I am going to try and explain how wrong you are, I hope you at least you have some basic knowledge about how a bitmapped matrix display and digital imagery actually works (though from your postings I do not hold out much hope )
Consider a single pixels data in a single frame of video after it has been decoded by the destination box mpeg decoder.
You have a minimum of 24 bits of digital data (3 bytes), each 8 bits represents the colour of each of the sub pixels Red Green and Blue of the full pixel.
8 bits in binary represents the numbers 0-255 in decimal. Hence there are 256 variations in red green and blue for each full pixel. Each pixel can only display a single colour, The variations possible are 256 x 256 x 256, which equates to over 64 million individual variations in colour (from pure back to pure white). Simple examples Red 255 Green 0 Blue 0 gives pure RED, Red 255, Green 255 and Blue 255 gives pure white. If you do not get this then you don't have the basic physics of a O level physics student, nor the understanding of physics discovered a long time ago about the nature of light.
Each pixel on a LCD/Plasma display is presented with 3 - 8 bit binary numbers for a single frame (1/25 second for 1080i).
To simplify things lets consider the data a square matrix of just 4 pixels might be presented with during a single frame.
Here's the purely arbitrary data presented by row and column order.
Top Row
Red 240, Green 0, Blue 127.
Red 0, Green 50, Blue 60
Bottom Row
Red = 60
Green = 60
Blue = 60
(Actually a grey scale monochrome image)
Red =256
Green = 0
Blue = 0
(A pure Red pixel).
Now lets consider how to display this rudimentary fragment of a video frame in a 4 by 4 matrix of pixels (stretching it 2 times in both axes by your very weird ideas).
Interested to see what the 4 rows in each row for each column might contain, if stretched (I have no idea what this means, in a digital environment it has no meaning.
Digital has an absolute and precisely defined set of data for the specific time period of the data. It represents a single point in time, You cannot as you put it stretch it, you can guess what the intermediate values might be based on a second data reading taken a bit later (sample rate).
Only analogue provides a continuous value for any quantity, the resolution being given by the carrier frequency used to transmit the data.
Basically until you at least demonstrate at least a schoolboy level of physics as it relates to digital imaging. It's clear you have no idea how to manipulate digital images even for a simple digital still camera (Which in themselves are compressed in a lossy format (normally .jpeg), let alone the complexity of compressing multiple frames in a video.
Here's a simple quiz for you, until you can answer it don't bother to post such rubbish.
Rather take a O Level Physics course.
How much data will it take to transmit (Bytes or Bits) a single frame of 24 bit colour video with every pixel exactly represented ?
At 25 fps what bitrate would a broadcaster need to use in Mbps (Megabits/second) ?
What roughly would a 1920 x 1080 broadcast recorded with no compression require in the way of storage in GB (Gigabytes) on a PVR hard disk ?
I am going to try and explain how wrong you are, I hope you at least you have some basic knowledge about how a bitmapped matrix display and digital imagery actually works (though from your postings I do not hold out much hope )
Consider a single pixels data in a single frame of video after it has been decoded by the destination box mpeg decoder.
You have a minimum of 24 bits of digital data (3 bytes), each 8 bits represents the colour of each of the sub pixels Red Green and Blue of the full pixel.
8 bits in binary represents the numbers 0-255 in decimal. Hence there are 256 variations in red green and blue for each full pixel. Each pixel can only display a single colour, The variations possible are 256 x 256 x 256, which equates to over 64 million individual variations in colour (from pure back to pure white). Simple examples Red 255 Green 0 Blue 0 gives pure RED, Red 255, Green 255 and Blue 255 gives pure white. If you do not get this then you don't have the basic physics of a O level physics student, nor the understanding of physics discovered a long time ago about the nature of light.
Each pixel on a LCD/Plasma display is presented with 3 - 8 bit binary numbers for a single frame (1/25 second for 1080i).
To simplify things lets consider the data a square matrix of just 4 pixels might be presented with during a single frame.
Here's the purely arbitrary data presented by row and column order.
Top Row
Red 240, Green 0, Blue 127.
Red 0, Green 50, Blue 60
Bottom Row
Red = 60
Green = 60
Blue = 60
(Actually a grey scale monochrome image)
Red =256
Green = 0
Blue = 0
(A pure Red pixel).
Now lets consider how to display this rudimentary fragment of a video frame in a 4 by 4 matrix of pixels (stretching it 2 times in both axes by your very weird ideas).
Interested to see what the 4 rows in each row for each column might contain, if stretched (I have no idea what this means, in a digital environment it has no meaning.
Digital has an absolute and precisely defined set of data for the specific time period of the data. It represents a single point in time, You cannot as you put it stretch it, you can guess what the intermediate values might be based on a second data reading taken a bit later (sample rate).
Only analogue provides a continuous value for any quantity, the resolution being given by the carrier frequency used to transmit the data.
Basically until you at least demonstrate at least a schoolboy level of physics as it relates to digital imaging. It's clear you have no idea how to manipulate digital images even for a simple digital still camera (Which in themselves are compressed in a lossy format (normally .jpeg), let alone the complexity of compressing multiple frames in a video.
Here's a simple question for you, until you can answer it don't bother to post such rubbish. Rather take a O Level Physics course.
How much data will it take to transmit (Bytes or Bits) a single frame of 24 bit colour video with every pixel exactly represented ?
At 25 fps what bitrate would a broadcaster need to use in Mbps (Megabits/second) ?
What roughly would a 1920 x 1080 broadcast recorded with no compression require in the way of storage in GB (Gigabytes) on a PVR hard disk ?
I give up you clearly do not understand a word of what you are posting. The posting is entirely on topic, because you do not have the education to understand the relevance is not my fault.
This is now getting to the stage where your posts are clearly at the best ignorant and at the worst best discebided as a troll.
Ignore this poster until at least he indicates at least some basic ideas as to what he is on about.
Answering the pretty basic questions I posted would be a good start.
You are saying the original comment is describing stretching not upscaling, simply because upscaling involves more than just merely fitting the screen.
The complexity of the upscaling is not in question, its sole purpose is to make images fit the screen accordingly.
You are saying the original comment is describing stretching not upscaling, simply because upscaling involves more than just merely fitting the screen.
The complexity of the upscaling is not in question, its sole purpose is to make images fit the screen accordingly.
No I'm not as it wasn't the OP. When some states that upscaling is "merely" used to fill the screen, this is wrong in definition of what the word merely means. Merely fitting the screen would be stretching the image. Upscaling is totally different to a stretched image, that's why I said it.
I give up you clearly do not understand a word of what you are posting. The posting is entirely on topic, because you do not have the education to understand the relevance is not my fault.
This is now getting to the stage where your posts are clearly at the best ignorant and at the worst best discebided as a troll.
Ignore this poster until at least he indicates at least some basic ideas as to what he is on about.
Answering the pretty basic questions I posted would be a good start.
Getting back on topic would be an ever better start.
Merely fitting the screen would be stretching the image. Upscaling is totally different to a stretched image, that's why I said it.
What exactly is this "stretched image" of which you speak? It's not a technical term that I'm aware of, all it does is describe in layman's terms a change in aspect ratio from what was intended by the broadcaster.
There is 1) scaling and 2) aspect ratio change. Both involve re-mapping the input to fill different areas of the screen from what was broadcast. This would if left alone potentially leave gaps between 'lit' screen pixels, or create jaggies, but these are of course filled or smoothed by interpolation, in the case both of 1) and 2). They probably use the same scaling chips in all cases (upscaling, downscaling, aspect ratio change).
Your "stretching" example (aspect ratio change) is just an example of scaling on one axis instead of two. But aspect ratio change can be done on either axis; vertical, horizontal or indeed both and the end result won't necessarily look stretched. It's all done by scaling - upscaling or downscaling on one or both axes, usually using the same scaling chips. You are creating a distinction between "stretching" and scaling (upscaling or downscaling) that in reality doesn't exist: upscaling is not "totally different to a stretched image" as both use the same technology.
What exactly is this "stretched image" of which you speak? It's not a technical term that I'm aware of, all it does is describe in layman's terms a change in aspect ratio from what was intended by the broadcaster.
There is 1) scaling and 2) aspect ratio change. Both involve re-mapping the input to fill different areas of the screen from what was broadcast. This would if left alone potentially leave gaps between 'lit' screen pixels, or create jaggies, but these are of course filled or smoothed by interpolation, in the case both of 1) and 2). They probably use the same scaling chips in all cases (upscaling, downscaling, aspect ratio change).
Your "stretching" example (aspect ratio change) is just an example of scaling on one axis instead of two. But aspect ratio change can be done on either axis; vertical, horizontal or indeed both and the end result won't necessarily look stretched. It's all done by scaling - upscaling or downscaling on one or both axes, usually using the same scaling chips. You are creating a distinction between "stretching" and scaling (upscaling or downscaling) that in reality doesn't exist: upscaling is not "totally different to a stretched image" as both use the same technology.
A stretched image you describe known as "fattyvision" is totally different to upscaling. The point was with the word "merely" to fill the screen. That is a job of a stretched image to merely fill the screen, hence it will look distorted to how it should. That's not upscaling. Upscaling wouldn't distort the picture to "fattyvision". Yes, upscaling fills the screen. But it doesn't just merely fill the screen as stretching an image does. It uses technology as to not distort to "fattyvision". My main gripe was with the word merely.
A stretched image you describe known as "fattyvision" is totally different to upscaling. The point was with the word "merely" to fill the screen. That is a job of a stretched image to merely fill the screen, hence it will look distorted to how it should. That's not upscaling. Upscaling wouldn't distort the picture to "fattyvision". Yes, upscaling fills the screen. But it doesn't just merely fill the screen as stretching an image does. It uses technology as to not distort to "fattyvision". My main gripe was with the word merely.
You aren't related to Stanley Unwin are you ? Never read such a load of rubbish. Did you actually look up the relationship between Pixel Aspect Ratio and Display Aspect Ratio ?
The broadcast video data does not have a shape, it's just component data (YprPb) for the required number of pixels compressed using H264/AVC for HD content or mpeg2 for SD content. . The correct aspect ratio to display the data is specified by the Pixel Aspect Ratio specified in the data stream.
A 4:3 and a 16:9 DVD both have the same number of Pixels. 720 x 576 for a PAL version and 720 x 480 for a NTSC version. In both cases mpeg2 is used for the compression codec.
You aren't related to Stanley Unwin are you ? Never read such a load of rubbish. Did you actually look up the relationship between Pixel Aspect Ratio and Display Aspect Ratio ?
The broadcast video data does not have a shape, it's just component data (YprPb) for the required number of pixels compressed using H264/AVC for HD content or mpeg2 for SD content. . The correct aspect ratio to display the data is specified by the Pixel Aspect Ratio specified in the data stream.
A 4:3 and a 16:9 DVD both have the same number of Pixels. 720 x 576 for a PAL version and 720 x 480 for a NTSC version. In both cases mpeg2 is used for the compression codec.
Rubbish? Everything I've written there is 100% correct.
Rubbish? Everything I've written there is 100% correct.
I have been creating DVD's and latterly Blu-rays for years. You really are either stupid or more likely a Troll. With a user name like yours suspect the latter.
I'm sure many of you have heard of Cleartype.
Well with 4K being exactly double 1080p in both horizontal and vertical then obviously you can use Cleartype principals to make a 1080p picture look better on 4K.
And it seems the BBC is showing a few world cup games in 4K at the BBC centres.
(I hope my comments sounds like Plain English, much of the thread fails me.)
I'm sure many of you have heard of Cleartype.
Well with 4K being exactly double 1080p in both horizontal and vertical then obviously you can use Cleartype principals to make a 1080p picture look better on 4K.
And it seems the BBC is showing a few world cup games in 4K at the BBC centres.
(I hope my comments sounds like Plain English, much of the thread fails me.)
Got a bit lost in all that to be honest, but, they seemed to be having fun. Reality is that HD content upscaled on a 4K TV does look amazing, whatever the theory says..;-)
And probably caused [in the case of upscaling] by a bigger screen, higher bitrates, better compression algorithms and better internal electronics, given the price differential.
It is in my experience, feeding a 4K set and a normal HD set from the same HD source produces pretty much identical pictures - despite the huge difference in price. But that's the same with anything, scalers don't increase resolution, merely make the picture fit the screen whilst trying to produce as few artefacts as possible.
BD players generally have good scalers though, usually producing a slightly better picture than a DVD player on the same disk - and if you have a poor TV, then possibly quite a worthwhile improvement?.
I'm sure many of you have heard of Cleartype.
Well with 4K being exactly double 1080p in both horizontal and vertical then obviously you can use Cleartype principals to make a 1080p picture look better on 4K.
I don't see how ClearType -- a technology designed to smooth out the appearance of aliasing in a low-resolution screen -- can be compared with direct interpolation of an n*n image to 2n*2n (the latter, at its most basic, is a simple process of creating an 'invented' pixel by means of averaging its neighbours).
Anti-aliasing would be more useful when displaying a 4K image on a 1080p screen, surely?
They are both general terms for smoothing the between pixels.
Cleartype really being just a trademarked description for it.
Though anti-aliasing is far more a general term, having many a meaning.
But there is no "smoothing the between pixels" in a TrueType font, which is by definition a vector-based object. That's the point -- ClearType (which works on TrueType fonts) is a mechanism of compensating for a lack of pixels -- it would have no purpose in a scenario where there are a surplus of pixels, which is what we're talking about here.
Comments
Surely the SD 4x3 image has already been upscaled - if it wasn't upscaled it would be a little box in the middle of the screen...
Those who stretch/zoom images to full screen generally do it to get rid of the black bars, not because the image hasn't been upscaled.
Please go back and reread the argument.
I am going to try and explain how wrong you are, I hope you at least you have some basic knowledge about how a bitmapped matrix display and digital imagery actually works (though from your postings I do not hold out much hope
Consider a single pixels data in a single frame of video after it has been decoded by the destination box mpeg decoder.
You have a minimum of 24 bits of digital data (3 bytes), each 8 bits represents the colour of each of the sub pixels Red Green and Blue of the full pixel.
8 bits in binary represents the numbers 0-255 in decimal. Hence there are 256 variations in red green and blue for each full pixel. Each pixel can only display a single colour, The variations possible are 256 x 256 x 256, which equates to over 64 million individual variations in colour (from pure back to pure white). Simple examples Red 255 Green 0 Blue 0 gives pure RED, Red 255, Green 255 and Blue 255 gives pure white. If you do not get this then you don't have the basic physics of a O level physics student, nor the understanding of physics discovered a long time ago about the nature of light.
Each pixel on a LCD/Plasma display is presented with 3 - 8 bit binary numbers for a single frame (1/25 second for 1080i).
To simplify things lets consider the data a square matrix of just 4 pixels might be presented with during a single frame.
Here's the purely arbitrary data presented by row and column order.
Top Row
Red 240, Green 0, Blue 127.
Red 0, Green 50, Blue 60
Bottom Row
Red = 60
Green = 60
Blue = 60
(Actually a grey scale monochrome image)
Red =256
Green = 0
Blue = 0
(A pure Red pixel).
Now lets consider how to display this rudimentary fragment of a video frame in a 4 by 4 matrix of pixels (stretching it 2 times in both axes by your very weird ideas).
Interested to see what the 4 rows in each row for each column might contain, if stretched (I have no idea what this means, in a digital environment it has no meaning.
Digital has an absolute and precisely defined set of data for the specific time period of the data. It represents a single point in time, You cannot as you put it stretch it, you can guess what the intermediate values might be based on a second data reading taken a bit later (sample rate).
Only analogue provides a continuous value for any quantity, the resolution being given by the carrier frequency used to transmit the data.
Basically until you at least demonstrate at least a schoolboy level of physics as it relates to digital imaging. It's clear you have no idea how to manipulate digital images even for a simple digital still camera (Which in themselves are compressed in a lossy format (normally .jpeg), let alone the complexity of compressing multiple frames in a video.
Here's a simple quiz for you, until you can answer it don't bother to post such rubbish.
Rather take a O Level Physics course.
How much data will it take to transmit (Bytes or Bits) a single frame of 24 bit colour video with every pixel exactly represented ?
At 25 fps what bitrate would a broadcaster need to use in Mbps (Megabits/second) ?
What roughly would a 1920 x 1080 broadcast recorded with no compression require in the way of storage in GB (Gigabytes) on a PVR hard disk ?
And meanwhile, back on topic....
I give up you clearly do not understand a word of what you are posting. The posting is entirely on topic, because you do not have the education to understand the relevance is not my fault.
This is now getting to the stage where your posts are clearly at the best ignorant and at the worst best discebided as a troll.
Ignore this poster until at least he indicates at least some basic ideas as to what he is on about.
Answering the pretty basic questions I posted would be a good start.
No need to....
You are saying the original comment is describing stretching not upscaling, simply because upscaling involves more than just merely fitting the screen.
The complexity of the upscaling is not in question, its sole purpose is to make images fit the screen accordingly.
No I'm not as it wasn't the OP. When some states that upscaling is "merely" used to fill the screen, this is wrong in definition of what the word merely means. Merely fitting the screen would be stretching the image. Upscaling is totally different to a stretched image, that's why I said it.
Getting back on topic would be an ever better start.
What exactly is this "stretched image" of which you speak? It's not a technical term that I'm aware of, all it does is describe in layman's terms a change in aspect ratio from what was intended by the broadcaster.
There is 1) scaling and 2) aspect ratio change. Both involve re-mapping the input to fill different areas of the screen from what was broadcast. This would if left alone potentially leave gaps between 'lit' screen pixels, or create jaggies, but these are of course filled or smoothed by interpolation, in the case both of 1) and 2). They probably use the same scaling chips in all cases (upscaling, downscaling, aspect ratio change).
Your "stretching" example (aspect ratio change) is just an example of scaling on one axis instead of two. But aspect ratio change can be done on either axis; vertical, horizontal or indeed both and the end result won't necessarily look stretched. It's all done by scaling - upscaling or downscaling on one or both axes, usually using the same scaling chips. You are creating a distinction between "stretching" and scaling (upscaling or downscaling) that in reality doesn't exist: upscaling is not "totally different to a stretched image" as both use the same technology.
A stretched image you describe known as "fattyvision" is totally different to upscaling. The point was with the word "merely" to fill the screen. That is a job of a stretched image to merely fill the screen, hence it will look distorted to how it should. That's not upscaling. Upscaling wouldn't distort the picture to "fattyvision". Yes, upscaling fills the screen. But it doesn't just merely fill the screen as stretching an image does. It uses technology as to not distort to "fattyvision". My main gripe was with the word merely.
You aren't related to Stanley Unwin are you ? Never read such a load of rubbish. Did you actually look up the relationship between Pixel Aspect Ratio and Display Aspect Ratio ?
The broadcast video data does not have a shape, it's just component data (YprPb) for the required number of pixels compressed using H264/AVC for HD content or mpeg2 for SD content. . The correct aspect ratio to display the data is specified by the Pixel Aspect Ratio specified in the data stream.
A 4:3 and a 16:9 DVD both have the same number of Pixels. 720 x 576 for a PAL version and 720 x 480 for a NTSC version. In both cases mpeg2 is used for the compression codec.
Rubbish? Everything I've written there is 100% correct.
I have been creating DVD's and latterly Blu-rays for years. You really are either stupid or more likely a Troll. With a user name like yours suspect the latter.
Random links
http://www.animemusicvideos.org/guides/avtech3/theory-videoaspectratios.html
http://en.wikipedia.org/wiki/Pixel_aspect_ratio
http://helpx.adobe.com/premiere-pro/using/aspect-ratios.html
Which part of my post do you reckon is incorrect.
Well with 4K being exactly double 1080p in both horizontal and vertical then obviously you can use Cleartype principals to make a 1080p picture look better on 4K.
And it seems the BBC is showing a few world cup games in 4K at the BBC centres.
(I hope my comments sounds like Plain English, much of the thread fails me.)
Got a bit lost in all that to be honest, but, they seemed to be having fun. Reality is that HD content upscaled on a 4K TV does look amazing, whatever the theory says..;-)
I'd imagine any improvement is marginal.
And probably caused [in the case of upscaling] by a bigger screen, higher bitrates, better compression algorithms and better internal electronics, given the price differential.
It is in my experience, feeding a 4K set and a normal HD set from the same HD source produces pretty much identical pictures - despite the huge difference in price. But that's the same with anything, scalers don't increase resolution, merely make the picture fit the screen whilst trying to produce as few artefacts as possible.
BD players generally have good scalers though, usually producing a slightly better picture than a DVD player on the same disk - and if you have a poor TV, then possibly quite a worthwhile improvement?.
I don't see how ClearType -- a technology designed to smooth out the appearance of aliasing in a low-resolution screen -- can be compared with direct interpolation of an n*n image to 2n*2n (the latter, at its most basic, is a simple process of creating an 'invented' pixel by means of averaging its neighbours).
Anti-aliasing would be more useful when displaying a 4K image on a 1080p screen, surely?
Cleartype really being just a trademarked description for it.
Though anti-aliasing is far more a general term, having many a meaning.
But there is no "smoothing the between pixels" in a TrueType font, which is by definition a vector-based object. That's the point -- ClearType (which works on TrueType fonts) is a mechanism of compensating for a lack of pixels -- it would have no purpose in a scenario where there are a surplus of pixels, which is what we're talking about here.