Video Revision
Content
Video revision
Contents
Technical ................................................................................................................................................. 2 Digital Compositing ............................................................................................................................. 2 Traditional compositing method..................................................................................................... 2 Processed foreground method ....................................................................................................... 3 Blending modes............................................................................................................................... 4 Gamma ................................................................................................................................................ 5 Slice Graph .......................................................................................................................................... 7 Matte creating & manipulation .......................................................................................................... 9 Rotoscoping .................................................................................................................................... 9 Luma keying .................................................................................................................................... 9 Chroma keying .............................................................................................................................. 10 Colour difference keying ............................................................................................................... 10 Matching the light space................................................................................................................... 11 Essay...................................................................................................................................................... 13 Digital revelation ............................................................................................................................... 13 History of 3D ..................................................................................................................................... 15 Motion graphics ................................................................................................................................ 16
Technical
Digital Compositing
Traditional compositing method
The traditional compositing method is used to create combine two images together The formula for compositing is: C = (m x fg) + ((1-m) x bg) C = Composite m = Matte fg = Foreground bg = Background
The composite operation consists of three steps : The foreground is scaled (multiplied) by the matte The background is scaled my the inverse of the matte Then the results are summed together to create the final composite (add)
Flow graph for the traditional compositing method:
BG
multi Scaled BG add
invert
matte
Composite FG multi Scaled FG
2
Processed foreground method The process foreground method of compositing takes a different approach: A separate plate, known as the ‘backing plate’, is formed which is the same colour as the backing colour of the foreground image. This is multiplied by the inverted matte to form the scaled background plate which is then subtracted from the foreground to create the processed foreground image.
The processed foreground method is used for better edge quality o But a downside is that it is vulnerable when there are variations in luminance levels of the background colour. o So it the background needs to be uniformly lit so it is all the same colour
FG
sub Pull matte invert Process foreground
multi Scaled backing plate Backing plate
Sub is the processed foreground because you are subtracting to get the final result Subtracting the scaled backing plate from the foreground image to get excellent edge quality Multi = is the scaled backing plate: o Multiply is just: O=AxB
3
Blending modes
Maximum Maximum operation chooses the brightest pixel between the 2 images This is a very clean way to create beautiful composites. Most useful when you are combining the bright areas of one image to the dark areas of another
Minimum The minimum operation is the opposite of the maximum operation. It picks the darkest pixel between the 2 images Again a great way to create digital composites Most useful when compositing the dark areas of one image onto the light areas of another.
4
Gamma
This is the formula for gamma:
O = (i) g
O = output pixel i = input pixel g = gamma factor
This is the inverse formula for gamma
O = (i) 1/g
O = output pixel i = input pixel g = gamma factor What is gamma? o o o o o o o A mathematical operation that is performed on image pixels that alter the brightness of the mid-tones without affecting the black and white points. In the early days display devices had a non-linear display characteristic Studies have shown that humans like a little bit of gamma in our image display The amount of gamma depends on local lighting, this means that the amount of gamma changes from video to film Gamma is a power function that is applied to pixel data in order to lighten or darken an image. In other words, the new pixel value is the original pixel value raised by the power of gamma. For example if a pixel value of 128 represented as 0.5 with a gamma value above 1 will darken an image and below 1 will lighten it.
Most software packages have a gamma operation and this implies using the inverse of the gamma, so that when the values is above 1 it will brighten an image, o This is done to conform to what the user would expect
Gamma changes are used rather than scaling operations to alter the brightness of an image for 2 reasons: o o Firstly gamma changes effect an image in a way similar to the non-linear response of the eye, so it seems more natural. Secondly they do not clip the image.
5
It is cheaper for TV studios to broadcast a non-linear signal that was gamma corrected than to build gamma correction into TV sets. Monitor gamma o o o o Has a non-linear characteristic and is an inherent feature of a cathode ray tube Increasing the voltage to the electron gun increases the stream of electrons to the phosphors on the screen which in turn glow brighter. However, the stream of electrons does not increase linearly with gun voltage Instead monitor gamma has a characteristic corresponding to a gamma value between 2.35 and 2.5 depending on the monitor’s construction.
Gamma Correction o o o o o o Gamma correction is used to compensate for the monitor gammas non-linear characteristic Most systems refer to this as ‘Display Gamma’ The correction is applied by using a look up table (LUT) for the monitor, with values being determined by using the formula 1/gamma. The original image is first loaded into the frame buffer for a work station. The Look up table is used to convert the input values to output values. In practice we tend to use a gamma correction value of 2.2 instead of 2.5
End-to-End gamma o Display devices are improved if they have a slight amount of gamma to them, because of this gamma that is applied that does not cover the gamma range of the device. For example, if a device has a gamma value of 2.5 and a gamma correction value of 2.2 may be applied to give an overall value of 1.1
o
o o
This overall value is known as end-to-end gamma, and this is the most important as it determines how an image will look on a display device. Formula for end-to-end gamma is: End-to-end gamma = monitor gamma / gamma correction
6
Slice Graph
The slice graph is a very important tool for analysing digital images The graph is created by using a slice tool to create a line across a region of interest in an image. Then the pixel values from under the line are plotted onto a graph This graph shows the colour levels relative to each other and provides insight for pulling mattes and despill operations
This is the slice graph for a mid-grey circle on a green background
Formula for the raw colour difference mate is: rm = G – max(R,B) rm = Raw matte G = Green R = Red B = Blue
7
The flow graph for colour difference mate is:
GS
R G B
Sub
Max (R,B)
Raw matte does not have enough density for the final matte, As the final matte will need a backing region of 100% white and a foreground region of 0% black. This is done by scaling the density of the raw matte It is scaled using a colour corrections curve. Scaling the white for the matte must not be overdone as it can harden the matte edges if over scaled.
If non-zero black pixels start to become lights (go grey), then this again is dealt with in the curve. To get the foreground region to 0% back for the final matte
8
Matte creating & manipulation
Rotoscoping To draw mattes frame by frame, it can manually but it is very time consuming. In the past mattes were drawn by hand, o An excellent example of this is the scape station, orbiting Earth in 1968 “2001: A Space Odyssey ”
Now there are software packages that aid this process and increase productivity. The tools in these packages often include interpolation routines to work out the evolution of travelling mattes between key frames. Essentially it is a matte that changes its shape frame by frame
Luma keying When a matte is created based on the original RGB images luminance information.
Using a single threshold value can create very hard edged mattes, this is more than often not wanted So in a lot of luma-keyers a second threshold if offered to introduce a softer edge to the matte Luminance + 0.30R + 0.59G + 0.11B For a single threshold, a threshold value is set and pixel values greater than the threshold are set to 100% white and pixel values below are set to black A second threshold is added to soften matte edges, o So this is an inner density edge of 100% o And an outer density edge of 0% With a gradient in between them, this gradient allows for softer edges in the matte
9
Chroma keying This is a matte which is based on colour, it takes an RGB image and converts it into HSV (Hue, Saturation and value(or colour, saturation and value))
HSV colour model A range of colours are selected and this is used as a basis on which to define the matte Also it is possible to select a brightness range to give more control.
Colour difference keying Most popular keying method as it gives superior edge characteristics to the matte It relies of the background being one of the 3 primary colours (Red, Blue, Green) The idea is that the difference between the green record and the other 2 records is relatively larger in the green backing (green screen) region but is zero or very small in foreground regions The dark grey partial density is then scaled to 100% white If there are some non-zero pixels in the foreground region of the matte then they are pulled down to zero to give a solid black matte Simple to invert the matte if needed.
10
Matching the light space
Matching the light space is a process of making two images that were shot separately look like they were shot together in the same light space. The method is to approach matching the light space in a logical way. First by making a luminance version of the composite to correct the brightness and contrast, then the colour aspects of the image are addressed, following that light direction, quality and interactive lighting effects are explored to refine the basic colour correction Shadows are examined for edge characteristics also a task to handle atmospheric haze
Stages of matching the light space
Brightness and contrast are matched using the luminance (greyscale) version of the image. o This is done by looking at the black, white and mid-tones of both images o Then adjusting the brightness and contrast until the foreground and background are matching o This is done using the curve tool. o This stage dramatically improves results when matching the light space Example with numbers: Foreground 0.15 0.83 Background 0.08 0.94
Blacks Whites
The background levels are measured and the foreground levels are adjusted to match the background So if the background black levels were 0.08 black and the foreground was only 0.15 black the curve tool would be used to push the black levels in the foreground to 0.08 black so they would match the levels in the background. When increasing contrast an issue of clipping the blacks and whites can occur, when using a simple colour curve, o Instead an “S” curve is more elegant as it gently pushes the blacks and whites Mid-tones Even though the blacks and white points have been matched, mismatches in the midtones can still arise due to the dissimilar gamma in the two images However this can be fixed easily without disturbing the blacks and white points, because their control points have already been set. So it is just a matter of bending the middle of the colour curve to adjust the mid-tones.
11
Colour issues o o o Matching the colour can be difficult depending on the content of the foreground and the background layers Using Grey cards can help us match colours as they have equal levels of R, G and B Matching flesh tones, if possible, is also a good technique as these will be similar in humans. Adjusting foreground colour bias to match background colour bias.
Light direction, quality and interaction o Light direction is largely important if there are cast shadows, as the direction of light should obviously be the same direction in both the foreground and background to create believable shadows. Light quality refers to the nature if the light source For example, hard or soft. So if the light source is casting soft shadows, it would make for a unbelievable composite if there were hard shadows in the background and soft shadows in the foreground. Interaction between layers can take a verity of forms, For example, a foreground object casting shadows on a background layer. It can be used as a means of embedding a foreground object into a composite For example, if a person in a white shirt walks next to a brightly lit red wall, the shirt will pick up a red tint.
o
o
o
Shadows o o Their edge characteristics depend on light source, the distance between the light source and the object. Shadows cast from a coloured light source have colour themselves
Atmospheric haze o o Has the effect of blurring objects in the distance and turning them bluer Associated with light scattering
12
Essay
History of video and film & The Digital Revolution
History of video and film Until persistence of vision was understood it was difficult to consider the idea of moving images being possible Photographers such as Eadweard Muybridge helped with the understanding the nature of motion Muybridge had cameras activated by trip wires to capture the motion of a horses gallop His findings are still used by animators today The principle behind being able to understand and visualize moving objects is known as persistence of vision The development of celluloid film provided a convenient way of stacking series of images, which could be played to the eye at a suitable frame rate The kinetoscope was one of the first devices to enable images to be stacked Lumiѐre brothers were the first to project moving images in 1896 Special effects have been used in film making from the early days, for example stop motion and split screen techniques All early compositing techniques were optical, and this remained the case until the start of the digital revolution in the 1990’s Some optical composting techniques like blue screen colour separation and blue screen colour difference process These early techniques were improved with more direct techniques, one of these being rear projects This involved foreground action being filmed against a screen where the background action was being projected o It was important that the cameras shutter speed matched up to the camera o And that the action in the foreground was lit the same as the rear projection
13
These were known as process shots Hitchcock used process shots in “Vertigo” 1958 Rotoscoping was another technique that was used instead of process shots, it was tedious to do. o But results could be spectacular, as shown in Kubrick’s “2001: A Space Odyssey” 1968 With the rise of television in the 1950’s experiments were made with film formats and projector methods In 1952 CinemaScope was patented and had an aspect ratio of 2.35:1 o Was used in “The Robe” 1953 Larger film print sizes such as 70mm were used to improve the quality of film
Digital Revolution One of the most important breakthroughs in film technology was the digital revolution By scanning celluloid film using lasers, it was possible to convert picture information into data, which could then be manipulated using a computer After the film had been manipulated, the final print was created by imprinting the data onto raw film using a laser The next step was to remove the film out of the process, by shooting using a digital video camera and projecting the final result in a digital cinema This opened up a whole host of possibilities Computer based-editing packages allowed editing to take place without any cutting of film Optical techniques were replaced with digital compositing techniques which allowed for composites to be so realistic it is hard to see the joins in most cases o This can be seen in the “Lord of the Rings” films Colour grading is now much easier and more versatile o As well as grading whole frames, selective colour correction can be used on small elements Special effects are easier to use and more versatile, match moving can be used it composite computer generated images onto real life footage (or vice versa) Collaborative post production become easier, with the advancement in the internet and communication This caused a slight negative impact to storytelling however, as an emphasis on the visual spectacles became more apparent.
14
History of 3D
In 1838 Charles Wheatstone published a paper on stereoscopic vision He made an important point that by presenting each eye with a slightly different vision of a scene, the illusion of 3D can be achieved In 1844 Brewster created a practical stereoscopic viewing device
Oliver Wendell Holmes improved the design by adding convex lenses; the new lenses enhanced the depth perception. In 1862 Holmes and Joseph Bates worked together to create a cheap stereoscopic viewer, that forms the basis for many modern designs. Some magic lantern shows projected two images on a screen that were colour coded to create a stereoscopic effect. o They needed to be viewed using glasses o This method is known as an Anaglyph image The earliest film that used this technique was “The power of love” 1922 Hammond & Cassidy invented the teleview system, it had to be synced with the projector, but did not last as it was uncomfortable for people watching the film Died out in 1930 In the 1950’s there was a boom in 3D and this was due to polaroid technology o The screen had to be silver to protect the polarized light o Two projectors needed to be used with polarizing filters One of the most famous and successful 3D films was “House of wax” 2953 By 1954 3D films had ceased again, this time because of the rise in popularity of CinemaScope 3D films were no longer making enough money.
Present day The success of “Avatar” 2009 has revived 3D in present day
15
3D movies require 2x the amount of frames then a 2D film so there is a rise in production costs Also need specialist 3D cameras to film footage. Investments needed to be made in cinemas in order to play the films Many films make a large amount of their income from home distribution, so development and rollout of low cost 3D TV.s is vital for the success of 3D in the future It will be interesting to see if 3D will die out again, like it has done before in the past
Motion graphics
In the early days film titles were developed using techniques from early animation as they showed the graphical potential for moving images Early animations tried to mix animation with live action Max Fleisher pioneered this and also invented Rotoscoping These techniques were not used straight away in film titles and they were dominated by text o Romantic films had swirly text o Horror films would have straight edge font Titles would only be used to introduce a list of characters to start with The potential of using the title sequence to set the mood of the film had not yet been explored. The Charlie Chaplin film “City Lights” 1931 made a huge step forward in motion graphics o As it made clever use of super imposition by means of optical printers In 1940 “Spook Sport” began to use the early animation techniques in a title sequence o Remarkable for its time Saul Bass Revolutionized title sequence design in the 1950’s As he saw the potential in establishing a mood and atmosphere for the film in the sequence He designed the titles for many popular films such as: o “The man with the golden arm” 1955 Where white lines move around the font in different patterns until the final image of a twisted arm is viable o “Anatomy of a Murder” 1959 Synced a silhouette of a body with music o “Vertigo” 1958 Considered his best work These title sequences were impressive at the time but primitive it todays standards In the 50’s creating title sequences took a lot of time as it has done by hand and cost a lot of money 16
Animation was further used in title sequences in the 60’s by Friz Freleng o Who created the title sequence for “The pink panther” 1963 Moris Binder further developed the field in James Bond title sequences, as did Robert Brown John, who created the “From Russia with Love” title sequence. o Using a projector to project text onto dancers o However the dancers had to dance accurately for this to work. Binder further improved this technique using an optical printer to project colours onto a tank in “ThunderBall” ALL of the Bond title sequences perfectly reflected the theme of the films In 1995 Kyle Cooper produced the ground breaking title sequence for the movie “se7en” o In contrast to the time when Saul Bass was designing titles, Cooper had access to technology which allowed him to work faster with much less constraints o Cooper is considered one of the best motion graphics artists in the world at the moment due to his ability to depict the nature and mood of a film
17
Contents
Technical ................................................................................................................................................. 2 Digital Compositing ............................................................................................................................. 2 Traditional compositing method..................................................................................................... 2 Processed foreground method ....................................................................................................... 3 Blending modes............................................................................................................................... 4 Gamma ................................................................................................................................................ 5 Slice Graph .......................................................................................................................................... 7 Matte creating & manipulation .......................................................................................................... 9 Rotoscoping .................................................................................................................................... 9 Luma keying .................................................................................................................................... 9 Chroma keying .............................................................................................................................. 10 Colour difference keying ............................................................................................................... 10 Matching the light space................................................................................................................... 11 Essay...................................................................................................................................................... 13 Digital revelation ............................................................................................................................... 13 History of 3D ..................................................................................................................................... 15 Motion graphics ................................................................................................................................ 16
Technical
Digital Compositing
Traditional compositing method
The traditional compositing method is used to create combine two images together The formula for compositing is: C = (m x fg) + ((1-m) x bg) C = Composite m = Matte fg = Foreground bg = Background
The composite operation consists of three steps : The foreground is scaled (multiplied) by the matte The background is scaled my the inverse of the matte Then the results are summed together to create the final composite (add)
Flow graph for the traditional compositing method:
BG
multi Scaled BG add
invert
matte
Composite FG multi Scaled FG
2
Processed foreground method The process foreground method of compositing takes a different approach: A separate plate, known as the ‘backing plate’, is formed which is the same colour as the backing colour of the foreground image. This is multiplied by the inverted matte to form the scaled background plate which is then subtracted from the foreground to create the processed foreground image.
The processed foreground method is used for better edge quality o But a downside is that it is vulnerable when there are variations in luminance levels of the background colour. o So it the background needs to be uniformly lit so it is all the same colour
FG
sub Pull matte invert Process foreground
multi Scaled backing plate Backing plate
Sub is the processed foreground because you are subtracting to get the final result Subtracting the scaled backing plate from the foreground image to get excellent edge quality Multi = is the scaled backing plate: o Multiply is just: O=AxB
3
Blending modes
Maximum Maximum operation chooses the brightest pixel between the 2 images This is a very clean way to create beautiful composites. Most useful when you are combining the bright areas of one image to the dark areas of another
Minimum The minimum operation is the opposite of the maximum operation. It picks the darkest pixel between the 2 images Again a great way to create digital composites Most useful when compositing the dark areas of one image onto the light areas of another.
4
Gamma
This is the formula for gamma:
O = (i) g
O = output pixel i = input pixel g = gamma factor
This is the inverse formula for gamma
O = (i) 1/g
O = output pixel i = input pixel g = gamma factor What is gamma? o o o o o o o A mathematical operation that is performed on image pixels that alter the brightness of the mid-tones without affecting the black and white points. In the early days display devices had a non-linear display characteristic Studies have shown that humans like a little bit of gamma in our image display The amount of gamma depends on local lighting, this means that the amount of gamma changes from video to film Gamma is a power function that is applied to pixel data in order to lighten or darken an image. In other words, the new pixel value is the original pixel value raised by the power of gamma. For example if a pixel value of 128 represented as 0.5 with a gamma value above 1 will darken an image and below 1 will lighten it.
Most software packages have a gamma operation and this implies using the inverse of the gamma, so that when the values is above 1 it will brighten an image, o This is done to conform to what the user would expect
Gamma changes are used rather than scaling operations to alter the brightness of an image for 2 reasons: o o Firstly gamma changes effect an image in a way similar to the non-linear response of the eye, so it seems more natural. Secondly they do not clip the image.
5
It is cheaper for TV studios to broadcast a non-linear signal that was gamma corrected than to build gamma correction into TV sets. Monitor gamma o o o o Has a non-linear characteristic and is an inherent feature of a cathode ray tube Increasing the voltage to the electron gun increases the stream of electrons to the phosphors on the screen which in turn glow brighter. However, the stream of electrons does not increase linearly with gun voltage Instead monitor gamma has a characteristic corresponding to a gamma value between 2.35 and 2.5 depending on the monitor’s construction.
Gamma Correction o o o o o o Gamma correction is used to compensate for the monitor gammas non-linear characteristic Most systems refer to this as ‘Display Gamma’ The correction is applied by using a look up table (LUT) for the monitor, with values being determined by using the formula 1/gamma. The original image is first loaded into the frame buffer for a work station. The Look up table is used to convert the input values to output values. In practice we tend to use a gamma correction value of 2.2 instead of 2.5
End-to-End gamma o Display devices are improved if they have a slight amount of gamma to them, because of this gamma that is applied that does not cover the gamma range of the device. For example, if a device has a gamma value of 2.5 and a gamma correction value of 2.2 may be applied to give an overall value of 1.1
o
o o
This overall value is known as end-to-end gamma, and this is the most important as it determines how an image will look on a display device. Formula for end-to-end gamma is: End-to-end gamma = monitor gamma / gamma correction
6
Slice Graph
The slice graph is a very important tool for analysing digital images The graph is created by using a slice tool to create a line across a region of interest in an image. Then the pixel values from under the line are plotted onto a graph This graph shows the colour levels relative to each other and provides insight for pulling mattes and despill operations
This is the slice graph for a mid-grey circle on a green background
Formula for the raw colour difference mate is: rm = G – max(R,B) rm = Raw matte G = Green R = Red B = Blue
7
The flow graph for colour difference mate is:
GS
R G B
Sub
Max (R,B)
Raw matte does not have enough density for the final matte, As the final matte will need a backing region of 100% white and a foreground region of 0% black. This is done by scaling the density of the raw matte It is scaled using a colour corrections curve. Scaling the white for the matte must not be overdone as it can harden the matte edges if over scaled.
If non-zero black pixels start to become lights (go grey), then this again is dealt with in the curve. To get the foreground region to 0% back for the final matte
8
Matte creating & manipulation
Rotoscoping To draw mattes frame by frame, it can manually but it is very time consuming. In the past mattes were drawn by hand, o An excellent example of this is the scape station, orbiting Earth in 1968 “2001: A Space Odyssey ”
Now there are software packages that aid this process and increase productivity. The tools in these packages often include interpolation routines to work out the evolution of travelling mattes between key frames. Essentially it is a matte that changes its shape frame by frame
Luma keying When a matte is created based on the original RGB images luminance information.
Using a single threshold value can create very hard edged mattes, this is more than often not wanted So in a lot of luma-keyers a second threshold if offered to introduce a softer edge to the matte Luminance + 0.30R + 0.59G + 0.11B For a single threshold, a threshold value is set and pixel values greater than the threshold are set to 100% white and pixel values below are set to black A second threshold is added to soften matte edges, o So this is an inner density edge of 100% o And an outer density edge of 0% With a gradient in between them, this gradient allows for softer edges in the matte
9
Chroma keying This is a matte which is based on colour, it takes an RGB image and converts it into HSV (Hue, Saturation and value(or colour, saturation and value))
HSV colour model A range of colours are selected and this is used as a basis on which to define the matte Also it is possible to select a brightness range to give more control.
Colour difference keying Most popular keying method as it gives superior edge characteristics to the matte It relies of the background being one of the 3 primary colours (Red, Blue, Green) The idea is that the difference between the green record and the other 2 records is relatively larger in the green backing (green screen) region but is zero or very small in foreground regions The dark grey partial density is then scaled to 100% white If there are some non-zero pixels in the foreground region of the matte then they are pulled down to zero to give a solid black matte Simple to invert the matte if needed.
10
Matching the light space
Matching the light space is a process of making two images that were shot separately look like they were shot together in the same light space. The method is to approach matching the light space in a logical way. First by making a luminance version of the composite to correct the brightness and contrast, then the colour aspects of the image are addressed, following that light direction, quality and interactive lighting effects are explored to refine the basic colour correction Shadows are examined for edge characteristics also a task to handle atmospheric haze
Stages of matching the light space
Brightness and contrast are matched using the luminance (greyscale) version of the image. o This is done by looking at the black, white and mid-tones of both images o Then adjusting the brightness and contrast until the foreground and background are matching o This is done using the curve tool. o This stage dramatically improves results when matching the light space Example with numbers: Foreground 0.15 0.83 Background 0.08 0.94
Blacks Whites
The background levels are measured and the foreground levels are adjusted to match the background So if the background black levels were 0.08 black and the foreground was only 0.15 black the curve tool would be used to push the black levels in the foreground to 0.08 black so they would match the levels in the background. When increasing contrast an issue of clipping the blacks and whites can occur, when using a simple colour curve, o Instead an “S” curve is more elegant as it gently pushes the blacks and whites Mid-tones Even though the blacks and white points have been matched, mismatches in the midtones can still arise due to the dissimilar gamma in the two images However this can be fixed easily without disturbing the blacks and white points, because their control points have already been set. So it is just a matter of bending the middle of the colour curve to adjust the mid-tones.
11
Colour issues o o o Matching the colour can be difficult depending on the content of the foreground and the background layers Using Grey cards can help us match colours as they have equal levels of R, G and B Matching flesh tones, if possible, is also a good technique as these will be similar in humans. Adjusting foreground colour bias to match background colour bias.
Light direction, quality and interaction o Light direction is largely important if there are cast shadows, as the direction of light should obviously be the same direction in both the foreground and background to create believable shadows. Light quality refers to the nature if the light source For example, hard or soft. So if the light source is casting soft shadows, it would make for a unbelievable composite if there were hard shadows in the background and soft shadows in the foreground. Interaction between layers can take a verity of forms, For example, a foreground object casting shadows on a background layer. It can be used as a means of embedding a foreground object into a composite For example, if a person in a white shirt walks next to a brightly lit red wall, the shirt will pick up a red tint.
o
o
o
Shadows o o Their edge characteristics depend on light source, the distance between the light source and the object. Shadows cast from a coloured light source have colour themselves
Atmospheric haze o o Has the effect of blurring objects in the distance and turning them bluer Associated with light scattering
12
Essay
History of video and film & The Digital Revolution
History of video and film Until persistence of vision was understood it was difficult to consider the idea of moving images being possible Photographers such as Eadweard Muybridge helped with the understanding the nature of motion Muybridge had cameras activated by trip wires to capture the motion of a horses gallop His findings are still used by animators today The principle behind being able to understand and visualize moving objects is known as persistence of vision The development of celluloid film provided a convenient way of stacking series of images, which could be played to the eye at a suitable frame rate The kinetoscope was one of the first devices to enable images to be stacked Lumiѐre brothers were the first to project moving images in 1896 Special effects have been used in film making from the early days, for example stop motion and split screen techniques All early compositing techniques were optical, and this remained the case until the start of the digital revolution in the 1990’s Some optical composting techniques like blue screen colour separation and blue screen colour difference process These early techniques were improved with more direct techniques, one of these being rear projects This involved foreground action being filmed against a screen where the background action was being projected o It was important that the cameras shutter speed matched up to the camera o And that the action in the foreground was lit the same as the rear projection
13
These were known as process shots Hitchcock used process shots in “Vertigo” 1958 Rotoscoping was another technique that was used instead of process shots, it was tedious to do. o But results could be spectacular, as shown in Kubrick’s “2001: A Space Odyssey” 1968 With the rise of television in the 1950’s experiments were made with film formats and projector methods In 1952 CinemaScope was patented and had an aspect ratio of 2.35:1 o Was used in “The Robe” 1953 Larger film print sizes such as 70mm were used to improve the quality of film
Digital Revolution One of the most important breakthroughs in film technology was the digital revolution By scanning celluloid film using lasers, it was possible to convert picture information into data, which could then be manipulated using a computer After the film had been manipulated, the final print was created by imprinting the data onto raw film using a laser The next step was to remove the film out of the process, by shooting using a digital video camera and projecting the final result in a digital cinema This opened up a whole host of possibilities Computer based-editing packages allowed editing to take place without any cutting of film Optical techniques were replaced with digital compositing techniques which allowed for composites to be so realistic it is hard to see the joins in most cases o This can be seen in the “Lord of the Rings” films Colour grading is now much easier and more versatile o As well as grading whole frames, selective colour correction can be used on small elements Special effects are easier to use and more versatile, match moving can be used it composite computer generated images onto real life footage (or vice versa) Collaborative post production become easier, with the advancement in the internet and communication This caused a slight negative impact to storytelling however, as an emphasis on the visual spectacles became more apparent.
14
History of 3D
In 1838 Charles Wheatstone published a paper on stereoscopic vision He made an important point that by presenting each eye with a slightly different vision of a scene, the illusion of 3D can be achieved In 1844 Brewster created a practical stereoscopic viewing device
Oliver Wendell Holmes improved the design by adding convex lenses; the new lenses enhanced the depth perception. In 1862 Holmes and Joseph Bates worked together to create a cheap stereoscopic viewer, that forms the basis for many modern designs. Some magic lantern shows projected two images on a screen that were colour coded to create a stereoscopic effect. o They needed to be viewed using glasses o This method is known as an Anaglyph image The earliest film that used this technique was “The power of love” 1922 Hammond & Cassidy invented the teleview system, it had to be synced with the projector, but did not last as it was uncomfortable for people watching the film Died out in 1930 In the 1950’s there was a boom in 3D and this was due to polaroid technology o The screen had to be silver to protect the polarized light o Two projectors needed to be used with polarizing filters One of the most famous and successful 3D films was “House of wax” 2953 By 1954 3D films had ceased again, this time because of the rise in popularity of CinemaScope 3D films were no longer making enough money.
Present day The success of “Avatar” 2009 has revived 3D in present day
15
3D movies require 2x the amount of frames then a 2D film so there is a rise in production costs Also need specialist 3D cameras to film footage. Investments needed to be made in cinemas in order to play the films Many films make a large amount of their income from home distribution, so development and rollout of low cost 3D TV.s is vital for the success of 3D in the future It will be interesting to see if 3D will die out again, like it has done before in the past
Motion graphics
In the early days film titles were developed using techniques from early animation as they showed the graphical potential for moving images Early animations tried to mix animation with live action Max Fleisher pioneered this and also invented Rotoscoping These techniques were not used straight away in film titles and they were dominated by text o Romantic films had swirly text o Horror films would have straight edge font Titles would only be used to introduce a list of characters to start with The potential of using the title sequence to set the mood of the film had not yet been explored. The Charlie Chaplin film “City Lights” 1931 made a huge step forward in motion graphics o As it made clever use of super imposition by means of optical printers In 1940 “Spook Sport” began to use the early animation techniques in a title sequence o Remarkable for its time Saul Bass Revolutionized title sequence design in the 1950’s As he saw the potential in establishing a mood and atmosphere for the film in the sequence He designed the titles for many popular films such as: o “The man with the golden arm” 1955 Where white lines move around the font in different patterns until the final image of a twisted arm is viable o “Anatomy of a Murder” 1959 Synced a silhouette of a body with music o “Vertigo” 1958 Considered his best work These title sequences were impressive at the time but primitive it todays standards In the 50’s creating title sequences took a lot of time as it has done by hand and cost a lot of money 16
Animation was further used in title sequences in the 60’s by Friz Freleng o Who created the title sequence for “The pink panther” 1963 Moris Binder further developed the field in James Bond title sequences, as did Robert Brown John, who created the “From Russia with Love” title sequence. o Using a projector to project text onto dancers o However the dancers had to dance accurately for this to work. Binder further improved this technique using an optical printer to project colours onto a tank in “ThunderBall” ALL of the Bond title sequences perfectly reflected the theme of the films In 1995 Kyle Cooper produced the ground breaking title sequence for the movie “se7en” o In contrast to the time when Saul Bass was designing titles, Cooper had access to technology which allowed him to work faster with much less constraints o Cooper is considered one of the best motion graphics artists in the world at the moment due to his ability to depict the nature and mood of a film
17
