Can someone please explain to me RAW?

[lukeROB]

Like all that Alcina said, basically RAW file recorded everything from the sensor in 16 bits and write it uncompressed to the card so you can get maximum data for the picture. Resolution does not really change between RAW and JPEG. However this is not true of Nikon RAW file from current Nikon such as D70 which is 12 bits compressed (eventhough Nikon clamined that its compression is lossless, whatever that is).

Most RAW files would use some form of lossless compression. It would seem silly to save a larger file for no reason. Also most RAW files are 12bit - reason being the sensors are usually 12bit - I know the canon compacts are the same. There are I'm sure 16bit sensors around though. A lossless compression is something like a .zip file. All information is retained however file size is decreased. This is done through mathematical formulas which allow certain data to not be saved because this data is retained from the knowledge about other parts of the data. A very simple example is in a black and white image, you only have to save the information about black pixels, the compressed file when opening knows that if the pixel has not been marked then it must be white. Simple and in a perfectly half/half black white image you have applied a lossless compression to achieve a file size 50% smaller. Colour standard images are more complex and so compression is not really that great but you get the idea.

The real advantage of RAW is in post processing as the others have said. Think of it this way. A RAW file allows you to apply software filters to the image with the raw information recorded by the sensor. You can control what the software does opposed to the camera selecting and doing it for you. Since in camera software is basically fixed once manufactured and it must fit in small amounts of memory it is never going to be as good as external software. This external software can change and be built on and as it improves so you will be able to apply more sophisticated filters to your image. RAW is often thought of as a 'digital negative' because of this. All files start out as RAW but you are deciding whether your camera makes standard decisions to produce an image or whether you control the decisions to be made.

Sorry this is badly written and confusing but I thought I would just write it down as it was coming into my head. I might do a better job when I'm not so tired so if it isn't understandable just let me know and I'll try explain it better.

 

[ssra30]

opps thanks for the correction on 12 bits vs 16 bits. I think I got it confused with PSD file. I have no idea how the Nikon lossless compression works but D70 NEF file is quite a bit smaller than the older D100 and it is even a bit smaller than ORF file from Oly C5050, so my impression was that most RAW file are uncompressed, may be it is just a less aggressive compression than the new Nikon NEF file.

 

[Kim]

If I compare RAW from the C5050 - just over 7Mb/picture - with TIFF around 21Mb for the same picture saved from Photoshop - it really makes me wonder about the compression!

 

[alcina]

skip tiff - too big and there are options that are just as good - psd in photoshop for one and RAW on camera...

 

[lukeROB]

opps thanks for the correction on 12 bits vs 16 bits. I think I got it confused with PSD file. I have no idea how the Nikon lossless compression works but D70 NEF file is quite a bit smaller than the older D100 and it is even a bit smaller than ORF file from Oly C5050, so my impression was that most RAW file are uncompressed, may be it is just a less aggressive compression than the new Nikon NEF file.

What I was saying before applies to all RAW, NEF, ORF any file that is saved in a raw format is likely to use some form of lossless compression. That is all information is retained without saving data for every single pixel. I explained a simple example of how it can work with a black and white image. Colour is more complex this means that smart people at Nikon, etc come up with better mathematical formulas, that determine what data needs to be retained in order to have complete knowledge if every pixel. This is why you need a specific RAW converter, the converter needs to know the equation used so it can apply the opposite in order to recreate all the information. Using the word aggressive is not really appropriate as no format reduces image quailty or alters the information in any way. A better word to use would be that the NEF format has superior algorithms which lead to smaller file size.

TIFF is a totally uncompressed, pixel for pixel representation of the image. However it is only an uncompressed version of the image after the internal camera filters have been used. This often means only an 8bit file - far less information is retained than the 12bit raw. You can see from a TIFF file size that using a direct pixel representation uses lots more space. Therefore don't use TIFF as you get most of the disadvantages of using a JPG (minus the compression) and not much of the advantages of RAW.

Hope this helps.