Anyone playing with the DV Rack Demo?

Discussion in 'Professional Video Production' started by Loren Amelang, Sep 4, 2004.

  1. I've been experimenting with the DV Rack Demo, and have run across
    several questions that lead beyond my knowledge and experience. Would
    be interested in any comments or theories...

    The "Spectra60" component measures the brightness of any single pixel
    in the active image, in any of several formats. Using my personal
    Elura2 camera (which is all I have access to at the moment), the
    reported brightness values are constantly shifting up and down, on a
    time scale of a half-second or so. Looking at a live image from the
    camera, the extremes of variation are around +-5%. Looking at a tape
    or DVR recording, the variation is maybe half that - obviously more
    stable. Looking at a played back memory card still image, the numbers
    are rock solid.

    I assume this is the AGC hunting. Unfortunately I can't disable it on
    this little camera to prove the theory. Even that wouldn't explain why
    the variation seems reduced in a recording, whether made on tape in
    the camera, or by the DVR using the computer. I could imagine that
    something changes in the camera when the tape is running, but if I'm
    just recording the firewire signal to an external computer, the camera
    can't know anything is different.

    I thought it might relate to the image stabilization moving what is in
    front of whichever pixel is being measured, but turning off
    stabilization makes very little difference if any. I can't visually
    see the brightness variation in the monitor, but I can see the shifts
    in the DV Rack Waveform Monitor. All of the various levels seem to
    "shiver" up and down together. I guess when exposure is measured in 2X
    stops, 5% is insignificant, so maybe this is all normal.

    Speaking of levels, DV Rack immediately answered a question I'd seen
    raised here about consumer cameras. My Elura2 never puts out anything
    below 7.5 IRE, and typically runs the white limit out to 107 on camera
    images. When playing back a jpg of bars, OIRE comes out at 7.5, but
    curiously, the whitest white is limited to exactly 100.

    Watching numerically what the auto white balance does has been very
    interesting, as has comparing manual wb to the presets. What a
    wonderful learning tool!

    Oh, one more question... Playing back my camera's jpg of bars hits
    exactly the same target points in the DV Rack Vectorscope as their
    supposedly extremely precise internally generated bars. The difference
    is the connecting lines between their corner points are perfectly
    straight, and the jpg lines are slightly bowed. I always thought it
    was the corners that mattered. Do straight lines matter?

    Loren
     
    Loren Amelang, Sep 4, 2004
    #1
    1. Advertisements

  2. No, the lines don't matter. they just represent the transition between
    the saturated colors which are supposed to be centered in the target
    boxes. It doesn't seem all that surprising to me that we don't seee
    the kinds of gross colorimetry problems with DV that used to be
    common in analog formats. When you really analyze it, "playback"
    of a static image of color bars is not that much different than what
    most modern test signal generators do.
     
    Richard Crowley, Sep 4, 2004
    #2
    1. Advertisements

Ask a Question

Want to reply to this thread or ask your own question?

You'll need to choose a username for the site, which only take a couple of moments (here). After that, you can post your question and our members will help you out.