PF Photography: Tips, Tricks, & Photo Sharing

[Andre]

I see your point Jonathan, however imo the body shape of the bird looks sufficiently free from the edge, arguably. Anyway with that suggestion, maybe we do a cropping exercise, everybody posting his/her own preference.

Here is the orginal small JPG direct from the camera, crop as you like and reduce to maybe 25%.

http://dl.dropbox.com/u/21940023/IMG_1352.JPG

 

[turbo]

Here you go.

The bird's eye is still about on the center-line, the right-most third of the frame is "blank" except for the beak, and the diagonal line of the birds' posture lends a little dynamic. Edit: I would like just a little more background at the top of the image, but it wasn't in the original. Still, it would make the composition a little more ideal for me if I could have 5% or so more background at the top.

 

[Jonathan Scott]

I'm happy with turbo's version.

 

[Andre]

Thanks,

As I infer from the paper, the test persons preference peaked at an offset of 42/58%

I used the "optical center of gravity" like this:

Turbo put it on 43% and I used 39% whereas the rule of thirds would demand 33%.

I have to say that my choice was also influenced by my desire to get close to a common rectangular frame ratio, which I did not specify before.

 

[Andy Resnick]

Here's what I came up with:

[PLAIN]http://img231.imageshack.us/img231/6255/img1352crop12.jpg

 

[Borek]

Late to discussion...

1. Thirds rule seems to me to be a poor's man version of golden ratio. Could be that's because I have learned golden ratio rule many years ago, could be that's because I have seen on too many occasions people using bastardized rules instead of the right ones, when the right ones are beyond their comprehension. (And if you were ever trying to help in homework sections you know that ratio of ratios IS beyond comprehension of Joe Average). In practice the difference is not that large - it is enough to move the object about 5% to the side to switch between both rules.

2. I like the first picture most (intuitive crop) but I agree with Jonathan that there is not enough space of the left. Perhaps the best approach is to change the width of the picture, leaving starling eye in the exact center? Interestingly, seems to me that would be the same as cropping original picture symmetrically.

Great shot BTW. Spring is coming, I have heard a lot of singing birds yesterday.

 

[Andy Resnick]

One of the papers Fuzzyfelt posted (#979/#980) specifically mentions that the rule of thirds could be a simplified version of the golden ratio.

Something to consider as well- film formats (and now sensor formats) are not square- I haven't sat down and characterized the aspect ratio of them all, but they all appear to be golden-ratio-ish. Which is odd if you think about it- lenses are circular; why standardize a non-symmetric film shape?

 

[Andre]

Thanks for your interesting crop Andy, looks very powerful that way.

Borek, the centered eye is logically explained as I used the center focus point on the eye to focus and then I recomposed by moving the center straight down before shooting, so the eye stayed in the vertical centerline.

Maybe I'll also upload the full processed Raw later, to demonstrate the difference. But chores first

 

[fuzzyfelt]

Yes, the ratio was mentioned, but there are so many related names. Arnheim's "centre" was also mentioned. The ratio and thirds are close, but the thirds is restricted to one grid, whereas the golden ratio is about proportions, which allows flexibility, and amongst other things, can be applied to other aspects of composition more, too.

I see I was wrong about the level of difficulty the ratio involves, sorry. I’m not in the position to compare, it and others like it, being the maths I've studied and spent time on, and I had the wrong impression.

There is lots of research on the topic of the ratio, and tests, and also, I thought there was one especially good discussion I read here some years back, I'll keep looking for it.

Nice to see so many opinions. Scrolling down gave me a cropped bird itself, which I liked, too, as well as Turbo’s.

I liked more of Andy’s opinions of film, too.

 

[Andre]

Meanwhile http://dl.dropbox.com/u/21940023/IMG_1352-2.jpg is the full processed raw, for Borek to see that the basic 70-300mm zoomlens doesn't have to disappoint too much when hooked onto the 7D.

Also, I found http://www.usa.canon.com/dlc/controller?act=GetArticleAct&articleID=3167 very useful for a first reconnaissance of the dazzling capabilities of the 7D

 

[Borek]

Meanwhile http://dl.dropbox.com/u/21940023/IMG_1352-2.jpg is the full processed raw, for Borek to see that the basic 70-300mm zoomlens doesn't have to disappoint too much when hooked onto the 7D.

Thanks. It looks great in the center - but is hard to judge on the borders, as they are blurred for a purpose. But the shot is really great, one of those you remember you took them for the rest of your life

Also, I found http://www.usa.canon.com/dlc/controller?act=GetArticleAct&articleID=3167 very useful for a first reconnaissance of the dazzling capabilities of the 7D

Yes, they are quite good. I have already seen some of them, they explained at least one mystery.

At the moment I already know what I am missing - A-DEP mode. Not that I ever used it as designed, but it was quite useful.

I still can't decide if I could afford the camera, but at least I will save on the birthday gift for Junior in April Sadly that will mean parting with 28-105 lens, which was my main workhorse for years.

 

[Andy Resnick]

Thanks for your interesting crop Andy, looks very powerful that way.

Borek, the centered eye is logically explained as I used the center focus point on the eye to focus and then I recomposed by moving the center straight down before shooting, so the eye stayed in the vertical centerline.

Maybe I'll also upload the full processed Raw later, to demonstrate the difference. But chores first

Heh.. you didn't think I'd stick to convention, did you? :)

 

[Andy Resnick]

Meanwhile http://dl.dropbox.com/u/21940023/IMG_1352-2.jpg is the full processed raw

Oh, good! I was running out of pixels before. I wanted to try and do something like this:

[PLAIN]http://img535.imageshack.us/img535/9168/img13522.jpg

I was also drawn to the strong lines in the wing.

 

[Andre]

Heh.. you didn't think I'd stick to convention, did you? :)

Most amused that you didn't

 

[Andy Resnick]

I've also been remiss- I want to thank you for giving me permission to cut up your image. It's quite excellent "as is"!

 

[turbo]

Meanwhile http://dl.dropbox.com/u/21940023/IMG_1352-2.jpg is the full processed raw, for Borek to see that the basic 70-300mm zoomlens doesn't have to disappoint too much when hooked onto the 7D.

Also, I found http://www.usa.canon.com/dlc/controller?act=GetArticleAct&articleID=3167 very useful for a first reconnaissance of the dazzling capabilities of the 7D

Nice detail in that shot, Andre, and I'm impressed by the lack of noise. Maybe a 70D is in my future. I'd love to have a 100 mm L macro, too. I'd better find a money tree.

 

[Andre]

You're welcome Andy, the pleasure is in sharing.

Turbo, between the old 100mm macro and the 100mm L macro of Borek, when t L first came, I wondered if I made a wrong choice, especially since the L has image stabilisation.

However the folks at dpreview reassured me. No doubt the L is the better lens, but, a big but, while the image stabilisation is great under normal conditions (4 stops) in marcro conditions the testers could not get better than one stop.

Moreover the image quality of the L is clearly superior to the old version, at large apertures, that is, smaller than 5.6 the image quality is virtually identical. You can all see that in that test. However the old lens is slightly better at chromatic abbaration.

So as a telelens, the L is clearly better, as a macro lens using small apertures, you're basically left with only one stop image stabilisation effectivily.

So even if I'd had the choice today (not needing the 100mm as a tele) I'd probably still go for the old 100mm at a considerable better price point, using it predominantly for macro.

 

[Borek]

This can be a little bit tricky. In my experience lens quality is not only a property of a model, but also property of a particular piece of hardware. At least in theory L is built to better standards, so should be better on average.

Doesn't have to. We are waiting for spring, to see how it will work in field. At home it works OK, although focusing is often tricky.

 

[Andre]

True, I'm merely reflecting that http://www.dpreview.com/lensreviews/widget/Fullscreen.ashx?reviews=67,69&fullscreen=true&av=3,3&fl=100,100&vis=VisualiserSharpnessMTF,VisualiserSharpnessMTF&stack=horizontal&&config=LensReviewConfiguration.xml%3F2 , if you move the aperture past F5.6/8 you'll see that the difference diminishes and for macro, I rarely work below F11.

But indeed I have the impression that my 70-300 performs better than this one here

 

[turbo]

So as a telelens, the L is clearly better, as a macro lens using small apertures, you're basically left with only one stop image stabilisation effectivily.

So even if I'd had the choice today (not needing the 100mm as a tele) I'd probably still go for the old 100mm at a considerable better price point, using it predominantly for macro.

And that is a big part of my quandry. I like to shoot candids of friends and family at gatherings. When I was using Olympus OM-series cameras, my very favorite portrait lens was the Zuicko 135mm prime at f:2.8. Long enough to let me shoot without crowding the subject, and long enough to avoid facial distortion. With the crop factor, the 100mm would get me right into that range, and give me enough aperture to get a shallow DOF, PLUS image-stabilization, so I could get some extra stops for portraiture. Can I justify the extra money for the L? I don't know.

 

[Borek]

Can I justify the extra money for the L? I don't know.

That's when wife sharing hobby comes in handy

 

[turbo]

That's when wife sharing hobby comes in handy

You are WAY too crafty, Borek! I gave my wife my old Olympus C-3040, but she only uses it for kids' birthday parties, etc. We have been greatly expanding the flower gardens, so maybe some macro flower/insect shots will tip the scales.

 

[Andy Resnick]

That's when wife sharing hobby comes in handy

You share your wife? Zounds! :)

 

[lisab]

You share your wife? Zounds! :)

That's what I thought too !

 

[Borek]

 

[Borek]

Spring is here

No, it is not Junior with his GF, this is some random pair I have seen today at Warsaw Old City.

 

[AlephZero]

Sorry if this isn't the right place, but I guess the PF photographers will see it here:

I want to make digital images from about 1000 sheets of A3 size paper (= USA 11x17). This is a "one off" task, so I don't want to buy an expensive A3 size scanner. The paper originals are clean high contrast black-and-white images (actually music scores, hand-written in ink). They are separate single sheets, not bound in a book.

I have a digital camera that is good enough to handlle this. The question is, is it possible to buy (rather than make) a fairly cheap camera stand for this? The easiest way would probably be to have the paper flat on a table and the camera pointing vertically down, so a normal camera tripod won't work. I'm thinking of something like an L shaped arm that would clamp onto the edge of a table, with a mount for the camera pointing downwards. Do such things exist ready made?

My camera has a 10:1 optical (not digital) zoom, so the stand wouldn't necessarily need to be height adjustable.

I'm not bothered about color balance etc so long as the images are "readable", so lighting is not an issue.

 

[Borek]

Can't you put the sheets on the table tilted 45 deg? Resting on a stool or something? Then simple tripod will work. To keep sheets in place you can glue slat of wood to the table, water soluble glues should be later not difficult to remove.

Alternatively you can put the tripod horizontally ON the table - just use some heavy books on the legs.

In my experience makeshift solutions for such tasks are the best

 

[turbo]

You could clamp the camera onto the edge of a coffee table and photograph the pages with the sheets lying on the floor. If you have a relatively bright, sunny room to do this in, it could go pretty quickly. Once you have the camera set to match the aspect of the first sheet, make some "Ls" out of tape to mark the corners of that sheet so you don't have to align every sheet one-by-one.

 

[Borek]

make some "Ls" out of tape to mark the corners of that sheet so you don't have to align every sheet one-by-one.

Even better - instead of tape use something sticking out (like two books for a corner, just on one side), so that you don't have to align each sheet to the marks, just push it till it stops touching the edge.

 

[turbo]

Even better - instead of tape use something sticking out (like two books for a corner, just on one side), so that you don't have to align each sheet to the marks, just push it till it stops touching the edge.

Good idea. That would work very well, and would eliminate the need for visual alignment with the tapes. 1000 sheets could go very quickly that way.

 

[Borek]

In 1998 I moved from bpp.vr.pl to bpp.com.pl. Problem was, I had several thousands copies of booklet to put into CD boxes with one of my programs, and the booklet was printed with the old domain name. What I did was I ordered a self inking rubber stamp "we inform you our new website is at bpp.com.pl" (in Polish). Then I took a piece of plywood, glued two slats to it - so putting the booklet always on the same position was not a problem, and the stamp was pushed against the same slats - which meant all the booklets were stamped in exactly the same position in exactly the same corner. Few hours and all were modified.

I love such witty solutions

 

[AlephZero]

You could clamp the camera onto the edge of a coffee table and photograph the pages with the sheets lying on the floor. If you have a relatively bright, sunny room to do this in, it could go pretty quickly. Once you have the camera set to match the aspect of the first sheet, make some "Ls" out of tape to mark the corners of that sheet so you don't have to align every sheet one-by-one.

That nudged my brain in the right direction. I've got an ironing board where the metal wires that make the stand for the iron are the right spacing to locate the camera lens down, without needing to actually clamp it.

Next task: recharging all the available batteries. I'm not sure how many shots the camera will take on one set, but I'm pretty sure the answer is "less than 1000".

 

[Borek]

That nudged my brain in the right direction. I've got an ironing board where the metal wires that make the stand for the iron are the right spacing to locate the camera lens down, without needing to actually clamp it.

And legs don't come into the sight?

Note that if the camera just lies, you will probably move it slightly with each picture taken (just by pressing the button). That in turn means each sheet will have slightly different position on the picture, which is quite inconvenient and may mean a lot of work later (work that can't be easily automated). Go for something more rigid. Perhaps even just tying camera with a strong cord will work.

 

[AlephZero]

And legs don't come into the sight?

The paper will be on a low table that overlaps the bottom end of the X-shaped legs.

Note that if the camera just lies, you will probably move it slightly with each picture taken (just by pressing the button). That in turn means each sheet will have slightly different position on the picture, which is quite inconvenient and may mean a lot of work later (work that can't be easily automated). Go for something more rigid. Perhaps even just tying camera with a strong cord will work.

It's easy to locate the camera in a fixed position pressed against the wire frame. An elastic band round the camera and the frame might be useful though.

I'm not too bothered about accurate matching of the positions of each image. The real task is typesetting the music with computer software (which will probably take a few years, given the quantiity and complexity of it). The images aren't going to be published as a facsimile of the original, they are just for displaying on a monitor while doing the transcribing. That will be much less risky than continual handling of the originals (no coffee spills, etc!).

Even if the originals were photocopied to avoid the coffee, physically searching through hundreds of large sheets of paper to find two or three sheets to compare some details side by side, etc, would be much more hassle than working on a computer screen.

Sample page, done to check that the music notation software could handle the job, attached...

 

[Borek]

If the simplified approach works for you - that's great. I just did similar things in the past and I know from experience where the possible problems are, if they are not important for you - don't bother.

 

[Andre]

I got inspired by all that reproduction work and started to digitize my pix of before the digital era. I used a little wooden box and the 45 degrees angle. Allows to stack the pix quickly into the right position. Wow decades of memories.

Oldest daughter, 15 years ago:

 

[Jonathan Scott]

I'm not too bothered about accurate matching of the positions of each image. The real task is typesetting the music with computer software (which will probably take a few years, given the quantiity and complexity of it). The images aren't going to be published as a facsimile of the original, they are just for displaying on a monitor while doing the transcribing. That will be much less risky than continual handling of the originals (no coffee spills, etc!).

There is software available (Neuratron PhotoScore) which will take scans or photos of music and convert it to computer music notation, typically requiring a bit of manual fix-up, but much less work than starting from scratch. I've successfully used the free version that comes with Sibelius, both with photos and with PDFs of old music, but not for anything as ambitious as your example.

You would definitely need a higher resolution than your sample for such software to be able to read the score very successfully.

 

[Andy Resnick]

The new multi-volume book set "Modernist Cuisine" has some *amazing* photos. There's a lot of 'cutaway' shots- a grill cut in half to show what goes on inside, for example- but lots of amazing photos throughout.

http://modernistcuisine.com/
http://modernistcuisine.com/docs/ModernistCuisine_About%20the%20book_spreads.pdf

 

[physics girl phd]

The new multi-volume book set "Modernist Cuisine" has some *amazing* photos. There's a lot of 'cutaway' shots- a grill cut in half to show what goes on inside, for example- but lots of amazing photos throughout.

http://modernistcuisine.com/
http://modernistcuisine.com/docs/ModernistCuisine_About%20the%20book_spreads.pdf

http://www.nytimes.com/2011/03/09/dining/09modernist.html" . If only it weren't so expensive... and didn't make me so hungry! What eye candy!

 

[AlephZero]

There is software available (Neuratron PhotoScore) which will take scans or photos of music and convert it to computer music notation, typically requiring a bit of manual fix-up, but much less work than starting from scratch. I've successfully used the free version that comes with Sibelius, both with photos and with PDFs of old music, but not for anything as ambitious as your example.

I have Sibelius and PhotoScore.

I don't think there is much chance of Photoscore doing anythng useful with this sort of handwriting. Actually this clip is pretty neat compared with some of it!

 

[Andy Resnick]

As I mentioned, I got some good news recently- my NIH proposal was approved for funding- so I decided to spend some of my (non-grant) startup money and get two 'big boy' lenses. They just came this morning, and so rather than do what I'm supposed to be doing, I wandered around campus playing with the new toys.

The first lens is a 15mm f/3.5 rectilinear lens- an ultrawide angle lens. This lens is really good for indoor shots as it distorts the volume of spaces- here's my palatial-sized office:

[PLAIN]http://img848.imageshack.us/img848/8488/dsc5698.jpg

And an indoor commons area:

[PLAIN]http://img600.imageshack.us/img600/8067/dsc5699v.jpg

I also took a shot inside of an elevator- it's not very good, but it shows the distortion effect fairly well. The elevator can hold at most 5 people, it's like 3' x 3'.

[PLAIN]http://img600.imageshack.us/img600/1513/dsc5702j.jpg

This image shows some of the challenges of this lens- the blue blobs at the bottom are flare from the overheard lights, and the vertical lines aren't quite vertical- the huge angle of view makes alignment super critical- both pitch and yaw angles, but also centering. But, I got this lens both because of the huge field of view/depth distorition and that it is a *rectilinear* lens- there is no distortion (in the lens aberration sense).

The other lens is a little intimidating, frankly- it's a 400mm f/2.8. I found an older manual focus version, so I could afford the freaking thing- the front element is about 6" in diameter, and the lens is *really* heavy. I also picked up a 2x teleconverter for it to also use it as a 800mm f/5.6- why not? It close-focuses to 10', so it's practically a macro :)

It's an outdoor lens, and since it's raining today I won't be able to use it. So, I have time to check that my tripod can handle the weight...

In other news, I've been carefully stripping down the crackberry's LCD display- there are some really weird lighting elements that I've been trying to understand, and when I do, I'll post those pics.

 

[Andy Resnick]

Finally- we have sun outside. I took my 'new' lenses out for a test drive- here's the 15mm, full frame and then 100% crop:

[PLAIN]http://img856.imageshack.us/img856/1071/dsc57451.jpg

[PLAIN]http://img196.imageshack.us/img196/1316/dsc5745.jpg

And here's the 400mm, full frame and 100% crop:

[PLAIN]http://img718.imageshack.us/img718/9855/dsc57461.jpg

[PLAIN]http://img856.imageshack.us/img856/1184/dsc5746.jpg

The nikon-sony adapter has some slight vignetting at long focal lengths (which I knew about, it's the cost of doing business...), but even so, the image is quite impressive- the thermal lensing is clearly visible.

Adding the 2x teleconverter increases the vignetting considerably, but even so:

[PLAIN]http://img251.imageshack.us/img251/9561/dsc57471.jpg

[PLAIN]http://img690.imageshack.us/img690/4407/dsc5747u.jpg

"lens shake" is extreme with this lens- except it's not the lens exactly- it's the camera flopping up and down. I used the 10-second timer delay to get (reasonably) crisp images. If the weather hold up, I'll try taking some images of the moon tonight.

 

[Andy Resnick]

Got a really nice shot of the moon last night:

[PLAIN]http://img233.imageshack.us/img233/4696/dsc58161.jpg

[PLAIN]http://img8.imageshack.us/img8/523/dsc58162.jpg


The settings were 800mm f/5.6, ISO100, 1/125s exposure.

The star photos were not acceptable. Seeing was mediocre at best, and even though I got excellent separation of Alcor and Mizar and may even have resolved Mizar as a quadruple, it's not clear due to vibration issues. At extreme blow-up, there is some purple fringing (spherochromatism), but I'm reserving judgment until my technique is 'up to spec'- getting tack-sharp images with this lens is not easy, even when the view is spectacular through the viewer.

After working with the lens for a while, the main residual problem seems to be due to 'mirror flap'- I'll try locking the mirror up first, waiting a few seconds, and then use a remote shutter to trigger- I ordered the remote last night. On the plus side, the lens has almost no flare- here's the sun, about 30 minutes prior to sunset:

[PLAIN]http://img684.imageshack.us/img684/3091/dsc5806g.jpg

The settings were 800mm f/22, ISO 100, 1/2500s exposure, HDR enabled and set to maximum.

I ordered a couple large sheets of polaroid- it would be nice if I could see sunspots.

 

[Borek]

The settings were 800mm f/5.6, ISO100, 1/125s exposure.

Have you tried higher ISO and shorter times?

 

[Andy Resnick]

Have you tried higher ISO and shorter times?

I usually stick to ISO 200 and slower because the Sony Exmor is very susceptible to amplifier noise. The noise is noticeable at ISO 400, and annoying at ISO 1600. That's one of the reasons I prefer fast lenses- it's easier to throw photons away than it is to scavenge every last one.

That said, I could take an image at ISO 6400 and after re-scaling it to 650x480 you would never know how much noise there is; I can work in the ISO 400-1600 range and smooth the image later to get acceptable results at 100% crops.

Now that the weather is nicer, I can spend some time outside taking nighttime shots with the goal of improving (reducing) camera shake. I'm hopeful that using a remote shutter will fix the majority of the problem.

 

[Andre]

Yes the noise is indeed an issue, but using high ISO is just another trade off, if you like to shoot tele or macro handheld in a dynamic environment of moving subjects and you like low ISO's for the lack of noise, you run out of options even in fair light conditions. Then it would nice if say F16 and 1/1000 sec could be done, but you might need ISO3200 to accommodate that.

In that case, one would currently like the Nikon D3S best, closely followed by the quicker Canon 1D mkIV

But you could also buy a decent car for that money,

 

[Borek]

Now that the weather is nicer, I can spend some time outside taking nighttime shots with the goal of improving (reducing) camera shake. I'm hopeful that using a remote shutter will fix the majority of the problem.

It is either the shake or atmosphere, we have discussed it earlier. I was surprised several years ago by the sharpness of Moon images taken at very short times.

 

[Andy Resnick]

But you could also buy a decent car for that money,

Too true.

 

[Andy Resnick]

It is either the shake or atmosphere, we have discussed it earlier. I was surprised several years ago by the sharpness of Moon images taken at very short times.

Based on the star images, I suspect my main culprit is mirror flap- the stars consistently got distorted into short lines, even for 1/10 second shutter speed- and the lines were always the same length and orientation (of course, I didn't change the orientation of the camera to check this last night).