One of our illustrious Photo Community members asked me this question and I thought it interesting enough to post here. I'll keep him nameless since he did not give me permission to use his name, but he can chime in if he wants: --- Replying to message from XXX: "If you ever find the time Kort, please can you answer this question for me. I have asked this of others but never had a reply." Q: "If you took a photograph using RAW format you would see the superior quality on screen or on a printout. If you then change this image to a Jpeg for the gallery does that not take the RAW image quality away. In other words, would you not have ended up with the same image as if you had shot in Jpeg format in the first place?" Me: VERY intersting question. What you have to think about is the process. When you shoot in JPEG, you let the camera make all the decisions for you about what info is kept or left out of the image. It captures it at 24 bits then gets rid of half that info when going to JPEG. That is what you download and make changes to. Kept RAW, you have the full 24 bit image to adjust as you see fit. When adjusting the levels or hue/saturation of a RAW image, you are playing with a lot more data and the quality is that much higher. Nothing has yet been lost and you can save all that data to your hard drive to play with at will. There are no JPEG artifacts or lost data in RAW format. When you finally get the image where you want it, you do save it as a JPEG for online use. However, this is the first time it has been saved this way and you control the compression ratio. You do lose a tiny bit of quality, but not as much as if you saved it as a JPEG initially and then saved it as a JPEG again after fine tuning it... --- Much of what I wrote I picked up from listening to some of our more vociferous RAW proponents. :) If they would like to add anything to this, I welcome it. Here is another photo comparison using my C8080, since the one I originally did with my C5050 is long gone. The top JPEG is a bit smoother, but you can see the detail that is lost when looking at the RAW on the bottom. I'd also be curious to see other RAW/JPEG comparisions from other cameras.

Kort Kramer - Kramer Kreations

Hmm. I think this one is ultimately going to be down to either personal taste, or what is required at the time. To me, the red file in the RAW is 'noisier' but the paper is more 'papery'... in the JPG it's just white. Using that comparison on the rest of the image, I personally would use the RAW version for 'picture' use, but the JPG version for manipulation... the smoothness of the colours would make it a lot easier to select/work with... Just my opinion though ;-) (",)

Yep it is just a case of - do you think you can do better than the standardised settings in the cam ? or will you be able to in the future ? The JPEG looks best if you don't manip the RAW because it has had noise control applied to it. The RAW will give you many more options for processing today and later though.

And every one said, 'If we only live,
We too will go to sea in a Sieve,---
To the hills of the Chankly Bore!'
Far and few, far and few, Are the lands where the Jumblies live;
Their heads are green, and their hands are blue, And they went to sea in a Sieve.

Edward Lear
http://www.nonsenselit.org/Lear/ns/jumblies.html

Mike, I may have been able to eliminate the noise when importing the RAW file, but kept it AS IS for the purposes of the test. I do agree that for people just shooting for online galleries, the JPEG could be the better way to go.

Kort Kramer - Kramer Kreations

Ahh, I see Onslow beat me to the reply there. :)

Kort Kramer - Kramer Kreations

There is another way of doing things: shoot in RAW and save in uncompressed TIFF. Without being scientific in any way it's what I have started to do when passing files on to others (by CD only as the TIFF files are even bigger than RAW).

Charles

One of the main reasons I shoot RAW is the ability to work with the white balance of the photo. If your white balance is off, it is much easier to correct when convert the RAW format, then working with a jpeg (or tiff) As stated before, if you shoot a high contrast scene, where if you lose detail either in the highlights or shadows, it is easier to correct when shot in RAW. Just convert the photo twice, one corrected for the highlights, one for the shadows. Then in photoshop it is fairly easy to use a layer mask to get correct exposure for the photo. (Highlights and Shadows)

You see, this is why idiots like me get confused... If CD there has taken the shot and it is in his camera, then he saves it in TIFF format, how does the file get BIGGER? Surely in the camera it's in the most basic format available, which I now gather is RAW. If it's bigger then something must be added to it, yeah? So... what does it add? Or does it just make pixel size bigger? But if it does any of this, then surely it's better than RAW? lol - now, I guess that ain't so, and that RAW is actually the best to work with. See? Confused. lol (",)

Liken RAW to a digital negative...unprocessed in cam...all you are doing is what the process labs do..brightness, contrst, white balance, sharpness, saturation...I could g on but you get the drift...which gives you far more scope initially over the control of the image...RAW is not a new format but its the way to go for results. I beieve that TIFF is 16 bit as opposed to 8 bit JPGS...hence the extra file size.

TIFF and Photoshop PSD can support 16 bit.

IMO if you convert a RAW file to Jpeg then you are back to what a top quality Jpeg would be like out of the camera. (Possibly a little better with RAW depending on what software you are using.)

Like Jim says, it is about the versatility you have during processing, and especially if you want to print. (You will be able to print lager good quality images from a RAW converted to TIFF/PSD than you would be able to from a Jpeg. Maybe even more so using a lab rather than an inkjet

If all you want is web images then RAW may not be of any real benefit except for the versatility

Bottom line is if you process your image and later decide you dont like it for some reason you can re process it again (a million times if you like) with different exposure values etc.
All images are captured in RAW by the camera, (even in Jpeg setting) the difference is that if the camera processes the raw you have no control over the processing parameters, if you do it yourself you have as much control as the software that you are using allows you.
when prosessing. Edit TIFF can be 8 or 16 bit that is 8 or 16 bits per pixel so times that by 3 8/16 bit per each colour channel RGB 8 bit = 24 bit image 16bit =48 bit image (double the amount of colours so there is little or no banding and the histogram is much smother), so your tone transition between colours is much smother. Jpeg can only have 8

Message edited on: 03/22/2005 18:02

The Truth is Out There

I shoot in JPEG since my cam doesn't have the RAW option :( so what I do is that I open my images in my programme and save them in TIF. Then I re-open them in TIF before I start doing anything on the pic! Does it change anything to do it that way?

Tibet Pascale, I am the last to qualify to answer your question...but, I often do the same thing. If in jpg from the cam, I always ALWAYS copy and save in tiff or the primary uncompressed file for the program, whether photoshop or paint shop pro or some other. All editing is done on a copy of the saved file in an uncompressed format. That way, the least damage is done to the original. In fact, the original for me is "saved" and the final output is made from one of the uncompressed formats, then saved in jpg or whatever is needed to print or go on-line. Tom ;)

same as TomDart but I forget some times and work with the jpeg. but the original is not touch at all. I save what I work with under a diff name and in a tiff. plus my program willgive it 300d instead of the 72d when it comes out the camera[ it big image like 31x32 inches big

what you see is not what you know; it in your face

joko, I agree. And, I generally work in higher than 72 pixels since some will be printed. The first effort I make is to copy the original to work with and save the original untouched. By forgetting this and messing up images, I slowly learned. : )

now when the image come in black from the camera I will work with it to get the image out of it out of photoshop then save it right on top of the original one. That the only time I do destroy the original image [ it black anyway]

what you see is not what you know; it in your face

Yea, I do that to and also convert the 72dpi to 300dpi. I save the original JPEG in a different folder and only work on the TIF/300dpi one! I'm glad to see that I'm doing the right thing! :) Thx guys!

I can only suggest conducting your own tests, a tripod mounted camera, a static subject and a period of stable weather to make the test even handed. My understanding of the jpeg in camera image compression function is that to make the image smaller the onboard computer looks for areas of similar tone or colour and when found it converts them the same value, this is why I feel that in the main parts of the image you may not see a vast improvement using RAW but it is perhaps the subtle graduations of tone you find within areas of the image, i.e sky that have been lost using jpeg. If a 10 Mb image has been in camera compressed down to a 2 Mb jpeg a considerable amount of valuable information has been discarded with no input or control from the photographer, opening this image and saving it as a tiff will not put back what has already gone. The range of extra adjustment functions within Photoshop as already been highlighted above. The downside of RAW is that the image is going to be large so extra memory cards are going to be needed but from this fine digital negative any amount of jpegs in whatever size and for whatever purpose can be made whilst still retaining an unadjusted RAW file, should you decide to revisit your image you will always be working with the best possible quality original. My personal approach is that at the time of taking an image I do not know what usage it will be put to but what is so important is that I have the potential to make a large paper print from any image I make, I am a hobby photographer so everything is self funded but achieving the highest quality image is the priority, from the selection of a viewpoint to give the required perspective ( remember your legs ! ) accuracy of exposure, selection of a suitable aperture/shutter speed combination to give the depth of field required to the post work ( limited in my case ) in Photoshop.

I think the most basic and understandable comparison between RAW and JPEG I've ever heard was by a photographer whose name I forget. He simply said:

Shooting in JPEG mode instead of RAW mode is like shooting with a film camera, taking the roll into the local lab and when it's ready you take the negatives, throw them in the trash and walk out with the prints.

I think the mistake most people make is they talk about RAW as an image format like JPEG or TIFF. RAW isn't an image format, it's a data format. It's the data that represents the light levels as seen by each individual cell in the imaging array. It's not an image until you actually process it through a RAW converter.

Two Pynts - The only thing I see "wrong" in your initial post is in the bit depth. JPEGs are 24-bit (8-bits per color channel). TIFF images can be (depending on where they come from ) 24-bit, 48-bit (16-bits/channel) or even 72-bit (24-bits/channel). Most contemporary digital cameras and all DSLRs use 16-bits/channel (actually most use 12-bits/channel but put them in a 16-bits/channel format). RAW holds 12 to 16-bits of data for each point on the array.

-=>Donald

Okay, you caught me Donald. I wasn't thinking about the bit depth clearly, you are right. I'm glad this has spurred discussion and totally agree with the comparison Donald presented. I'd like to see some more photo examples if anyone is willing. Here is the one I posted before. Again it was handheld and not exact, but you can see the difference. This was with the Olympus C5050. Thanks to everyone for voicing your viewpoints! 8]

Kort Kramer - Kramer Kreations

Mike - The reason the TIFF file is bigger is because it is uncompressed. RAW is a compressed file format all be it lossless compression.

randyrives - If you use what you have described as a method of broadening the dynamic range available to you surely you will lose a lot of detail. You would have to underexpose the shot to avoid any burn out where there is no data to recover and if you underexpose you are going to be using darker shades and therefore limiting the number of shades available to you, there being far more lighter shades than darker ones available. Message edited on: 03/23/2005 11:30

And every one said, 'If we only live,
We too will go to sea in a Sieve,---
To the hills of the Chankly Bore!'
Far and few, far and few, Are the lands where the Jumblies live;
Their heads are green, and their hands are blue, And they went to sea in a Sieve.

Edward Lear
http://www.nonsenselit.org/Lear/ns/jumblies.html

Well since a DSLR is like color reversal (slides) there is more latitude for underexposer than over exposure. So if you expose for the highlights, which will cause the shadows to be to dark, then during the Raw Conversion, you can make a 2nd conversion that will bring the detail back out. Then in Photoshop blend the two to give a much better exposed picture. This is like the technique where you take the same shot twice, one exposed for highlights and one exposed for shadows and blend them together. The problem with that method is unless you are using a tripod, you very seldom get an exact duplicate. I read about this technique on DPReview forums.

Thx for replying and explanation - yes I can see it would work that way it just doesn't sound to be as good as bracketing the exposure and using the higher range of tones available at the brighter end of the scale. I will have to conduct some experiments with both methods (when I get time) see if I can see any difference.

And every one said, 'If we only live,
We too will go to sea in a Sieve,---
To the hills of the Chankly Bore!'
Far and few, far and few, Are the lands where the Jumblies live;
Their heads are green, and their hands are blue, And they went to sea in a Sieve.

Edward Lear
http://www.nonsenselit.org/Lear/ns/jumblies.html

Don, thanks - that's what I thought RAW was, I was getting confused with all the different options/ terms/ possiblities. I didn't realise the information was compressed though Onslow, I thought it would be uncompressed..... cheers for clearing that up!! (",)

One last thing to add into the mix after reading the above on whether to bracket or underexpose. Remember that imaging arrays are linear, but light falloff is not. Hence, you don't get even distribution of bits (levels) across the dynamic range of array. We had a discussion about that about a year and a half ago. The jist being that if you underexpose too much, you throw away a lot of your data levels. It's best, in most cases, to expose so that the data is pushed as far towards the highlight side as possible without clipping and then using your converter to dial that back to a proper exposure. Of course, that's RAW only. Never tried it, but it should work for the expanded dynamic range trick too. You do one where you don't clip the highlights so you know you have highlight data, then do one where you clip the crap out of the highlights, which gets your midtones and shadows up into the areas where the most bits are used to store the information. Then dial them both back to proper exposure during conversion and combine the best parts of each. Here's what I wrote back then (2003) - tech speak ahead: Science behind it is that light falls off using the inverse square law, but CMOS and CCD arrays are linear devices. In practical terms, this means that if you assume that a CMOS/CCD camera has a 6-stop dynamic range, the first stop (the brightest) will contain half of the available data. By the time you get to the the sixth stop, the number of bits used is very small compared to the first stop. So, assuming a 12bit dataspace (4096 individual bits): 1st stop = ~2048 bits 2nd stop = ~1024 bits 3rd stop = ~512 bits 4th stop = ~256 bits 5th stop = ~128 bits 6th stop = ~64 bits (there's an additional 64 bits left over that are distributed in there somehow) So, if you get as much of the image data as possible into the first stop, you effectively increase the amount of data collected in those areas of the image. And in shifting the shadows away from the sixth stop (which has the least # of bits of data) towards the 5th or even 4th stop you increase the amount of data present in the shadows. -=>Donald

Thx Donald you have explained what I was trying to get to. My thought process was if you took a landscape shot say with the highlights not blown all the detail would be in the sky and the landmass would be darker therefore have less levels of tone in it to show even if it was lightened using the exposure compensation control in a RAW processor. If the shot was bracketed at shooting the landmass could be shot with brighter levels (the sky would be blown) so it would have more to incorporate in the combined image. Not sure that makes sense just seemed right to me.

And every one said, 'If we only live,
We too will go to sea in a Sieve,---
To the hills of the Chankly Bore!'
Far and few, far and few, Are the lands where the Jumblies live;
Their heads are green, and their hands are blue, And they went to sea in a Sieve.

Edward Lear
http://www.nonsenselit.org/Lear/ns/jumblies.html

Interesting Donald, I did not stop and think about CMOS sensor's being Linear. These does change my thinking some. Thanks for the info. I really need to go out and try some of this.