Is there really a big advantage of using raw formats?

[jito]

Honestly, being an electronics engineer who have studied how media image compression works, I read a lot of comments everywhere claiming that if you save the images in raw formats you have the possibility to further fix the picture than if you save them in jpeg.
Depending on how lossy a jpeg is, such statements don't really make much sense and that can be proved with not so esoteric mathematics.

I'm either missing a very important piece of information here or, at the risk of sounding arrogant, may I argue that almost everybody that preaches the advantages of raw formats doesn't really know how jpeg compression works?

This is not a rant, I am genuinely curious about this, I believe it's a myth. But then, I could be missing something. I can understand why an engineer or a scientist would use raw formats, but seriously, why would a photographer use raw formats?

[Attila]

I use it due sharpening works well with RAW format and much less good with Jpeg.

[Kram]

When I shot my first big digital job in 2005 (which was out of state) I realized on the first day that I was not going to have enough storage space to shoot in RAW so I shot high res JPGs. The final job comprised about 6000+ images. At the time you could not batch edit JPG's in Adobe Bridge. This put me under huge pressure to edit the work under a tight deadline. I made it but I'll NEVER do that again.

I am not an engineer but as a pro I do find certain things "saveable" in RAW but not so much in JPG (like blown highlights). I see no reason not to shoot RAW. I wouldn't buy a camera without RAW capability.

[Smoli4]

I use it because it captures a vider range of colors and shades, also because it stores all the information that the sensor had captures, for example i took a shot and the sky is overexposed and has a totaly white color, i know that if i shot raw, the can correct that by adjusting the exposure in the sky area and it will reveal the color that was there - blue, and in jpeg, if the sky is white there is nothing underneath it just stored it was white color,
same in shadow areas where you cant see the detail becuase it looks black, in raw, i can recover that detail (ofc it will increase the noise in that area) but it just gives you control that jpegs dont.

and about the color range that i can see in raw files vs jpeg files i see more natural transition, and more life like colors when im looking at a raw file then a jpeg file.

maybe im wrong so id like to hear more, this info i heard and read on the web and after that i experienced that when working on images, but maybe there an other side to the story.

but i ask, why shoot in jpeg and not in raw? just because it takes less space on the card?

[Esox lucius]

If you don't get why 14-bit RAW with a multitude of layered adjustment possibilities is better than a flattened 8-bit JPEG output file, you should probably shoot JPEG and be happy.

[Spotmatic]

A great analogy:

Shooting JPEG: print the pictures and throw the negatives away. Copies will never be better than the prints.

Shooting RAW: make unlimited prints from the negatives. Copies are at least as good as the original.

[stingOM]

For me it is simple - better dynamic range, custom temperature balance, better gradation of colours and tonality, and noise control!! Those extra "bits" thrown away really counts!!

[Attila]

One more important thing with RAW, you don't need to give away RAW just jpeg, if you need to prove evidence shoot is yours RAW can do it, JPEG not, but perhaps there is some software what is convert back JPEG to RAW...

[visualopsins]

raw -- 16-bit colors, easy to change white balance

jpeg -- 8-bit colors, not so easy to change white balance

[poilu]

if the jpeg have been well processed, it could make sense
but you can only fix colors, sharpening and noise from what you see on your lcd
the 60% jpeg compression look great on your lcd but how it will look on a better lcd
when you will get a better lcd, you will have some surprises
your pics will be over sharpened
artifacts that you didn't see on the old lcd will be visible
your colors will be dull as the srgb color space
your colors will be off and too much saturated
highlight and shadows will be clipped
noise reduction will be to much or not enough
you will not be able to use new noise reduction algorithm
pics will be blurred because the new monitor have bigger resolution that the jpeg pic
CA and bloom that you never see will appear
... and it will be too late to fix it

[jito]

Ok,
as I suspected, there was a an important piece of information I was missing. You guys are telling me that present time cameras capture more color detail than any current jpeg standard supports. This comes with a bit of a surprise to me.

Well, in that case the reason is simple. But to be fair, strictly speaking, the format has nothing to do with it after all.

[William]

RAW editing is non destructive. You create additional files with the editing information in and so do not modify the original file. So long as it is kept safe, as with negatives, you always have the original to come back to.

JPEGs created by the camera have a fair bit of processing applied straight after capture, some of which you may not necessarily want and may be destructive. If you want instant results then these JPEGs will usually look a lot better than the plain RAW files but after processing that changes and the processed RAWs look the way you choose. There's a lot more leeway in your processing than with JPEG. Also you do not get the artifacts when saving your images if you use a lossless format and normal practice is to only save in a lossy format at the very end when outputting.

It seems crazy to me to throw away a large amount of the data that your sensor has captured, especially as with the exception of super high resolution cameras RAW files are not that big. About 50,000 of mine will fit on a 1TB drive. Granted you could fit many more JPEGs but 50000 is several years worth even for a very heavy shooter.

[fuzzywuzzy]

Don't forget how ancient the JPG standard is, i.e. early 90s.

I kind of resisted shooting everything in RAW because of disk space concerns, but I'm now a total believer in RAW.

I think everyone has to discover it for themselves. Find a tricky light situation around the house this evening and take a shot in each format, and see how well each can be rescued in post.

[poilu]

[Sevo]

[ForenSeil]

I shoot 97% in JPEG format because I hate that editing of RAW files eats a lot time.

But under some situations when I see that the automatic white balance in the camera is not working correct and I have no time to set it manually I always set the camera to RAW as there it's no problem to set the white balance afterwards.

JPEGs gives sometimes ugly results when I try to change the white balance afterwards.

[ludoo]

From an engineer to an engineer, as someone already said above it's not a difference in format, but in quality and quantity of data: RAW encodes all the information received by the sensor, as raw data. Jpegs are the result of applying layered functions in a pipeline on that data, to convert it to a user defined color space (eg srgb), apply a contrast curve, apply color transformations (eg saturation), filter noise, etc. What you are left with is not the original data, nor a subset of it, but the result of applying all the functions in the above pipeline. You input something, you get out something else.

Another reason to shoot raw is that all the functions in the pipeline are perfectible: if someone devises a radically new noise removal algorithm, or sharpening algorithm, if you shoot raw you can apply it on the original raw data with all its bits of information intact even years later.

As an example, consider noise removal and sharpening, two functions which often depend one from the other: change noise filtering characteristics, and you have to change sharpening amounts to get the best possible result, or viceversa. If a new noise removal function comes up in the future, the sharpening you (or the camera) used to output a jpg image cannot be undone, and the new function might not work at its best. With raw data, you always input the original information to the whole pipeline.

Of course, you could argue that a flat jpeg with neutral contrast, no noise reduction, no sharpening etc is essentially similar, and that's in fact what some of the people who shoot jpg have their cameras configured to output. But even then, you are discarding a huge quantity of information (bits per pixel), and applying the camera's jpg engine version of "neutrality" for wb, colors, etc.

[Orio]

It's not only about the colours.
RAW file format records the luminance equivalent of four stops of image information.
If you can find a JPG image that records that much...

[yoyomaoz]

This is an old thread but I thought it was worth me having my say. I used to shoot JPEG for convenience and because I too was not convinced about the merits of shooting RAW. But I soon began noticing the deficits of jpg images processed in camera.

First, I noticed the presence of JPEG "artifacts" in my images. JPEG is a "lossy" compression algorithm that discards information in order to save space and reduce file size. But in doing so it will inevitably cause some blocking and loss of detail in the fine image structure. This is often evident as blotchiness in areas of single color such as the sky. If the JPEG images are saved at the highest quality level this may be less noticeable as there is less lossy compression and less image detail loss. But if you are a pixel peeper you will probably be unhappy with the result even then. (I do not class myself as a pixel peeper but I can say I certainly am unhappy with artifacting from JPEG compression).

Second, in many shots there were blown highlights. This happens naturally in outdoor shots due to the inability of sensors to cope with large variations in brightness in the same image. And it is caused by the way in which image sensors operate when too much light hits a sensor site - the data for that site goes abruptly to pure white (a value of 255). An analogy that works for me is that of a bucket sitting out of doors in the rain. As more and more rain falls, the bucket begins to fill up until it is full and overflows. Now imagine that bucket is an electronic sensor site, and the rain is comprised of photons of light. When the sensor site is overfilled with light it conveys that to the processor as having pure white value. Because the JPEG algorithm is lossy, some data is lost in translating the file to JPEG and this accentuates the above problem rendering it unable to be recovered by software. Meaning if you store the result as a RAW file, there is more scope to regain some detail in the blown areas (but not all as once it is gone, it is gone though some residual detail can still be present but occult / hidden) than you could if you saved the image as a JPEG. I find for example that Lightroom is about the best software there is to recover some detail which might otherwise be lost when this happens but only if some detail is still there in the file. And this will only happen with RAW files.

Finally I started post processing all of my images seriously as I acquired better skills to do so. RAW files are bigger files (and RAW images always have much bigger files than comparable JPEG images) meaning there is more data to manipulate and much greater scope to produce good looking final images. The more you process an image the more data is lost. Every time you open it in an editor and resave it as a new JPEG the resulting 2nd, 3rd, 4th etc generation image will lose more details. If you post process you really should save images as RAW files. Without any doubt whatsoever if you know what you are doing you WILL get far better results from your processing. Save the image as RAW in camera and on your computer, process it then save it finally once as a JPEG and I guarantee that other things being equal you will have far better results

The JPEG algorithm is useful in a couple of respects. JPEGS are ubiquitous. Every computer device can handle them. That is an advantage. They are also much more compressed / smaller than RAW files of the same image. This used to be a huge advantage 20 years back when computer storage was still comparatively expensive. Now days you can have a multi terabyte hard drive and gigabyte sized camera memory cards for cheap. Storage cost is just not an issue that is worth sweating over. Finally jpg images are useful for sharing. If you wish to email an image a small JPEG is better for obvious reasons though even there this advantage should erode over time as bandwidth grows. Finally if you are not all that "picky" about your images and especially if you do not post process or minimally post process you can get away with shooting in JPEG. In which case "go for it"

I cannot answer the OP's question as an Engineer, but take it from me, my real world experience is that I just find that if I ever do shoot only as JPEG, I am dissatisfied with the final result. I guess I have become more discerning about my images.

[cooltouch]

Since I got my first DSLR over 10 years ago, I have shot raw, in the beginning because I just figured that raw images were likely to hold all the detail there is possible to hold. But as my post-processing skils improved I began to appreciate one aspect of shooting raw, and that is the ability to control EV. With jpgs, I've found I can bump an image's EV value by *maybe* 1EV before all sorts of bad things begin to happen to the image. But when I manipulate raw image EV values, there is no sacrifice in quality.

About raw image sizes, I find it's all relative. Yes, you get smaller files at a price -- losing data -- with jpgs, but when I first got into digital imaging, it was converting slides to digital images, and my Epson scanner was defaulting to .tiff files, which I stuck with because they are not lossy. Ever saved to .tiff files? A file that may be 10MB raw and 3 or so as a jpg may be 80 or 100 MB as a .tiff file. I was filling up my hard drive fast back then and started burning my slide tiffs to CD, then DVD, and was then burning up huge numbers of disks. So it's all relative. And then, after filling up all those disks, I discovered the little used png format, which is not lossy, and which has file sizes that are comparable to raw files.

It's worth noting at this point that not all raw converters are equal. The one that came with my Canon DSLR was pretty good. The one that comes with Paint Shop Pro -- the photo manipulation software I use the most -- is just okay. The one that came with my Sony NEX was simply awful. But the one that comes with Photoshop is, hands down, the best I've ever used. In fact, Photoshop's raw converter is so full-featured that often when I finally convert the image over to jpg (most often), there's little if anything, left to do.

When it comes specifically to image exposure, there's one thing I've noticed I can do with slide images that I haven't been able to do with anything else. If I use my NEX or Canon DSLR to duplicate a slide, and let's say it's a severely underexposed slide, I can often extract very large amounts of detail from portions of an image that appear to be almost pitch black. It's quite remarkable how much detail I've been able to pull out of slide dupes. But that's only for shadow areas -- to use yoyomaoz's analogy, once that light bucket has gotten full up with photons, there's no room left for any shadow detail at all cuz it's all photons. All 255s. Well, it's exactly the same with slides. Once they're blown, that's it. There's no more information to be gleaned from a slide that has become transparent because of too much exposure. So, when I'm duping slides, I tend to be a big fan of dimmer slides because I can almost always extract a maximum amount of info from it. But only if I'm shooting raw and using Photoshop's raw converter.

I've looked at the possibility of saving my images as raw in Photoshop and Paint Shop Pro -- maybe I've just missed this capability, but I've never found it. However, both Photoshop and Paint Shop Pro have native, non-destructive formats (psd and pspimage, respectively) that even remember the editing that has been done to an image after it's been saved and called back up again, so these native formats are just as good as raw.

[tromboads]

As someone mentioned before.

JPEG = 8bit
RAW = 12, 14, 16bit these days

If you don't get why bits matter, it doesn't matter. If you think of it as stops, or shades, it might help you understand why it matters.

Remember;

8bit, 2^8 = 256
10bit, 2^10 = 1,024
12bit, 2^12 = 4,096
14bit, 2^14 = 16,385
16bit, 2^16 = 65,536

Now think of a black or white pixel. With an 8bit file you have 256 possibilities, (per pixel) from black to white. With a 12bit file you have 4,096 possibilities, or 4096 shades between white and black.

Now add colour, for a 12bit image, that 4,096 possibilities per colour channel (red, green, blue)

8 bit = 256 * 256 * 256 = 16,777,216 Colour possibilities
12 bit = 4096 * 4096 * 4069 = 68,719,476,736 Colour possibilities

That's a lot of information you can discard!

Depending on the image, it might not matter, but I hope now you can decide if it matters or not now

[tromboads]

Gah, this re bit depth is probably a bit clearer.

https://www.youtube.com/watch?v=6yXYxp0UiVg

[tromboads]

Oh... Sorry OP. I gather you understand all that from re reading your question

I agree with others whom comment about ease of colour grading, shadows, highlights, etc.

[yoyomaoz]

cooltouch wrote:

Since I got my first DSLR over 10 years ago, I have shot raw, in the beginning because I just figured that raw images were likely to hold all the detail there is possible to hold. But as my post-processing skils improved I began to appreciate one aspect of shooting raw, and that is the ability to control EV. With jpgs, I've found I can bump an image's EV value by *maybe* 1EV before all sorts of bad things begin to happen to the image. But when I manipulate raw image EV values, there is no sacrifice in quality. It's worth noting at this point that not all raw converters are equal. The one that came with my Canon DSLR was pretty good. The one that comes with Paint Shop Pro -- the photo manipulation software I use the most -- is just okay. The one that came with my Sony NEX was simply awful. But the one that comes with Photoshop is, hands down, the best I've ever used. In fact, Photoshop's raw converter is so full-featured that often when I finally convert the image over to jpg (most often), there's little if anything, left to do. When it comes specifically to image exposure, there's one thing I've noticed I can do with slide images that I haven't been able to do with anything else. If I use my NEX or Canon DSLR to duplicate a slide, and let's say it's a severely underexposed slide, I can often extract very large amounts of detail from portions of an image that appear to be almost pitch black. It's quite remarkable how much detail I've been able to pull out of slide dupes. But that's only for shadow areas -- to use yoyomaoz's analogy, once that light bucket has gotten full up with photons, there's no room left for any shadow detail at all cuz it's all photons. All 255s. Well, it's exactly the same with slides. Once they're blown, that's it. There's no more information to be gleaned from a slide that has become transparent because of too much exposure. So, when I'm duping slides, I tend to be a big fan of dimmer slides because I can almost always extract a maximum amount of info from it. But only if I'm shooting raw and using Photoshop's raw converter. I've looked at the possibility of saving my images as raw in Photoshop and Paint Shop Pro -- maybe I've just missed this capability, but I've never found it. However, both Photoshop and Paint Shop Pro have native, non-destructive formats (psd and pspimage, respectively) that even remember the editing that has been done to an image after it's been saved and called back up again, so these native formats are just as good as raw.

I quite like aspects of Paintshop Pro but as you say some aspects can be mediocre - I would argue that the RAW converter in particular is mediocre - it has what I would regard as very limited ability to fix blown highlights. Lightroom is much better in this respect. I used to use this software almost exclusively with the Nik Suite of plugins running under it. But I found that partly due to the above issue and partly due to how slow PSP can be to convert and process multiple images compared with Lightroom it was advantageous to swap over to Lightroom as my main editor. Lightroom is excellent in that you can run Nik plugins under it and also run PSP or Photoshop more or less as plugins as well. Thus, I normally start my workflow for any image in Lightroom, undertake basic adjustments including pulling back blown highlights and enhancing detail in shadows where needed then I will drop into a Nik plugin or into PSP to undertake more advanced processing should this be required. For example PSP has layers which Lightroom does not. This can be an advantage in some situations. Then when completed processing in PSP (or Nik) its simply a matter of clicking "save" and the program exports the edited file back into Lightroom for you to complete any further work that may be needed in that software. Sounds complex but its really simple and allows me to use the best processing software for the task in hand more or less seamlessly.

As regards blown highlights specifically one tool in Nik Color Efex I like to further help with this is the Tonal Contrast Filter. This filter works just like a "clarity" slider in Lightroom but it allows you to adjust Highlights, Midtones and Shadows independently of each other. This means that you may be able to use it to enhance some of the hidden detail in highlights by adjusting the highlight slider in this filter Or you can do the same for either of the midtone or shadow ranges that is needed. Just be careful not to overdo it as it can produce some unpleasant artifacts - and you end up with an image that looks like an overdone HDR image. This technique works best if you have already used Lightroom's highlight slider first as this gives the Tonal Contrast Filter something to work with and enhance.

As you say I am often also amazed at how much detail can be pulled out of shadows that appear to be black. Lightroom in particular also has sliders for adjusting shadows and blacks and using these sympathetically with its highlight slider can work miracles. As a result I now feel confident, when shooting in highly contrast situations, in shooting my RAW images perhaps a stop or even more under what is metered by the camera to minimize blown highlights, as so much information that would otherwise be lost can be made evident.

Of course, this technique really only works well with RAW images as the starting point.

[MaddisonAllen]

I always use the RAW format. This is the most high-quality and easy-to-handle format. Here you can see a good example of the design of the article using a quality background. Of course, such a three-dimensional image should not be displayed on the site, because the speed of the site will decrease, but in the design business you need to use only high-quality formats.