megapixels
- Thread starter -Paul-
- Start date
- Messages
- 9,277
- Edit My Images
- No
Only if you have it set to maximum resolution.
Steve.
Steve.
valb
Suspended / Banned
- Messages
- 1,589
- Name
- val
- Edit My Images
- No
IMO It is not essential to use all available pixels for every photo. A lot depends on the intended use of the photo. If, for example, it will be put on the web and you use all pixels then you will have to downscale the photo significantly. If printing then you want the full pixel count.
That said I always use all available pixels. You can always downscale if you have too many pixels. You cannot upscale if you have too few.
That said I always use all available pixels. You can always downscale if you have too many pixels. You cannot upscale if you have too few.
Absolutely, as above, only if you have your camera configured to use maximum resolution.
I tend to use a lower resolution when I am carrying out experimental shots or just working on a particular technique, or learning to use a new camera.
If I am looking to do more with the picture then usual shoot with full resolution.
I tend to use a lower resolution when I am carrying out experimental shots or just working on a particular technique, or learning to use a new camera.
If I am looking to do more with the picture then usual shoot with full resolution.
Instead of having a picture 2100 x 1400 pixels you can, using software, reduce it to 700 x 467. It will still show the full picture but with 1/3 the width and height. Good for reducing file size but you will only have 1/9th of the detail so no good for zooming in or printing large.
Once this detail is discarded you can't get it back by enlarging. You can increase the resolution again, but each pixel of the reduced image will then become 9 pixles of the same colour* rather than 9 different colours in the original full res shot.
(*Not exactly, they will be blured across adjacent pixels.)
I'd leave it in full resolution, unless there was a specific reason i wanted a low res shot. (e.g. Uploading a shot from my camera phone.)
It's too easy to forget to change it back again afterwards causing a problem when you get a shot you want in higher res.
Once this detail is discarded you can't get it back by enlarging. You can increase the resolution again, but each pixel of the reduced image will then become 9 pixles of the same colour* rather than 9 different colours in the original full res shot.
(*Not exactly, they will be blured across adjacent pixels.)
I'd leave it in full resolution, unless there was a specific reason i wanted a low res shot. (e.g. Uploading a shot from my camera phone.)
It's too easy to forget to change it back again afterwards causing a problem when you get a shot you want in higher res.
Last edited:
Teflon-Mike
Suspended / Banned
- Messages
- 1,076
- Name
- Mike
- Edit My Images
- Yes
Paul... are you confusing Mega-PIXELS with Mega-BYTES by any chance?
Sorry for going into very long winded explanation, but the possible confusion between Maga-Pixels and Maga-bytes, and the difference between IMAGE size and FILE size was causing very heated family arguments two-weeks ago, when my mum, bless her, was trying to get a photo-graph printed on a mug for a cousin's wedding :bonk:
So... if this is any help at all!
Pixels is the number of squares coloured in to make up your image, the 'dots'.
My camera is a 24Mp camera, and at Hi-Res 'setting' it makes pictures 6000 pixels wide, by 4000 pixels tall. 6000 x 4000 = 24,000,000 or 24 million pixels, or Mp. Though I can set lower pixel sixes in the camera's menu.
However... the FILE size, the amount of memory on the computer the picyure uses in Mega-Bytes or Mb, is variable, for any given pixel count or IMAGE size..
Technically, to define the colour and brightness of each pixel, you need to define three values; one for Red, one for Blue and one for Green light falling on the pixel, between 0 and 100%.
We then get into a thing called 'colour depth', and how many 'steps' the camera divided the range of brightness into.
Its a computer, it works on O's and 1's, so it doesn't count like people, 0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, etc, it counts, 0000, (0) 0001 (1), 0010 (2), 0011 (3), 0100 (4), 0101 (5), 0110 (6), 0111 (7), 1000 (8), 1001 (9), 1010 (10), 1011 (11), 1100 (12), 1101 (13), 1110 (14), 1111 (15)... ERROR.... because being a computer it needs to know how many numerals to expect in every number before you begin....... and with four 'bits' or potential spaces to load with a 0 or a 1... you can record up to 16 values.... 0-15.
So one pixel, wee need three values, and if we give each value just four 'bit' colour depth, that means 12 bits per pixel..... if we want 16 bit colour depth, we need 56 bits per pixel.
And we have on my camera... 24 million pixels... so we probably need, 56x24-million bits to store the data that define it.... that's 1,344,000,000 'bits' of data...
Actually probably a few more, because we probably need some sort of redundant code telling the computer whee the pixels belong, like whether the picture is 6000 pixels wide by 4000 tall, or 12,000 wide and 2,000 tall and possibly a bit more besides to record the 'embedded' data, like what camera it was taken on, what lens settings were, time and date, and 'stuff' like that.
So, 24 Mega-PIXEL image.... ought to make a data-file, perhaps 1,500 Mega-bytes.
It doesn't... well, it CAN and if you shoot in RAW or scan images as TIFF, they can produce file sizes in that sort of order; but most cameras use the JPEG encoding system, which handily contains a bit of 'data-compression'... which without getting over complicated, the computer uses a bit of short-hand, to save a bit of space.
And depending on how much compression is applied, it will pull a 100 Mega-BYTE image file down to something like 10Mb at low compression, perhaps 1 Mb at high compression.
BUT depends on what's in the picture.
If you have a picture that's got large blocks of colour,say a big blue sky with no clouds in it, then the computer can shorten a lot of big blocks of data down, using the computer equivalent of 'ditto' marks; while if you have lots of detail in the picture, say it's a shot of a wood, every time there's a big change in colour between two adjacent pixels, it has to record the full data-set for it.
I have one now; Tiff Scan, it has a pixel count of 5184x3360, that's 17.4Mega-Pixels. File-Size in relatively uncompressed TIFF format is 238 Mega-bytes. Remember, these FILE sizes are all for an image of the same PIXEL size.
- Saved in JPEG, with low compression is only 8 Mega BYTES.
- Saved at 'Normal' Compression, it's 1 Mega-byte.
- Saved at highest compression, it's to 0.3 Mega-Bytes
OK... now lets make the whole picture a single block of mid-tone grey. Still 24 Mega-Pixels, but ALL the same colour.
- Saved in uncompressed TIFF file is still 238 Mega-Bytes big
(remember its the same number of Pixels, and the computer is still saving the full long-hand colour brightness values for each pixel)
- Saved in JPEG, with low compression it is now, 1.1Mega-bytes, instead of 8
- Saved at 'Normal' Compression, it is now, 0.3 Mega-bytes, instead of 1.
- Saved at highest Compression, it is now, 0.2 Mega-bytes, instead of 0.3.
So... the PIXEL count is the same no matter what. That is fixed by the camera. If you have a 12 Mega-Pixel camera, it will probably be taking pictures with a pixel count something like 4000 x 3000 give or take.
LOWER pixel count 'resolution' may be possible set in the menu. to you may have the option to set the camera to, 'Large' 4000x3000, or 'Normal' maybe 2000 x 1500 or 'small' perhaps 1500x1000 or numbers in that order.
There may also be 'quality' settings as well, perhaps 'Fine', 'Normal' and 'low', which will probably set the colour depth or number of bits allowed for each pixel.
These settings will be suggested, in the manual, either for keeping FILE sizes small, if your camera only has internal memory or a small memory card, and you want to maximise the number of pictures you can take before having to clear down the camera or card, or for reducing 'cycle time', if you need to take a succession of photo's fairly quickly, and the cameras processing chip would get bogged down processing so much data in each picture.
BUT within those user-set limits, the pixel count of each photo, number of squares that make it up, will be constant from photo-to photo; and what the camera can achieve, or the lower value you set.
However the FILE SIZE, the number of bytes, kilo-bytes or mega-bytes, the data-file used to generate store that picture, will vary; depending on the PIXEL resolution set, the colour-dept or picture quality set on the camera, THEN for those fixed settings, the file format used, the amount of compression used within that format, and the complexity of the actual picture itself.
Right... I shall now consider the anomaly of Digital Zoom!
12 Mega-Pixel camera, Image sizes of 4000x3000. We might presume, that the sensor 'array' in the camera that looks at the scene we are taking a picture of, has one sensor 'cell' for every pixel in the image it makes.... (and we'd probably be wrong, but no matter!)
So, we have a lens in front of the sensor array, and it focuses the scene on the sensor array, and each cell gathers light, and is sampled by the camera's processor, and records a value for that pixel.
Now... we want to zoom in; make the subject bigger in our view-finder; so we increase the focal length of the lens, to increase magnification... and bit of mechanics, worked by screws and levers and cams and stuff, the lens is physically moved to 'zoom'.
Sensor array is still gathering light falling on it, and the processor sampling the levels to record the image file. And using every single cell in the array to do it.
But we have run out of 'zoom' on the mechanics of the lens... BUT we want to zoom a bit more... NOW what do you do?
Well, if you look at JUST the centre of the array, and sample the cells in the middle... you frame more tightly on your subject. THIS looks like more 'zoom'.
You have a frame, 4000x3000 pixels, right, so if, I don't know, a car that fills less than 1/4 of the frame, you could JUST record the data from the cells in the center of the frame; maybe 1000x750, ignoring rows of cells above and below and to wither side.
The subject is no bigger, and you haven't magnified the image any, you have merely shrunk the frame around it, so it looks bigger in it.
THIS is Digital Zoom...... BUT.... what your camera does with that data is any-one's guess!
It could store it at the sampled image pixel size, 1000x750... ie smaller than the set pixel count resolution.
But more likely, to sharpen it up a bit, and stop it being too pixilated, it will process the sampled image, and do a thing called 'interpolation' and looking at each pixel in turn, try and quarter it, and record four pixels for each one actually sampled, comparing each pixel to it's neighbor to decide whether the 1/4 pixels ought to be more or less the same of something between them.
It can now, the processor can fill in the 'guessed; data into the spaces from the cells it didn't sample, and create a file the same pixel count, and same file size as one that has been taken using only optical zoom.
I have a cheap 5 Mpix compact that does this, and it only offers two 'digital' zoom ratio's, 2x magnification (sampling half the cells) or 4x magnification, (sampling 1/4 the cells)
BUT if you have a camera that offers more graduated 'digital' zoom... it's likely that it wont be able to easily halve or quarter pixels for interpolation, and the processor might have to add a few pixels on the edge or drop a couple out to do its maths and make the zoomed image fit the pixel frame it is programmed to store... OR it might add or delete those interpolated rows or columns, possibly changing the image proportions slightly as well as the pixel count of the stored image file. though, this is, I understand, very uncommon.
Normally, Digital Zoom will create an image of the same pixel dimensions as the quality settings you set on the camera, BUT it is not IMPOSSIBLE that due to interpolation it might vary SLIGHTLY from pictures taken on the optical zoom, sampling the entire sensor without any interpolation.
Sorry for going into very long winded explanation, but the possible confusion between Maga-Pixels and Maga-bytes, and the difference between IMAGE size and FILE size was causing very heated family arguments two-weeks ago, when my mum, bless her, was trying to get a photo-graph printed on a mug for a cousin's wedding :bonk:
So... if this is any help at all!
Pixels is the number of squares coloured in to make up your image, the 'dots'.
My camera is a 24Mp camera, and at Hi-Res 'setting' it makes pictures 6000 pixels wide, by 4000 pixels tall. 6000 x 4000 = 24,000,000 or 24 million pixels, or Mp. Though I can set lower pixel sixes in the camera's menu.
However... the FILE size, the amount of memory on the computer the picyure uses in Mega-Bytes or Mb, is variable, for any given pixel count or IMAGE size..
Technically, to define the colour and brightness of each pixel, you need to define three values; one for Red, one for Blue and one for Green light falling on the pixel, between 0 and 100%.
We then get into a thing called 'colour depth', and how many 'steps' the camera divided the range of brightness into.
Its a computer, it works on O's and 1's, so it doesn't count like people, 0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, etc, it counts, 0000, (0) 0001 (1), 0010 (2), 0011 (3), 0100 (4), 0101 (5), 0110 (6), 0111 (7), 1000 (8), 1001 (9), 1010 (10), 1011 (11), 1100 (12), 1101 (13), 1110 (14), 1111 (15)... ERROR.... because being a computer it needs to know how many numerals to expect in every number before you begin....... and with four 'bits' or potential spaces to load with a 0 or a 1... you can record up to 16 values.... 0-15.
So one pixel, wee need three values, and if we give each value just four 'bit' colour depth, that means 12 bits per pixel..... if we want 16 bit colour depth, we need 56 bits per pixel.
And we have on my camera... 24 million pixels... so we probably need, 56x24-million bits to store the data that define it.... that's 1,344,000,000 'bits' of data...
Actually probably a few more, because we probably need some sort of redundant code telling the computer whee the pixels belong, like whether the picture is 6000 pixels wide by 4000 tall, or 12,000 wide and 2,000 tall and possibly a bit more besides to record the 'embedded' data, like what camera it was taken on, what lens settings were, time and date, and 'stuff' like that.
So, 24 Mega-PIXEL image.... ought to make a data-file, perhaps 1,500 Mega-bytes.
It doesn't... well, it CAN and if you shoot in RAW or scan images as TIFF, they can produce file sizes in that sort of order; but most cameras use the JPEG encoding system, which handily contains a bit of 'data-compression'... which without getting over complicated, the computer uses a bit of short-hand, to save a bit of space.
And depending on how much compression is applied, it will pull a 100 Mega-BYTE image file down to something like 10Mb at low compression, perhaps 1 Mb at high compression.
BUT depends on what's in the picture.
If you have a picture that's got large blocks of colour,say a big blue sky with no clouds in it, then the computer can shorten a lot of big blocks of data down, using the computer equivalent of 'ditto' marks; while if you have lots of detail in the picture, say it's a shot of a wood, every time there's a big change in colour between two adjacent pixels, it has to record the full data-set for it.
I have one now; Tiff Scan, it has a pixel count of 5184x3360, that's 17.4Mega-Pixels. File-Size in relatively uncompressed TIFF format is 238 Mega-bytes. Remember, these FILE sizes are all for an image of the same PIXEL size.
- Saved in JPEG, with low compression is only 8 Mega BYTES.
- Saved at 'Normal' Compression, it's 1 Mega-byte.
- Saved at highest compression, it's to 0.3 Mega-Bytes
OK... now lets make the whole picture a single block of mid-tone grey. Still 24 Mega-Pixels, but ALL the same colour.
- Saved in uncompressed TIFF file is still 238 Mega-Bytes big
(remember its the same number of Pixels, and the computer is still saving the full long-hand colour brightness values for each pixel)
- Saved in JPEG, with low compression it is now, 1.1Mega-bytes, instead of 8
- Saved at 'Normal' Compression, it is now, 0.3 Mega-bytes, instead of 1.
- Saved at highest Compression, it is now, 0.2 Mega-bytes, instead of 0.3.
So... the PIXEL count is the same no matter what. That is fixed by the camera. If you have a 12 Mega-Pixel camera, it will probably be taking pictures with a pixel count something like 4000 x 3000 give or take.
LOWER pixel count 'resolution' may be possible set in the menu. to you may have the option to set the camera to, 'Large' 4000x3000, or 'Normal' maybe 2000 x 1500 or 'small' perhaps 1500x1000 or numbers in that order.
There may also be 'quality' settings as well, perhaps 'Fine', 'Normal' and 'low', which will probably set the colour depth or number of bits allowed for each pixel.
These settings will be suggested, in the manual, either for keeping FILE sizes small, if your camera only has internal memory or a small memory card, and you want to maximise the number of pictures you can take before having to clear down the camera or card, or for reducing 'cycle time', if you need to take a succession of photo's fairly quickly, and the cameras processing chip would get bogged down processing so much data in each picture.
BUT within those user-set limits, the pixel count of each photo, number of squares that make it up, will be constant from photo-to photo; and what the camera can achieve, or the lower value you set.
However the FILE SIZE, the number of bytes, kilo-bytes or mega-bytes, the data-file used to generate store that picture, will vary; depending on the PIXEL resolution set, the colour-dept or picture quality set on the camera, THEN for those fixed settings, the file format used, the amount of compression used within that format, and the complexity of the actual picture itself.
Right... I shall now consider the anomaly of Digital Zoom!
12 Mega-Pixel camera, Image sizes of 4000x3000. We might presume, that the sensor 'array' in the camera that looks at the scene we are taking a picture of, has one sensor 'cell' for every pixel in the image it makes.... (and we'd probably be wrong, but no matter!)
So, we have a lens in front of the sensor array, and it focuses the scene on the sensor array, and each cell gathers light, and is sampled by the camera's processor, and records a value for that pixel.
Now... we want to zoom in; make the subject bigger in our view-finder; so we increase the focal length of the lens, to increase magnification... and bit of mechanics, worked by screws and levers and cams and stuff, the lens is physically moved to 'zoom'.
Sensor array is still gathering light falling on it, and the processor sampling the levels to record the image file. And using every single cell in the array to do it.
But we have run out of 'zoom' on the mechanics of the lens... BUT we want to zoom a bit more... NOW what do you do?
Well, if you look at JUST the centre of the array, and sample the cells in the middle... you frame more tightly on your subject. THIS looks like more 'zoom'.
You have a frame, 4000x3000 pixels, right, so if, I don't know, a car that fills less than 1/4 of the frame, you could JUST record the data from the cells in the center of the frame; maybe 1000x750, ignoring rows of cells above and below and to wither side.
The subject is no bigger, and you haven't magnified the image any, you have merely shrunk the frame around it, so it looks bigger in it.
THIS is Digital Zoom...... BUT.... what your camera does with that data is any-one's guess!
It could store it at the sampled image pixel size, 1000x750... ie smaller than the set pixel count resolution.
But more likely, to sharpen it up a bit, and stop it being too pixilated, it will process the sampled image, and do a thing called 'interpolation' and looking at each pixel in turn, try and quarter it, and record four pixels for each one actually sampled, comparing each pixel to it's neighbor to decide whether the 1/4 pixels ought to be more or less the same of something between them.
It can now, the processor can fill in the 'guessed; data into the spaces from the cells it didn't sample, and create a file the same pixel count, and same file size as one that has been taken using only optical zoom.
I have a cheap 5 Mpix compact that does this, and it only offers two 'digital' zoom ratio's, 2x magnification (sampling half the cells) or 4x magnification, (sampling 1/4 the cells)
BUT if you have a camera that offers more graduated 'digital' zoom... it's likely that it wont be able to easily halve or quarter pixels for interpolation, and the processor might have to add a few pixels on the edge or drop a couple out to do its maths and make the zoomed image fit the pixel frame it is programmed to store... OR it might add or delete those interpolated rows or columns, possibly changing the image proportions slightly as well as the pixel count of the stored image file. though, this is, I understand, very uncommon.
Normally, Digital Zoom will create an image of the same pixel dimensions as the quality settings you set on the camera, BUT it is not IMPOSSIBLE that due to interpolation it might vary SLIGHTLY from pictures taken on the optical zoom, sampling the entire sensor without any interpolation.