Epson R800--new twist

Discussion in 'Digital Photography' started by Mark Herring, Jan 16, 2004.

  1. Mark Herring

    Mark Herring Guest

    It just dawned on me that the new Epson R800 no longer uses "photo"
    cyan and magneta--i.e. lighter shades of the two colors to help get
    good tonal scale. They have also added red and blue.

    AND, they are advertising 1.5 pl drops.

    1. How do they get the same tonal scale? e.g. is the red and
    blue ink a lighter shade? Or is it the much smaller drop size that
    compensates for fewer colors?

    2. What motivated this seemingly radical change in the
    technology? I can't find anyone else that has done it. What
    performance issue were they trying to address?

    3. Now we have to readjust out thinking. If you thought you
    understood subtractive primaries in printing, how do you relate to
    mixing red ink and cyan ink? Sound like black to me.

    Interesting side note: The HP 7960 is staying with the old 6-color
    paradigm (CMYKphCphM) + the extra gray and black. Curiously, this is
    not immediately available on their website---I found it at Reinks.
    Thus is seems that Epson's move is not a reaction---unless they are
    just doing something to be different.
    Mark Herring, Pasadena, Calif.
    Private e-mail: Just say no to "No".
    Mark Herring, Jan 16, 2004
    1. Advertisements

  2. Mark Herring

    leon Guest

    check it out here:
    leon, Jan 16, 2004
    1. Advertisements

  3. Mark Herring

    Mark Herring Guest

    An amazingly complete review, but does not directly answer most of my
    questions. As I read it, the purpose of red and blue is to increase

    I find nothing to refute the postulate that the small drops offset the
    lack of diluted colors.

    Looks like a fabulous printer.
    Mark Herring, Pasadena, Calif.
    Private e-mail: Just say no to "No".
    Mark Herring, Jan 16, 2004
  4. Mark Herring

    David Chien Guest

    AND, they are advertising 1.5 pl drops.

    See full review here: R800/page_1.htm
    "Colours on the R800 are nothing short of superb. I have never had a
    better quality from the photo-i test print, skin tones are more lifelike
    than anything I have seen before and I have seen allot of printers over
    the years."

    "Would I buy one? - Yes,

    The R800 is for the creative photographer who wants the highest quality
    pigment ink printer currently available. I hope Epson don't leave it too
    long before they introduce a larger format printer using the new
    UltraChrome inks and Gloss Optimizer. Wouldn't it be nice if the delay
    on the 4000 printer was due to a re-design to incorporate the new inks.
    All we need now is the announcement of the Epson Stylus Pro 2300 - just
    wishful thinking."
    The R800 is basically the same as the Japanese PX-G900.

    See expanded color gamut picture showing you how the extra colors help
    and here:
    on this page:

    Basically, the Red and Blue help cover areas of the sRGB gamut that
    previously could not be covered by a 6+ color printer, and basically
    expands the print range to cover more of the standard sRGB gamut.
    What technology change? Smaller dots? That's been shrinking for the
    past 10+ years of color inkjet printing. More colors? That's already
    been done 10+ years ago in the press world with 6-color Hexachrome and
    other 6+color systems, so desktop printers are only 'catching up'. (And
    even they don't use lighter shades of a color.)

    Different ink colors? Nobody's agreed what's the 'best' set of
    colors to use beyond CYMK - every maker will disagree and every printer
    will have a different goal in mind.

    Encapsulated inks? Epson introduced this years ago with their
    widebody, archival printers (2200 series and older), so that's not new
    news. Of course, the 'heat' they received over the Epson 870/1270etc
    prints fading quickly to orange (
    probably was a big motivator to move to more archival ink sets.

    (Of course, even on Epson Japan's website, they toot how their latest
    printers with the new PX/PM series of 'improved' inks last far longer
    than their prior Epson printer prints - which they show fading in less
    than a year in their advertisements even!!.

    eg. see this comparison between new Epson and old Epson

    Sucks when you think all of those older Epson prints aren't that long
    lasting (and if you remember, they initially were tooting these older
    printer models to make long-lasting prints! - something they retracted
    after the orange-fading fiasco became well-known.)
    Actually, think crayola's and pointilism. Basically, if you have
    just 4-colors, there's no way you'd make a true blue or red just by
    mixing the 4-colors you have. You can get close, but never quite a pure
    blue or red. Adding these two colors allows prints to contain these
    shades, and you simply do what has been done before - dither and space
    out the dots to get lighter shades.

    On the page, the dots can be placed side-by-side, and when you use
    different colors beside each other, they'll combine together when viewed
    at regular viewing distances just like an impressionist painting or
    CRT/LCD monitor to produce an apparent shade of color in between.

    This apparent 'mixing' of colors when viewing small 'dots' of color
    at long distances is in play in CRTs, LCDs, magazine prints and inkjet

    That said, it works, and you get more shades of color and an expanded
    print color gamut with more ink colors used.

    Of course, one of the colors that they still would need to put in to
    truely make people happy is a good orange, one of the tougher colors to
    produce from other colors they're using today in inkjet printers. (ala
    Hexachrome, which I believe introduced a nice green and orange - correct
    me if I'm mistaken)
    HP is targeting the B&W print market with their extra gray/black
    colors, but for real - how many of us out there print B&W? Yes, a few,
    but a very small minority vs. the many other color users. (just like
    those who still use B&W film vs. color)

    While it's interesting to see HP target what is still Epson's market
    for B&W prints ( was selling the Piezography BW system
    for Epson printers many years before HP even got into this; still a
    great, gorgeous system for B&W enthusiasts if you're into making
    archival, museum quality B&W prints), they haven't tacked the other
    issues of having a large selection of archival, museum type paper
    choices quite like the Epson + Piezeography BW system has with 3rd party
    papers and archival ink sets.

    Full details of the Piezography BW system for the B&W print enthusiast here:

    History of Piezography BW system here:

    (See Yahoo! Groups for the Epson Inkjet forum for very lengthy,
    in-depth discussions on this B&W Piezography system if you're interested
    in making quality B&W, archival inkjet prints.)

    ( Of course, with 4 ink cartridges in the 7960 system, one of which
    you must swap out and keep unused in the spare storage compartment, it's
    a bit silly when you think about it -- why not simply have all four
    cartridges colors available all the time like the Canons and Epsons? A
    pain in the butt if you think about it - I want nice B&W prints - swap
    cartidige, I want nice color prints - swap cartridge, oh, I want nice
    B&W prints again - swap cartridge...

    Still, it is not a bad printer by far (see review): 7960/page_1.htm )
    David Chien, Jan 16, 2004
  5. Mark Herring

    Tony Guest


    Actually, David, the thing that struck me most was that even this new R800
    still shows signs of not being entirely accurate in the blue through purple
    color range. I've been using Epson 6 color printers for a long time, too
    (although have not yet replaced my 1270) and have used Colorvision's Print
    Profiler to make my own profiles (as well as using Epson's own profiles).
    One thing that has always bothered me about the Epsons is color reproduction
    through this range. Look at the spools on

    page 3. Granted, this is with an uncalibrated R800, but if you even look at
    the spools on page 10, those last two spools on the right are still (IMO)
    not as accurate as the Canon i950. A lot of my photos tend to have colors in
    this range, and I can never seem to get them accurately reproduced on the
    1270. Is the R800 better? It looks like it might be (for this issue), but in
    looking at this review I'm starting to think the Canon i950 has a better
    color accuracy across the range.
    Tony, Jan 17, 2004
  6. Correct. An interesting change.
    Ah. There's the problem. There is no "subtractive primaries" in Epson
    (or Canon) inkjets. The various ink colors are not "layered", mixed, or
    combined in any actual way. The appearance of a full range of output
    colors is accomplished by dithering many tiny drops of the available
    ink colors. If the dots are small enough, and the dot spacing is close
    enough, the appearance will be that of pure colors, to the unaided eye.

    Dye sublimation printers use a subtractive approach. That is also why a
    300 "dpi" dye sublimation print looks as good or better than a 1440 "dpi"
    inkjet print. The dye sub is producing a real 300 24-bit-color pixels per
    inch. The inkjet uses a much higher dpi pattern of its few colors to achieve
    the same effect.

    I tried an experiment where I printed 3 scans of the same original photo,
    where each scan was set to produce 150, 300, and 600 ppi of information
    at my desired print size. After printing each sample on my 4 color Epson
    880 printer at 2880 x 720 dpi, I found virtually no difference between the
    3 prints (to my unaided eye). Essentially, that says my 880 is only capable
    of printing an effective 150 Pixels Per Inch. Going about it another way,
    I took a rough stab at how big a block of pixels would have to be to represent
    24 bit color with 2880 x 720 dots per inch of 5 colors (3 colors, black,
    and the white paper). As I recall, my calculation came out in the ball park
    of 150 ppi as well. (Actually, I think the numbers actually came out to some
    thing like 18-20 bit color at 150 ppi.)

    Anyway, the upshot is that more colors (what ever they are) will allow higher
    resolution. I believe that the use of light cyan and light magenta were originally
    chosen to address large areas of light color with gradual tonal variations.
    Typically, this was an area where earlier 4 color inkjects were particularly
    I was under the perhaps mistaken impression that HP adopted the extra
    grey and black to improve grey scale printing.

    Given the high quality outputs I've seen from Epson's printers, I'm quite
    interested to see what their new idea does in the R800.

    Dan (Woj...) dmaster (at) lucent (dot) com

    "They took all the trees, and put em in a tree museum
    And they charged the people a dollar and a half to see them
    No, no, no / Don't it always seem to go
    That you don't know what you got 'til it's gone
    They paved paradise, and put up a parking lot."
    Dan Wojciechowski, Jan 21, 2004
  7. No, it says that something in the *system* has a resolution not much
    better than 150ppi - and that could be, and probably is, your unaided
    eye. A good adult eye ought to be able to resolve 250ppi when properly
    corrected, so you might see the difference between 150 and 300ppi, but
    you are unlikely to see any difference between 300 and 600ppi without
    some magnification.

    Also, your choice of source resolutions is not optimised for the Epson
    printer range, which should be integer divisions of the driver's native
    resolution. For Epson's current desktop range that is 720ppi,
    irrespective of the dot placement resolution, and for their wide body
    professional range it is 360ppi. ie. you should have used 180ppi,
    360ppi and 720ppi to avoid any resampling artefacts from being
    misinterpreted as resolution differences. (Any resampling artefact
    would reproduce as a coarser pitch effect, due to aliasing.)
    That would be fine if the printers used fixed half tone cells, so that
    each cell was capable of printing the full colour gamut of the printer.
    However that is not how these inkjet printers operate. They use a
    stochastic dither process, where the error between the colour that is
    required and the closest approximation the printer can achieve in the
    pixel is equally distributed into the neighbouring pixels. The upshot
    of this is that the printer achieves the full colour gamut across many
    pixels - roughly corresponding to the estimate that you made. However,
    where sufficient contrast is present in the image, the resolution can be
    much higher - up to the native resolution of the driver itself. Thus
    the printer automatically trades off colour precision for resolution.

    There is nothing new in that principle, it is similar to the process
    used in TV systems, where the luminance resolution can be as much as 4x
    that of the chroma resolution. They both exploit the fact that your eye
    is less critical of the precision of the colour at high resolutions and,
    provided the resolution is present in the image, the colour precision
    can be distributed over a much coarser scale.

    I regularly print contact strip pages of all of my films, with each
    image on the page measuring around 4.5 x 3cm. These images are normally
    viewed with a magnifier (approx. x4) and the images certainly contain an
    awful lot more resolution than 150ppi. I have printed test patterns
    which prove conclusively that an Epson desktop can achieve 360cy/in (ie
    720ppi equivalent) when the contrast in the image is adequate.
    Kennedy McEwen, Jan 22, 2004
  8. ....
    \> >880 printer at 2880 x 720 dpi, I found virtually no difference between the
    Actually, quite true. For the record, I used my unaided, corrected (to 20-20)

    Again, point taken.
    A 720 dpi Epson printer certainly can't display any more output resolution than
    720 ppi. Given a 4, 6, or 7 color printer, even this could only be achieved if the
    pixels were alternating uses of the colors available to the printer. (The 4, 6, or 7
    inks + white from the paper.) Hmmm, given your point above of error diffusion
    and human perceptual differences between chroma and luminance, I'm willing
    to grant a slightly greater color gamut to 720 ppi. Still, I see these cases as
    fairly unrepresentative of real world photographs.


    Dan (Woj...) dmaster (at) lucent (dot) com

    "They took all the trees, and put em in a tree museum
    And they charged the people a dollar and a half to see them
    No, no, no / Don't it always seem to go
    That you don't know what you got 'til it's gone
    They paved paradise, and put up a parking lot."
    Dan Wojciechowski, Jan 27, 2004
  9. A common enough concern, but of little substance. The purpose of using
    a test pattern is not to represent a real world image, but to address a
    particular aspect of real world images in a repeatable, controllable and
    measurable way. Nevertheless, and despite being exceedingly
    non-intuitive, it is a mathematically provable fact that any and every
    real world image can be decomposed into the summation of a suitably
    weighted series of sine and cosine intensity patterns. ie. test
    patterns. Similarly each and every image can be reproduced from the
    summation of a suitably weighted series of test patterns. This
    procedure is a fundamental part of the jpeg image encoding scheme, with
    the level of compression merely being determined by the precision and
    accuracy to which those test patterns and their relative weightings are
    computed. So the argument that resolution test patterns have no
    relevance to real world images is, in fact, completely groundless.
    Kennedy McEwen, Jan 27, 2004
  10. Mark Herring

    Burtron Guest

    Hello All-
    Here's a another website to review the R800: - It's
    very thorough & detailed - With lots of pictures for comparison. Of
    course, I posted this in an earlier post.
    Burtron, Feb 9, 2004
  11. Mark Herring

    devans Guest

    I heard that the R800 is a ink guzzler from a post on the
    . It makes sense but I haven't heard
    anyone else discuss the number of 8x10 prints from a fresh all full set
    of cartridges. Here is his reasoning:
    Can anyone shed more light on prints per ink comparasion? Also does this
    printer suffer from metamerism ? I'm looking for a high quality 8x10
    full facial photo printer w/ good facial tones
    devans, Apr 20, 2004
  12. Mark Herring

    David Chien Guest

    I heard that the R800 is a ink guzzler from a post on the
    1) Almost all inkjet printers made today (with exception of
    higher-end models with larger tanks), will produce about ~50 letter
    sized photo prints on a set of cartridges regardless of brand, make or

    I've tested Canon through HP to Epson, and all of the models I've
    tested do about that same amount no matter what. (These inkjet makers
    are smart! --- make you go out and buy lots of $$$ cartridges!)

    No, you won't find much variation from that in general, unless you
    specifically find a large capacity ink printer, or install a bottle ink
    feed system such as the CIS ink system sold by The
    CiS ink system will significantly reduce ink costs if you're doing lots
    of prints per month.

    2) Yes. at home inkjet photo printing is $$$. Expect about $0.30 -
    $0.50 cents per 4x6" print you make regardless of brand, make or model,
    in general.

    Heck, even paper alone costs you. eg. cheapest brand-name out of
    HP, Canon & Epson is Epson 4x6 100pk @ $15, or $0.15/4x6" print w/o cost
    of ink!

    3) See R800 detailed review here: R800/page_1.htm
    David Chien, Apr 21, 2004
    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.