First time poster so let me know if this is the wrong area.
First up primarily I'm a cinematographer (not a developer or VR wiz) and I've got a feature film script on my desk which I believe lends itself to the immersive/VR cinema idea perfectly. It has 3 defined acts, each act in 1 single small location (rooms within a house).
In my research though I'm coming up against a fair few issues I guess which are concerning and this seems to be the best hive of activity for all things VR related so thought I'd post and see what answers I can get.
1. I haven't come across any examples of live action narrative story telling using VR that I find engaging.
I've looked at some of the material from Jaunt but their use of the 360 degree camera seems to mostly be as a gimmick.
Can someone point me to an example of a production which has used high end equipment (red Epics perhaps?) which has followed through and has maintained its quality to distribution.
2. Speaking of distribution, I'm concerned that if making a feature film in any VR format it seems to be locked in to 1 distribution hardware/platform.
Can someone (if its possible?) talk me quickly through the process of taking footage from a multicam Red rig like this:
And then once it is stitched together... is it possible to output it so different platforms can view it (i.e. rift/vive/PSVR/cardboard) ? OR is it that you pick your platform first and from pre production to post production you are locked in to 1 avenue.
3. Speaking of Google Cardboard... and thinking about codecs... it would appear that h264 can not handle the data required for 360 video. Even with a 90 degree viewing angle... a 360video requires 4x the bitrate of a regular HD video. Now because compression rates are already set on youtube, it means the compression for each view is 4x greater so most of the videos you can watch on youtube seem extremely blocky/low res. The playback is truly awful. Does anyone have any insight in to codec support for the different VR systems and if anyone will have native h265 support which COULD potentially support the higher bit rates required for 360 video. Doing a rough calculation my 70minute feature film, if encoded for h265 and played on screens of similar resolution to a Rift, would require 11gb of video. Anyone have any discussion/contributions as far as this as being delivered viably? 4. It would seem the majority of 360 degree videos provide the viewer with an angle of view of between 90 and 110 degrees. For anyone familiar with cinematography/lenses/photography this equates to a 9mm lens on an APS-c (super 35mm motion picture format) or a 14mm lens on a full frame camera (for those who have shot 35mm film stills).
These are EXTREMELY wide lenses and don't offer very flattering images and also offer extremely exaggerated foregrounds and backgrounds.
Does anyone know if people have shot with tighter lenses and then also offered tighter viewing on the distribution end? Ideally I'd be aiming for around 50-55 degrees of field of view which is slightly closer to the humans field of view than the extremely distorted views we are used to in VR.
5. Audio! Is 50% of a film. And I'm very worried that anyone listening using headphones will become disorientated when they are turned 180 degrees from the action and hear sounds in the wrong ear.
This post from Story Studio covers these concerns but I'm wondering what would then happen when you try and output the video for a different VR headset. I guess these are standards that aren't accepted and every delivery would require a new soundmix?
Final thoughts
If anyone has read/thought about/might have answers to any of these questions I'd LOVE to hear any ramblings, thoughts or comments.
I guess in my ideal world, it is possible to shoot in a tighter field of view, with high end equipment... Then stitch everything together and output to multiple different delivery options (different VR headsets). The playback codecs will catch up with the intense amount of data required to output a high quality image (something I haven't seen yet in any example) while the audio is also aware of the direction of the viewer.
I believe these points are major hurdles blocking the progression of VR and Cinema combining.
We work on this problematic a long time. We had started with 3D stereo movie shooting nad we move now to VR formats. Some points: real word shooting in a good quality an VR was as far as I know not up to date done. 3D stereo shooting is complex, I do not thing anybody had made a non-gimmick 360 3D shooting up to now. We are in a process of teaching our self how to generate correct image by rendering. According playback - none of existing software offer quality fully utilizing Rift futures (FPS, resolution). We had developed our code based on GPU decompression and proprietary codec (DepthQ VR Player, see www.depthq.com) - this is not a solution for common users but for relay hi-end presentation. Just now I can not imaging a H264 based distribution in enough good quality (performance reasons).
Your best bet is to get even a simple VR video setup and shoot some tests or a test scene BEFORE making your dream feature film in VR. Changing lenses, etc. doesn't enter into real VR shooting. It's a lot of work to stitch well. Reds are great for resolution but terrible for close narrative in normal size rooms. Trade off between quality and distance required to stitch since Reds are so much larger than GoPros.