I need help getting clearer pictures

[tasha13662]

This came up in another post I had about my 28-135IS - I need help getting clearer pictures. It was suggested that I shoot f8-11 for sharper pictures, so I tried that.

Here are a couple of pictures I took yesterday. I think these are almost 100% crop - haven't been able to quite pin down what that means. I look at the 100% view and cropped it down to approximately what I could see when I did that.

I took one with and one without the IS on and I'm pretty sure this is the one without

http://i8.photobucket.com/albums/a8/nadina13662/468c49b6.jpg

and this is the one with

http://i8.photobucket.com/albums/a8/nadina13662/e9a410b3.jpg

but it doesn't really matter. Either way I'm not getting a focused picture and I would like to know what I can do to improve it.

I use AF because my MF skills are really bad.

Both shots are using a Rebel XTi with Canon 28-135IS, at f10, 1/80 sec and ISO 400, 135mm, handheld.

I have AF set up to only focus on the center spot (since it kept trying to focus on objects other than what I wanted.

Is my problem just with the focus or something else? The ISO is higher than I normally use for outdoors - but how much difference would it have made in clarity if I had used 100?

These may not be the best examples, but too many of my pictures are not as clear as I would like them - even when I use a tripod (however, I did recently get a remote).


Thanks for any help you can give this newbie. I have been reading through lots of posts here to learn more. It's a bit intimidating posting, but I'm trying to get over that.

[makphoto]

I was wondering why I was having such an issue with grain on macro and outdoor shots. Changed over to 100ISO - no issue anymore.

[RedHot]

Were those those in RAW? They look like they were under exposed and then pushed in RAW developing leading to a noiser picture. There could be JPG compression degrading the image more.

I suggest trying a simpler subject to try to get good sharp pictures before you go worrying about small subjects in the shadows. Try pictures in good sunlight (morning or late afternoon) of non moving objects such as a building face.

[DAMphyne]

Dragonflies are tough to take pictures of. Not to mention shooting them in the shadows, at 1/80 sec. I agree with RedHot about shooting a simpler subject, under a more controlled condition. Maybe a flower or even a butterfly. This stuff needs lots of light, and practice.
To me there looks like a fair amount of motion blur, even with the IS.
I suggest at least a shutter speed equal to the focal length, and f-8 on that lens is the "sweet spot".
Good Shooting technique is the Best way to get better photos.
Hold the camera correctly, Cradle the lens in your left hand and get your elbows in to your sides.
Stand with your feet apart, one ahead of the other.Lean-in slightly to get good weight distribution.
Squeeze the shutter-button gently Grasshopper.

[Robert_Lay]

So far as I can tell, this was shot with the exposure system indicating a scene brightness of EV=11, which is a fairly well lighted scene.

Further, it is a little under 1 Megapixel, and my guess is that it is the full frame.

For your information, a full detail 100% crop would be an arbitrarily chosen small portion of the image that is cut out from the overall image at its full resolution and is posted so as to show the detail in that small portion just as it would be seen by the person having the full size image at its full resolution. This is not a trivial concept to grasp. If you have PSCS or something equivalent, here is a procedure for making a 200 x 200 pixel full detail 100% crop, courtesy of Bobster:

Select the marquee tool - then in the options palette select Style->Fixed Size - key in 200 px X 200 px.

Click with the marquee tool anywhere in your image, and drag this around the screen as with a normal marquee.

Use Image -> Crop to complete the process.
-------------------------------------------------------
I recommend that you follow that procedure and post the results here for any image for which you need a critical review for sharpness.

I will also be providing an article in the next 24 hours, if possible, on the subject of evaluating the sharpness of an image using either the camera's built-in Playback with Zoom (which goes to 9X) or the image magnification capabilities of Photoshop (which goes to 16X). These procedures will provide the photographer with far better abilities for doing the evaluation himself than expecting other reviewers to make evaluations based on low resolution images such as those posted here.

[Robert_Lay]

Examining Image Sharpness
A tutorial on the use of both computer software and in-camera features for determining image sharpness.

Introduction
Quite often a photographer seeks an independent appraisal of the sharpness of an image in order to satisfy himself about the precision of his equipment and image capturing capabilities. Unfortunately, the reviewers are at a serious disadvantage because the image they are asked to review has been reduced in resolution from 5MP or even 10 MP to a mere 480,000 pixels in order to post it on this forum. Granted, there are other hosting options that allow the images to be posted at their full resolution, but in most cases the image will be posted on POTN at 800 x 600 pixels, which is not only a mere fraction of the original resolution, but has also been compressed to low or medium quality JPG. One solution to this dilemma is the full detail 100% crop. Unfortunately, the procedure for making a full detail 100% crop eludes most new contributors.

Magnification
Our first goal is to provide an understanding of what is meant by the "magnification" [1], of an image. Given an understanding of what is meant by the magnification level at which the image is displayed on the computer monitor, we should then consider the ability of the human eye to evaluate the sharpness of the image.

How Big Are The Pixels?
If we apply the maximum magnification possible in Photoshop to an image (16X or 1600%), how big would each pixel be? Based on a typical computer monitor [5] and the 1600% zoom capability of PSCS2, we obtain (16) (13.5")/1280 pixels = 0.168", which is quite large.

Given an image being reviewed on the LCD monitor of the Rebel XTi [4], which has 9X zoom capability in playback, how big is each pixel? Based on the maximum zoom capability of 9X or 900%, (9) (1.97")/3888 pixels = 0.00456", which is 0.325 mils at the viewing distance of 14". Here we have a much smaller pixel size - only 1/36th of that possible in Photoshop.

Comparing the two capabilities above, the in-camera capability is only about 1/36th of the capability of Photoshop). However, that is not as desparate as it sounds, because the magnification capabilities of the Photoshop software are much greater than needed for making critical evaluations of the sharpness of images. For example, a 5 MP Powershot G5 [3] provides a 2592 pixel by 1944 pixel image. On a typical video system having a resolution of 1280 x 1024, that image, at only 50% magnification, would fill the screen, and the detail would be essentially that of an 8 x 12" print. Using 100% magnification would present the viewer with detail equivalent to a 16 x 24 print, which is about as large as one should expect to get from a 5 MP camera.

If we consider the magnification range of 50 - 100% in Photoshop as adequate for a critical examination of image sharpness, then how does the in-camera zoom capability of the Rebel XTi compare? As we have shown above, the Rebel XTi provides a playback image on the LCD at maximum zoom such that the pixels are seen as 0.325 mils at a normal viewing distance of 14". This compares favorably with the size of a pixel in Photoshop at 50% magnification. (The eye sees a pixel on the computer monitor at 18" viewing distance of 0.29 mils when the magnification level is 50%)

Summary
Therefore, we are suggesting that the photographer has other attractive choices allowing him to make his own critical examination of image sharpness. Choices which are far superior to the customary 800 x 600 pixel rendition posted on POTN as a low to medium quality, low resolution JPG. Use of either the in-camera Playback with maximum zoom or using computer display software at 50 - 100% magnification will provde much more accurate reviews than are possible by posting an attachment on POTN.

End Notes
[1] The video system of the computer has a defined resolution. For example, it may be set for a resolution of 1280 x 1024 pixels. If the screen itself has dimensions of 13.5" by 10.75", then the center to center pixel spacing on the display is approximately 0.0105". Another way of looking at it is that the pixels are approximately 95 to the inch. Your video system and monitor may be slightly higher or lower.

An image is displayed on the computer monitor using the pixels of the video system. If the image happens to be exactly 1280 by 1024 pixels, and the video card and monitor are as described above, then that image will be displayed at 100%. However, if the number of pixels in the image were 2560 x 2048, then the video system would effectively display only about 1/4th of the image at a time and the magnification level would be100%. If you wanted to see the entire image on the screen all at once, the displaying program would have to down-sample the image (by throwing away 3/4ths of the pixels) so that it would have only 1280 x 1024 pixels, effectively, for display purposes, and the magnification would be at the level of 50%. Under those conditions, the video system is actually showing only one fourth as many pixels as in the original image in order to show the entire image within the dimensions of the monitor screen. Note also, that under these conditions the resolution of the original image is significantly reduced.

Conversely, if the displaying program is asked to display that same image of 2560 pixels by 2048 pixels at 200% magnification, it must use 4 pixels of the video system to display each pixel of the image, thus creating "fat" pixels that occupy 4 times as much area on the screen. The display software must effectively up-sample the image to 5120 x 4096 pixels, but will only display 1/16th of those pixels on the screen at one time. Note also, that when it "up-samples" the original image, it uses the information from one image pixel to define the value for 4 pixels. No new information is created. The effective resolution and detail is still as it was in the original image.

[2] The resolving power of the human eye is approximately 0.1 mil or about 20 arc-seconds.

[3] Powershot G5
LCD Monitor Screen: 1.47" x 1.1"
Maximum Playback Zoom: 9X
Viewing Distance: 14"

[4] Rebel XTi (400D)
LCD Monitor Screen: 1.97" x 1.3125
Maximum Playback Zoom: 9X
Resolution (size in pixels): 230,000 (Approx. 590 x 390)
Viewing Distance: 14"

[5] Typical Computer Monitor
Size: 13.5" x 10.75"
Resolution (Size in pixels): 1280 x 1024
Pitch: approximately 95 dpi
Viewing Distance: 18"

[6] Photoshop CS2 Display Characteristics
Maximum Magnification: 1600%, or 16X
Adobe Camera RAW: 400% or 4X

[oldradioman]

Bob, I’ll be the first to admit that I am totally confused after reading your “Examining Image Sharpness” tutorial. :confused:

The method I use is as follows:

In camera, I view the image on the LCD monitor and magnify it about as far as it goes to see if it still remains sharp.

On my 19” monitor I magnify the image to 100% to see if it remains sharp.

If the image passes the above two tests, it is sharp. If not, the amount of OOF is evaluated and, in some cases, can be corrected in PP.

Is this anywhere close to your procedure?

I have also evaluated some of my lenses in a very unscientific way at http://ourmnacres.us/sharpness.html

Bob, your comments and critiques are the greatest.

Thank you
Arvid
K0BCM

[Robert_Lay]

Bob, I’ll be the first to admit that I am totally confused after reading your “Examining Image Sharpness” tutorial. :confused:

The method I use is as follows:

In camera, I view the image on the LCD monitor and magnify it about as far as it goes to see if it still remains sharp.

On my 19” monitor I magnify the image to 100% to see if it remains sharp.

If the image passes the above two tests, it is sharp. If not, the amount of OOF is evaluated and, in some cases, can be corrected in PP.

Is this anywhere close to your procedure?

I have also evaluated some of my lenses in a very unscientific way at http://ourmnacres.us/sharpness.html

Bob, your comments and critiques are the greatest.

Thank you
Arvid
K0BCM
Dear OM Arvid,

What you describe is exactly what I suggest doing in my article. The difference is that I get a bit wordy at times - Hi!

Thanks for the compliment,
73, Bob, W9DMK

[Palladium]

your 2nd image

Camera Make: Canon
Camera Model: Canon EOS DIGITAL REBEL XTi
Image Date: 2007:07:16 06:58:57
Flash Used: No
Focal Length: 135.0mm
CCD Width: 6.67mm
Exposure Time: 0.013 s (1/80)
Aperture: f/10.0
ISO equiv: 400
White Balance: Auto
Exposure: aperture priority (semi-auto)

[bcap]

I didn't read any of the above comments, but I checked out your EXIF, and on both, you're shooting at 1/80th of a second. That's the answer right there. Try pushing up around 1/320th and raise your ISO and you should be good.

[StewartR]

Both shots are using a Rebel XTi with Canon 28-135IS, at f10, 1/80 sec and ISO 400, 135mm, handheld.1/80th ought to be OK, if the IS is switched on, unless you have very shaky hands.

These pictures were both shot at f=135mm. The standard rule of thumb would suggest that, bearing in mind the 1.6x "crop factor" of the XTi, you should have a shutter speed of at least 1/200th to 1/250th to avoid camera shake. But IS should give you an extra 3 stops, which means you should be OK as slow as 1/25th to 1/30th.

Two things to bear in mind here:
1. This is a rule of thumb that applies to people with decent technique and averagely shaky hands. If you've just had your 27th double espresso and you're shooting one-handed, you'll need faster speeds. If your technique is good you can go even slower.
2. IS will help to eliminate camera shake but won't do anything about subject motion.

[Trevorcjones]

Wow... I think ive done this all wrong from the beginning... I though Aperture determined how much of the image is sharp... lower number the better... danng

[StewartR]

Wow... I think ive done this all wrong from the beginning... I though Aperture determined how much of the image is sharp... lower number the better... danngAperture determines the depth of field (DOF), which is how much of the image (in front of and behind the focus point) is in focus. The higher the f-number, the greater the DOF. (That might be better or worse, depending on your requirements!) DOF also depends on the sensor size, the focal length and the subject distance, so it's not particularly straightforward. There are places on the Internet where you can calculate the DOF in specific situations: DOFMaster (http://www.dofmaster.com/doftable.html) is a good one.

But there are lots of factors that contribute to making your images sharp. These include:
- whether you have focussed accurately (using the auto AF point selection is a good way to screw up here)
- whether all of the subject is within the DOF
- whether the shutter speed is fast enough to freeze any motion of the subject
- whether the shutter speed is fast enough to cancel out camera shake

If you think you've been getting things wrong, why not post some photos (not here: start your own thread) and ask for advice?

[tasha13662]

I appreciate all the advice. I've been trying to take several shots of the same thing with different settings to see how the changes affect the shot. I am signed up for a photography class which starts next week and I think hands on help will be great.

There is so much to remember and even though things make perfect sense when I read them, I can't seem to put it all together when behind the camera. I know with some practice and patience it will get easier.

Stewart, would you please elaborate on this "- whether you have focussed accurately (using the auto AF point selection is a good way to screw up here)".

I have my camera set to focus just on the single center point. Can you be more specific about which focussing method you feel is not a good one?

[Goshawk]

This does not look like a focus problem. Actually I would say that the focus looks fairly good.
Camera shake and under exposure seems to be the culprit.
How I get sharp images:
Your lens is extremely important. Good zooms will give reasonably sharp images but not sharp enough for detailed ultra sharp images.
For that you need a good prime lens or a dedicated macro lens for macro work.
I always use a tripod or monopod if I am looking for sharp images.
If my shutterspeed is around 60 or lower I will use a remote shutter release. For macro work or very low shutter speeds I always use tripod/remote shutter release and then also Mirror Lockup.
Very few lenses is sharp wide open, best is an aperture around f5.6 to f11.
Correct exposure will give you clean images without noise.
I use expensive zooms and although they fairly sharp and superb for portraits, my 100mm Macro Prime lens that cost a fraction of the zooms gives much sharper images.

[StewartR]

Stewart, would you please elaborate on this "- whether you have focussed accurately (using the auto AF point selection is a good way to screw up here)".

I have my camera set to focus just on the single center point. Can you be more specific about which focussing method you feel is not a good one?Using the single centre AF point is generally fine.

If you have a really narrow depth of field, and the object on which you want to focus is a long way from the centre of the frame, then using the centre point can cause a slight mis-focus when you recompose to take the shot. Some people prefer to use the nearest AF point for this reason, and you'll see people saying that focus-recompose (which is what you do when you always use the centre AF point) is a bad technique. But switching the AF point takes time which you may not have, and centre point works most of the time - it's only in these extreme circumstances that you'd notice the issue.

The thing I was recomending avoiding is using all AF points, which is tantamount to letting the camera decide what to focus on. (How's it supposed to know?) We see lots of pictures here on POTN where someone isn't happy with the focussing, and relying on auto AF point selection is a common culprit.