8bit v 16bit TIF images

[dragonlady]

Hi all,

I originally started taking my shots in a 6.3 megapixel JPEG format - however, quickly realised that as soon as you do any modifications to the image, you loose quality each time you save the changes.

I quickly changed to RAW images and now save my photos as TIF files. Unfortunately, what I'm not clear on is the difference between 8bit and 16bit TIF images - can anyone advise if the difference is vast and worth worrying about (apart from the size of the file)?

Any help would be really appreciated.

Kind regards

Chris Berry

[w10d]

The difference is down to the 'bit depth', you can think of this as the number of steps between the lightest values and the darkest values. Until quite recently most digital imaging was done in 8 bit, 16 bit was just too large for the computers, Photoshop only offered limited support, etc. (In fact most digital camera's don;t capture the full 16 bit, being more often 12 bit).

Is it worth working in 16 bit? That's more difficult to answer. It's pretty much essential if you're using a really big Colourspace to edit in. It can be beneficial if you intend to do some drastic edits. Mostly, you probably won't see any difference in your final prints.

Try both workflows and make up your own mind which suits your needs.

[Longwatcher]

Technically there is a vast difference, in reality, there can be a lot, a little or none.

a 16 bit image has considerably more information available for each pixel then an 8 bit image. This is most likely to affect subtle colors in the image.

Part of the problem though is the 10D only provides a 12bit image, so you are not getting 16bits of data, but at the same time with 8 bits you may lose some color information that you could get by saving as 16 bit.

On to part 2 - depending on the subject/background you are shooting it may make little to no difference.

For example when shooting models against a plain backdrop, I have found no significant difference in the final results and in fact since it makes the model's skin a touch more even, using 8 bit can actually be beneficial sometimes.

On the converse, I took a picture of a duck a few months back and the subtle colors in the sheen of the wings got lost when I converted to 8 bit. So when shooting a rainbow of colors and you want the most subtlety 16-bit is best.

which brings to a reason you may need to convert an image to 8-bit. In order to have a local camera store print out my image at 16x24, I had to convert the duck image to 8-bit, since not all printers can handle 16 bit information. Also not all software can handle 16-bit images.

Lastly, a recommendation: when ever possible archive your images in the highest quality possible. For the 10D/D60/300D this will be either the original RAW data or the next step down a 16-bit Tif. If you can do all of your processing in 16-bit, then do so, then save the image as 16 bit for archive and 8-bit for printing if you have to.

My recommendation of course is biased by the fact that I spend 20 years as an USAF imagery analyst and don't like losing any data what so ever if it can be helped.

Hope that helped,

[maderito]

Chris,

I think everyone is assuming that you already know that a JPEG image (non-RAW) from the 10D can be saved as an 8-bit TIF image (but not 16-bit) in your image processing software.

You degrade the original JPEG by repeatedly saving it as JPEG (even if you don't manipulate the image). However, if you save the initial JPEG as a TIF file, that file will inherit any image degradation that resulted from the initial JPEG compression (performed in the camera), but you don't add any more degradation related to repeated cycles of JPEG compression (as long as you continue to re-save it as a TIF file).

I can see that my last sentence has the hallmarks of a run-on sentence. Hopefully, you understand. Besides, you probably know all of this already. If so, happy RAW processing :)

[iwatkins]

I think the only time you will really see the difference between an 8-bit and a 16-bit image is if you do any heavy image manipulation. Well, that is the case in my experience.

I do lots of landscape/skyscape shots so often have huge swathes of sky of almost all one colour.

With an 8-bit image if you do lots of work increasing/decreasing contrast etc. (I mean heavy usage here) you soon start to see banding in the colours. With 16-bit images you don't.

Needless to say, I stick to 16-bit TIFFs from the original RAW. OK, uses disk space, but hey ho....

Cheers

Ian