Epson P-2000/P-5000 head-to-head test results:

Discussion in 'Digital Cameras' started by Mark², Oct 30, 2006.

  1. Mark²

    Mark² Guest

    At the request of Roger Clark, I just did some time tests with the new
    P-5000 unit and my P-2000 unit.
    The results are impressive:

    2GB SanDisk Extreme III CF (completely filled with RAW+Large JPEGs from
    Canon 5D)
    -Full copy to each device:

    P-2000 time: 12:49
    P-5000 time: 4:35 (!!) -A second copy test rendered the EXACT same time

    4GB SanDisk Extreme III CF (completely filled with RAW+Large JPEGs from
    Canon 5D)
    -Full copy to each device:

    P-2000 time: 26:34
    P-5000 time: 9:27

    *These numbers actually EXCEED the claims by Epson (of 250% copy-speed
    This speed means not only less time, but also more battery life.

    Now view image viewing times:
    Scrolling and view times are even more impressive:

    Rendering the standard screen of 12 thumbnails:

    P-2000: 10.5 seconds
    P-5000: 1.2 seconds (approximate)

    Rendering 64 small thumbnails:

    P-2000: N/A
    P-2000: 3.2 seconds (and you do NOT have to wait for the screen to can keep scrolling quickly to subsequent pages.

    *Note: The 64-Thumbnail view is surisingly useful due the amazing
    high-definition 4" screen running a full 24bit color scheme.

    Rendering full-screen individual images:

    RAW files (Canon 5D, 12.8MP files):

    P-2000: 8 seconds (painful)
    P-5000: 1.2 seconds (!!)

    Rendering full-screen JPEG images were the same as above.

    *Epson claimed a 500% improvement in rendering speed, but again...they've
    EXCEEDED their promise in my tests.

    The bottom line:
    After viewing with the new model, you'll be at pains to ever want to browse
    with the old one.
    It's simply a TOTALLY different experience.

    That alone may be reason enough for many to upgrade, but the looong list of
    other improvements are extremely significant as well. The timing figures
    speak for themselves. No ocmparison. Battery issues...customization...and
    on and on.

    If anyone was on the fence, take my word for it: You won't be disappointed.

    Mark², Oct 30, 2006
    1. Advertisements

  2. ??? I don't understand what you are doing here. On my P2000, displaying an
    image full screen takes minimal time (about 1 second). ???

    David J. Littleboy
    Tokyo, Japan
    David J. Littleboy, Oct 30, 2006
    1. Advertisements

  3. Mark²

    Mark² Guest

    If you're getting 5D images to render that quickly, then it is almost
    assuredly due to having already viewed them once previously. RE-viewing
    images (after the initial rendering view) will be decent on the 2000/4000,
    but that's hardly a help when you want to quickly view newly-shot images in
    the field.. I'm clicking on a newly-created thumbnail of a full-quality 5D
    jpeg or RAW image for the *first* time.

    My numbers are accurate, and very closely reflect the numbers claimed by
    Epson (actually better than Epson's claims).

    Mark², Oct 30, 2006
  4. Mark²

    just bob Guest

    I'm interested to know what Epson claimed for this operation on P-2000. I
    doubt you will find "8 seconds" anywhere on their literature and I would be
    pissed if I had bought one.

    Thanks for your work on this.
    just bob, Oct 30, 2006
  5. Mark²

    Mark² Guest

    According to's review data, a 7MP image-to-image display time
    was 3.2 seconds.
    When you nearly double that to my 12.8MP 5D images, (adding the possibility
    of the resulting decrease in operating memory "head-room" left for system
    operations when the image itself eats memory) isn't too hard to believe
    a time of 6-8 seconds for shot-to-shot views based on dpreview's times. My
    P-2000 *could* be a bit slower due to many cycles of use due to Epson's lack
    of defragmentation functions...but my numbers are too far off of the
    extrapolated guess for a 12.8MP image compared with their tested 7MP image

    See here:

    As they note...once images have been cached, the view-time is much faster.
    But again, this is of little assistance when your purpose is making an
    initial assessment in the field.

    On the same chart, their card-copy time for a 1GB Sandisk CF card (Ultra II)
    is 469 seconds (7 minutes 49 seconds). My P-2000 times for **2GB was
    12:49...which is a bit better over-all, but my card was Ultra III.

    Mark², Oct 30, 2006
  6. Mark²

    Mark² Guest

    See Bill Hilton's comments within my other thread, "Epson P-5000 a winner!"
    on this forum.
    He's happily used one in addition to his P-4000 (which is the P-2000, but
    with 80GB instead of 40).
    Mark², Oct 30, 2006
  7. Mark²

    Mark² Guest

    Can you verify that it's a cache issue on your device?

    See my other post about dpreview's numbers regarding pre-cached
    image-to-image viewing.
    Here's a link:
    Mark², Oct 30, 2006
  8. Mark²

    Mark² Guest

    Also notice that my times for thumbnail page creation are right on par with
    dpreview's cited number (They cite 11 seconds for initial render...mine was
    10.5 seconds)
    Mark², Oct 30, 2006
  9. I just snapped off five shots, dowloaded them to the P-2000, and went
    through them one at a time. The first time through, each image displayed in
    under 2 seconds, the second time a bit faster. So there does seem to be a
    cache phenomenon. And who knows how long it would take per image if there
    were 150 images in the directory.

    My P-2000 was purchased in Japan, and may be different from overseas models.

    David J. Littleboy
    Tokyo, Japan
    David J. Littleboy, Oct 30, 2006
  10. Mark²

    ASAAR Guest

    If the P-2000 can fully power off (instead of just faking it), you
    might try sending another set of five shots to it, powering off,
    powering on and then displaying them to try to eliminate any
    possible caching. I don't know how much memory the P-2000 has, but
    presumably it wouldn't have enough to effectively cache copied files
    if you tried displaying them after copying 100 files to the P-2000.
    Another thought. If the P-2000 creates any kind of thumbnails on
    the fly, are they created as the files are being downloaded, or as
    they're being displayed for the first time? I wouldn't think that
    this would take very much time per shot, but it might be noticeable
    if the goal of the P-2000's designers was to first get it out the
    door and then optimize the code for future versions. If they're
    created prior to first display, that could also explain the slightly
    longer first display time you noticed.
    ASAAR, Oct 30, 2006
  11. I don't ever remember my P-2000 being painful to actually display an image.
    (I only save RAW, so it's displaying the tiny jpeg embedded in the RAW,
    which it won't zoom.)

    The thing my P-2000 is painful for is displaying the thumbnails. It shows 12
    per screenful, and does seem to cache them, but the cache size is limited,
    and as soon as it gets to the next page, it gets glacial.
    Again, there seems to be a cache, and they may be caching the first page of
    thumbnails for each directory, but by the time you get to the second page,
    it gets slow. But displaying an image full screen isn't bad at all, even if
    you ask it to display an image it hasn't displayed the thumbnail for.

    Again, this P-2000 was purchased in Japan, and may act different from those
    purchased outside Japan.

    David J. Littleboy
    Tokyo, Japan
    David J. Littleboy, Oct 30, 2006
  12. Mark²

    Mark² Guest

    Perhaps that's true, though I can't imagine Japan would use two different
    If what you're saying is correct, then you're also differing greatly from
    dpreview's unit.

    In any event...the 5000 is miles ahead of *my* 2000.
    Mark², Oct 30, 2006
  13. SNIP
    Maybe Raw+JPEG is different (perhaps it could try to downsample the
    JPEG instead of extracting the Raw thumbnail)?
    Bart van der Wolf, Oct 30, 2006
    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.