RAW file sizes for D50, D80, D200

Just been looking at some different RAW file sizes

D50 (6MP) approx 5MB

D80 (10MP) approx 8MB

D200 (10MP) approx 15MB

Can anyone explain why the sizes are so much bigger for similar megapixels, especially between the D80 and D200. Is there file compression or something else?

The megapixel-rating of digital cameras has been adopted as a selling-point by the manufacturers. Everyone asks how many megapixels your camera is.

How can I explain that the quality of the pictures taken with my wife's Olympus mju 800 8 megapixel camera are not as good quality as pictures taken with the 6.1 megapixel Fuji S2 Pro?

In fact, I would go as far as to say that once interpolated to the same size file, there is little to choose between sharpness of the S2 Pro and the D200. I would even go so far as to say, that without adjustment in CS2 the colours are much more film-like on the Fuji than with the Nikon.

It just ain't that simple!

another question for you John regarding the same subject. Why is it when I take a RAW image with the D100, in CS2 it shows the resolution to be 3008 x 2000 @ 240dpi and a JPEG image has a resolution of 300dpi. I've looked around the net and can't find a sensible answer, I'm sure there is one. Peter

The D50 and D80 (and D70) all shoot compressed RAW as standard, and AFAIK don't even offer an uncompressed RAW. The compression uses Nikon's lossless compression - and while I struggle to understand how compression can be lossless - it is certainly almost impossible to see in the final images. In fact when I have tested it I cannot.

The D200 (and D2X and D100) all shoot uncompressed RAW as default. Although you can set them to shoot compressed RAW. The files are about twice the size (for the same megapixel sensor).

Alex

another question for you John regarding the same subject. Why is it when I take a RAW image with the D100, in CS2 it shows the resolution to be 3008 x 2000 @ 240dpi and a JPEG image has a resolution of 300dpi. I've looked around the net and can't find a sensible answer, I'm sure there is one. Peter

I am going to set myself up for derision here as, being a technophobe, I have no business even trying to answer the question!

As I understand it, the difference between a jpg and an uncompressed file is not the number of dots per inch. It is the type of dots. Jpgs cluster them if they are similar whereas an uncompressed RAW file shows each dot as read by the sensor. So the loss of quality caused by converting to jpg is very much down to the subject matter and whether these clusters form artifacts (often visible in areas of blue water background).

Now I hope someone who really knows will give an answer we all can understand.

The other answer is: I dunno... Ask Paul Kay!

another question for you John regarding the same subject. Why is it when I take a RAW image with the D100, in CS2 it shows the resolution to be 3008 x 2000 @ 240dpi and a JPEG image has a resolution of 300dpi. I've looked around the net and can't find a sensible answer, I'm sure there is one. Peter

Acutally, it's just a default setting in the RAW converter. It doesn't affect picture quality at all. Resolution expressed in dpi is purely related to what happens when you print. You can change that in the Image Size dialogue box without affecting the image, it will just change how large it prints.

The same holds for JPEGs at 300dpi. It's actually more common to see JPEGs at 72dpi (for the Web), not 300dpi. The former relates to a standard numer of points per linear inch on a computer screen, such that if you display the file at 72dpi on a screen, it will display at 100% zoom. I hope that last sentence makes sense <_<

The bottom line, I guess, is that resolution doesn't affect your image, just how large it prints...

Question for Alex; does compressed RAW increase the number of shots one can get on a card? I was playing with my D200 menus the other night and it did not seem to increase the number of RAW pics on my $GB card (240) when I changed to compressed RAW.

Question for Alex; does compressed RAW increase the number of shots one can get on a card? I was playing with my D200 menus the other night and it did not seem to increase the number of RAW pics on my $GB card (240) when I changed to compressed RAW.

The "Remaining Shots" indicator doesn't change, but you can take more shots. The reason for this is that compressed RAW vary in size, and the remaining capacity is estimated on-the-fly, but by default, the camera considers 0% compression efficiency (i.e. same file size as uncompressed RAW).

To understand this, take 5 or 6 shots in compressed RAW. The indicator will go down by 5 or 6, then go back up by 2 or 3, depending on how much compression was achieved...

The compression uses Nikon's lossless compression - and while I struggle to understand how compression can be lossless - it is certainly almost impossible to see in the final images. In fact when I have tested it I cannot.

Lossless compression works on the basis of mathematical algorithms that regroup patterns of similar information into a more efficient way of storing repetitions.

In simple terms, good examples of lossless compression are Winzip or WinRAR (don't know the Mac equivalent). When you zip a file (e.g. a text document), you certainly expect to get it back looking exactly the same when you uncompress it, not just approximately the same, with some added typos or formatting errors . The same holds for compressed RAW files.

Following comments are all rough!

As has already been said the output resolution is simply a setting which tells you how large the image will be at that resolution. I suppose something has to be set by default and 240dpi is common enough (for inkjet/laser printing purposes its difficult to see any changes by increasing this figure). As for 300dpi, well it is like 72dpi for screens (which aren't always or often 72dpi now), its how publishers 'used' to require scans to be presized at I think and its still there as a default figure. Its all pretty nominal!

Lossless compression works on identifying IDENTICAL pixels and somehow grouping them together (as opposed to jpeg which works on identifying similar pixels within a given area). Given 12 bit colour depth, it is impossible for a camera to pre-determine either the compressed size which will vary or even an estimate of it, so it shows how many uncompressed raw files can be shot - but actually more can be. Result is as Mattdiver states.

I'm spending too much time image processing today. Must switch the computers off!!!

Thank you guys. Very nicely explained. So is there any disadvantage to shooting compressed RAW, sounds like the way to go?

The D50 and D80 (and D70) all shoot compressed RAW as standard, and AFAIK don't even offer an uncompressed RAW. The compression uses Nikon's lossless compression - and while I struggle to understand how compression can be lossless - it is certainly almost impossible to see in the final images. In fact when I have tested it I cannot.

 

The D200 (and D2X and D100) all shoot uncompressed RAW as default. Although you can set them to shoot compressed RAW. The files are about twice the size (for the same megapixel sensor).

 

Alex

The usual definition of lossless is that you can take a file and compress it and uncompress to a file that is identical to the original. Programs like zip does just that. As a demo you can take your uncompressed raw file and run zip on it, note the change in file size and unzip and check to see that you recovered an identical file.

There was some controversy on whether Nikon should claim their RAW compressed files for the D70 as lossless.

http://wetpixel.com/forums/index.php?s=&am...ost&p=34321

I think it's true that you cannot generate an uncompressed file from a D70 compressed file that is indentical to one that was never compressed to begin with. I don't know if anything has changed in the newer models.

It appears from that thread (funny how fast digital changes since that 2004 thread) that NEF Compression is slow. (Craig Raux) However in the D200 handbook page 31 it states that recording time is reduced. I'll have to check that out. It appears then that the potential advantage are faster write time which with the D200 does not appear to be an issue, and more images on a card, which has to be balanced with slight loss of image quality.

Many thanks for the comments, understand now, I don't know why! But thought the JPEG file was creating a larger output image when compared to an uncompressed RAW image of the same resolution, been back into CS2 and compared like for like and they are exactly the same, as they should be, 'nanna <_<

It appears from that thread (funny how fast digital changes since that 2004 thread) that NEF Compression is slow. (Craig Raux) However in the D200 handbook page 31 it states that recording time is reduced. I'll have to check that out. It appears then that the potential advantage are faster write time which with the D200 does not appear to be an issue, and more images on a card, which has to be balanced with slight loss of image quality.

Recording in compressed RAW is much slower than uncompressed RAW on a D100. This shouldn't be an issue on a D200 in most cases, thanks to the large buffer. That aside, there really is no compromise, quality-or otherwise.

The D200 manual states that it actually writes faster. I will test it this weekend