Blu Vs. DVD screenshots... POTC 2 (i got bored)

[syncguy]

Nyquist theorem indeed applies to the digital still photography and it determines the spatial sampling limits of digital cameras. Also, it applies to the still film cameras, however, in this case the size of the film grain is needed to determine the limits of the film.

I will try to explain using a crude example. Say the still camera sensor is 100x100 array (i.e. 10K pixels), and use this camera to take a still photo of a 1080p projected image. The outcome can be easily visualized. The photo will be pixilated and will not be a good representation of the original. If the sensor resolution of the still camera is increased to 200x200, the picture will get better. So, it can be easily visualized that by increasing the pixel resolution of the still camera, a better picture of the projected image can be obtained. Now what would happen if the pixel resolution of the still camera is identical to the projected image (i.e. 1920x1080). In this case, if you could exactly map the pixel grid of the still camera on to the pixel grid of the display device, the you could capture the state of each pixel of the image. However, what would happen if the two grids are not exactly aligned (this is the case in reality). Then a single pixel of the still camera may photograph up to 4 partial pixels of the projected image. In this case, the still camera cannot capture the projected image correctly and aliasing of the edges of the projected image will occur. This is a well known phenomenon in digital cameras and many cameras use anti-aliasing (blur) filters to blur the edges to hide aliasing art-effects when the camera’s sampling ability (or the resolution) approaches the Nyquist limit in relation to the photographed image.

This is where the Nyquist theorem helps to work out the minimum amount of still camera pixels needed to truthfully capture the projected image. As I said in my previous post, this would be about 8 mega pixels for HD. This is the Nyquist limit. Some people may see an improvement to the picture if the resolution of the still camera is further increased (i.e. 16 or 32 mega pixels). At least this is the case for audio sampling – some say they hear the difference between 44 and 96kHz sampling.

Since the Nyquist limit is a function of the original image resolution, DVDs would benefit when compared with HD if the still camera resolution is below the Nyquist limit for HD. So, about 10 mega pixel still camera is suitable to carryout an unbiased comparison of HD and DVD images.

[maxmcleod]

Originally Posted by syncguy

Nyquist theorem indeed applies to the digital still photography and it determines the spatial sampling limits of digital cameras. Also, it applies to the still film cameras, however, in this case the size of the film grain is needed to determine the limits of the film.

I will try to explain using a crude example. Say the still camera sensor is 100x100 array (i.e. 10K pixels), and use this camera to take a still photo of a 1080p projected image. The outcome can be easily visualized. The photo will be pixilated and will not be a good representation of the original. If the sensor resolution of the still camera is increased to 200x200, the picture will get better. So, it can be easily visualized that by increasing the pixel resolution of the still camera, a better picture of the projected image can be obtained. Now what would happen if the pixel resolution of the still camera is identical to the projected image (i.e. 1920x1080). In this case, if you could exactly map the pixel grid of the still camera on to the pixel grid of the display device, the you could capture the state of each pixel of the image. However, what would happen if the two grids are not exactly aligned (this is the case in reality). Then a single pixel of the still camera may photograph up to 4 partial pixels of the projected image. In this case, the still camera cannot capture the projected image correctly and aliasing of the edges of the projected image will occur. This is a well known phenomenon in digital cameras and many cameras use anti-aliasing (blur) filters to blur the edges to hide aliasing art-effects when the camera’s sampling ability (or the resolution) approaches the Nyquist limit in relation to the photographed image.

This is where the Nyquist theorem helps to work out the minimum amount of still camera pixels needed to truthfully capture the projected image. As I said in my previous post, this would be about 8 mega pixels for HD. This is the Nyquist limit. Some people may see an improvement to the picture if the resolution of the still camera is further increased (i.e. 16 or 32 mega pixels). At least this is the case for audio sampling – some say they hear the difference between 44 and 96kHz sampling.

Since the Nyquist limit is a function of the original image resolution, DVDs would benefit when compared with HD if the still camera resolution is below the Nyquist limit for HD. So, about 10 mega pixel still camera is suitable to carryout an unbiased comparison of HD and DVD images.

i see your point, makes more sense now

but this is just a comparison, so im not too worried about all that for this.
Still a better comparison than the king kong example, which doesn't truly reflect a dvd on a large screen... I've never watched a movie that was blurry like that, because it doesn't exist, its still clear, just less details

[syncguy]

Originally Posted by maxmcleod

i see your point, makes more sense now

but this is just a comparison, so im not too worried about all that for this.
Still a better comparison than the king kong example, which doesn't truly reflect a dvd on a large screen... I've never watched a movie that was blurry like that, because it doesn't exist, its still clear, just less details

No worries. Thanks for your comparison.

I made those comments as I have seen many people using digital cameras to capture HD images and may assist when choosing a camera for this purpose.