Page 11 of 11 FirstFirst ... 7891011
Results 101 to 109 of 109
  1. Collapse Details
    Senior Member
    Join Date
    Aug 2006
    Location
    NYC
    Posts
    219
    Default
    Don't get too excited. You're not getting HD-SDI with A1, but HDV and uncompressed analog only ;-)

    As far as keying is concerned, with HDV deartifacting applied and excellent luma resolution, I dare to say Canon rivals the competition, to say the least.


    Reply With Quote
     

  2. Collapse Details
    Senior Member Elton's Avatar
    Join Date
    Dec 2005
    Location
    UT
    Posts
    2,633
    Default
    Quote Originally Posted by Bogdan
    As far as keying is concerned, with HDV deartifacting applied and excellent luma resolution, I dare to say Canon rivals the competition, to say the least.
    Good point. With greenscreen, that very high luma sampling of HDV actually helps a great deal. It worked great on a friend's film, and that was recompressing HDV 24F to DVCPRO HD.


    Reply With Quote
     

  3. Collapse Details
    Senior Member MovieSwede's Avatar
    Join Date
    Nov 2003
    Location
    Sweden, Europe
    Posts
    917
    Default
    I was always on the impression high chroma helps with greenscreen...


    Reply With Quote
     

  4. Collapse Details
    Senior Member Elton's Avatar
    Join Date
    Dec 2005
    Location
    UT
    Posts
    2,633
    Default
    The Y (luma) signal is mostly green, so a high luma sampling using greenscreen is almost like pulling a luma key, rather than chroma.


    Reply With Quote
     

  5. Collapse Details
    Senior Member
    Join Date
    Aug 2006
    Location
    NYC
    Posts
    219
    Default
    Quote Originally Posted by Elton
    The Y (luma) signal is mostly green, so a high luma sampling using greenscreen is almost like pulling a luma key, rather than chroma.
    Yes, that's true. G takes most of the weight in Y signal and with hi rez to start with, keying results can be very good with Canon's HDV. Sony V1 surely promises to deliver very nice results also.


    Reply With Quote
     

  6. Collapse Details
    Senior Member
    Join Date
    Sep 2005
    Posts
    162
    Default
    Elton,

    Thanks for the 1080i / 24f comparison. It's really quite impressive. I'd like to see some wide angle scenic comparisons. Trees really do seem to tell the story.

    The CA or "fringe" I see, do you think is the price we pay for such resolution from 1/3" chip cameras with consumer lenses? In other words, an unavoidable trade-off. The problem doesn't seem compounded by using 24f. Have you noticed any different?

    I have a question regarding the motion of 24f. I've heard that the f mode only field doubles + voodoo sometimes, leaving some of the frames or parts of the picture still interlaced. If true, this might explain why there is so little difference in the grabs. I also remember someone saying 24f has an interlaced feel. If this IS the case would a PAL model with just a straight 50i to 25f get rid of this effect? Could you please clear this up for me and others?

    Thanks a lot.


    Reply With Quote
     

  7. Collapse Details
    Senior Member Elton's Avatar
    Join Date
    Dec 2005
    Location
    UT
    Posts
    2,633
    Default
    John,

    There's no doubt that the stock lens has some CA/color fringing, particularly in different combinations of focal lengths and apertures. It's amazing how much people do fixate on the issue once it's been raised. I've played clips to many professionals numerous times and nobody ever talks about CA. It's when you pause and examine still frames that it becomes something even worth mentioning.

    I think it's an unfortunate reality in the entire category. I've seen the same fringing issues in some V1 grabs and had to laugh a little. It's a curse because I believe the reality is it just won't matter to any audience.

    There are some NLE plug-ins and even a Shake script that will massage away CA to some degree, but yes, I think this is just a reality of the category. Remember though, I posted a grab from Star Wars ROTS and it had obvious color fringing from a CineAlta and the best Fujinon HD lenses money can buy. To me, it's much ado about nothing. It's unfortunate that we can't have absolute perfection, but it is what it is.

    About 24F: I don't think it actually field doubles. You saw the Z1U/XL-H1 comparison pics, right? The Z1U DEFINITELY field doubles in CineFrame modes, and it shows. I think Canon has more clever voodoo, and as Bogdan had mentioned (phase-clocked CCD's), I think they have a different way of scanning the CCD to render very progressive-looking results.

    No interlaced feel at all, to my eyes.

    It just works.


    Reply With Quote
     

  8. Collapse Details
    Senior Member
    Join Date
    Aug 2006
    Location
    NYC
    Posts
    219
    Default
    John -

    Actually the detail is quite impressive in Elton's examples. If you look at the fence in the background, you will notice a pattern of fine diagonal lines. XL-H1 resolved them very nicely, even though the focus was set on girls. Other cameras in this class would be in big trouble in this situation, not resolving enough resolution and suffering from aliasing.

    For examples of some of the most stunning and beautiful images of nature made with XL-H1, look for posts submitted by disjecta, visit xlcinema.com forum or his website at http://www.pinelakefilms.com.


    Reply With Quote
     

  9. Collapse Details
    Senior Member Elton's Avatar
    Join Date
    Dec 2005
    Location
    UT
    Posts
    2,633
    Default
    Quote Originally Posted by John Trent
    I have a question regarding the motion of 24f. I've heard that the f mode only field doubles + voodoo sometimes, leaving some of the frames or parts of the picture still interlaced. If true, this might explain why there is so little difference in the grabs.
    I've never seen deinterlace artifacts other than the slight res drop. It just doesn't happen, and that's why I think it's something else besides field doubling or even "smart deinterlacing". Even the best deinterlacing algorithms will have an interlaced hiccup here and there and show the temporal displacements. The F modes have never had this problem, and that's where I think the phase clocked theory has some credence. Bogdan could elaborate on this idea, I guess.

    Maybe what you're referring to are pulldown frames?


    Reply With Quote
     

Page 11 of 11 FirstFirst ... 7891011

Posting Permissions

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