처음 처음 | 이전 이전 | 1 | 2 |다음 다음 | 마지막 마지막

http://www.redhax.net/wiki/Understanding_Rolling_Shutter_Artifacts

 Understanding Rolling Shutter Artifacts

The image is scanned acrossed each row of pixels in order from top to bottom. 

The image is scanned acrossed each row of pixels in order from top to bottom.On digital cinema cameras that have a rolling electronic shutter, strobes and other similar phenomenon can cause the image to be split across a horizontal line. This happens during two brief periods where the camera is either begining an exposure or reading an image from the sensor chip. The split is called a rolling shutter artifact. There is no actual shutter involved, the shutter refers to a process occuring on the sensor chip that simulates the behavior of a mechanical shutter while accomplishing the neccessary task of transmitting the image data from the sensor chip to the rest of the camera.
Rolling shutter artifacts also exist in cameras that have a mechanical rotary shutter. However because the edge of the mechanical shutter is between the back element of the lense and the film plane it is very out of focus and creates a gradual line. Digital shutters create a sharp line that is only slightly softened by how gradually the flash comes up to power.

How a Digital Rolling Shutter Works

A digital camera sensor is composed of a grid of individual light sensor sites called photosites. Each photosite site corresponds to a pixel or sample of the image. As light particles bombard each photosite they knock electrons from one part of the sensor to another and create a small pocket of electrical charge. The sensor chip has the power to:

 


  • Release collected charge from each photosite and start fresh.
  • Measure collected charge on each photosite and thus measure the amount of incoming light.
  • Transmit that measurement to the other circuitry of the camera so that it can be reconstructed as a digital image.
The activity of the entire sensor chip can be divided into four parts.

  1. The sensor scans across the image and resets each individual sensor so that it begins freshly collecting charge.
  2. The sensor waits as charge is collected. The length of this delay determines the shutter speed.
  3. The sensor scans across the image and reads each individual sensor and transmits the information.
  4. The sensor waits until the appropriate time to begin a new exposure.
These four actions effectively simulate the action of a mechanical shutter. The time and order in which pixels are reset is the same as the that of reading them and transmitting their data. This ensures that the exposure time of each pixel is equal. In reality the chip is always actively collecting charge. It is only that information collected during the inactive periods is ignored.

 

Unlike a film camera, it is possible to keep a digtal shutter open for the entire duration of each frame. This is sometimes called a 360 degree shutter. This is achieved by combining steps 1 and 3 from above into a single step and eliminating step 4.

Shutter Artifacts


Rolling shutter artifacts occur because the camera must transmit the measurement of light from each photosite one at a time. This process happens so instantaneously that it is normally imperceptable. Strobes however, create a burst of light that can start and stop faster than the sensor can transmit an image. When a strobe occurs during those short periods that the sensor is reading or reseting itself (stages 1 or 3 from above) the light from the strobe will have a different effect on pixels that have been read than it will on those that have not. This difference creates a visible horizontal line, a rolling shutter artifact.

Hybrid digital mechanical cameras such as the Dalsa Origin and the Arri D20 do not suffer from rolling shutter artifacts because the readout process is hidden behind a mirror.


This diagram shows three examples of strobes interacting with a rolling shutter. These images were recorded on a RED One camera in September 2007. Images are from Die Mannequin used by permission of Marc Bachli and Fini Films Inc. 

This diagram shows three examples of strobes interacting with a rolling shutter. These images were recorded on a RED One camera in September 2007. Images are from Die Mannequin used by permission of Marc Bachli and Fini Films Inc.
Because the readout time of the sensor is fixed changing the shutter speed does not change the probability that a shutter artifact will occur (it only changes the probability that the strobe will occur during the active or inactive period. Changing the frame rate, does change this probability. Lower frame rates have a lower likelyhood of shutter artifacts.

A 360 degree shutter does not eliminate rolling shutter artifacts but instead splits them evenly between consecutive frames.

Using the sensor in a windowed mode (such as 2K Windowed RAW or 720p RGB) can reduce the probability of catching a shutter artifact. In windowed mode the sensor must only read or reset one quarter of the pixels. The readout time is therefore one quarter of the time of the full chip.

Syncing Strobes with the Shutter


This can be achieved by connecting the Genlock output of the camera to a strobe controller.

댓글(0) 먼댓글(0) 좋아요(0)
좋아요
북마크하기찜하기
 
 
 

http://www.juzaphoto.com/eng/articles/five_steps_to_improve_your_photos.htm

Five Steps to Improve Your Photos

You have just purchased your first DSLR, and you are not happy by the result that you get? Even tough it takes some time to improve, there are some things that can immediately help your photography - in this article, I have listed five suggestions that I consider very important to improve your results!

 


Read the camera manual

Many photographers - sometimes even professionals - don't fullfill the potential of their cameras because they don't know how to use properly some features, or they don't even know that some features exists. When I buy a new camera, I always read the entire manual - of course, I'd be able to use the camera even without reading the manual, but I'd miss many useful things.

Other than that, if you don't know well your camera you may think that it does not work or it has defects, while actually you are not using it properly. Some common complaints are:

The autofocus is not precise: even though there are few defective cameras that actually have focus problems, generally the 90% of focus errors are due to the photographer. Read carefully the pages about AF and learn how to use it; pay attention to center the subject (or the area where you want to focus) into the selected AF point. Other than that, remember that AF is not perfect - you can not expect to get always sharp photos, in particular if you photograph a fast moving subject. 

The images have color cast (or: my friend's camera gives better colors): usually the color cast are a consequence of white balance errors; personally, I always use Auto White Balance and I don't care at all about color casts, because when I convert the RAW files I can adjust the white balance to get perfect colors. Another complaint that I hear often is "that Canon/Nikon/ect camera gives much better colors than mine". All DSLRs currently in production gives quite good colors; what really matters is not the camera, but the post processing - if you process well your RAW files, you can get excellent colors with any camera.


The photos are soft: many photographers that have switched to a DSLR from a compact cameras think that the photos are soft. The reason is that compact cameras often use a strong in-camera sharpening, while DSLRs generally apply less processing to the photos; this is a positive thing, since a less processed photo has less artifacts and it is more "customizable" by the photographer. You can get a perfect sharpness by shooting in RAW and with good post-processing techniques; of course, you need also good lenses and good techniques (the photo must be properly focused, and you have to use a tripod or a fast shutter speed to avoid blur).

The photos are underexposed: some cameras actually tend to underexpose by 0.3 or 0.7 stops, if you use them in complete automatism. I recommend to spend some time learn the basis of exposure, in particular the use of histogram. The histogram is a very useful tool, once you know how to read it you will be able to get always properly exposed images.

A DSLR is a very versatile and powerful tool, and the time that you spend reading the manual will be repaid by great images.

 

Understand composition, light and background


To master the technical and the artistic side of photography, I recommend to read the articles in the Nature Photography section. But if you don't know where to begin, try to remember there essential guidelines:

Composition: don't place the subject (or the horizon line) right in the center of the frame: centered images often have a "static" look. Instead, you should try asymmetrical compositions: if your subject is a flower or an animal, put more room in front of it than behind it; if you are photographing a landscape, frame 2/3 of earth and 1/3 of sky (or vise-versa); and so on.

Wildlife - angle of view: one of the most common errors of beginners is to take photos of animal and plants from the eye level of the photographer, that results in a "high to low" perspective and a very amateurish look. If you want to improve your photos, you have to get low: the camera should be at the same height of the eye of the subject, or even lower.

Light: the light changes during the day. At early morning or at late afternoon there is the better light for photos; a warm, "sweet" light. The central hours of the day, instead have a white, harsh light: they are far from ideal.

Background: the background is a very important element of an image: usually, it should be as clean as possible, otherwise it distracts the attention from the subject. The easiest way to get a pleasing background is to use a long focal length and a wide aperture.


 

Learn from other photographers

One of the best ways to improve your photos is to receive suggestions and critiques from other photographers, and to analyze their images. I suggest to joyn a nature photography forum, where you can post your images and receive useful feedback; you can also view and comment the images of other photographers. This is a fantastic opportunity to learn and to know other nature lovers! If you are not a member yet, I highly recommend to joyn the Juza Nature Photography Discussione Forum!

 

Learn how to post-process your photos

Good post-processing techniques are essential to get the best results. I know many photographers - even professionals - that have very poor post-processing techiques, so they are not able to optimize images that otherwise would be fantastic. I suggest to get Adobe Photoshop CS3 (or Photoshop Elements if you have a tight budget) and to learn how to use it for nature photos - many PS techniques are described here on JuzaPhoto, and there are many other websites that explains how to use PS.




Before and after: a good post processing makes a big difference. 


 

Use the right equipment

If you have a limited budget, I recommend to invest in lenses, instead of buying an expensive camera: a cheap Canon 450D with an excellent lens gives better results than a Canon 1DsIII with a poor lens!

Which are the "good" lenses? A 70-300mm zoom lens costs nearly $300 and it is lighter and more versatile than a 300mm f/2.8 lens, that costs ten times as much ($3,000-$4,000). Nevertheless, a professional is willing to spend that much to get the bulky, fixed focal 300 f/2.8 : why? Both the lenses reach the focal length of 300mm, so you get exactly the same magnification, but the images taken with the 300 f/2.8 are sharp and rich of details, while the images taken with the 70-300 are much softer.

Of course, this is an extreme example; you don't have to spend an exaggerate amount of money, there are also some affordable high-quality lenses. The following table can help you to choose the right lens.


  Description Examples
 Cheap (and crappy) lenses  These lenses are quite cheap, but they have poor image and built quality. Avoid them! Many 18-200 or 28-300 f/4.5-6.3 zooms; the cheapest 70-300 f/4-5.6; etc.
 Cheap (but quite good) lenses  These lenses are quite cheap, but they have relatively good image quality for their price. Recommended if you have a tight budget. Many 18-55 f/3.5-5.6 image stabilized zooms; 50mm f/1.8; 70-300 f/4-5.6 image stabilized lenses; etc.
 Semi-Pro lenses  These lenses are not flawless, but they give a very good quality. Often, they have fast ultrasonic motors and image stabilization. They are an excellent combination between price and performances, and I recommend them to amateurs. The Canon 17-85 IS, Nikkor 16-85 VR and other stabilized standard zooms; Sigma 10-20 and similar; etc.
 Cheap Pro lenses  Some professional lenses are not too expensive; their quality is very good for the price. Highly recommended! Sigma 180mm macro; Canon, Nikkor and Sigma 300mm f4; 70-200 f/4 IS, etc.
 Top Pro lenses  The best of the best, but at a very high price.  14-24 f/2.8, 300 f/2.8 image stabilized, 600 f/4 image stabilized and other superteles, etc.

Other than lenses, another thing that can make a big difference is the tripod. The tripod is a great help to get sharp images. If you  handhold the camera, many times you need to use wide apertures or high iso to get a shutter speed fast enough to freeze hand shake; often resulting in insufficiend depth of field or poor image quality due to noise.

With the tripod, instead, your only concern are the subject movements; other than that, you have the freedom to choose the shutter speed and the aperture that you prefer; you can use the lowest ISO setting, too. The result are sharp, noise free images.

 

Do you have comments or questions?

If you have comments or questions about this article, feel free to ask in the Juza Nature Photography Discussion Forum!

 


댓글(0) 먼댓글(0) 좋아요(0)
좋아요
북마크하기찜하기
 
 
 

http://www.juzaphoto.com/eng/articles/exposure_and_histogram.htm

The Exposure

One of the great advantages of digital photography is the possibility to check the exposure immeditaly after the shot, thanks to the LCD preview and the new tools as the histogram and the highlights screen. To take a perfectly exposed photo, there is no need to to complex calculations: I take a test shot with the settings suggested by the camera, I check the histogram and, if necessary, I set the exposure compensation on the proper value and I repeat the photo. Learning to use the basic exposure modes and the histogram requires just a little of patience and practice, but it gives complete control on the image.




With the histogram and other powerful tools of modern digital cameras, it is relatively easy to expose properly even a difficult subject as white birds against a dark background.

 

Exposure Modes

Every serious camera has four essential exposure modes: P, S, A, M (consumer cameras have also many useless "scene" modes).

P is the Program mode. The camera chooses both the aperture value and the shutter speed; you can choose an equivalent combination by turning the main dial. This is the simplest exposure mode, but I'd not recommend to use it, unless you have very little experience.

S ( Tv on Canon cameras ) is the Shutter Priority mode. You choose a shutter speed and the camera choose the appropriate aperture. Personally, I don't like the S mode because it doesn't give control on the depth of field; moreover, if you select an excessively fast shutter speed and there is not enough light, it is easy to get underexposed photos, in particular with "slow" lenses (e.g. f/5.6 lenses) that doesn't allow to use bright apertures to counterbalance the fast shutter speed.

A ( Av on Canon cameras ) is the Aperture Priority mode, and it is by far the most useful exposure mode: I use it for the 90% of my photos. You choose an aperture value and the camera choose the appropriate shutter speed. The A mode allows to control the depth of field; moreover, you can use the A mode to control even the shutter speed. If you need the fastest shutter speed that you can get at a given ISO sentitivy for a given scene, you just have to select the widest aperture: the camera automatically selects the fastest shutter speed. (if the shutter speed is still too slow, you have to raise the ISO sensitivity: if you switch to S mode and you select a faster shutter speed, the photo will be underexposed, because the lens is already at the widest aperture and the camera is not able to counterbalance the shutter speed).

M is the Manual mode. It allows to choose both shutter speed and aperture; it can be very useful in particular situations, but it requires some experience. I use M mode for the 10% of my photos: it is very useful when the flash is the main light source and in particular situations as night photography. The Canon cameras have a limited exposure compensation range (+/- 3 stops, while Nikon cameras have +/- 5stops), so in some rare situations the only way to get the correct exposure is to switch in M mode. Other than that, the manual mode is useful when the subject is moving against a changing background. 

 

Metering Modes

There are mainly three types of metering modes: Evaluative, Center-weighted and Spot. Spot mode uses only a very small area, usually correspondent to the selected AF area, to determine the exposure; Center-weighted, instead, uses the central area of the frame.

Evaluative metering analyzes the entire image to determine the exposure. I use always Evaluative; Spot and Center-weighted were the best choice with film, but with the advent of digital they are no longer so important. In fact, using Evaluative metering and checking the histogram you can determine the right exposure much faster than using spot metering.

 

Exposure Compensation


The exposure compensation is essential to get correctly exposed photo. Even though the modern cameras have very sophisticate exposure meters, I rarely use the "default" exposure, I often use more or less exposure compensation to get the best results.

If you set the exposure compensation of a positive value, e.g. + 0.3 or +2, you increase the exposure and you brighten up the photo; if you set the EC on a negative value, e.g. - 0.3 or -2, you decrease the exposure and you darken the photo. Some cameras have a quite limited compensation range; for example, the Canon 350D allows to underexpose or overexpose just up to 2 stops. To overcome the limits of the camera, you have to switch to manual mode and choose the combination of shutter speed/aperture that gives the desired results.

Usually, I set the exposure compensation after reading the histogram: if the photo is overexposed I set the EC on a negative value, and vise versa. An alternative technique that I use when I photograph white subjects in difficult light condition, to preserve the highlights, is to set the EC on -1 and to lower the ISO sensitivity by 1 stop. This way, I get the same shutter speed, but there are less chances to overexpose the highlights: for example, for given scene and aperture both ISO 400 EC 0, or ISO 200 EC -1 gives the same shutter speed, but the latter setting helps more to avoid overexposed highlights. Of course, by underexposing by 1 stop



A white subject against a dark background: here, the exposure compensation of -1 helped to preserve the highlights. Since the -1 setting gives a 1 stop faster shutter speed, I lowered the ISO by 1 stop. In other words, you get the same shutter speed both at ISO 400 and ISO 200 @ -1, but the latter setting is a better choice in difficult situations, because it avoids the overexposure.
the whites are properly exposed, but the midtones and the shadows might be quite dark.

 

 

The Histogram

The LCD screen allows to see the image, but it is just an approximate preview that does not allow to judge precisely the exposure: to determine if the image is exposed properly, you have to use the histogram. The histogram is a graph that shows the brightness of the photo; every SLR camera offers the possibility to view the luminance histogram of the photo, and some cameras have even RGB histograms. In Canon cameras, the histogram is shown in the main photo review screen, together with a small preview of the photo and the exposure data.



The horizontal axis represents the level of brightness, from 0 (pure black) to 255 (pure white). The vertical axis represents the number of pixels that have a determined brightness.

Learning to read the histogram displayed by your camera is quite simple: you just have to remember that the histogram should match the subject. For example, if you take a photo of a white subject, the histogram should be in the right area (brighter values); if it is in the left area the photo is underexposed because the white subject has been recorded  as dark tones by the camera. For the same reason, a dark subject should create an histogram located mainly in the dark side; an subject of average brightness should create a centered histogram, and so on.

When you look the histogram, it is very important to check if the graph touches an extremity (0 or 255): in this case, the image is severely underexposed or overexposed and it has lost detail. Even when you photograph a very light or very dark subject, the histogram should never tough the edges.

I think that the best way to learn using the histogram is to observe the graph in various situations: here you can see some examples.




    
This photo shows a subject with average brightness. The photo is well exposed histogram, as expected, is relatively centered; the tonal range goes from black to white, but the majority of the image is composed by midtones. This is not an easy subject to expose - it is enough to overexpose or underexpose by 1 stop to lose some detail in the black or in the withes. Here, instead, the graph is close to to both the sides, but it does not touch any of them.

 





This is a completely overexposed photo. The dark areas are recorded as midtones, and the highlights are so overexposed that they have become a white area, without detail. The histogram, clipped on the right side, shows the loss of image data: this photo can not be recovered. If you get an histogram like that, set the exposure compensation a negative value (e.g. -1) and repeat the photo. If the dynamic range of the camera is not enough to record the scene, you have to use special techniques as blening exposures with Layers and Layer Mask.

    


 




    
This photo has two main areas, a very dark one and a midtone area; there are very few intermediate values between the teo areas. The dark area is very close to the black side, but it is not clipped; the second, lower, peak of the graph represent the sky and the other midtones area of the photo. This is the correct exposure when you want to create a kind of "siliouette".

 

 





This is the classic "layer" image. In the histogram, there are four main peacks that represents the four layers of the photo; each layer has a slightly different brightness. Other than that, the histogram has a lot of empty room on both sides: it mean that the photo has low contrast. It would be pretty easy to increase the contrast with PS by dragging the white and the black point to the edges of the histogram, but here I preferred to mantain the low constrast of the scene, to create a soft, relaxing mood.

    


 




    
This photo is taken during the night: the light on the landscape is the Moon light, not the Sun. It would have been very easy to get a brighter photo, but the scene was dark so I preferred to preserve the original darkness. The histogram is on the dark side, and here it means that the photo is correcly exposed, since it shows that a dark scene was recorded as dark tones.

 

 





In this photo, I set the exposure to get the Great White Heron properly exposed: the exposure compensation was set on -1. If I had used the defaul exposure, the background would be brighter, but the white feathers of the bird would be overexposed. Here, instead, the background is almost completely back (some areas are actually pure black, as you can see by the histogram clipped on the dark side), but the subject is properly exposed. When you have a scene like this one, don't worry about the clipped blacks; you have to expose for the subject.

    


 




    
During my workshops, I'm often asked if a peak like this one represents a bad exposure. The answer is no: this photo is well exposed. You have to worry only when the graph is clipped on the sides; if it is clipped on the top, it just mean that the large majority of the photo has a certain level of brightness, and the peak would be too high to be displayed entirey in the small histogram window. In this case, the majority of the photo is constituted by light midtones.

 





This is a clearly underexposed photo. When you get an histogram like that, it mean that the photo is very underexposed, unless you are photographing a very dark subject. Here, instead, the butterfly and the leaves were midtones, but they had been recorded as dark tones. It is still possible to recover the exposure, since the histogram is not clipped, but you would increase a lot the noise: in these situations, if it is possible, I recommend to set the exposure compensation on a positive value and to repeat the photo.

    


 




    
This photo looks "muddy" and flat because it lacks of contrast. The exposure is not bad - in these situations, you can do nothing to enhace the photo in-camera, you have to adjust the image with Photoshop, by dragging the white and black points close to the edges of the histogram, "trimming" the empty areas. It is easy to get histograms with empty areas (= lack of contrast) in foggy days, or when you photograph a subject in shade or soft light. By increasing the contrast with PS you get excellent colors and perfect exposure.

 





This photo has a wide tonal range, from pure black to almost white, so the histogram is spread on the entire graph. Note that the biggest peack of the histogram represents the dark blue sky, not the mountain - the black silouette is a completely black (clipped) area, and it is represented by the thin and high peack on that is clipped the left side. This is the correct exposure to get the mountain as a silouette; if you want to get some detail in the foreound in such contrasted scenes, instaed, you have to use the blended exposures technique.

   


 




    
This photo is overexposed because the midtones of the subject had been recorded as light tones, but it can be easily rescued with Photoshop, since the histogram is not clipped. Actually, this exposure could be considered the "perfect exposure" for this subject - it is the brightest exposure that you can get without clipping the highlight. By correcting it with PS, you get less noise and more dynamic range than a photo that has been already "exposed properly" in-camera.
 





This photo, instead, is too overexposed to be corrected with PS. The pink petals of the flowers have become completely white, without detail. The histogram shows clearly the overexposure, since it is clipped on the right side (highlights). In the Highlights screen, the petals would blink as overexposed areas.

You have to set the exposure compensation on a negative values and to repeat the photo.

    


 




    
This photo is exposed properly, since the bright subject (a white bird against a light grey background) has been recorded as light tones.

The high peack represents the uniform background; as already said, it is not a problem if the histogram is clipped on the top, while it would be a prolem if it was clipped on the left or right side.

 

The Highlights screen

The Highlights screen integrates the histogram. This screen shows a preview of the image: if there are blinking areas, the image is overexposed, and some areas are completely white and devoid of detail. If it is possible, you have to set the exposure compensation on a negative value and to repeat the shot.

 

Do you have comments or questions?

If you have comments or questions about this article, feel free to ask in the Juza Nature Photography Discussion Forum!

 


댓글(0) 먼댓글(0) 좋아요(0)
좋아요
북마크하기찜하기
 
 
 

http://www.dpreview.com/learn/?/Glossary/
- Glossary about camera system, digital imaging, exposure, optical etc.
- Digital Photography Review

http://cpn.canon-europe.com/content/infobank.do
- Exposure settings, white balance, autofucus, lenses (image stabilisation) etc.
- Canon Professional Network - Infobank

http://www.redhax.net/wiki/Main_Page
- Optical (optical low pass filter), digital pathway, metering (zebras), camera, image control (rolling shutter artifacts), post production (Bayer demosaicing), finishing

http://www.juzaphoto.com/eng/articles_photography.htm
- Image quality, exposure, autofocus, image stabilisation etc.

* Pixel Correction (bad, defect, dead)
(1) Agilent application note
http://cp.literature.agilent.com/litweb/pdf/5988-2371EN.pdf

(2) Image gradient technique
http://www.rad-icon.com/pdf/Radicon_AN03.pdf


댓글(0) 먼댓글(0) 좋아요(0)
좋아요
북마크하기찜하기
 
 
 

http://www.avchd-info.org/format/index.html

AVCHD format specification overview

Recording Media 8cm DVD media/
SD Memory Card/"Memory Stick"/
Built-in Media
V
i
d
e
o
Video Signal 1080/60i
1080/50i
1080/24p
720/60p
720/50p
720/24p
480/60i 576/50i
Pixels (horizontal x vertical) 1920×1080
1440×1080
1280×720 720×480 720×576
Aspect ratio 16:9 16:9 4:3, 16:9 4:3, 16:9
Compression technology MPEG-4 AVC/H.264
Luminance
sampling frequency
74.25MHz
55.7MHz
74.25MHz 13.5MHz 13.5MHz
Sampling structure 4:2:0
Quantifying bit number 8 bit (luminance/color contrast)
A
u
d
i
o
Compression technology Dolby Digital (AC-3) Linear PCM
Bit rate after
compression
64 ~ 640kbps 1.5Mbps (2 channels)
Audio channels 1-5.1 channels 1-7.1 channels
System MPEG-2 Transport Stream
System bit rate ~24Mbps
(~18Mbps for DVD)

 


댓글(0) 먼댓글(0) 좋아요(0)
좋아요
북마크하기찜하기
 
 
 
처음 처음 | 이전 이전 | 1 | 2 |다음 다음 | 마지막 마지막