TIFF vs RAW

[Ultra Magnus]

I just read the long thread about RAW files, and I was just wondering what the difference was between TIFF and RAW? They both produce really large file sizes, and from what I understand, TIFF is an uncompressed format?

Thanks.

[Lionheart]

I haven't played with RAW format much, but I think RAW is an uncompressed format that seems to be proprietary to the camera brand that the image was captured with. For instance, Canon RAW format will not be the same as Nikon's RAW format, but both are uncompressed images that needs software to be interpreted and viewed correctly on your computer. TIFF is also an uncompressed format, but one that is universally accepted and viewable with the various photo browsers available, such as ACDSee. The various RAW formats are not necessarily viewable on such browsers unless they have specific plugins or have the ability to view the various RAW formats built into their code. The software that comes with the cameras may or may not have a conversion utility to save the RAW images into more universally accepted formats such as jpg, bmp etc.. There are programs available for converting most RAW formats regardless of brand, such as BreezeBrowser. Any one out there who knows more, please correct me if I'm out in the Mutara nebula on this, but that's my understanding of how this works.

[mikishots]

When you make a RAW capture from a digital camera, the data exists in a "quantum" state that is NOT an image file. It's not a TIFF or a JPEG: it's just raw data (some cameras offer a medium rez image along with it, so you can view it immediately). The camera simply takes this info and sets it up in a propreitary format. The RAW file is like an unprocessed latent image on film.

Here's where it takes a turn away from the common notion that it's "just like a negative": The RAW file can be "processed" many times over as a first-genenration image. If you save the RAW file you can come back to it in the future and reprocess the data in any way at any time. Can't do that with film, and all you can get from a negative is a second-generation image.

It's important to note that the RAW converters make changes to your image files PRIOR to processing. Essentially you're sending the image data through an algorithm - your photo is still a first-generation image file. Again, as long as you have a copy of the RAW file, you can make changes without hurting the integrity of the original pixels.

As far as the difference between TIFF and RAW, The TIFF image format has been around for a long time, and will likely be around for quite a while more. The TIFF format was the cross-platform image format by choice until PNG was introduced, and according to some still is. TIFF's are a developer's nightmare. Last I counted there was over 50 different styles of TIFF images. You never know if an application will be able to read a TIFF image or not. The main benefit of TIFF is that it offers the least amount of loss when converting photos for your use. My route for working with a photo is the following: Take it in RAW, make adjustments in RAW, convert to 16-bit TIFF mode, make any color corrections, then convert to JPEG once I am completely happy with everything. Then I go out and buy another 120 GB drive.

[Lionheart]

My understanding is now much clearer. I've shot RAW before and processed them in BreezeBrowser, but haven't really thought about how it all works before.

[another view]

A lot of good info here. I'd say that a TIFF is a lot like a JPEG, just with no compression. That's why TIFF files are so large. When shooting either, they both will have a white balance, contrast and sharpening settings applied to them. RAW files don't have any of those settings applied - it's done later. So however you shoot, do your postprocessing (including RAW conversion if shooting that way), color balance, etc - and save the final image that's ready for printing as a TIFF file. You could also save it as a JPEG but you're throwing some info away.

I save as a TIFF file, and that goes to the lab for making a print. If I'm having a print done on a Fuji Frontier minilab (really nice cheap prints) then I save a copy of that same TIFF file as a JPEG, with the quality set at 12 in Photoshop (maximum, for the least amount of compression). The reason for this is that the Frontier (at least the one I use) only takes JPEGs.

[Ultra Magnus]

Thanks! I think I have a decent understanding now. I took some shots over the weekend in tiff and jpeg of the same object, at the same distance, etc. and noticed that zoomed way in the tiff had sharper more defined edges. I think for now I'll take close up photos of still objects in .tiff because I've noticed in some pictures of flowers that I've taken in jpeg there's that jpeg glow around objects that have a lot of contrast with whatever is next to them, in this case the dirt backround around the orange flower petals.

[yaronsh]

Some answers to my own questions from the other thread:

Conversion from RAW format to another format (be it compressed or uncompressed) is not just a matter of rearranging the bits from a proprietary storage format to a standardized one. A TIFF does not contain the same information as a RAW.

I hold these truths to be general knowledge: That sensor pixels, unless it's a Foveon sensor, are monochromatic, and capture either a red, blue, or green value; that the pixels are generally arranged in a checkerboard pattern, with the green pixels comprising 50%, and red and blue 25% each; that a RAW file essentially contains the red, green, or blue value of each pixel (and so, can be thought of as three monochrome layers).

In the TIFF (or whatever) file generated from the RAW file, each pixel is not just a red, green, or blue pixel, but can have any color in the color space. If it's an 8-bit format, the color space is broken up into 256 gradations; if 16-bit, 64k gradations. The "raw" R/G/B pixels are transformed to colors from a continuous space to create a smooth, non-dithered image (within limits).

How does a red, blue, or green monochrome pixel become a continuous-space pixel that is not necessary red, blue, or green? Well, each RGB pixel is interpreted in the context of the other pixels in its vicinity, as well as some general assumptions about how humans see. The white balance, color tone, and other parameters influence this interpretation. Different cameras and s/w have different algorithms for interpreting the data, and can come up with different results.

That is why, when you work from a non-RAW file (compressed or not), you're working from already-interpreted data. You may still be able to change color temperature, tone curves, etc. However, at that point you are essentially modifying an initial interpretation that is completely automated, rather than guiding that initial interpretation yourself. At that point, decisions have already been made regarding what information from the RAW file is more vs. less important, what can be discarded, and what to make of the information.

So the decision, I guess, comes down to how you like the interpreted formats coming out of your camera or whatever conversion s/w, how much control you want over the interpretation, and how much trust you're willing to give the automated tools.

- Yaron