PDA

View Full Version : I dont understand what it means



kmf165
12-26-2006, 01:56 AM
ok, I am trying to see if my photo pass the "uality control" from professional website. After that they returned and say too many noise (i believe). Then I use noise remover to remove it. At the end, still same comment. Could some expert please guide me how to produce a good and acceptable photo. I am using Canon S3 and quality set to max. What is the undue artifacts means? Any good site with example?

Comments I got
==================
Unfortunately this image contains undue artifacts when viewed at full size. Artifacts are most commonly caused by over-compression but may be a result of other factors. Be sure your camera is at its highest quality setting and remember to also save your JPEGs at the highest possible quality (level 12).
==================

Thanks

toriaj
12-26-2006, 09:39 AM
kmf165, I really like that first photo! What settings did you use? For example, in a night shot like that, often a high ISO (800 or more) is used. It helps capture the picture, but adds a lot of noise. Have you ever noticed little dots of color in a photo that don't blend in with the other colors? Like little colored spots? That's noise. I don't see noise in your picture at this size, but noise is usually there, especially in high-ISO shots. Noise-reduction software, such as the one you used, can reduce noise. But it's a tradeoff, because it also introduces artifacts. I'm no expert here, but I believe artifacts are other changes in the image created by reducing the noise. When I've noticed them in my shots, they've looked like little squiggles.

This particular shot might not be salvageable, at least for the pro website. But if you could go back to that spot, you could take the shot at many different times of day with differing settings and see if you could get a "cleaner" shot.

I like the colors on your second shot, but the focus looks a little off. When taking pictures of people or animals, focus on the eyes. You can try both auto and manual focus and see which works best. You might also try using a smaller aperture (larger F-number) to help get the whole lizard into focus.

These are good shots. Keep practicing, and I'm sure they'll accept one of your shots soon.

Gopher
12-26-2006, 10:54 AM
PHoto #1 - Even on this reduced image, that sandy graininess is clearly visible. Little white dots next to darker ones next to sandy colored ones. You need to look carefully. Fiddling with settings could help.

Imagine what a good film photographer with Ektachrom 64 film and a good tripod would produce - most likely after taking about 2 or 3 rolls at different angles and exposures and maybe different film, then with expert developing technique. A Digital SLR could produce similar results, but a P&S (particularly for night photography) will undoubtably have artifacts.

Try getting a tripod, set ISO as low as possible, compose the photo per rule of 3rds (http://photoinf.com/Golden_Mean/), and do a timed exposure. See if you notice an improvement.

Suggestion: When posting a photo for analysis; do a 100% crop (small unreduced portion maybe 300x300 pixels).

CMYanko
12-26-2006, 08:25 PM
What site? Those are great shots "."

sla
12-27-2006, 12:15 AM
"Unfortunately this image contains undue artifacts when viewed at full size. Artifacts are most commonly caused by over-compression but may be a result of other factors. Be sure your camera is at its highest quality setting and remember to also save your JPEGs at the highest possible quality (level 12)."
==================

Hi
It looks like they suggest that you set too high JPEG compression. You know what JPG artifacts look like, don't you? Did you set JPEG quality to max in your Canon AND in your post-processing software (if you use any)?
Images look nice, but technically they are hard to judge, because you uploaded your photos heavily resized down.
Seems strange to me, that a photo is rejected because of "artifacts". Unless you uploaded it for a specific reason.
regards and good luck

Gopher
12-27-2006, 10:54 AM
"Artifact" is an appearent object introduced by the camera or lens.

The long answer ...

Imagine what happens in-camera after the lens drops the light image onto the CCD sensor. The CCD converts it into a digital representation of pixels, colors, and luminance values, then transfers it off to some in-camera processors that do things like boost up the color saturation, find patterns that appear to be "noise" and change their values, find patterns that appear to be blur and change their values (sharpening), boost or reduce luminance, etc.

Then it compresses it down in size to form a JPG file. For this, it does things like taking surrounding pixels and averaging their values, finding surrounding pixels that are nearly identical and writing them as "10,000 pixels in a row value x" rather than writing each of the 10,000 pixels. Note that they might not be exactly identical.

If the "ISO" were boosted up, then the pixels function to increase the "gain" (boost output), and that causes "noise".

There can also be "hot pixels" and "dull" or "dead" pixels that don't record the same, and this also causes a false representation. The in-camera processors attempt to even these out.


So...
The result is an imperfect photo. Often the imperfections are often not viewable unless you "pixel peep". Of course; a pro site's editor will "pixel peep". His client might want to use it for a magazine glossy cover, who knows.

What can you do?


Use a tripod for low light landscapes.
Use a monopod or lighting for shots like the lizard. Your S3 lacks much lighting capability, so get a monopod.
Turn off the in-camera saturation and sharpening. Or; shoot in "RAW", where these settings become reversable on your computer.
Be very careful to use correct exposure to begin with. Probably "bracket" the exposure.
Use your lowest ISO setting - without sacraficing desired shutter speed and aperture. This means finding or making adequate lighting and/or using a tripod/monopod, and doing on-computer post processing.


So; now you may begin to see why DSLR types like the bigger sensor (with bigger pixels that have less noise natively and greater dynamic range - that means how many stops of dark-to-light can be on the same photo), advanced flash capabilities, far better optics, greater weight (to further counteract camera shake), more effective in-lens IS, less in-camera monkeying around with image quality, etc. etc. It's just an advantage. You can make nice photos with an S3, but a DSLR will be pro-quality more often.

All this is nothing if you can't compose properly, hold the camera still, and have suitable subjects.