noise said:
…Take an old fashion camera image, a photograph. It's not going to be such and such mega pixel's, it is a photo and there seems to be a difference. If I understand mega pixel's (MP) correctly for example a 1 MP image there are 1 million pixels. A camera photograph can be enlarged and enlarged and enlarged and it never begins to pixelate….
Perhaps I can help clarify some of the differences between analog images and digital images. First, in a digital image (even one that was created by scanning an analog photograph) there is the optical resolution (in the case of digital original we may call it native resolution) and color resolution. In the case of optical resolution, this means the number of pixels (picture elements - basically small color squares) in each dimension, height and width, determines the overall optical detail "resolving capability." A digital image does not have a fixed "size," that is to say, you can change the "final output size" by changing the resolution - pixels per inch or centimeter or whatever - and redistributing the pixels: if you increase the pixels per inch (ppi) of the image (but keep the total number of pixels the same in each dimension) it will decrease the "final output size;" and if you decrease the ppi (keeping the total number of pixels the same) it will increase the "final output size."
Now a 35mm still analog film camera image has a relatively high optical resolution compared to many consumer digital cameras. However, 35mm has relatively low resolution compared to medium (6 x 6 cm) and large (4 x 5 inch & 8 x 10 inch sheet) professional film formats that have enormous resolving power that is hard to compare to any digitally originated image on even professional digital photography equipment commercially available.
To put into perspective the problem of digital images pixelating, you need to understand simply that all digital images are made of pixels, whether originally captured digitally or photographed on analog film and then scanned into a digital version of the analog image. So a relatively low optical resolution digital image, say 1200 pixels by 1500 pixels (1.8 mega pixels), will have a limited amount of photographic detail. This image can be set to 300 ppi for example and get a 4 inch by 5 inch image; this will have relatively large pixels because there are only 300 pixels in one inch. If you compare this to a relatively high optical resolution digital image, say 6000 pixels by 7500 pixels (45 mega pixels), you can set the resolution at 1500 ppi and get a 4 inch by 5 inch image. You will notice that the pixels will be 5 times smaller at this resolution because there are 1500 pixels in one inch. So if you magnify the lower resolution image enough you will begin to see the pixels much quicker because the pixels are relatively large to begin with, whereas with the higher resolution image you can enlarge it 5 times bigger before you begin to see the pixels. Now all of this oversimplified explanation applies much more if you were to print the images than viewing it on a computer monitor where the monitor’s technical specifications (including size, maximum screen resolution, etc.) as well as the software the image is being viewed in and its settings also influence how the image is rendered.
Now, there is also interpolated resolution (as opposed to optical resolution) that cannot add any detail to a digital image but does increase the number of pixels. In other words whether originally digitally captured or photographed on analog film and then scanned, there is the optical resolution of the equipment that is digitizing the image (hardware) and then there is software interpolation which adds pixels by using different math techniques that the user can usually set and adds pixels by averaging values of surrounding pixels. Again this adds to the total number of pixels but cannot add any photographic detail that’s not in the original optical resolution of the image. So while, strictly speaking, the interpolated increase of resolution (adding pixels by software computations) will give the impression that the image may be enlarged more before it starts to pixelate, there will be other loss of quality noticeable in the lack of photographic detail.
The very expensive, highest quality professional digital systems are not cameras but digital backs that attach to a professional film camera where the film back (in the case of medium format roll film - 6 x 6 cm, 6 x 4.5 cm, etc.) or film holder (in the case of view cameras {those that have accordion-like bellows} for large format sheet film - 4 x 5 inch & 8 x 10 inch) would be. These digital systems can theoretically give the same quality as analog film, but there are many practical problems that usually prevent one from obtaining the quality of shooting on film (particularly the larger format films and color reversal film stocks where there is an enormous amount of information captured) and scanning on high quality equipment at high optical resolution and high bit (color resolution) to get a really high quality final image. This is due to the fact that the higher the quality (optical and color resolution) of a digital image is, the larger the file size will be.
While an analog film image will not pixelate when over-enlarged, what happens is the film equivalent: you will start to see film grain and also notice a loss of sharpness when enlarged a great amount.
So now let me explain briefly color resolution measured in bit depth (and there are other aspects that affect image quality but we’ll leave those for now). An 8 bit per color channel RGB image (24 bit image) gives a pretty high color resolution raw image. As you probably know RGB stands for Red, Green, Blue, the primary colors that give the most possible combination of colors (as opposed to CMYK - Cyan, Magenta, Yellow, and Black, for example which is used in printing, which are the complementary colors plus black because you can’t get true black by combining the other three inks and there is a limit in how much ink you can put on the paper before it becomes a soggy mess). There’s also one other important advantage to RGB besides being a much larger color space than CMYK; it is a combination of colored light which comes
through transparent media to created the patterns of the image rather than images on paper where a light source reflects off the paper as some colors are absorbed and some are reflected back.
We can only output to whatever device at 8 bits per channel (RGB is a 3 channel color space whereas CMYK is a 4 channel color space). As I said, a 24 bit RGB image is good as a raw scan, however all images need adjustments. Some of the adjustments are due to correcting some things that you don’t like in the original capture, others, in the case of shooting on film and then scanning, are because the scanning equipment added some bias to the image, etc. But whenever you manipulate a digital image in terms of traditional photographic darkroom techniques (adjusting contrast, color shifts, darkness/lightness aka density, etc.) you lose information in the original raw image that you cannot get back. So the basic thing about color resolution is this. In an 8 bit per channel image (24 bit RGB which gives around 16.8 million colors) you have 256 levels (steps or shades) between black and white in each color channel. When you manipulate the image to get it to look the way you want just in the adjustment stage (later you will also target it to the output method) you lose information so that you will be left with something less than 256 levels between black and white in each color channel (so you’re losing color information that was in the raw image by making adjustments). However, if you scan or capture digitally with professional equipment (those professional digital backs I mentions are just high-end scanners that attach to the back of the camera where the film would normally be) the equipment gives you the option to scan at a higher bit than 8 bits per channel - up to 16 bits per channel.
Now to give you an idea of how much color information is added by adding bits per channel, a 14 bit per channel (42 bit RGB) image gives you 16,384 levels between black and white in each color channel (as opposed to 256 for 8 bit) and a 16 bit per channel (48 bit RGB) image gives 65,536 levels in each color channel. Now as I said, you can’t output these high bit images to any commercially available devices - you have to down sample the high bit image to 8 bits per channel after all the adjustments are done. But as you’re doing the adjustments, there are a lot more bits to work with (levels between black and white in each color channel), so that when you are done there are still 8 good bits to choose from when you tell the image editing software (such as Photoshop) to downsample the image to 8 bits per channel. So the new adjusted and downsampled image still has a perfect 256 levels between black and white in each channel.
The significance of this is that if you lose enough levels, the image becomes "posterized." This means that there will be visible abrupt steps between color / shade values losing the photographic details and nuances. It is different than "pixelating" but also creates quality loss. Lose enough levels and it stops looking like a photograph.
I hope this explanation was not too confusing.
Coming to analog film motion picture verses electronic digital video motion picture, we have to add bit rate to the mix which also affects the "smoothness of motion" besides the other aspects of quality discussed above. So we can imagine everything explained above applied to 24 frames per second, the standard "sync sound speed" of film. This means that there are 24 still frames for every second of screen time. Now we are talking about truly colossal digital file sizes to get the same overall quality (amount of information) as film. To get the full quality in every way of a 90 minute finished film onto a video transfer, it would take a very high bit rate and would fill up several DVD’s. Also most consumer DVD players would be unable to read that high of a bit rate. But keep in mind that watching the video on a TV screen, even a very large flat panel, you don’t need to have the highest bit rate (containing all of the quality) of the original film. You can get away with a considerably lower bit rate without noticing a quality loss simply because the original image area size on the film is not being enlarged to the size of the biggest theater screens in a cinema.
Finally, coming back to the possibilities of control and manipulation of the viewers watching a motion picture program, I will say this: no one knows the vulnerabilities and weaknesses in human perception - its capacity to be fooled - better than a filmmaker. On the most basic level, even analog film projection completely fools the viewer, tricks him into seeing many things that are not actually there in reality. First, as mentioned, there are 24 still photos per second of screen time that fool the viewer into seeing motion. This illusion of motion is created in the viewers mind by having a certain frames per second rate that is passed through a film projector. Each frame is pulled into position in the projector, held steady while the projector light shines through it and the lens projecting it onto the screen. Between each frame there is a very small fraction of time when there is nothing on the screen. The viewer fills in these gaps in his mind and creates the illusion of motion from still photo frames.
Then the whole film is shot all out of sequence and then the takes are chosen and edited together to give the illusion of a sequence and continuity of action that is totally unreal. But the viewer basically follows this action in a state of dissociation and suspension of disbelief completely being engrossed by the film. The filmmaker is actually constructing artificial time and space relationships in the viewers’ minds that are completely manufactured and don’t exist in reality. If the film were to break or the projector otherwise malfunctioned or was turned off, the viewer would suddenly snap out of his dissociated state and become much more aware that he’s watching a film. Now if we consider electronic signals creating the motion picture, whole other worlds are opened up to manipulate people’s perceptions, awareness, senses, feeling, etc. Truly scary when you really thing about it.
