Enlarged digital images with more detail than the original

Discussion in 'Photoshop Tutorials' started by Douglas, Jan 22, 2006.

  1. Douglas

    Douglas Guest

    For most of last year I endured insults defamation, jibes and threats as a
    few idiots ( being a main offender) attempted to coerce me into
    releasing industrial secrets of my process to people who never even bothered
    to identify themselves and probably wouldn't know what machine code looked
    like in the first place, even if I had. I can only presume their motives
    were to steal my work and profit from it themselves. Why else would they
    continue with their defamation and lies about me?

    These crackpot drop outs and wannabe photographers who even today, have
    refused to accept evidence or proof that everything I have said is true and
    not just possible but highly repeatable and available today, from my digital
    print service, continue with their taunts and stupid beliefs it can't be
    done.

    Some Norwegian mathematician seems to think I'm in the charity industry with
    his challenges and then admits he is not qualified to judge me when offered
    the chance.

    Next month Sally Walker-Brown's Exhibition of historic Boggo Road Jail
    photographs, shot with a Nikon D70, goes on display. All the pictures have
    been enlarged by my company and printed on canvas. All the pictures are
    larger than the (300dpi) print size the original images would print at. All
    are as sharp as the original image and many, have more detail - as defined
    by the crackpots ( and Tacit) - than the originals. Sally has graciously
    allowed the use of her copyright image for this demonstration of proof.

    Pictorial evidence that it is indeed possible to enlarge a digital image
    which has a normal print size of 6.5" x 10" at 300 dpi, to 24" x 36" poster
    print with 720 dpi and still maintain the same sharpness and detail - even
    adding detail which was never there in the first place.

    There is a brief description here: http://www.tecphoto.com.au along with
    pictures of the results and clips at the actual print size to prove once and
    for all that this process of mine not just works but works better than any
    commercially available software running under Windows or Apple operating
    systems.
     
    Douglas, Jan 22, 2006
    #1
    1. Advertisements

  2. Douglas

    isnot Guest

    What in heavens name does machine code have to do with anything?
    Nobody wants your 'work'. And how would defamation help them take anything?
    We can't see your prints except over the net. They may as well be postage
    stamps.
    Hot air. We can't see the images. And on CANVAS? What's with that? Arty?
    Textured?
    More air.
    None of the links on the left work. Convention says they should be links.
    Cut in machine code, are they?

    Anyway, keystoned, distorted pictures sell there? I'd say that's a pretty
    uncritical market. The EXIF data says it was shot with flash. Oh my, erred
    data or clumsy photographer?

    The "perspective corrected" image is incorrectly corrected. There is just so
    much one should even try to do with a poorly made original. The "gates of
    hell" section is far to small to show a darned thing. There should be no
    noise at that light-level, exposure anyway. The greater detail one is no
    surprise... and by the way, it's not properly perspective corrected either.

    Sorry. I simply see nothing extraordinary whatsoever. So you have a
    nominally decent digital print made with a fairly large picture. Routine.
     
    isnot, Jan 22, 2006
    #2
    1. Advertisements

  3. Douglas

    isnot Guest

    ....but I can appreciate coding on an Alpha system. :) My last Alpha 400 had
    1gb boards that were about 24" x 24". :) Fast bus, though.
     
    isnot, Jan 22, 2006
    #3
  4. Douglas

    Mike Russell Guest

    ....
    I've looked carefully at your web page, and have a specific question about
    the letter G as it appears in the last two images in the series:
    gates-of-hell.jpg and finished.jpg.

    In gates-of-hell the G in the word GAOL is about three pixels wide, and so
    little structure is visible that I could not guess what letter it might be.
    In finished.jpg, the G is 7 pixels wide, and it is completely obvious that
    this is the letter G: the outline of the letter is very clear, and even the
    bar in the center of the G is clearly visible.

    My question is this: was the G reasonably clearly visible in the original
    captured image or not?
     
    Mike Russell, Jan 22, 2006
    #4
  5. Douglas

    HornBlower Guest

    Not to mention shot in JPG. What kind of Pro photographer shoots pictures
    for an exhibition in JPG. Not impressed with the samples. As isnot said
    until we can see them printed and side by side nothing to be impressed with.

    R
     
    HornBlower, Jan 22, 2006
    #5
  6. Douglas

    Douglas Guest

    Are you expecting an air ticket and accommodation to see the exhibition?
    Just how precise do you think an Internet display actually is?
    If you care to tell me how you post a 273 megabyte file - or even a pixel
    depth part of it you can identify on a web page, I'll try to accommodate the
    wishes of an anonymous critic *NOT*.

    If you want some dialogue on the topic, identify yourself and stop making
    silly comments.

    FWIW, Newspaper Photographers - which Sally Walker-Brown is a very respected
    one with at least 6 awards to her credit... Routinely shoot in jpg so their
    images are instantly accessible for publication. The notion everyone goes
    around shooting in RAW and has the time to developing each individual image
    afterwards is quite at odds with the real world of a working Photographer.

    Besides this... Uncompressed or lightly compressed JPG images "pre
    Processed" in a modern DSLR's onboard computer are pretty much free of
    artefacts and quite easy to interpolate.

    Any half smart Photographer who understands their gear, can set their camera
    to produce the features they want and have the JPGs ready to use, the minute
    they are on the flash card. I don't have any problem with JPG images, unlike
    the plethora of RAW specification floating around in the wild.
     
    Douglas, Jan 22, 2006
    #6
  7. Douglas

    Douglas Guest

    No it's not mike, this image was a throw away for this reason. The
    photographer agreed to let me manipulate it on the understanding I could
    publish to results as I have done. I knew I could make major improvements to
    it. Even if it were not to be enlarged, just creating a vector mask could
    have achieved the same thing.

    This area is one randomly chosen by my program during the search for
    portions suitable to convert to vector data. It's still early days (after 3
    years development!) in this application. Eventually I hope to have a
    forensic application law enforcement agencies might use to do what the TV
    shows seem able to perform now; Enhance obscure detail to make it readable
    and identifiable. For the present I have to eat so I use the application to
    make a living enlarging digital images.

    Probably what prompted your question, is the conversion of any definable
    outlines to a vector mask which makes the outline predictable and therefore
    enhancable. - adds detail, in Tacit's words - I have had the most difficulty
    with this function. To my knowledge, 3 other developers are working on this
    feature for their own enlargement applications and none have managed to
    master it.

    You might notice a 'bend' in the horizontal of the image, this is
    unfortunately a result of a Nikon lens anomaly which is magnified with the
    vector mask. It is not an issue for direct prints. Only when I try to plot a
    path using flawed image information does it rear it's ugly head.

    Today, I can't fix that. Maybe in another few months things will be
    different. Maybe new lens manufacturing techniques might surface soon.
    Ideally every outline should be a mathematical projection, filled with
    interpolated bitmaps. Fantasy at the moment but it's getting closer. This
    picture is probably representative of the state of development in this area
    today. Certainly it's the state of MY development today.
     
    Douglas, Jan 22, 2006
    #7
  8. Douglas

    Mike Russell Guest

    If this works as shown you are indeed in possession of something major that
    will transform the world of photography and image analysis.

    I suggest a further test, consisting of a picture of my bookcase, with some
    letters larger and others smaller than those in the jail image. If you run
    your software on it, and get something resembling the result shown on your
    web page, I will be thoroughly convinced of the your claims.
     
    Mike Russell, Jan 22, 2006
    #8
  9. Douglas

    chrlz Guest

    (Copied from aus.photo)
    OK, that's the claim, so let's look at exactly what Douglas is
    displaying, shall we?

    First, the D70 has an image size of 3008 x 2000 pixels. At 300 dpi,
    that does indeed give a very sharp print at 10" x 6.7". Douglas is
    correct.

    But let's look at the (presumably full-frame) image that he displays
    here:

    http://www.tecphoto.com.au/interpolation-2.htm

    and compare it to the 'detail' crop here:

    http://www.tecphoto.com.au/interpolation-4.htm

    It can be clearly seen that the crop is actually about *one-third* of
    the entire image. OK, so it's not exactly a highly enlarged section -
    but then, just how enlarged *is* it?

    Well, if the original image is 3008 pixels wide, then that section
    *should* be approximately 960 pixels wide if displayed at actual pixel
    size. But it has been presented at less than 800 pixels wide (786 to
    be exact, but let's not quibble). In other words, it is not quite an
    'actual pixels' view, it is slightly reduced. OK, no great problem
    there, and the letters *should* be reasonably sharp - they are about
    what I would expect. They don't represent particularly fine detail in
    such a large image file.

    But hang on.... Douglas claimed that crop was supposed to be from an
    *enlarged* file (in his words - "a clip at the actual print size") for
    a 24"x36" print. That means it should be 2.4x bigger again. So not
    only is the detail slightly short of actual size, even if it was.. it
    would still be 2.4x short of the enlargement he claimed to be showing.

    (And it's already showing a bit of fuzz... (O:)

    I would invite anyone else with too much time on their hands on a
    Sunday evening (in Oz..) to check my measurements and mathematics, and
    I will apologise humbly if wrong..

    Please note, you may need to get in quick, as Douglas tends to pull
    images if they don't get the reception he wants.

    I repeat, this is not meant to indicate that Douglas cannot do fine
    enlargements, and I'm sure Sally's exhibition will be wonderful - but
    the claim and the demonstration do not match, by a long shot.

    By the way, on the 'perspective correction' image:
    http://www.tecphoto.com.au/interpolation-2.htm
    ...it is quite clear that Douglas has taken the pillar to the left of
    the entrance, and cut and pasted it (very clumsily) over the right
    side.. That is why so many of the brick layers don't align, there is a
    repeated bike stand (?), and the window support has eerily vanished
    halfway down. He has used it again on the lefthand pillar - note the
    bikestand and shadow, *again*.. (This was initially spotted by Graham
    Fountain in aus.photo)

    I'm guessing he initially used PS's perspective tool, but when he found
    that didn't quite work, out came the clone tool...

    Either that, or it must be a *very* interesting algorithm... (O: ..
     
    chrlz, Jan 22, 2006
    #9
  10. Douglas

    isnot Guest

    That kind of manipulation is unacceptable under the auspices of a show done
    by a journalist. First, the image was stupidly distorted in the camera, and
    it's an image that was clearly not taken under duress - it could have been
    done properly in the first place.With all the Digital Magic applied to that
    image, and the time taken to do it, the photographer could have shot it
    properly in the first place with camera perspective controls on a larger
    format. (and how embarrassing that the photographer used flash at a great
    distance, for gosh sakes.)

    Finally, it would be interesting if this were a photo exhibit of an object
    that one could just walk down the street and see in person.

    Regarding the creation of vectors from raster images - yes, in special,
    simple cases such as rendering type, and much simpler geometry it can be
    made to work, but making it work on so-called natural things like trees,
    building, people won't happen within the complexity accommodated by any
    digital computer. That's where the myth of Genuine Fractals was created
    twenty years ago under the misapplied hype of Chaos theory and imaging
    (that, for example, natural, organic objects could be represented as sets of
    fractals). You see, such cases that work require first that the object BE
    vector-susceptible and be defined algorithmically - which is the definition
    of compressibility. Big scientific names have screwed up in this case of
    problems. It will turn out that with an image of sufficient complexity to
    be worthwhile that there will not be enough atoms in the universe to serve
    as bits to solve the problem. It's even beyond NP-Complete. To put it
    another way - human beings who become satisfied with any outcome that CAN be
    accomplished will have to re-define the terms of the problem, be happy with
    the outcomes even when the outcomes drastically fail to meet the objectives.
    (The same reason so many people are happy with poor quality.)
     
    isnot, Jan 22, 2006
    #10
  11. Douglas

    Harry Limey Guest


    Sounds fair and a good test for the software and might obviate all the
    invective flying about (I notice it even includes the unfortunate
    photographer now)

    I think Mike is of more than sufficient standing in this little community to
    be accepted as a fair and impartial judge of the outcome! although no doubt
    we will all be able to see the finished results!
    What say you Douglas?
     
    Harry Limey, Jan 22, 2006
    #11
  12. Douglas

    HornBlower Guest

    You just made our point. Posting those images on the web and expecting
    people believe that they were enlarged and had increased detail due to this
    new method is bull.

    There is no proof in those images what-so-ever.

    R
     
    HornBlower, Jan 22, 2006
    #12
  13. Douglas

    chrlz Guest

    Just to save anyone's time going to dead links...

    Douglas:
    Oh no there isn't. As predicted, Douglas has pulled all evidence of
    this fiasco.

    I'm so suprised

    yawn
     
    chrlz, Jan 23, 2006
    #13
  14. Douglas

    MetaMorph Guest

    I got this when i tried to open the web pageThis domain you are seeking is currently mothballed

    It is highly unlikely it will ever be populated again.
    For sale if you would like to buy it.

    Feel free to Email the owner from this link
    <<
     
    MetaMorph, Jan 25, 2006
    #14
    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.