File Size vs. Printed Photo Size

Discussion in 'Digital Photography' started by Richard Belthoff, Mar 19, 2005.

  1. I finally bought the Lumix FZ20k. It includes a variety of photo size
    options, from TIF to JPG, from 2560 to 640, from find to now to fine.
    Are there general recommendations about file size vs. the type of photos
    one might want to print? For example, if I will never exceed 5x7, can I
    reduce from, say, 2560 to 1280, and not worry about it?


    Richard Belthoff, Mar 19, 2005
    1. Advertisements

  2. Richard Belthoff

    rafe bustin Guest

    See the vast thread about "Big Megapixels."

    Bottom line... yes, of course you can make you
    file size smaller, but that limits your options
    in the future.

    *Why* do you want smaller files? More pictures
    per memory card? OK, but cards are cheap.

    It's up to you... that's why there are choices.
    That instruction manual may explain some of
    the tradeoffs, if they're not obvious.

    rafe b.
    rafe bustin, Mar 19, 2005
    1. Advertisements

  3. A general recommendation is to print at 300ppi or thereabouts.
    Following the general recommendation, 2560 allows to print 8.5 inches,
    and 1280 allows 4.3 inches. Printing larger will result in visible
    loss of sharpness, but some people tolerate that.

    Bart van der Wolf, Mar 19, 2005
  4. It's a commonly made recommendation. That doesn't mean one has to follow

    (As an aside: I find it comic that people discussing the latest and
    greatest technology persist in using inches. It's like measuring the
    cargo capacity of the Space Shuttle in bushels.)

    I suggest you run a few tests and determine for yourself what you need.
    I can't see the difference between 120 pixels per cm (300 pixels per
    inch) and 100 without a magnifying glass. I suspect that if you took a
    dozen photographs of a dozen different subjects and printed some after
    resampling to 120 and some to 100, practically no-one could tell you
    which was which on naked-eye inspection.

    At 80 ppc you still have to look closely to see any unsharpness - you
    can get very good prints at 80 ppc. Even 60 ppc is in my experience good
    enough for large prints of most subjects (though pixellation is visible
    if you have *very* high-contrast lines.)

    So 1280 will give very good results on a 16cm photograph (approx 7

    Having said that, I would recommend taking most photographs at maximum
    resolution. It gives you room for cropping, for example. Once you've got
    them on the computer you can decide to leave the best ones in original
    form and trim down the so-so photos to a smaller size if you wish to
    economise on disk space or make backing-up easier. (There are a few
    people around here with money coming out of their ears who say you
    should just go off and buy a few more hard disks, but that isn't
    practical for everyone.)

    As for TIFF/JPG, I'd suggest that medium-compression JPG is likely to be
    good enough for almost all purposes. You might need finest JPG if you
    are cropping very heavily (i.e. printing a small area out of the photo)
    or are doing a lot of post-processing. I wouldn't bother with TIFF at
    all unless and until you know precisely why you need it.
    Stephen Poley, Mar 20, 2005
  5. Richard Belthoff

    Bob Williams Guest

    You paid extra bucks to get 5MP. Why not use all those MP?
    You may THINK you won't print larger than 5 x 7 but as you gain
    experience you will be making more and more 8 x 100s
    The only reason to shoot smaller is to conserve memory card space.
    That was a valid concern a few years back when a 1GB card cost $300.
    But today you can find a 1GB card for $75. Get one and don't look back.
    Shoot at highest resolution and lowest compression.
    That is what you paid for .....use it.
    Bob Williams
    Bob Williams, Mar 20, 2005
  6. I finally bought the Lumix FZ20k. It includes a variety of photo size
    As a general rule, you want 300dpi in a final print. (400dpi looks
    better, but beyond that you don't get much improvement. A 200dpi
    print doesn't look all that bad, either). So for a 5x7, you want
    1500x2100, or about 3.1Mpix. With JPEG encoding, you can get the file
    size down, but you're really better off using compressed TIFF. Some
    pictures look fine after they've been JPEG encoded; some look

    Dr. Joel M. Hoffman, Mar 21, 2005

  7. I don't believe that you (or anyone else) can tell the difference
    between a 300 dpi print and a 400 dpi print.
    Stephen Poley, Mar 21, 2005
  8. Richard Belthoff

    Scott W Guest

    A printer like the Fuji Frontier only prints at 300 dpi so there is not
    point in having the photo at higher dpi. I can only see the difference
    between 300 and 400 dpi because I am very nearsighted, so for me
    looking at a print without my glasses is like using a pretty strong
    magnifying glass, most people would never see the difference.

    On a different point Joel has this thing against jpeg files that goes
    against everybody else's experiences, they should not be a problem as
    long as the compression is not too high.

    I am still waiting for a tiff photo that Joel has that he believes
    can't be turned into a good jpeg, I will be glad to post the jpeg as
    well as the tiff files for all to see.

    Scott W, Mar 21, 2005
  9. Richard Belthoff

    Bubbabob Guest

    If you have fine vertical lines or parallel diagonal lines, you'll see
    the difference. You need to pick a print resolution that has a
    relationship to the actual pitch of the nozzles. For all Epson and Canon
    printers it's 360 ppi. 720 works but there's not much point in it. Other
    values will cause moires.
    Bubbabob, Mar 21, 2005
  10. The human eye has incredible resolution. Most current
    printer are limited to good quality, but the printer is the
    limit, not the eye. See:

    Notes on the Resolution and Other Details of the Human Eye

    Here is a test where people can tell the difference between
    300 and 600 ppi prints:

    Save your images at maximum resolution, highest quality (if jpeg)
    to preseve the detail for the future when technology to print
    them gets better.

    Roger N. Clark (change username to rnclark), Mar 21, 2005
  11. Roger N. Clark Commented courteously...
    Thanks for the links. Interesting reading. Viewing
    distance and quite a few other factors enter into this
    issue, of course.

    Dr. Kris Zaklika of Corel's Paint Shop Pro division
    (formerly, Jasc), has published a pretty convincing
    treatise that suggests that the human eye.

    My personal opinion at this point is that the OP may have
    been more interested in the practical minimums for PPI
    rather than perhaps what might be termed the theoretical
    minimums. But, that's just my opinion.

    Here's what Dr. Zaklika had to say:

    The way you eye works is roughly like this. There are
    light sensitive cells in the back of your eye which have a
    particular size and spacing. Your eye can't resolve two
    objects that, when projected on the back of your eye, fall
    on just a single light sensor. How big an object will be
    when it is projected onto the back of the eye depends on
    the angle it subtends at the eye.

    Ignoring failures of the eye lens and the need to wear
    corrective glasses, there is a minimal subtended angle for
    an object to be resolved. Objects that subtend a smaller
    angle are perceived as one indistinct object. It's this
    angle that is critical. What this angle means is that
    small details can be seen when they are close to your eye
    but only large details can be seen when they are far from
    your eye. Stand next to a dog and you can see each hair of
    its fur but look at a dog 100 meters away and you cannot
    see the hairs (even though you can see the fur in some
    general way). All this means that the largest image you
    can print with acceptable quality depends on how far you
    will be from this image when you look at it.

    The rough rule of thumb is that at normal viewing distance
    of around 24 inches, an object subtending the minimal
    angle at the eye is about 1/200th of an inch. This means
    you need 200
    pixels per inch to show all the detail. If you place more
    pixels in every inch your eye will simply be unable to see
    them or, more accurately, will be unable to discern the
    detail they represent. Some really pedantic people examine
    their images very closely and claim that they need 300
    pixels to capture all the detail. Other people believe 150
    pixels per inch give them an acceptable picture. I would
    use 150 pixels per inch as a lower limit for images seen
    at normal distances.

    This would mean that your 3072 by 2048 image could be
    printed at 3072/150 by 2048/150 inches or about 20.5 x
    13.6 inches. However, you have to consider this. The
    bigger you print your image, the less likely it is that
    you will be very close to it when looking at it. This is
    because you won't be able to see all of the image when you
    are too close. Therefore, if you need a large image to
    hang on a wall that will typically be viewed from a
    distance of 6 feet and not 2 feet, you could print it at
    50 to 100 pixels per inch (instead of 150 to 300). This
    would allow you to make a picture as big as 3072/50 by
    2048/50 inches or about 61 x 41 inches. Of course, if you
    stood close to such a print, you would see the distinct
    pixels. This, then, is why the answer to your question
    depends on the circumstances and on what you need. The
    attached diagram is an attempt to explain some of what I
    have said in a visual way.
    All Things Mopar, Mar 21, 2005
  12. Richard Belthoff

    Bubbabob Guest

    If you want to find out what the appropriate rez for your printer is, go to
    this site, download the PDF files and print them. It will be obvious. A
    little empirical science beats a whole lot of anecdotal rot.

    Bubbabob, Mar 21, 2005
  13. This is pretty close to the actual data, but he misses a critical
    factor. His number of 1/200 inch at 24 inches is 0.43 arc-minutes,
    close to the published visual acuity of 1.7, which
    means resolving objects 0.6 arc-minutes separation. But this
    number refers to RESOLVING two objects as separate. You need to
    pixels to resolve something (there must be a pixel in between
    the two spots/lines). Thus his numbers should be doubled.
    If one takes the published visual acuity data for the human eye,
    then 0.6 arc minutes at 24 inches is about 1/240. But needing
    two pixels, that means in the print you need 480 pixels per inch.
    The above numbers are not in conflict; meaning the lower
    limit, though subjective, is generally accepted as about
    150 pixels per inch. However, this doesn't mean you can't
    make larger prints from interpolated images. I made a
    stunning (if I do say so myself) 16 x 18 inch prints from
    a 3 megapixel original (6.3 MP camera, cropped). In original
    pixels that is about 114 dpi in original pixels per inch
    interpolated up to 305 ppi for printing on a Lightjet.
    The image won in the Natures Best International photo contest:

    Subject and image impact is more important than pixels per inch!
    BUT a truly spectacular image, when printed large draws people
    in close. In a large print, say 40 x 50 inches, watch people's
    mouths drop as they approach a really sharp print, especially
    when it is sharp up close, 10 inches away. You only get that
    with large format cameras. 40 inches * 300ppi x 50 inches *
    300 ppi = 180 Megapixels. Large format 4x5 Fujichrome Velvia
    delivers such images (and more).

    Photos, digital info at:
    Roger N. Clark (change username to rnclark), Mar 22, 2005
  14. OK, we may be talking at cross-purposes. If we're talking about the
    actual printing process, I agree with your comments. There can be
    reasons for having a printing pitch finer than the resolution of the
    image being printed. But the OP was talking about photo image size.

    It's also true that my statement required some qualifications - but I
    decided to get a reaction first and then add them. ;-)

    See my other post.
    Stephen Poley, Mar 22, 2005
  15. Having made a sweeping statement, I guess I'd better add the

    Firstly: I have frequently seen it said that the major commercial labs
    do no better than 300 pixels per inch, and I have seen no evidence to
    the contrary. And based on the evidence of what I have seen friends and
    acquaintances producing, one isn't going to improve on the quality of
    commercial labs with the run-of-the-mill printers that most people have.
    One is going to have to expend considerable money and effort. In that
    context the bald statement to an apparent newbie that 400 dpi (I gather
    pixels per inch was intended) looks better than 300, with no further
    qualification, is simply silly. Hence my rather sweeping reaction.

    Having said that, I remain rather skeptical about the claims of the
    megapixelphiles as far as real photographic situations are concerned.
    Yes, in testcard-like situations there may well be a visible difference.
    But no-one I know wants to look at photographs of testcards. Yes, if one
    is photographing fine black and white print, there may be a difference.
    I use a scanner/photocopier, not a camera. The line resolution tests you
    refer to apply only to pure black / pure white line pairs. Replace them
    with medium brown / medium green line pairs and you won't get even
    approximately the same results.

    In any case, for 80% of the camera users I have met (note that I didn't
    use the word photographer) striving for anything over 300 ppi would
    certainly be a complete waste of time.
    Stephen Poley, Mar 22, 2005
  16. Rubbish. The pack of paper lying in front of me now announces its
    dimensions as 210x297mm. Nothing about inches. My local processing
    service offers 13cm photos, 20cm photos etc. Nothing about inches.
    Who said anything about 100 dpi? I didn't.
    Stephen Poley, Mar 22, 2005
  17. Richard Belthoff

    Bubbabob Guest

    Noritsu 3101 processors, pretty much the top of the line, use 320 ppi.
    Fuji Frontiers operate at 400 ppi.
    Bubbabob, Mar 23, 2005
  18. Richard Belthoff

    Bubbabob Guest

    The accomodation for metric papers is in the printer driver software. The
    ink nozzles are laid out in inches.
    Bubbabob, Mar 23, 2005
  19. Richard Belthoff

    Confused Guest

    I don't know about ink nozzles, but when this Californian wrote
    printer and plotter drivers for Epson and Lotus back in the dot-matrix
    and ribbon days I don't remember having to do any metric conversions.
    Internationalization other than accommodating Japan and the USA was
    considered, yet. The important criteria was microcomputer based
    systems that worked 100% of the time... until the paper jammed. :)

    Technology was changing so fast very few people could keep up, let
    alone worry about silly things like which measurement system was used.
    The bulk of the inventions and programs came from the USA and there
    was no need to change "metrics".

    That was then, this is now, and a lot has changed. Maybe we need an
    entirely new measurement system so everyone has to change and the "It
    must be metric!" camp won't be so uppity. ;^)

    Confused, Mar 23, 2005
  20. Richard Belthoff

    chrlz Guest

    Yes, indeed A4 is a metric size of 210 x 297. Nice round figures..

    My point was, visit ANY website, talk to any printer or photographer or
    publisher, and ask them to think in pixels per cm. You will get blank
    stares. Can you point us to any sites that discuss these issues and
    use dpc and/or ppc?

    And no, I see that you *didn't* say that you couldn't spot the
    difference between 100 and 300 dpi - on re-reading your post, you said
    I apologise - because I am *so* used to talking in dpi and ppi, I read
    that as 100 *ppi*. You didn't label it, nor did you put the conversion
    in, as you did for the first figure. And it didn't occur to me that
    you would be making a point of not being able to tell the difference
    when you increased linear resolution by only 20%...
    chrlz, Mar 24, 2005
    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.