PF Photography: Tips, Tricks, & Photo Sharing

[Andy Resnick]

The vignetting is due (most likely) to the lens adapter; I have a full-frame camera (Sony a850) and the lens is a Nikkor 400mm f/2.8 ED-IF. The Nikon-Sony adapter has a small lens to line up the image planes correctly, and (so I'm told) use of long focal length lenses can cause vignetting. I don't get vignetting with either the 15mm or my old 50mm Nikon lenses.

 

[Borek]

This is not exactly photography, but close.

I did some testing to check what is quality of 1080p video shot with EOS7D - and results are disappointing. Basically my line of thinking was - if the sensor and lens are able to produce sharp images at the over twice higher than needed resolution, HD video shot with the camera should be crisp sharp. But that's not the case. Either it is a matter of the scaling algorithm, or of the codec used, but the image is very soft and looks like 720p scaled up (or even worse).

I put the camera on the tripod, shot some video and took a picture from exactly the same place, not changing anything (camera focused before the test and AF switched off to avoid any surprises). Then I used avidemux to export a first frame from the video and rescaled - without sharpening nor any other tricks - the single picture taken with the camera. See below. Video on the right, picture on the left, this is crop from the middle down part of the view.

This is whole frame, just to give you some idea:

Nasty surprise I must say. I understand it is NOT a camera, but I hoped for more

 

[Andre]

I wonder about the performance of other cameras and not sure if you expect too much. Consider the writing speed and the amount of the data for instance. Maybe they needed to compromise on the compression algorithm causing a reduced amount of data.

Maybe I shuld try that too.

 

[Borek]

I wonder about the performance of other cameras and not sure if you expect too much. Consider the writing speed and the amount of the data for instance. Maybe they needed to compromise on the compression algorithm causing a reduced amount of data.

Could be - but if my calculations are right video is saved at about 44-45 Mb/s (4GB, which is a maximum file size supported, gave 12'35.6" or 755.6 sec video, 4*1024³*8/755.6/10⁶ = 45.5 Mb/s), which is more or less in the range of the Blue-ray bitrate. If anything I suspect the chip is not capable of fast compression, as it wasn't optimized for that.

 

[AlephZero]

I put the camera on the tripod, shot some video and took a picture from exactly the same place, not changing anything (camera focused before the test and AF switched off to avoid any surprises). Then I used avidemux to export a first frame from the video and rescaled - without sharpening nor any other tricks - the single picture taken with the camera. See below. Video on the right, picture on the left, this is crop from the middle down part of the view.

This is just speculation (the last "serious" video stuff I did was long ago on 16mm film) - but there are different requirements for "good video" compared with stills. For example you want relatively long exposure times to get realistic motion blur, not pin-sharp images. Animated movies with pixel-sharp images tend to create "flicker" when things move.

On film-based cine cameras the shutter speed was fixed at half the frame rate - i.e. 1/48 sec for standard movies - hence the hassles with movie lighting, because the f-stop was the only camera setting you had to play with. Changing to different speed film stock could introduce color matching problems.

Compare the reflections of the trees in the water, which is obviously moving - it's not so clear if there was any wind moving the vegetation.

I would guess that digital video cameras try to emulate the same behavior as cine film, since human vision works the same way whatever the source of the image.

 

[khemist]

Here is a recent shot from my trip to Arizona, north of Phoenix.

air by Hoxie Photography, on Flickr

The vignetting is due (most likely) to the lens adapter; I have a full-frame camera (Sony a850) and the lens is a Nikkor 400mm f/2.8 ED-IF. The Nikon-Sony adapter has a small lens to line up the image planes correctly, and (so I'm told) use of long focal length lenses can cause vignetting. I don't get vignetting with either the 15mm or my old 50mm Nikon lenses.

God damn that's like a 6000 dollar lens!

 

[Borek]

Compare the reflections of the trees in the water, which is obviously moving - it's not so clear if there was any wind moving the vegetation.

At this particular moment there was almost no wind, the branches were completely still. The single frame could be blurred by accident, but lack of sharpness here is not accidental. Actually I selected this particular place/crop as it nicely shows the general feeling.

EOS7D let's me select almost any shutter speed (as long as it is below that defined by frame rate) I want, and in fact I did more testing - it doesn't look as quality changes with shutter speed selected. At least not for the type of movies I shot - mostly still landscapes.

I plan to do some testing with the moving objects, but that will be next stage.

 

[AlephZero]

Some more thoughts on the movie experiments:

The image degradation may be coming from the playback/frame grabbing codec, not from the recording.

Also, the first frame of any video is by definition "special" because it has to contain the full image. Later frames that only contain the changes may converge to a higher quality image.

You can often see that "convergence" effect with video conferencing (using professional quality video links and equipment, not webcams!) where the picture quality reduces if there is a some rapid movement, and then recovers over the next few frames (of the order of 10 frames, I would guess).

 

[Andy Resnick]

Here is a recent shot from my trip to Arizona, north of Phoenix.

God damn that's like a 6000 dollar lens!

Worse- new it's a $10k+ lens. Mine's been gently used and was gotten for 20% of that. So I don't mind losing 5-10% of the frame... :)

 

[Borek]

Taken (and a little bit photoshopped) today. Some random girl taking pictures at Old City in Warsaw.

Well, her randomness is definitely skewed

Edit: that's not good, I am in love with a picture I took. Call me Pygmalion.

 

[fuzzyfelt]

Great photo!

 

[Borek]

Thanks

 

[Andy Resnick]

I don't know if the idiom "losing your marbles" is strictly American, but the meaning is similar to "becoming frustrated or angry".

My semester just ended, so perhaps it is fitting to show these. The overall context is "photographing glass objects", and it's not easy to get a decent image of the colorful bits inside the clear glass sphere.

[PLAIN]http://img651.imageshack.us/img651/2568/dsc6776s.jpg

[PLAIN]http://img717.imageshack.us/img717/8026/dsc6784u.jpg

[PLAIN]http://img860.imageshack.us/img860/530/dsc6770p.jpg

The trick is not only precise lighting, but also tiny apertures (the smaller the better). A clean sensor is good to- mine's gotten dirty...

 

[drizzle]

MARZENAAAA! :!) :!) ... Huh, don't get me wrong people, but that's MARZENAAAA! :!)

Thank you so much Borek, I love that shot

Edit: Ops, it's not Marzena. :shy:

 

[Andre]

Well this is how it looks now;

This time I "exposed to the right" and then processed the RAW back to normal. That should take care of some of the noise

And this is from today.

Striking that the orchid flowers open so slow and last for months.

 

[Andy Resnick]

Pixel-peeping my 15mm images, I discovered that the images are only sharp in the central 20% of the frame, and out around the periphery the images are horribly blurry. The adapter has a lens element in order to place the focal plane of the Nikon lens at the Sony camera sensor plane, so I wondered if the adapter needed some optimization- the lens is not cemented in place and the lens position can be adjusted slightly.

Here's my test setup- all these images were taken at full aperture (f/3.5):

[PLAIN]http://img231.imageshack.us/img231/4569/dsc70671.jpg

and here is the extreme upper left corner of the image, taken with the adapter lens element seated all the way in, then moved about 0.5 mm toward the 15mm lens, and then a final position 1mm toward the 15mm lens:

[PLAIN]http://img9.imageshack.us/img9/4910/61801809.jpg

[PLAIN]http://img694.imageshack.us/img694/5508/midk.jpg

[PLAIN]http://img535.imageshack.us/img535/8254/outx.jpg

There's a dramatic improvement at 1mm. Moral of the story- don't be afraid to tweak your equipment.

 

[Andy Resnick]

As I mentioned above, tweaking the position of the lens adapter made a huge difference in performance of the 15mm lens: this should not be too surprising, as 1mm represents an 8% shift in focal length. Not only was sharpness improved, but chromatic aberration *vanished*.

So, I was pleasantly surprised to notice the same dramatic improvement with my 400mm- I was not expecting much, since I was only shifting the adapter lens 0.25% of the focal length (or 0.12% when operating at 800mm).

Here's images I took previously with the lens wide-open (f/2.8)

[PLAIN]http://img90.imageshack.us/img90/2923/dsc6872.jpg

[PLAIN]http://img59.imageshack.us/img59/8051/dsc68721.jpg

I wouldn't say the performance is bad- I figured it's a 40 year old lens, modern technology has surely improved things, so I just stopped it down to f/8 and the CA went away. However...

Here's an image I took last night, now with the adapter lens optimized (again, f/2.8):

[PLAIN]http://img6.imageshack.us/img6/3579/dsc7380o.jpg

[PLAIN]http://img692.imageshack.us/img692/8084/dsc73801.jpg

What a difference!

 

[Borek]

Wow, that's surprising.

 

[Andy Resnick]

I know! Thinking more, I should have realized that since lens aberrations scale as the f/# (or worse), the correction would be *more* pronounced for the 400mm than the 15mm. Now, if I can just get someone to loan me a Noctilux...

 

[Andre]

...

And this is from today.

Striking that the orchid flowers open so slow and last for months.

It's past high noon now. The last two buds opened the other day, while the first flower (lower left) withered. The total flowers were 12 at the lower and 10 at the upper branch.

Shot is back lit by the sun.

 

[Borek]

Thunderstorm over Warsaw and Marki (I can still hear it). I have heard it early enough to prepare to shot a video from the attic, these are just selected frames:

 

[Andy Resnick]

Thunderstorm over Warsaw and Marki (I can still hear it). I have heard it early enough to prepare to shot a video from the attic, these are just selected frames:

Awesome! I was getting ready to ask about photographing lightning- how did you do it (e.g. exposure times, method, etc)?

 

[Borek]

I suppose you more or less already guessed, but just to to clarify: video, with each frame shot at 1/30 sec. At 25 frames per sec "lost" time is about 17% of total, so it is hard to miss anything. But that also means pictures are shot at 1920x1080, and - as I signaled in some other tread - their quality is much lower than expected. This is 1:1 crop of one of the images exported from the video (using Avidemux):

and 1:1 crop of the same view, the same lens, but picture, not video:

Difference is huge. Night pictures (with long exposure times) are much better.

I just realized I have neutral filter ND8 somewhere, so I could try long exposure times. Unfortunately, at the moment thunderstorm is away and it is quite nice.

 

[Andy Resnick]

How much video did you need to shoot in order to get those few 'good' frames? I guess I'm asking about the efficiency... I can imagine shooting minutes of video and then having to extract out 3 or 7 frames.

 

[Borek]

Much less. For sure it depends on the thunderstorm. I got 11 OK pictures out of six minutes, plus there 2 or 3 that were technically unacceptable (artifacts - like only half of lightning visible in one frame, and other half in the next frame). Thunderstorms in Poland are not that active when it comes to lightnings. I have a feeling lightnings were much more frequent about 15 minutes later; unfortunately at that time thunderstorm was already on the other side of the house and we don't have a window on this side of the attic that would allow tripod use.

 

[Andre]

Must have been years since I saw my last good thunderstorm

Anyway, this decent fellow, upon observing my activities, enquired if he could be of any service.

So I mentioned that I would not discourage him from giving an aerial display by any means. So he wasn't:

I should not have zoomed in that much though, as he did not fit in the frame.

 

[Triple_D]

The http://www.lightningtrigger.com/index.html#home . I've never used one, but a lot of people say they work great.

 

[Borek]

Arrgh.

Few months ago I bought two Kingston Elite Pro Compact Flash 8GB 133x memory cards for my camera. They work perfectly. I decided before going for vacations I am going to buy another two. I bought two Kingston Elite Pro Compact Flash 16GB 133x cards (same line, same specification, just twice the size) from exactly the same, trusted source (one of the largest hardware distributors in Poland). And they don't work as expected - they are too slow.

It turns out those bought earlier work much better than the specification, they work at at least 266x. That's why I have not realized 133x is too slow. Now I have two useless 16GB cards and around $75 in the hole. One I will probably sell, as it is unopened, but the other...

 

[Andy Resnick]

random change of subject:

Last night I re-watched Gaspar Noe's film "Irreversible"

http://www.imdb.com/title/tt0290673/

and was again astonished at the opening cinematography. I'm not I could 'recommend' watching the movie (if you haven't seen it, it's rather disturbing in places), but if you choose to watch it, the opening 5-10 minutes are really impressive. I was completely disoriented; I had no idea if I was looking up, down, upside down, rightside up, sideways, backwards... I would have *loved* to see whoever physically moved the camera around do their 'dance'. Noe did the cinematography as well as direct- it's a fine piece of work.

The opening credits are disorienting as well- some letters (E, R, K, N) are reversed, and the credits sort of go off-kilter before completely overloading your brain.

Again, I think this crowd would appreciate the technique- the way an image can take something obvious and ordinary and make it confusing and unrecognizable.

 

[Andre]

Arrgh.

Few months ago I bought two Kingston Elite Pro Compact Flash 8GB 133x memory cards for my camera. They work perfectly. I decided before going for vacations I am going to buy another two. I bought two Kingston Elite Pro Compact Flash 16GB 133x cards (same line, same specification, just twice the size) from exactly the same, trusted source (one of the largest hardware distributors in Poland). And they don't work as expected - they are too slow.

It turns out those bought earlier work much better than the specification, they work at at least 266x. That's why I have not realized 133x is too slow. Now I have two useless 16GB cards and around $75 in the hole. One I will probably sell, as it is unopened, but the other...

Apparantly the optimum choice seems to be the Transcend 400x CompactFlash Card, 32GB goes for some 65 euro. I'm just ordering one.