Skip to main content
Up and Coming
June 26, 2014

360 S3D camera for the OculusRift

Greetings!

Here is a bit of info on the 360 S3D camera setup that I am using to capture movies for the OculusRift. The setup is composed of 3 cameras each with a fisheye lens. Each camera captures 24 fps (12-bit colors) at 2k x 2k pixel resolution. The cameras are synchronized within 0.1 ms.

The big advantage of this design is that it allows a real-time image stitching method based on epipolar geometry which minimizes misalignments due to parallax. In fact, points along the epipolar lines align perfectly horizontally regardless of their depth. There remains small vertical misalignments vertically but you can remove them for a chosen depth.

You can see a sequence I shot with the camera here:
http://www.iximage.com/street-view-sequence/?lang=en

You can also find more info on the stitching method here:
http://nasv3d.iro.umontreal.ca/chapdelv/pdf/360-s3d-camera.pdf
or the camera itself here:
http://www.iximage.com/?lang=en#technology

Thanks for your feedback.
Cheers.
-Vincent

    36 replies

    Up and Coming
    March 12, 2015
    They look like these 250 degree fisheyes
    http://www.shtjx.cn/en/index_mcplook_t4_x4_s4_i45.html

    Entapano will selling a 250 degree fisheyes soon but I am not sure if full circle fisheye is possible with their focal length
    250 lens (it is 250 degree on the image circle but the image circle is cropped vertically I think even on Gopro 4:3 formats
    http://www.reddit.com/r/oculus/comments ... heye_news/

    The Entapano 280 degree fisheye is full circle capable I think -- but only on a Backbone modifed Gopro I think -- as otherwise, and maybe anyway, the camera body will be imaged in the reverse view coverage. Also the 280 degree lens is very large physically with associated stitching complications. I think the Entapano 250 would probably be ok if you oriented each camera differently so you are using sections of the image coverage where full circle is visible (each camera provides a total of 120 degrees to the stitched L/R panoramas -- 60 degrees sector in one direction and 60 degrees in the opposite direction).

    The SHTJX lens looks very compact so that is an advantage but there is a lot of optical technology in the Entapano lenses (ED glass etc).
    joan
    Up and Coming
    March 12, 2015
    "mediavr" wrote:

    Entapano will selling a 250 degree fisheyes soon but I am not sure if full circle fisheye is possible with their focal length
    250 lens (it is 250 degree on the image circle but the image circle is cropped vertically I think even on Gopro 4:3 formats
    http://www.reddit.com/r/oculus/comments ... heye_news/


    Maybe I'm misunderstanding but I don't think it's really related to the focal length per se. The image circle projected by the entapano 250° on the focal plane has a diameter of 5.5mm according to their spec sheet. We can directly compare this to the physical size of the sensor in the GoPro, depending on the crop/downscaling used for a given shooting format. So the lens is always capable of full circle fisheye, it's the sensor that may or may not be large enough to cover it.

    I'm making this distinction because for my own experiments I'm trying to use smaller camera modules with USB streaming and interchangeable lenses, but some of these have smaller sensors (1/3" format) so the crop would be even more pronounced.

    And whatever we do, including for chapdelv's upward-facing-multi-camera setup, we cannot have even 180° FOV clean of visibility of the other lenses. (Fig. 5 in the paper, depends on lens curvature and baseline distances).

    "mediavr" wrote:
    I think the Entapano 250 would probably be ok if you oriented each camera differently so you are using sections of the image coverage where full circle is visible (each camera provides a total of 120 degrees to the stitched L/R panoramas -- 60 degrees sector in one direction and 60 degrees in the opposite direction).

    Good idea, orienting the camera around their optical axis to have the dead area on an unused part of the image.
    Although I think each lens provides a total of 240° to the stereo panorama, 120° for each eye (as seen from above), at least in the 3-camera setup. That leaves very little room to position the part of the circle that lands outside the sensor, but maybe it's still doable.
    Up and Coming
    March 12, 2015
    "joan" wrote:
    Good idea, orienting the camera around their optical axis to have the dead area on an unused part of the image.
    Although I think each lens provides a total of 240° to the stereo panorama, 120° for each eye (as seen from above), at least in the 3-camera setup. That leaves very little room to position the part of the circle that lands outside the sensor, but maybe it's still doable.

    Less wasted sensor if cameras point radially outward from the sphere center. Smaller sphere if camera bodies overlap moving lenses using off radials. Smaller cameras, but then smallers sensors crop lens coverage. Compromises emerge in designs. But S3D needs NPP camera alignment, so non-radial camera alignment must be dealt with anyway. The devil is in the details, but this is an excellent topic for discussion, and Paul Bourke gives us lots of hints along the way.

    EDIT: What's up with the funky GMT timestamps? I do not recall moving to Greenwich... Did y'all reset my profile preferences? Okay. Set my local timezone in preferences AGAIN. "More better" now. I wonder if others got their profile preferences reset to defaults too...
    This work is licensed under a Creative Commons Attribution-ShareAlike 3.0 Unported License.
    Up and Coming
    March 28, 2015
    I would love to see what the footage from this rig?
    Do you feel you can get better results with this verse Kc Lai 3 camera Z3X
    The cost of Z3X is much less of course.

    Thanks
    momay
    Up and Coming
    March 30, 2015
    You need the Z6X3D with 6 modified cameras to get 360° in stereo.

    With the 3 250° fisheyes in the zenith orientation it captures stereo to 35° below the horizon.
    Up and Coming
    April 7, 2015
    Vincent,

    Once you are using lenses that have a FoV larger than 200°, why can't you tilt the lenses down? One-third of the image captured by omnipoloar arrangement is not used. Tilt each lens away from the other two. That nadir hole could become zero even with 250° lenses if allowed to go mono.