File sizes and cameras

And your camera doesn't produce 16 bits of data anyway. It's easy to meet the 50 Mb requirement if you don't care how many of those bytes actually contain data.
 
IanC_UK said:
Yes the 1DS is the only camera that will produce an image of that resolution with no manipulation, but thousands of people use software to upscale their images for photo libraries, no reason to not use that method instead of blowing 5k on a camera. If you need live view look at a 40D and just use software to increase the file size.
Click to expand...


Ids MKII produces 50mb tiffs (just) if they are 8bit uncropped.
 
JumboBeef said:
It's not so much as the thing falling (together with the remote head) than the wind. I have a complex chart of head-weights/wind speed/height: the winder it is, and the greater the head-weight, the lower the maximum height is. So, a 450D can go higher on a blustery day than a 1Ds can.

(and yes cost does come in to it!)
Click to expand...

I'd suggest the lens is going to weigh significantly more than the 450D if your using some high quality glass?
 
afaik alamy requires a jpeg that is over 48mb when uncompressed.

Open a raw file in photoshop, resize the image to around 140% (depending on original mps, 400d) Save as (not save for web) a jpeg at maximum. The saved file will be around 5mb, when you open this file in photoshop it will uncompress to around 50mb :)

I may be wrong but the company you are dealing with will be talking about uncompressed file sizes.


A 400d shoots at 3888x2592 which is around 30mb so would need upscaling to around 140% to reach 50mb uncompressed.

A 1ds mk3 shoots at 5616x3744 which is around 60mb. Saving the file as an 8bit jpeg will be around 6mb compressed, when opened back up it will still be a 60mb uncompressed image.
 
JumboBeef said:
Is the 57Mb TIFF really going to be better than the 18Mb RAW?
Click to expand...
As others have said, you can't really compare a RAW file and a TIFF. A RAW file is just an unprocessed dump of the sensor data. It's like an undeveloped negative. Whilst there is software to view RAW files, all such software has to decide how to "interpret" the data, including critical decisions such as what White Balance to apply. By comparison a TIFF is more like a finished print - it is what it is, and there is only one way to interpret it.

JumboBeef said:
What is the difference between 8 and 16 bit?
Click to expand...
8-bit means that the file uses 8 bits (1 byte) to store the colour information for each of the R/G/B channels for each pixel. That gives you 2^8 = 256 levels for each channel, and 256^3 = 16.8 million different colours. 16-bit means 2^16 = 65,536 levels for each channel and 65,536^3 = an unimaginably vast number of different colours.

It's believed that the human eye can distinguish about 5 million different colours, but that's not quite a subset of the 16.8 million available in 8-bit land; we're particularly good at distinguishing different shades of green. However, for most practical purposes, 8-bit colour is virtually indistinguishable from real life.

Digital cameras have progressed from recording 8 bits per channel in the old days up to 14 bits per channel for some of the newest high-end DSLRs (e.g. Canon 1D series). The reason they do this, even though 14 bits produces far more colours than we can see, is so that there's a bit of headroom to process the data in the camera or on a computer - e.g. recovering highlights - whilst not producing artefacts at the 8-bit level. Kind of like making sure they don't get any rounding errors. Some people say that you should do all your PP in 16-bit mode to maintain the highest possible quality and only convert down to 8-bit when you're finished. But I'm not aware of any evidence that it's possible to tell the difference in "blind" trials, except perhaps for images which have had utterly extreme amounts of PP done to them.

Any more questions?
 
WanderingMind said:
I may be wrong but the company you are dealing with will be talking about uncompressed file sizes.

...

A 1ds mk3 shoots at 5616x3744 which is around 60mb. Saving the file as an 8bit jpeg will be around 6mb compressed, when opened back up it will still be a 60mb uncompressed file.
Click to expand...
Ooh, this opens up another can of worms.

Everyone should know that JPEG compression is destructive. Compress an image and decompress it, and whilst it returns to its original size (e.g. from 6Mb compressed to 60Mb uncompressed) it will *not* return to its original quality.

When people who are commissioning images specify a 50Mb TIFF, what they really want/need is an image that has never been degraded by JPEG compression. The way to do that is to shoot RAW and always save as PSD or TIFF, never JPEG.

I feel a rant coming on.

I bet that many/most of the people asking for 50Mb files don't know why that's the specification, or what it means. And I bet most so-called "photographers" don't either.

They probably don't know and don't care that the specification was meant to restrict submissions to medium format, where you get 50Mb of real data from the sensor. They probably don't know and don't care that decompressing and upscaling an image from a 400D gives you 3Mb of data and 47Mb of junk. But the end product is a 50Mb TIFF, so that's OK, right?

This really is such basic, elementary stuff. I expect a competent butcher to know the differences between different cuts of meat and to be able to advise me which one will make a good casserole. I expect a competent tailor to know how cloth is made and to know which types will wear well or resist stretching. So why is it acceptable for photographers and other people in the imaging industry not to understand the basics of their profession?
 
I agree it won't return to it's original quality but for the likes of Alamy it works, can you imagine loads of people trying to upload 1000's of 50mb files everyday lol :) and for a lot of print purposes like small magazine ad's etc.. the difference is hardly noticable.
 
OK, looking at this another way...........

If you have a 400D/450D and you wish to work for a company who requests the files to be around the 50Mb mark (they did say 'or close'), how would you do it? I'm not going to send them a 4Mb file and tell them it is the same as a 50Mb one (even if it is). Right or wrong, 50Mb is their rule and if I want to work with them I need to do it their way.
 
JumboBeef said:
OK, looking at this another way...........

If you have a 400D/450D and you wish to work for a company who requests the files to be around the 50Mb mark (they did say 'or close'), how would you do it? I'm not going to send them a 4Mb file and tell them it is the same as a 50Mb one (even if it is). Right or wrong, 50Mb is their rule and if I want to work with them I need to do it their way.
Click to expand...

Shoot raw and do your normal processing. In photoshop image resize, alter the pixel dimensions to about 140% or more with resampling set to 'bicubic smoother' until it reads near enough 50m. Save as an uncompressed tiff :)

Iv'e not used genuine fractals so can't comment on it, but I believe it's better than photoshop for resampling?
 
Stuart, I feel your pain. You are probably right. I suspect this request comes from the "old" days when the image was digitised on a drum scanner. The agencies had to specify a spec so the 50Mb rule was formed. It's about the same size as a 4 x 5 trannie being scanned at 1000dpi or a 6 x 6 cm one at 2000 dpi. (roughly).

Also don't forget if you are using Photoshop's ACR you can specify the size of your final Tiff files. I can get a 72Mb from my 5D. You can also specify output size in Lightroom as well
 
Explanation of Digital Image File Sizes
v1.0, 2005.10.04

It can be difficult to determine the quality of a digital image based on its file size. Digital services like scanning are often priced by file size in megabytes. If you have a 30MB file, how many megapixels is that? What is the difference between compressed and uncompressed file sizes? How many megabytes do you need to make a large print? This page attempts to answer these questions.

Image file size formula:


width * height * color / 1,024,000 = Megabytes (MB)




Values for width and height must be pixels. For color, use the Bytes/pixel number from this table:

Color Depth Bytes/Pixel
8 bit Grayscale (each pixel = 0 to 255) 1
16 bit Grayscale (each pixel = 0 to 65536) 2
24 bit RGB (8 bits for each R,G,B) "8 bit color" 3
32 bit CMYK (8 bits for each C,Y,M,K) 4
48 bit RGB (16 bits for each R,G,B) "16 bit color" 6

Examples
Each calculated file size is the uncompressed size.



Nikon D70 camera, High Quality JPG, 8 bit color:
3000 * 2000 * 3 / 1024000 = 18 MB (17,578 KB) Nikon D70 camera, Raw file, 16 bit color
3000 * 2000 * 6 / 1024000 = 35 MB (35,156 KB) 35mm transparency film scan, 4000ppi, 16 bit color
5600 * 3700 * 6 / 1024000 = 121 MB (121,406 KB) Flatbed scan, 8" x 10" @ 300dpi, 16 bit color
(8 * 300) * (10 * 300) * 6 / 1024000 = 42 MB (42,187 KB) Flatbed scan, 8" x 10" @ 600dpi, 16 bit color
(8 * 600) * (10 * 600) * 6 / 1024000 = 169 MB (168,750 KB) Flatbed scan, 4" x 5" @ 1200dpi, 16 bit color
(4 * 1200) * (5 * 1200) * 6 / 1024000 = 169 MB (168,750 KB) Printed Poster, 24" x 36" x 300dpi, 8 bit color
(24 * 300) * (36 * 300) * 3 / 1024000 = 227 MB (227,812 KB)


Compression
When discussing the quality of digital files based on file size, comparisons should only be made based on uncompressed sizes. Compression algorithms will modify each image differently depending on the subject matter of the image. Therefore it is impossible to accurately compare the file size of two digital images once they have been compressed.

There are two types of file compression, "lossy" and "lossless". Lossy compression actually changes some of the original pixels and some details are lost. The most common type of lossy compression format is JPG. While the original JPG image out of a digital camera is fine, every time the file is saved again, detail is lost. If the same file is saved as a JPG a few times, significant quality is lost and cannot be recovered. Valuable originals should always be saved in a lossless format, like TIFF or RAW. TIFF files can be edited and saved any number of times without loss of detail because the compression does not alter any pixels. The trade off is that TIFF files do not compress as well as JPG.


Does this help or hinder
 
JumboBeef said:
OK, looking at this another way...........

If you have a 400D/450D and you wish to work for a company who requests the files to be around the 50Mb mark (they did say 'or close'), how would you do it? I'm not going to send them a 4Mb file and tell them it is the same as a 50Mb one (even if it is). Right or wrong, 50Mb is their rule and if I want to work with them I need to do it their way.
Click to expand...
Absolutely! I could not agree more. The client sets the rules and you have to comply with them. The fact that the clients are idiots and haven't got a scooby about the whole subject is totally irrelevant.
 
You must log in or register to reply here.
Back
Top