2k CinemaScope from next gen DSLRs?

[TrueIndigo]

Following this years' video trends in DSLRs, the base level for video implementation in most cameras in the new year will probably be 1080p (rather than 720p). But, if a manufacturer wanted to stand out with an unexpected video feature, which would capture the imagination of the indie film makers, my suggestion would be 2k CinemaScope. If video-enabled DSLRs become as widespread in 2010 as live view was in 2009, a camera maker like Pentax (who need to work hard for market share) could really stir things up by offering something other than 16:9 video. If this wide frame size/shape could be windowed/binned from the sensor, it would be a dramatic and playful acknowledgement of many indie aspirations.

Would you like to see such a video option? Any comments and discussion on this welcome. Thanks.

Happy New Year to you all.

 

[TrueIndigo]

Some initial thoughts:

1. Is it even possible?
I don't know, perhaps someone can chime in here. The whole thing rests on whether it is easy and therefore cheap to implement (through firmware) and also whether it is technically allowed by the hardware (the sensor and processing pipeline). So, if it is possible and if it is easy, why not do it? Why not include it as a welcome addition to a camera's video modes: 1080p (16:9), 858p (2.39:1), 720p (16:9).

2. Isn't the data rate of 2k too high?
I believe 1080p video should become base level for many DSLRs in 2010. And if they can do 1080, they should be able to do 2k CinemaScope. Why? Because the lesser height of the wide frame area is actually less than 1080. Disregarding compression schemes for simplicity of comparison, a full 16:9 HD uncompressed frame (1920 x 1080) is around 2.073 million pixels, while a DCI spec 2k 2.39:1 frame (2048 x 858) would be around 1.757 million pixels. Even using current engine performance and current codecs, recording video at 24/25 fps instead of 30 fps should easily make the datarate of 2k CinemaScope possible. And since every year manufacturers generally increase the performance of their engines to improve still shooting fps speed, video processing next year will benefit from this too.

3. Aren't indie film makers a niche market?
From a marketing point of view, DSLR video is already here, and will become an expected feature on any DSLR spec sheet in this competitive arena. My point is that in implementing video, a camera maker could think out of the box and use video mode as an opportunity to stand out from the crowd by offering the wider frame of CinemaScope. These are the things that would set the film making forums alight with comment.

4. What if the camera makers fail to see the attraction?
If there is no enthusiasm for it on the inside, perhaps it could be done on the outside: the Magic Lantern effort generated a lot of interest by bringing extra video usability to the Canon, so maybe there could be a future opportunity for an unofficial hack to produce a wide frame? I've also seen a website (http://pentax-hack.info/index.html) dedicated to Pentax firmware hacks. It's a non-commercial open project, though it was not clear to me what is do-able, just some pretty arcane programming information! An email I sent asking about their progress, particularly any connection with changing the existing video capability, went unanswered however.

5. Why CinemaScope and why 2k?
CinemaScope is a very evocative frame ratio for indie film makers and 2k is a very aspirational figure! This is not to say that a £1,000+ DSLR can realistically impact the RED Scarlet market or the intended user of an uncompressed camera like the Drama. Just as a prosumer camcorder has better resolution and ergonomics than this years' video DSLRs. Yet many people still bought a DSLR for video (I was one of them) to see what they could do with the shallow DoF, a large sensor aesthetic, good low light performance and interchangeable lenses. It was a good learning experience. Having CinemaScope in 2k takes things to the next level (rather than just more of the same) by providing an informal cinema camera. It's an exciting idea. It won't be perfect (using the same current codec/datarate), but it might be good enough for no budget just as current DSLR video is forgiven because of it's other benefits. CinemaScope itself won't be suitable for every project, and nor is 2k required for all projects, but it is undeniably an exciting potential.

6. If you want CinemaScope why not just crop 1080?
Yes, if you want CinemaScope on a low budget that's what you have to do at the moment, but with a huge sensor in a DSLR it seems a crime to have to do this. As shown previously, the image area of 2k CinemaScope is actually less than full HD, so I am advocating a more efficient use of the video frame for CinemaScope, rather than crop 1080 and then enlarge it slightly to 2k. I know that in the real world, most no budget indie productions are unlikely to find a theatrical window, so 2k seems unnecessary (just crop 1080 and you have your CinemaScope for a Bluray disc). But almost every film maker starts out on their project wondering if this is finally the movie that gets picked up by a sales agent for a limited run somewhere (especially as we start to go over more to digital projection, which does not require the expense of show prints). At the very least, it's very good for morale if the lone film maker can offer 2k CinemaScope for film festival projection. Saves using an anamorphic lens if you can shoot 2.39:1 directly in-camera. Down-sampling a 2k CinemaScope image for 1080 Bluray disc output might help smooth out any aliasing issues with the codec?