Click here to go to the first RED TEAM post in this thread.   Thread: Red One genlock tests for 3D....

Reply to Thread
Page 2 of 10 FirstFirst 123456 ... LastLast
Results 11 to 20 of 96
  1. #11  
    Junior Member
    Join Date
    Feb 2010
    Posts
    22
    Well this is certainly bad news, and IMHO the term genlock should be removed from the feature-list if it isn't working correctly.

    I'm dissappointed that EX3 is still the best option for low-budget 3D work, I can't really consider anything else than truely genlocked cameras as an option for stereo work.

    Nevertheless I'm shooting stereo with Red's in a couple of weeks, we'll see how it looks... I hope it's not a big difference, and doesn't drift.
    Riku Naskali
    Director of Stereography / Stereographer / Stereo post-production
    www.stereowindow.us
    Reply With Quote  
     

  2. #12  
    Senior Member Pedro Guimaraes's Avatar
    Join Date
    Apr 2009
    Location
    Burbank, CA
    Posts
    611
    Nobody complained after watching My bloody valentine, U2 3D ...... I've even seen the RED footage we shot filmed out to IMAX at the California science center and I didn't notice any issues.

    If your sticking with 24fps 180 shutter you will probably never notice it.

    That still is no excuse for it not working. Especially if you want to go with faster framerates.
    ---------------------------------
    Pedro Guimaraes, SOC
    IATSE Local 600 - Operator

    Steadicam-Gimbals / Underwater / 3D cinema
    C-Motion Cvolution Wireless lens control 8 motor system for rent.
    Handsfree Segway for rent
    Ronin/Movi gimbals for Rent
    www.pgfilms.tv - Bio&Reel
    www.steadi3d.com - BLog
    www.cinemilled.com - Gimbal accessories & other cinema products
    Reply With Quote  
     

  3. #13  
    Senior Member mikeburton's Avatar
    Join Date
    Sep 2007
    Location
    Sacramento CA
    Posts
    1,742
    I have to agree with Pedro here. We present these issues publicly not to bash RED or to scare anyone away from working with the cameras in 3D but to help educate so you are aware of some of the temporary issues that remain. Don't be afraid of it, the images look fantastic and like Pedro mentioned, there are plenty of us making it work in the field. Just be aware and prepare accordingly.
    Reply With Quote  
     

  4. #14  
    Senior Member Andrew M.'s Avatar
    Join Date
    Jan 2007
    Location
    Virginia EPIC #2441
    Posts
    1,885
    Pedro, what S/N of the cameras are you using?
    Are you using MX sensor on it?

    Andrew
    Reply With Quote  
     

  5. #15 Out of step 
    Senior Member Tom Gleeson's Avatar
    Join Date
    Nov 2007
    Location
    Sydney
    Posts
    1,309
    I can confirm Pedro's results. I tested at a range of speeds and the cameras was often very close in terms of sync but not spot on.
    Again we felt that while not ideal it would not prevent us using Red Cameras in a rig. If we were to do high speed car and stunt work we would more than likely not be using the large Red rig anyway.
    Tom Gleeson
    Reply With Quote  
     

  6. #16  
    Senior Member Pedro Guimaraes's Avatar
    Join Date
    Apr 2009
    Location
    Burbank, CA
    Posts
    611
    Quote Originally Posted by Andrew View Post
    Pedro, what S/N of the cameras are you using?
    Are you using MX sensor on it?

    Andrew
    serial#'s are on my sig

    but I actually tried the test on different cameras as well!

    and no none of them were teh MX sensor.
    ---------------------------------
    Pedro Guimaraes, SOC
    IATSE Local 600 - Operator

    Steadicam-Gimbals / Underwater / 3D cinema
    C-Motion Cvolution Wireless lens control 8 motor system for rent.
    Handsfree Segway for rent
    Ronin/Movi gimbals for Rent
    www.pgfilms.tv - Bio&Reel
    www.steadi3d.com - BLog
    www.cinemilled.com - Gimbal accessories & other cinema products
    Reply With Quote  
     

  7. #17 frame grabs, 
    Senior Member Pedro Guimaraes's Avatar
    Join Date
    Apr 2009
    Location
    Burbank, CA
    Posts
    611
    whew,

    I have the weekend off back at home then it's back on a plane on monday, shooting 3D for a upcoming 3D network all next week.....si2k and Red rigs. Steadicam, handheld and some technocrane thrown in for good measure.....should be cool footage.

    Since I had a few moments and looked thru my Raid at home for the test footage.

    Here is a good example.

    4k 16:9
    23.98
    aja -1-2-7
    No phase adjustments.
    T3 on RED zoom
    1/2000 shutter
    Genlock Icon on the monitor.

    Look at the amazing disparity!

    This is 1 frame before the slate clap.

    Look at the position of the white tape on the fan blade!
    Look at the disparity of the LED display numbers. There is no first digit on one eye and on the other....
    (please ignore the color disparity, I was messing with redcineX settings)



    SAME shot as above, but 1 frame after......the slate clap.

    Look at the last digit....not even there in one eye and on the other...



    I didn't spend too much time looking for a good frame.....I have lots of this boring test footage!

    I can tell you that even when there is motion blur you can see a difference in the motion blur between eyes!

    As poster above noted, this is NOT to bash RED in any way. I love the RED. I've shot some of the most beautiful 3D footage on it. But they NEED to fix this somehow.

    Just to show the more normal shooting condition during the same test.

    Same settings as above EXCEPT.
    180 degree shutter or 1/48th

    Not much of a noticable difference. Note the slate TC is the same, the clap is the same......the motion blur is almost exactly the same.



    So this is more of a problem during high speed shots as I outlined in my original post. Because then you might be using faster shutter speeds and your eye have more time to notice the disparity of the location of objects on screen cause retinal rivalry.....

    Goodnight all.......
    ---------------------------------
    Pedro Guimaraes, SOC
    IATSE Local 600 - Operator

    Steadicam-Gimbals / Underwater / 3D cinema
    C-Motion Cvolution Wireless lens control 8 motor system for rent.
    Handsfree Segway for rent
    Ronin/Movi gimbals for Rent
    www.pgfilms.tv - Bio&Reel
    www.steadi3d.com - BLog
    www.cinemilled.com - Gimbal accessories & other cinema products
    Reply With Quote  
     

  8.   Click here to go to the next RED TEAM post in this thread.
  #18 Stereo capture 
    Red Team Stuart English's Avatar
    Join Date
    Dec 2006
    Location
    Irvine, Ca
    Posts
    3,643
    Based on the difference in tape position, 1/2000th shutter speed and rpm of the fan,

    have you worked out what the actual difference in start time between the two frames is ?

    And this is with no shutter phase trim correct?
    Workflow Wizard
    Reply With Quote  
     

  9. #19 Lissajous figure 
    Senior Member Dan Hudgins's Avatar
    Join Date
    Feb 2007
    Location
    San Francisco, CA USA
    Posts
    1,369
    http://en.wikipedia.org/wiki/Lissajous_curve

    I use Lissajous figure (like many others do) in our magfilm transport system, there is an optical encoder used for one axis and the pilot tone goes onto the other axis, you can see very slight phase errors this way.

    If the cameras had a sine wave output you could use an oscilloscope to tune the phase between the cameras to have it hold true, and you could look at the oscilloscope on each camera start to check if the camera's are in phase, that was one way to phase two 35mm movie cameras for shooting 3D as well.

    Even better the cameras could have their own Lissajous figure display superimposed over the monitor image. You can put a small pip in the wave to see if it is 180 off better as well, and such.

    EDIT: You may be able to pick up the frame rate from the current draw into the camera, you put a small power resistor in series with the camera power input and send that through AC couple to a preamp then a low pass filter to cut everything above the frame rate, then into the oscilloscope, one camera's signal on the X and the other camera's signal on the Y, then you can see the phase between the cameras. There may be some bounce between the two waves do to the cameras not doing the same things all the time, but if you have them both setup the same it might work?
    Dan Hudgins is developing "Freeish" 6K+ NLE/CC/DI/MIX File based Editing for uncompressed DI, multitrack sound mixing, integrated color correction, DIY Movie film scanning, and DIY Movie filmrecorder software for Digital Cinema. RED (tm) footage can be edited 6K, 5K, 4.5K, 4K, 3K, 2K, or 1080p etc. see http://www.DANCAD3D.com/S0620200.HTM (sm) for workflow steps.
    Reply With Quote  
     

  10. #20  
    Senior Member
    Join Date
    Jan 2007
    Posts
    291
    Thanks for the extra detail Pedro.

    This might be nothing but:

    Looking at the last two frames in Pedro's post, the 180º shutter example, as mentioned there seems to be a noticeable difference in the motion blur between the two frames. Yet, judging by the reflections on the blades the fan seems to start (assuming clockwise rotation) at the same time/angle yet appear to end ever so slightly out and, in the right hand frame, with what appears to be a little more skew than the other. (In the right hand frame the bright fan blade ending at the 7 o'clock position appears to have more pronounced skew than the left frame)

    Being 3D this may simply just be a camera+lighting+subject geometry issue but the thought does occur to me that this genlock issue maybe a chip readout/skew issue.

    ie that the cameras may actually be synced (the blades at the start of frame appear straighter and at the same angle) but there is some minor difference in the chip readout time that leads to this offset over the duration of the frame.

    Any thoughts?

    The 1/2000th sec shots appear to blow the above theory out of the water - the offset appears enormous! The fan blade is around 10º out.

    Assuming a 1000rpm fan blade: (a fairly standard speed but this is a massive variable)
    1000rpm/60secs=~16.67 revs per sec
    16.67rps/23.98fps= 0.69516263552961 revs per frame
    0.69516263552961 revs per frame*360º = 250.25854879066º fan blade rotation per frame

    10º offset between cameras/250.25854879066º rot per frame = 0.03995867493168
    0.03995867493168*100 = 3.995867493168 percent offset between cameras

    Almost 4% out between two synced cameras?

    That can't be right? Has someone accurately measured the time difference between two synced REDs yet???

    JohnF

    P.S I'm surprised other RED users who have shot synced multicam with strobes haven't noticed this before
    Reply With Quote  
     

Tags for this Thread

View Tag Cloud

Posting Permissions
  • You may not post new threads
  • You may not post replies
  • You may not post attachments
  • You may not edit your posts