Finally, as After Effects is listed as the only effects editing package for this film, the logical assumption is that the included Keylight chromakey package was used for this film. Keylight is an awesome chroma-keying package that has been used in a number of films. Unfortunately, it’s designed to pull clean keys from progressive footage shot in the 4:4:4 color space, like the footage used in Sin City and Star Wars: Episode III. As such, it was not a kind taskmaster to the footage that had already gone through two serious technological problems to get there in the first place.
So, now that we know where the problems likely came from, how does one prevent them in the future?
First, if it’s possible to get a progressive camera, you will end up being able to pull cleaner keys. If you can get one that can do 24 P (or “24 Progressive frames per second”), you will get even better keys, because you can record more information per frame when you’re recording 24 fps than when you’re recording 30 (or, more commonly, 29.97). The DVX100 is one SD camera that can do this, while the HVX200 and the HD1 are HD cameras that can do this. However, if you can’t get ahold of a different camera and must shoot in NTSC, then I would recommend running the footage through Red Giant’s Magic Bullet Suite or new Frames package, as the de-interlacing features in these are extremely good and can provide you with clean progressive footage. (Additionally, they can convert your footage to 24 fps, if you want it to look more like film.)
From here, I would recommend using a plugin that’s actually designed for keying DV footage. One that was recently recommended to me (which is available for After Effects and a few other programs) is a $200 program called dvMatte Pro. (They also have a more basic version called dvMatte Blast which works in Apple’s Motion 2 for $99) This is a chromakeying program that samples color data, but, to give the most accurate results, it also samples luminance. Since DV records luminance sampling every single pixel, combining the information it gets from that with the color key gives much more accurate results with DV footage. While I’ll have more thoughts on this when we look at it officially this summer, I’ve been really impressed with the results of this program on green-screen footage.
I applaud all the things the director and his team attempted with their greenscreen efforts and I look forward to seeing what he will use this growing technology for in the future!
The overall cinematography in
this film is very impressive...
...with shots that mirror some
very
difficult shots in LOTR.
Use of Audio
Unlike many low budget films, there aren’t very many issues with the actual intelligibility of the words being spoken. This is because the entire film was redubbed. As such, the filmmakers had maximum control over everything. However, the unique thing is, it does not appear as though much effort was taken to make any of the voices sound natural to the settings they are in. As such, all the voices sound very much like they did in the recording studio, with an even volume that sounds like it’s right next to you, regardless of where the character is on the screen. Despite how surreal this is, it works because the folks behind Dork of the Rings keep this consistent. The lip sync is far from perfect, with probably a 20% of the lines being noticeably off sync. However, because it happens so regularly, it also becomes something that doesn’t distract you from the story and becomes strangely acceptable.
Unfortunately, one element of the dubbing which does not become strangely acceptable is the excess sibilance in the recorded voices. “Sibilance” is where the sounds of S’s and T’s are excessively pronounced. This is usually either due to brighter mics being used (many inexpensive condenser mics have a rise in the high frequency response which can be problematic for ADR) or mic placement that is less than optimal. If it’s the first issue, then, for future ADR sessions, I would suggest using a shotgun mic like you would use on a shoot, such as a Sennheiser ME-66 or ME-67. If it’s the latter issue, then I would suggest angling the directional mic at the person’s mouth at approximately 45 degrees from above or below, rather than recording directly in front of them, as the sounds produced by the S and the T tend to proceed in a fairly straight line from the mouth of the speaker. By putting the mic at an angle, the prominence of these sounds can be decreased. (For additional suggestions on future projects that need redubbing, please check out the article we wrote on setting up an ADR studio and the considerations therein.)
