Linckinn Loc: Okatie, SC and Edgartown, MA

If I convert a RAW file to a tiff, and then export/ import it into software that would not otherwise have supported the camera from which the original file came, do I understand correctly that the tiff is "lossless", meaning I can edit it just as I would have the original raw file?

I suppose of necessity the first conversion would have done some developing/ processing of the RAW.

Thank you for your help. ( I am looking at Capture One, which seems terrific in many ways, but it doesn't support one of my lesser, but nevertheless somewhat frequently used, cameras.)

nospambob Loc: Edmond, Oklahoma

You are correct.

kymarto Loc: Portland OR and Milan Italy

Lossless and RAW are different animals. RAW is the numerical data from the sensor, which is not an image file, and cannot be displayed until certain values are chosen from multiple possibilities and encoded in an envelope that is recognized by image display programs. The easiest way to envision this is to regard the RAW as a negative, and the image file as a print. The tiff "print" is better than the jpg "print", but it still doesn't contain nearly the information of the negative, which can be used to make a very different looking print.

Lossless refers to the fact that no pixel information is discarded. Jpgs use tricks to record pixel information based on surrounding pixel information, so that values are approximated in various areas of the image. Tiffs record the information for each pixel with no reference to the surrounding pixels, so the files are much larger.

In addition you can have either 8 or 16 bit tiffs. 8 bit files, whether they are tiffs or jpgs or pngs or whatever, record luminance and chrominance values of each pixel in only 256 possible values, whereas 16 bit files (of any image type) record them in 65535 possible values. The larger steps in 8 bit start to become significant when you post process.

So first, I strongly suggest saving and editing in 16 bit, which is nearly as flexible as RAW but not quite. FYI, when you open a RAW in Photoshop, it is automatically converted into an 8 bit file or a 16 bit file, depending on what you have set in the ACR preferences. LR is like editing in ACR--it is only making adjustments on the fly to the values it has temporarily encoded into a viewable file, while the original RAW values remain untouched. Once you save the file as a TIFF or JPG, those values are set in stone, so to speak, but the potential values available if you reopen that image file for further adjustments is many orders of magnitude greater if the file is 16 bit rather than 8 bit.

When you work in ACR, like LR, the RAW data is preserved intact, but a sidecar file is created that tells the software what values you have chosen for various parameters like contrast, saturation, black, white, etc. Those values do not change the original data. In image files, whether tiff or jpg, values are encoded in the image, because they are universal formats that can't be depending on reading and interpreting RAW values for every different kind of camera and model out there. So there is definitely some loss of data in ANY image file format, whether tiff or jpg or psd or png or bmp.

http://www.diyphotography.net/8-bit-vs-16-bit-color-depth-use-matters/

Apaflo Loc: Anchorage, Alaska

The RAW file contains unprocessed raw sensor data encoded using a Bayer Color Filter Array. It is not technically even an image because when the raw sensor data is "demosaiced" it can produce nearly an infinite number or very correct images, all different! The image data that results is not something that is merely the same information encoded in a different format. In fact it is impossible to convert a TIFF image back into the RAW format from which it came. The RAW file and a generated TIFF file are not the same in any useful way. It's a little like tossing a huge rock into a pool of water and grabbing one drop of water from the splash. It is not the entire splash. And from that drop the entire pool cannot be recreated. All you have is one small part of what was there.

The TIFF formated image may or may not be "lossless". That refers to compression of the data. It does not have anything to do with conversion from a RAW file. The RAW file is demosaiced, which means a matrix of data locations is used to calculate image values at each pixel location, and those pixel values constitute an "RGB" image. Each pixel in the image has a set of three values, one each for Red, Green, and Blue. When that RGB data is formated to write to a disk it can be compressed to save disk space and reduce data transmission times too. If the compressed data can eventually be restored to exactly the same original RGB data, then it is called "lossless compression". If it can only be restored to something close enough to not be visibly changed, it is called "lossy compression".

JPEG images are compressed with a lossy compression (which can be adjusted to very little loss or to a great deal of loss, depending on how much smaller the file size needs to be). The TIFF format can use any of a fairly large number of compression types, including the exact same type used by JPEG. But typically TIFF images are compressed with one of two or three lossless compression types.

Linckinn Loc: Okatie, SC and Edgartown, MA

Thank you both for your detailed explanations. Makes sense, and I am so glad I asked. Clearly my work-around would be no good.

For my purposes, that clearly means either don't buy the software, or buy it and process the unsupported camera images (5 to 10% of what I shoot) in other software.

Probably best to go with Lightroom CC and be done with it. Guessing Apple's new operating system (coming up soon) will kill my Aperture for good, and it is time to step up to a better processor anyway.

SBW

The question you ask is a loaded and very complicated question that is really best kept simple and the way you asked the question the answer is both yes and no. Ever since the PC became the photographer's darkroom this question as been asked over and over.
Again, attempting to keep it simple: the "yes" part of the answer to your question is when you convert from RAW to TIFF and you save the file as a TIFF file there should be no loss of detail in information of the TIFF version of the file you are working with provided your SW is using lossless compression or no compression at all. Keep in mind that most but not all TIFF save routines use "lossless" compression or no compression. Now for the "no" part of the answer. When you convert a RAW image to TIFF you are throwing away some image information the moment you do that conversion and save as a TIFF. Another way to understand it is to say you are a painter, you are going to start your painting using a "RAW" canvas that is huge and is almost limitless in what you can paint on the canvas. Once you have started working as a painter you decide that your RAW canvas is really too big and you do not need a canvas that large so you convert the canvas to a TIFF canvas. You still have a very, very large canvas to work with but it is just not as large as the RAW canvas. With the TIFF canvas you now have some self imposed limits but you would rarely if ever notice them or meet those limits. Yet another way to look at it is to use the analogy of Chrome or slide film versus regular negative film. Anyone that is or was a film photographer knows that negative film has a much greater exposure latitude than slide film. The exposure latitude in negative film can be up to 5-7 stops whereas in slide film the exposure latitude is usually no more than 2 stops. In other words back in the film days if you screwed up the image in the camera you had a much better chance of saving it or getting something you could use with negative film as opposed to slide film. That is why almost all chrome or slide film photographers would bracket their exposures. In this analogy the negative film would be "RAW" and the slide film would be the "TIFF" saved conversion. To be clear that analogy STOPS there.

I hope this helps answer your question. There are literally thousands of web sites out there that may provide a better or simpler answer than I have here if you Google the question.

Bottom line: for the digital photographer it is always good practice to save all of your important RAW images and use your TIFF conversion as your working image. Should you need it again, the original RAW image is always there to use or save in any format or extension you wish.

nospambob Loc: Edmond, Oklahoma

Yep, I, too, had been mislead by previous pundits. The affect on raw data makes perfect sense. Thank you gents. OkBob

BebuLamar

Do you notice that even the 8 bit uncompressed TIFF file is larger than the 14 bit uncompressed RAW file? That is because the TIFF file contains RGB values for each pixel while the RAW file only has 1 value for each pixel.

Linckinn Loc: Okatie, SC and Edgartown, MA

Thanks again everyone. UHH is so wonderful for getting insights and ideas that one just wouldn't with Google searches and articles.

kymarto Loc: Portland OR and Milan Italy

You really want to go crazy save as 32 bit floating point TIFF, which is an option when combining images for HDR. A raw file from my D800 camera is around 40MB. When converting that and saving as a 16 bit TIFF the file size jumps to around 200MB. 32 bit TIFFs are twice that size--400MB each.

zigipha Loc: north nj

Two definition of lossless
1. Any loss in converting from one format to another (raw to tiff example). I think this has been covered in this discussion
2. Any loss in the "open/edit/save" cycle. If you open a JPG edit and then save, the saving of the working file to JPG will degrade the image (beyond the editing that you did). With 16-bit TIFF, this degradation is substantially/imperceptibility small; 8 bit would be noticeable.

LR is a lossless editor because it does not modify the source image when you edit it.

Gene51 Loc: Yonkers, NY, now in LSD (LowerSlowerDelaware)

Some Nikon cameras have a few options for recording raw files - 12 bit/14 Bit uncompressed, 14 bit lossless compressed, and 14 bit lossy compressed. uncompressed files are correspondingly huge. 14 bit lossy compressed are smaller, but do discard data. Lossless compressed is effectively the same as uncompressed in that it is unlikely that the data lost in compression will be missed.

As far as TIFF files are concerned, the other two responses totally apply.

AzPicLady Loc: Behind the camera!

My current software doesn't read RAW images from my newest camera. So I convert the RAW files to DNG before importing them into the software. Would that option work for you? It's another step, but you're doing that step already. I've found that the DNG files can be edited the same as RAW files. I'm not great at PP, so take what I saw with several grains of salt. I've found that some of the editing possibilities for RAW files can't be done on TIFF or JPEG files. Converting to DNG allows those editing points to work.

Apaflo Loc: Anchorage, Alaska

There is nothing that can be called lossless about demosiacing data encoded with a Bayer Color Filter Array. First, it cannot be reversed. It is impossible to create a RAW file from the RGB data produced by demosiacing a RAW file. Second the RGB file that is produced is one of a nearly infinite number of possible correct RGB files that can be produced. The information for all of the others is not contained in any of the individual RGB files. And third the process of demosiacing is one of interpolation, which is to say that non-existent data values are calculated (estimated) to fit between the known value points; which is to say that all colors in the RGB file are approximations.

That does describe where the term "lossless" is correctly used, with the exception of the reference to 16 bit TIFF supposedly being less loss than 8 bit, which absolutely is not true.

But that description does obfuscate what lossless is. Simply, in digital photography, it is when a data compression algorithm makes a smaller data set (by increasing entropy) and then the resulting compressed data is uncompressed to generate a copy of the original. If the second data set is an exact copy the original data set, it is a "lossless" compression algorithm. If even a single bit is changed, it is to some degree a "lossy" compression scheme.

This is very significant for image data, but in fact the lossless/lossy designation has nothing to do specifically with image data or photography. In real life it does not have to be compression, and can be any different encoding of the same data set whether it is larger or smaller.

kymarto Loc: Portland OR and Milan Italy

While I certainly agree that it is hard to quantify "loss" when demosaicing, I would not say that an 8 bit tiff file is equal to a 16 bit file except in basic structure. It clearly contains more data, and useful data, than an 8 bit file, which--like raw data--is discarded when changing to 8 bit and which cannot be subsequently recovered if reconverting to 16 bit.