Capturing back off a DVD

Discussion in 'Video Cameras' started by Paul, May 4, 2004.

  1. Paul

    Tony Morgan Guest

    Opps ! :~(
    Sorry, that should have read 'BTW this is _not_ intended as a flame.'.
    [/QUOTE]
    Heaves big sigh of relief :)
     
    Tony Morgan, May 11, 2004
    #61
    1. Advertisements

  2. Paul

    Jukka Aho Guest

    This has been bugging me too. I understand that generating
    a perfect, standards-abiding test card with correct levels,
    useful features etc. for any video format requires a great
    deal of care and knowledge (more so than ordinary people
    looking at the pretty bars would usually think.) I also
    understand that people with the necessary tools and knowledge
    to create these decent and useful test cards often do _not_
    want to part from their creations without having some monetary
    compensation for their work - and that seems
    to be why there are not too many free offerings on the web.

    What I do _not_ understand, however, is why there are so
    many _hideous_, _bad_, _unusable_ test cards out there on
    various web sites. OK, some of the sites featuring these
    broken test cards are just tv buff nostalgia shrines
    (galleries of old test cards from days long gone) but not
    all of them are! Every now and then you find a site with
    "test cards for free download", and they almost invariably
    turn out to be mushy JPEG renditions with hideous artefacts,
    often with a non-standard size, and the colours are not
    calibrated to any standard. Totally useless, unless all
    you want to do with a test card is "display that colourful
    bars picture", which isn't very useful.
    I just recently found a test pattern generator - though
    not specifically for the DV format - that is actually
    free, from a respectable company, and is touted to be
    (video) standards compliant.

    <http://www.synthetic-ap.com/products/tpm/index.html>

    The blurb at the top of the page tells us that this tool
    has been made available for a free download specifically
    because there are so many non-standard, useless test
    cards floating around!

    The catch? It is available for PPC Macintosh systems only. :(

    I have not had a chance to test it at all since I do not
    personally know anyone with a PPC Mac system. If someone
    on this group actually _can_ try it out, I would love to
    hear some comments, and maybe see some test cards generated
    with it as well.

    * * *

    Of course, even if it works and (supposedly) produces decent
    quality test card stills in some useful uncompressed bitmap
    format, there is still the problem of "unknowns" - which you
    already mentioned - when encoding the cards to DV or to any
    other format.
     
    Jukka Aho, May 11, 2004
    #62
    1. Advertisements

  3. Paul

    Jukka Aho Guest

    Note that "test1A.jpg" <http://www.camcord.info/test1A.jpg>
    - despite supposedly having been captured from the analogue
    outputs of a consumer DVD player, using the analogue inputs
    of a consumer DV camcorder - does not actually display any
    of the typical tell-tale signs of an analogue capture.

    1) There are no blanked half lines (nor line 23 WSS
    signalling, which most of the modern DVD players should
    be inserting to the signal they output) at the top or
    bottom. All lines are complete!

    2) Additionally, the left and right edges are not falling
    to black at all but retain their brightness perfectly, all
    the way to the very edges of the image.

    It is just a little bit too perfect to be an analogue
    capture. :)

    Also note that the image is not in 720x576 DV pixel format,
    but in 768x576 square-pixel format, which is suspicious as
    well.
     
    Jukka Aho, May 11, 2004
    #63
  4. Paul

    John Russell Guest

    I'm willing to accept that Tony has done what he has described elsewhere.
    But this comes from someone who started this thread believing that vob's
    couldn't be directly used, and hadn't heard of lossless editing thorugh
    smart rendering, believing that the MPEG had to be rendered from the vob,
    not extracted or "stripped". We shouldn't be surprised if any novice in DVD
    ripping screws it up at first.If he didn't generate the MPEG from the vob in
    a matter of a few mins then he wasn't ripping it. If the software he used
    took longer to generate the MPEG than it would take to capture it then the
    software was obvioulsy re-rendering.

    Of course it's obvious a mistake has been made becuase to accept the results
    one is having to accept that a vob as bad as image B) can produce an image
    as good as A) when played on a DVD player. Hardware decoing is good but not
    that good! The logic of "rubbish-in , rubbish out" has to hold true.
     
    John Russell, May 11, 2004
    #64
  5. Paul

    Tony Morgan Guest

    Snipped a lot more of interest...

    When I started my "testing" - at that time simply to get a "standard"
    clip to use for comparison of various video editors (or more accurately
    the codecs that came with them), I actually started to hand-produce an
    image that could be used to produce a clip. I gave it up partly because
    it was too fiddly, and partly because I didn't know enough about
    "television standards".

    Having said that, I still feel that a useful tool for the sort of
    comparisons that I was considering could be produced that would give:

    1. Vertical line resolutions between (say) 200 lines and 550 lines.

    2. Grey scale from white to black with (say) 10% increments.

    3. R, G and B strips going from 100% hue to white for each.

    I suspect that the traditional "test card", being intended to set up and
    test analogue monitors/TVs, would not be appropriate for testing
    domestic miniDV camcorders or the software for digital video
    editing/manipulation - no doubt the experts here will give an opinion
    with due regard to this declared purpose of such a "tool" - i.e.
    comparisons not absolute measurements.

    Providing that it could be printed with your reasonable-quality PC
    printer on photo-quality paper large enough to avoid being in "macro
    mode" then I believe it would also be useful in assessing lenses
    (especially the WA and Tele add-ons.

    Naturally, there are a lot of "ifs and buts" - the one which immediately
    springs to mind is the conversion of a still image to a video clip is
    dependent on the video editor doing the initial conversion - but at
    least it would give a "level playing field" for any comparisons that it
    would be used for by individuals with what equipment/software they had
    available.
     
    Tony Morgan, May 11, 2004
    #65
  6. Paul

    Jukka Aho Guest

    The first obstacle on the way of designing a perfect test
    card image is that it should be constructed directly,
    natively in the YUV (YCbCr) colourspace, instead of using
    the RGB colourspace as a starting point. Ordinary paint
    packages generally do not let you work in the YCbCr
    colourspace so the raw image data must either be
    constructed programmatically or by treating each of
    the channels as a separate grayscale image and later
    recombining them to a single image.

    After having done that, the image should be encoded to DV
    (or MPEG-2 or whatever) format using an encoder that does
    not muck about with the colours, and the consistency of
    the colours should be checked and double-checked. (Some
    DV codecs, for example, rescale and/or clamp the colour/
    brightness values - any such modification is unacceptable
    for a test card.)

    Of course, you could get even more technical than that and
    actually _hand-craft_ raw DV or MPEG-2 data on the DCT block
    level (creating something like colour bars or frequency
    gratings at the desired levels should not actually be too
    hard to figure out if you can manage with the other things
    involved), but that would require even more obscure tools
    and quite an intimate (and impressive!) knowledge of the
    standards and formats.

    I guess there is a reason why properly designed, clean test
    images (and DV or MPEG-2 test streams) are usually not freely
    available...
     
    Jukka Aho, May 11, 2004
    #66
  7. Paul

    Tony Morgan Guest

    Thanks for that Jukka. All way out above my knowledge, capability and
    "tool" availability. I was looking towards something more "KISS" and
    pragmatic.

    Anyway, notwithstanding your very informative post here, I will perhaps
    go ahead to produce something for my own comparative purposes - but not
    until later in the year. It's the "Movie shooting" time of the year -
    and not the "playing around on the PC" season.

    Added to which I'll be "out of the frame" from next week for a little
    while since I go into hospital for eye-surgery - John Russell will no
    doubt be able to make some capital out of that :)
     
    Tony Morgan, May 11, 2004
    #67
  8. Paul

    Jerry. Guest

    Sorry to hear that Tony, hoping it's not to serious, that all goes well and
    your posts once again grace Usenet ASAP.

    Take care :~)
     
    Jerry., May 12, 2004
    #68
  9. Paul

    Tony Morgan Guest

    Thanks. By today's standards and technology - not all that serious, just
    a new lens. Main problem is psychological - they break the old one up
    with an ultrasonic hammer, suck it out, then put a new plastic lens in -
    all while I'm wide awake :-(
     
    Tony Morgan, May 12, 2004
    #69
  10. Paul

    Mark A Guest

    No problem, just look the other way.

    Regards

    Mark
     
    Mark A, May 12, 2004
    #70
  11. Paul

    Keith Laws Guest

    I have a video of my father's lens replacement. It might be worth taking
    a tape along if you also want to see it from the surgeon's POV. The
    microscope they use to do the op is plumbed into a CCTV and VCR, so it
    can be used for teaching presumably. They seem to use S-VHS VCRs.

    Back on topic, kinda ;-)
    --
    Keith Laws

    What's my solution?

    .....NOISE POLLUTION
     
    Keith Laws, May 13, 2004
    #71
  12. Paul

    Jerry. Guest

    Should be able to use a VHS tape though, the VCR should auto switch
    depending on the inserted tape.

    HTH.
     
    Jerry., May 13, 2004
    #72
  13. Paul

    Keith Laws Guest

    Yep
    --
    Keith Laws

    What's my solution?

    .....NOISE POLLUTION
     
    Keith Laws, May 14, 2004
    #73

  14. if you visit the webboard at www.dvdoctor.net you wll find a thread by
    alan roberts who has high quality test cards for download free of
    charge.
     
    Gary MacKenzie, May 21, 2004
    #74
  15. Paul

    Jukka Aho Guest

    Could not find the correct thread (Mr. Roberts seems to have written
    quite a lot of messages there), but a little bit of free googling
    revealed that the man has a web site here <http://www.computervi
    deo.net/video-test-cards.html> and hosts the test cards there.

    On a related note, I have already begun working on a test card
    of my own, constructing it directly in the Y'CbCr colourspace
    (instead of the RGB colourspace - I do not trust the R'G'B' <->
    Y'CbCr conversions that various codecs do in different ways,
    as it seems.).
     
    Jukka Aho, May 21, 2004
    #75
  16. Paul

    Jerry. Guest

    No, that is the web site of Computer Video magazine !
    Do a search on the uk.tech.broadcast group, Matt Overton has already done
    much work on various cards, and he did have them available for download at
    one time.
     
    Jerry., May 22, 2004
    #76
  17. Paul

    Jukka Aho Guest

    So? It is still Alan Roberts' page - even if hosted by CV
    magazine - and on that page he specifically mentions the
    discussion about test cards he participated on the DVdoctor
    message board and offers his cards for free download.
    I have seen them, and even participated in that discussion
    myself. However, the last time I paid attention to it, he
    was having problems with his levels (there was some confusion
    about whether to use ITU-R BT.601 ranges or full ranges, and
    I recall someone jumping in and pointing out this error) and
    the geometry of the circle (it was initially shaped wrong,
    and he corrected it by redesigning the card in square pixels,
    then decimating it back to ITU-R BT.601 pixels, which does
    not sound like the best way to do it.) Also, he is working
    in the R'G'B' colourspace while my card is drawn directly
    in the Y'CbCr space, which should ensure the correct levels
    and "legal" colours - without having to rely on the colorspace
    conversion algorithms of some 3rd party program.
     
    Jukka Aho, May 22, 2004
    #77
  18. Paul

    Jerry. Guest

    <pedantic>
    So if you are implying that it is a redirect, and not a page on the CV
    'server' you would be wrong, the page content might well be by Alan Roberts
    but it most certainly is not his web site...
    Yes, I remember now - doh !..

    For now, which card(s) would your run with, those of Matt Overton or those
    of Alan Roberts ?
     
    Jerry., May 22, 2004
    #78

  19. if you look at a few of his threads you will see where he worked and
    what his job was.

    i think you will find he really knows his subject.
     
    Gary MacKenzie, May 22, 2004
    #79
  20. Paul

    Jukka Aho Guest

    My apologies, of course you're right. I mistook the "Alan Roberts
    () writes:" line for a byline, and
    thought he was an employee of the CV magazine, hosting some half-
    informal-but-still-related personal stuff on his employer's web
    site.

    Anyhow, the main issue here was not who owns the computers the
    page is hosted on but where to find the test cards - and you
    can find them from that address.
    Roberts seems to have a more complete collection of cards, but
    I have not studied them in any detail yet. The interesting
    question, as always, is whether they retain the correct levels
    and legal colours when you run them through your toolchain and
    codecs - and if they do not, the problem, as always, is finding
    out whether it is a fault in the card itself or a fault in the
    colourspace conversion algorithms some of the tools in your
    toolchain use. (Which is one of the reasons why I feel it is
    necessary to have some reference in the Y'CbCr colorspace, too,
    instead of R'G'B' alone.)
     
    Jukka Aho, May 22, 2004
    #80
    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.