Well, it’s been a long time since part 1 of this article, but a lot has been going on. We ran tests on different scenarios, looked at different technologies and workflows, and of course, reworked the script several times. So at last, here’s a breakdown of the things we decided.
The Shooting Format
Undoubtedly one of the most important aspects, deciding on a shooting format proved a lot more challenging than I’d anticipated. Originally, I was sold on using HDV for its winning combination of good quality and low cost. The problem then was trying to decide whether a higher frame rate would be better, how the depth of field would affect the look and so on. The reality proved quite different. I set about to test Sony’s HDR-Z1U camcorder. The aim was to use it in conjunction with Serious Magic’s HDV rack, to enable us to capture material direcly onto a laptop with an attached USB2 disk drive. Unfortunately we ran into performance issues with this configuration, sometimes the footage would be ok, sometimes it would fall over. I suspect that the problem lies with the specfic configuration of the laptop rather than any particular fault in the software, which had proved very robust until now. (Note: at the time of writing, Serious Magic’s support forums are down, so I am currently unable to get an update on the status of this problem). So we opted to do without HDV Rack for this production, as we don’t want to take any chances. So I proceded to shoot some test material using the Z1U to tape, and then digitize the footage afterwards. I shot a very simple set-up, a garden scene with a slow pan and zoom, and repeated this for both the 1080i50 and 1080i60 formats supported by the camera, as well as several of its “colour correction” modes, to see how it looks. You can see the actual unmodified Z1U footage for yourself, if you have the ability to download using the bittorrent protocol. Here’s the bittorrent file (approx 500MB)… I then went on to try something with more action, to get a feel for the visual quality (see the images below). All in all, I have to say that I was very impressed. The picture quality was great, the camera was very user-friendly and portable, and it seemed that we’d be able to use it to get some great shots very quickly. But then everything started to go wrong.
The Test Edit
Avid recently announced an HDV update to their popular Xpress Pro editing software. Since we qualified for the update, I installed the new version to try it out. After a spending a day waiting for their support team to tell me how to update the dongle, I loaded it up to discover that I couldn’t use it. (I’ve given Avid enough of a battering in a previous review, so rather than regurgitating a rant, let me just say that the problem I encountered was that the application window was trying to load outside the screen area, and the only option was to force it to close.) At this point, someone recommended that I should take a look at Sony’s Vegas system, which also has HDV support. Sony kindly gave us an evaluation version, and I have to say I was impressed, it’s a very well designed piece of software. I was able to capture the HDV material in its native format with no real problems, and load it for viewing. This led to the next stumbling block. HDV is motion-compressed, using MPEG-2. When you transfer HDV footage, you typically end up with a series of .m2t (MPEG-2 Transitional) format files. There are a few applications available that can play these files directly, but the problem is not with playback, it’s that the performance is just not good enough to edit with. There’s a lag between going to a frame in a clip, and the frame actually being displayed. I preseumed the problem was that my hardware was simply not up to the required spec, but after going through Sony’s support forums, I discovered that this simply was not the case. It seems to be an inherent limitation of the format. To make things worse, there aren’t really any applications that can convert m2t into something more useful. There are a couple of options available, namely buying an add-on that creates proxies that can be substituted for the full-resolution footage during offline editing, but really none of these seemed to be any good and it just seemed like it would be an additional thing to go wrong down the line. I found myself having to think in terms of film digital intermediate pipelines, with generating EDLs and doing an online edit, and suddenly the benefits of using the HDV format seemed to be greatly diminished. It was around this time that (co-director) Mark Sum suggested that we “just shoot on film”. And I have to say, that sounded like a solid approach.
The Digital Intermediate for Film
I revised the plan to accomodate shooting on film, rather than digitally. The plan became (and is currently) to shoot 35mm, develop, scan the neg, and create proxies for output to different formats. This would serve two distinct purposes: we could use the proxies as digital dailies, and we could import them directly into our editing system for the offline edit. One of the things we still need to test is how to ensure that the timecode will be sync’d during the import process, but this is really a matter of using the right conversion format. Because of the way we’re working (and because we’ll have complete control over everything), we can essentially create a single virtual source reel for all the source footage, which will simplify things greatly. After the offline edit, we can produce an EDL and use that to conform the scans during the online edit. As a final note, we will have a video camera on-hand during the shoot, so it may well be that we end up mixing some video footage in with the film footage.
The Finishing Process
We plan to use Assimilate’s Scratch for the whole finishing process, from the conforming to the final grade (using the Scaffolds option). We’ll take the digital source master and use that to generate several output formats, including HD video (as our output master), DVD, Xvid or QuickTime (for web distribution via bittorent), Flash Video (for a version viewable directly on the website), and Mpeg-4 (for PlayStation Portable devices).
Pre-visualisation
Previsualisation has involved the use of Antics 3D Pre-Viz software, for expermenting with camera placement on shots with complex choreography, and traditional storyboards for other scenes. We’ll be posting the results of that on the website soon.
Production Database
The production database we’ve created (dubbed “Synasthaesia”) is taking a lot of work to put together, but so far, it’s proving very useful. We hope to have a public beta version that everyone can try out some time after we’ve completed the shoot.
The Production
Everything mentioned applies to our upcoming production, The Toilet Guy, which is to be shot very soon. For more details on that, including the option to read the script and look at some of our pre-production artwork, visit The Toilet Guy website…
Read on for part 3…
Posted: January 24th, 2006
Categories:
Articles
Tags:
digital production,
tests,
toilet guy
Comments:
No comments
