97FPS, whilst the HDV was shot at 25FPS (the HDV frame-rate was selectable, while the DVCAM was not). The intention was to later perform a frame-rate conversion on the DVCAM, but in retrospect it might have been much simpler to instead shoot the HDV
at 60i.
All in all though, the shoot was a success, you can read more about it at Mark Sum’s production diary…