Western Railroad Discussion > Digital Editing and Monitors (Need Technical Assistance

Need a little help here...

I have a pretty good feel for Adobe Photoshop for PC and have a pretty descent assortment of processed photos but occasionally I come across a few mysteries which I need to put to rest.

Most shots are taken set between 3.3 and 8 Megapixels
(3504 X 2336 low compression).
which is the highest resolution, largest size and lowest compression.
JPEG for now - haven't played with the RAW setting yet.
Monitor is a quality(?) Samsung 191T SyncMaster 19" Digital.
Resolution for Windows is set at 1152X864 anything more and my nose would be pressed against the screen.

1) Why is it that when I am editing in Photoshop, some photos only look sharp at certain magnification - 25% to 50% perhaps. In order to show the entire shot while editing, my photos need to be "shrunk" to 16.7% or 12.5%. At the lower magnification the details (number boards and logos for instance)look chopy and lines jagged and wavey yet magnifying to 25% or 50% they become razor sharp. Going to 100% some begin to get blurry (newspaper printlike).

2) What does the % indicate? Viewing a 12" X 8" image at 100% clearly does not represent viewing at "full size" 12" X 8" as my 15" X 12" monitor can only show perhaps a quarter of the shot.

3) Regarding monitor calibration - when calibrating my monitor with Adobe Gamma or "Natural Color" the correct colors appear way to dark and contrasty - I find it difficult to work with unless lightened - I think there may be a photograph setting available, but I then get a message in Photoshop stating my monitor profile appears to be defective.

4) For submitting photos over the Net for processing onto film what resolution is optimal versus overkill for 8"X12" or less? 600dpi? 300 dpi? Obviously for Trainorders photos 72 dpi is sufficient.

Thanks for any advice!

> 2) What does the % indicate? Viewing a 12" X 8"
> image at 100% clearly does not represent viewing
> at "full size" 12" X 8" as my 15" X 12" monitor
> can only show perhaps a quarter of the shot.


At 100%, one pixel as displayed by your monitor is the same as one pixel of your image. Other percentages are ratios of that.

I've seen recommendations to turn your monitor brightness as high as you can stand, then back off on notch.

For trainorders pictures: 75dpi, 12 X 9, JPEG compression to fit below the 300kb cut off. Make sure to View as Print Size or 100% for best viewing.

For keeper pictures: 100-600dpi but if you increase the size of the photo and keep the resolution the same you'll get pixelated (the choppy effect). Decreasing the size of the picture by Resampling the Image (check out Image Size option under the Image menu, then click the check box for Resample Image) will increase the dpi as your size decreases. 600dpi is typical for images to be printed in magazines, which is why you can have a larger image from a camera with more pixels.

Answers to #1 / #2:

There is only one true measure of a digital image - its size in pixels. Each pixel can be as large or small as it needs to be, as long as it's square. Any conversion into a conventional unit of length (cm, mm, inches, etc.) is dependent on the resolution of the output medium. Let's say, for example, that you have a 2400x1800 image. A monitor might have 72 dpi (dots per inch, essentially pixels per inch). So, displaying at 100% means that each pixel in the image corresponds to a pixel on the monitor. Thus, you'd need a 33.3" x 25" monitor to fully display the image. A high resolution color laser printer might have a native resolution of 600 dpi, so "full size" there would be 4" x 3". That's oversimplifying a bit, since printer dpi is a bit of a marketing number, and somewhat off from the true, effective dpi.

So, if you display at 25%, that means that you're displaying one pixel for every four vertically and every four horizontally - smashing 16 of the original pixels down into one for display purposes. As you smash them further, you can no longer distinguish the gradients between areas, making the image look sharper. Likewise, when not done on even divisions (say, displaying at 63%, which would put 2.51 pixels into every one), you'll often get strange image artifacts as a result of the way a program averages pixels together. These are the zigging and zagging and jagged edges you see - aliasing from having downsampled the image.

3) No idea...

4) DPI is a measure of the output device, but yes, a reasonable assumption is that most monitors are in the 72-100dpi range. (72 is oft-quoted, but the LCD I'm using to write this is 100dpi) If you're processing onto paper or film, your best bet is to have some guess as to what the output device's resolution is... Ideally, if you have a 300dpi (effective) printer, you want an image that's 300*(height in inches) high by 300*(width in inches) wide. For photographic prints (real ones on photo paper, not inkjet or dye sub), I find that staying around 300 gives a nice, sharp image, but 200 is more than acceptable. Most consumer prints really probably don't have an effective resolution much over 150 anyway... Obviously if you have a 600dpi printer and want an 8"x12" print, there' no need to blow up your original to match the DPI of the output device. Blowing up an image with a photo editor doesn't add any information to the image, just bloats up the size a lot. Best to just submit the largest thing you've got if the desired pixel size is larger than the original image.

Most computer users are now running screen resolutions between 800x600 - 1600x1200. My general rule of thumb is small enough to fit on the lower end resolutions and to be downloaded by modem users, but large enough to be appealing and to allow details to be picked out. Usually around here I set the longest dimension to 650-700 pixels on my picture posts, since that will almost fill an 800x600 monitor given TO's sidebar.

My other advice is *ALWAYS* save the unmodified originals. Once you resize an image, or even resave a JPEG, you can never go back - information is lost in the process. Having the high resolution original comes in quite handy in the future when printing technology gets better, or we all get computers with 52" monitors with multi-mega-super-broadband connections and can pull down 8 meg images like nothing... Trust me, I started shooting digital nearly seven years ago with a camera that did a whopping 1280x1024 (1.3 Mpixel). I didn't save some of the originals, just the resized 800x600 images. Oops.

Nathan

Not sure I'm clear on each of your issues. But on the jagged edge thing... The smaller the images is, the less pixles you have creating a curved or angled line. The larger it is, the more you have, and the better the pixels blend together to create the curved or angled line. It's kind of like the dot count in a printing process... The more dots you have, the more blended the colors and lines will be. There's also an issue when you make the window smaller to work with the image. The app. compacts that image quickly so you don't have to wait as long, and when it does it that way, it doesn't put all the pixels in a smooth transition on curved and angled lines. It makes them more stepped, because that takes less time to process.

When you want to print off a file, there's more to it then just the resolution. You also want to use the least compression possible, or none at all ideally. The dpi will depend on what kind of printer will be used. 300 will be OK on an average home printer, higher can be used on a pro printing system.

Not sure what the calibration problem is... Sounds like a profile might be corrupted or something. But you also want to make sure PS is set to the same gamma as you have the PC set to. Otherwise there can be a difference between how they look. A way to tell is open a photo in PS, then open the same one in a browser, and see if there's a difference. They should look identical.

Dave
-DPD Productions - Featuring the TrainTenna LP Gain RR Scanner Antenna-
http://eje.railfan.net/dpdp/

Thanks for the advice to date.

I realize as you shrink a displayed image you loose information and detail which can make the displayed image less accurate or smooth. In other words a curve becomes triangular or jagged....

My confusion, which may require a bit of research, is why some pictures I have taken look fine in smaller display windows (16.7% for instance) while others at the same resolution and size appear fuzzy. Those same fuzzy looking shots are razor sharp at 25% or 33.3%. What is the reason for this?

One frustration is in using the Unsharp Mask - I'm concerned that a razor sharp photo may be incorrectly adjusted in an effort to sharpen it when in fact it is the monitor that is really the culprit in the percieved sharpness of the photo.

Is is plausable that shots taken at a very wide angle (18 to 24 mm) are more susceptible to this condition - looking for a link, it appears most of my 50mm shots and especially the zoomed shots are the ones that look sharp throughout the magnification zones.

Mgoldman Wrote:
-------------------------------------------------------

> My confusion, which may require a bit of research,
> is why some pictures I have taken look fine in
> smaller display windows (16.7% for instance) while
> others at the same resolution and size appear
> fuzzy. Those same fuzzy looking shots are razor
> sharp at 25% or 33.3%. What is the reason for
> this?
>
> One frustration is in using the Unsharp Mask - I'm
> concerned that a razor sharp photo may be
> incorrectly adjusted in an effort to sharpen it
> when in fact it is the monitor that is really the
> culprit in the percieved sharpness of the photo.

Beware sharpen/unsharpen. Use only when viewing at 100%......my rule of thumb is to view at 100% when doing any editing other than crop and resize.

On #3, I'd offer the suggestion that the calibration issue may come from monitor settings - the system is sending a calibrated signal to the monitor but your monitor settings may be out of adjustment, things like brightness, color temperature, contrast, etc. may not be set optimally to deal with the signal it's receiving. Now it may be that those can't be controlled at the monitor or somehow it's drifted, or even room lighting, flourescent versus incandescent bulbs, too bright or too dim, bright window light, etc., can impact viewing.

On the "jaggies" and similar issues, the others have covered it pretty well. You are taking a rectangular array (of pixels) and squeezing it onto another rectangular array and often at odd multiples, then consider that the subject matter varies in "texture," rates of color or contrast changes, different shapes and curves, etc. So that "fit" isn't always going to be that good. Jaggies are most apparent where "lines" intersect at shallow angles and the gradation between pixels on a monitor is a lot coarser than that of a print.

The dpi "tag" isn't that important, some cameras "save" at 300 dpi, some at 72 dpi but that's generally irrelevant in actual use. Whether saved at 300 or 72 dpi, when displayed, it's displayed at the monitor resolution, like 1024x768, say the pictures's horizontal resolution is 2048, you'll get half the picture displayed - on a 17" monitor, a 13" monitor or a 21" monitor. If you adjust for a "full width" or similar setting, it squeezes the 2048 into 1024, not to bad a squeeze but edges and things can go off.

Something else to consider: the native resolution for this monitor is 1280x1024- running at other resolutions may impact the quality of the display, admittedly a problem for text and icons, you may need to reset them to larger sizes, but it might be diddling the viewing of pictures?

pacificeclectic Wrote:
-------------------------------------------------------
> Something else to consider: the native resolution
> for this monitor is 1280x1024- running at other
> resolutions may impact the quality of the display,
> admittedly a problem for text and icons, you may
> need to reset them to larger sizes, but it might
> be diddling the viewing of pictures?

CRT monitors generally don't have problems displaying one resolution vs. another. LCD monitors, such as the one on this laptop, tend to look absolutely horrible at anything but their native resolution. It's a tradeoff, as LCD monitors can be sharper but also often have problems with color and brightness shifting when viewed at different angles.

Just a couple of questions to help diagnose your problem:
What version of Photoshop are you using?
How much memory does your computer have?
Is your monitor set to hi-color? ie 16, 24,or 32 bit?
Are you using a DSLR or P&S camera?
Does your computer have a seperate hi-end graphics card?
tom

Generally, when resampling an image chose Bicubic. Use Bicubic smoothing when going to a larger image to smooth out all those jaggies and bicubic sharpening when making a smaller image. Older versions of Photo shop and some other programs work best if you resample for larger images in steps of about 10% enlargement each time until you get to the size you want. If you want the image to be 50% bigger then you resample it 5 times. The latest Photo Shop programs, CS and CS2 do not seem to need that. I do not know about Photo Shop Elements 3 and 4 in that regard.

Some screens have a 96 dpi resolution as opposed to 72 dpi. Read your monitor instructions to see which is best for yours. Using 75 dpi in a 72 and 96 dpi world will create all sorts of problems with the screen dispalys.

ALWAYS, I repeat ALWAYS save the original file from your camera on a CD-R just as it came out of the camera. Then if you do not like changes you have made you have the original to go back to. ALWAYS use the Save as function after you have made changes to save the changed file under a new file name so you do not corrupt the original file.

Have fun. Get to know the Levels tools, you will be amazed what is in the dark and light parts of the image you do not see at first. Do not be afraid to move the sliders all the way to the ends of their range for a look at what you have. You have saved the original on CD so you cannot hurt the original, explore the image.

But this is an lcd monitor.