Raw vs Jpeg

[leprechaun]

I don't shoot RAW as I am not really into a lot of PP.

http://www.michaelfurtman.com/jpeg_myths.htm

Hmm, not getting a hyperling so you have to copy and paste to get to the site.

The Real Truth About JPEG images by Michael Furtman

I came across this web site that had a very good explanation to the difference between RAW & JPEG.
Some interesting reading.

[Black Pearl]

Whenever I read something like 'The Real Truth...." I am immediately sceptical of anything the author follows it with. So far I yet to be proved wrong in that assumption and this article isn't going to change my run of luck.

[Mike Buckley]

I have often considered starting a thread that debunks the many myths of photography. (There are almost as many of them about photography as about wine.) Naturally, the myths debunked in Furtman's article would be included. I never started the thread for fear that the discussion would become emotional to so many and, thus, would become extremely dissatisfying.

[RandyRFJP]

Thanks, Pat, for posting the link to this article. I enjoyed reading it.

[leprechaun]

All your comments are welcome.
It is not intended to be a discussion on raw vs jpeg as this article is just one mans opinion. However I found it interesting comparing the examples between the raw and jpeg photos. I was trying to understand a little more on the raw image side. I understand that raw is like the old film neg and can be copies and mucked about without loss or affecting the original. Same can be done wit jpeg.
When I got my first digital P&S (Konica/Minolta)years ago it did not do raw and when I progressed to a dslr the basic editing programs I had did not convert raw images so I stayed with jpeg. Of the thousands of photographs I've taken over the years, vety few see the light of day and the few that I have printed was on a 8x10 format and as long as they were sharp, in focus and looked good to me I was happy.
Now that Adobe Photoshop CS5 is available from Adobe as a free download (no support of updates though) I may download it and play around with raw.
Just trying to educate myself on the various formats in digital photography. I just wish I was as good and as creative as some of the contributors to CiC are, well maybe some day!
Have a fun time shooting everyone.
Pat

[ajohnw]

It's a reasonable link really but omits the advantaged of shooting raw. Some of them are real but a lot depends on what the camera manufacturer has decided to offer in it's conversion process and also if those suite.

I can't really imagine how people could think that they loose significant info all of the time but would guess that goes back to times when the quality levels were set rather low. My jpg's sometimes get bigger when they are edited and saved so I must be inventing information.

John
-

[Shadowman]

I shoot both and my only "concern" with jpeg is the loss of quality (author mentioned and suggested it is minimal) and the problem with WB adjustments, other than that there is no debate. I will say that I often try to minimize any loss by converting the edited jpeg file to TIFF format, don't know if that eliminates the quality loss but it makes me feel comfortable.

[DanK]

IMHO, the author has some valid points but misses the boat in several respects. And I share Robin's response to "The Real Truth."

He's right on the money when he says that film:slides is somewhat similar to raw:jpeg. As someone who has shot literally thousands of slides, the problem with slides is precisely what he refers to: virtually no ability to edit. That's no argument for shooting jpeg, it seems to me. I would have loved to have had more control when shooting slides. I shot them in part because I couldn't afford to set up a color darkroom, which is much more demanding than the B&W darkrooms I used.

To some extent, he is shooting at a straw man. The biggest disadvantage of jpeg is not the loss of detail, which he takes great pains to show isn't large. The main issue is the reduction in flexibility in editing the image because of the loss of information. In many cases, it won't matter. In many cases, it will. The article would be more informative if he explained conditions under which each of these is true, rather than simply brushing aside most of the ones in which raw is superior.

Personally, I think this is much ado over very little. When I first converted to digital, I was apprehensive about shooting raw because I thought it would be difficult. So, I shot raw+jpeg for some time until I realized that I never used the jpegs. While sophisticated postprocessing is difficult to learn, the basics aren't all that hard, and the processing algorithms from which one chooses in selecting a picture style and shooting jpeg are not sophisticated. If I shot weddings or sports and needed huge numbers of photos quickly, I would probably shoot jpeg, or raw+jpeg. However, for my uses, raw isn't enough additional work to make it worth thinking about shooting jpeg.

Your mileage may vary.

[proseak]

If you use third-party lenses, shooting jpeg may not include the corrections for distortion, CA etc.

[Shadowman]

One other comment I'll add that goes with the tenet of "the best camera is the one you have with you" and sometimes that camera only shoots jpeg; so sometimes we are forced to shoot the format and often we are happy with the results.

[ajohnw]

Several comments about jpg's miss the point in some ways in my view.

When they are shown on a screen or in an editor or via any other software they are no longer jpg's. They are 3 separate colour images and can be manipulated just as any image can be that is split into 3 colours. What raw offers is what might be called hidden bit depth that can be bought into the colour space of the output at the expense of no longer being able to see what was there. That may or may not matter as it just adds a closer degree of control but is still stuck with the same output space. The original jpg will normally be left untouched - or is in my case so can be reworked again. Editors should save processing steps so work can be back stepped just as it can with raw. That depends on how a particular editor works. GIMP native files can be absolutely enormous due to that aspect. Rawtherapee keeps processing steps. PS should be the same even working on jpg's.

Loss of quality is another aspect. I always save a 95%. I think my cameras now work to a similar level. I could save at 99%, some I think allow 100%. All I ever see if I re edit an image is the changes I have made and as mentioned the files can get bigger. Last time I was around I think Manfred mentioned that he had re edited the same jpg a number of times and didn't notice any problems at all.

When people use raw they often use a camera profile. The camera does too. This is what sets just how the colour space from the sensor gets distributed into the output space. It's much like curves except the axis have different bit counts, say 14bit depth on one and 8 or 10 on the other. Adobe offer several for just about all normal cameras. Cameras these days all offer several of them as well. I wonder how many people have tried them? My favourite for jpgs for editing is one that stuffs more or less the lot in - rare. Jpg's do sometimes need a different pp approach. Some raw converters appear to produce better results than others. Cameras differ too.

Raw for some people is just a case of setting up a batch flow and walking away. That can often be fine for a professional for a lot of their work as they have already established the adjustments that they need for particular types of shot. In some cases with the right settings their cameras might be able to achieve the same thing and another batch flow might turn out identical results. Maybe even with the cameras default as supplied settings. Even when it's set in auto.

Digital camera colour fidelity is generally crap compared with even the monitor settings a cheap colorimeter will achieve so what white is white etc and in many cased fashion and taste sets just what the final results look like and that is seldom reality. Increasingly so as digital gets older and older.

Each to their own in my view. There are arguments both ways but often the reasons for not working in jpg are rather over stated.

John
-

[george013]

All your comments are welcome.
It is not intended to be a discussion on raw vs jpeg as this article is just one mans opinion. However I found it interesting comparing the examples between the raw and jpeg photos. I was trying to understand a little more on the raw image side. I understand that raw is like the old film neg and can be copies and mucked about without loss or affecting the original. Same can be done wit jpeg.
When I got my first digital P&S (Konica/Minolta)years ago it did not do raw and when I progressed to a dslr the basic editing programs I had did not convert raw images so I stayed with jpeg. Of the thousands of photographs I've taken over the years, vety few see the light of day and the few that I have printed was on a 8x10 format and as long as they were sharp, in focus and looked good to me I was happy.
Now that Adobe Photoshop CS5 is available from Adobe as a free download (no support of updates though) I may download it and play around with raw.
Just trying to educate myself on the various formats in digital photography. I just wish I was as good and as creative as some of the contributors to CiC are, well maybe some day!
Have a fun time shooting everyone.
Pat

I know many people here will be angry at me showing this diagram again. It's just helpful to de-mystify the question raw vs jpeg.

Important is to keep in mind that all your work, except the conversion, is done on a RGB-raster image. When you view an image, you view the RGB-raster image. The raw-file is an image of the sensor data. In the converter a RGB-raster image is created. When saving to disk you have some choices, one of them is the jpeg format.
The first time you can see a raw file is in the converter which creates a RGB-raster image in pc memory and sent to the monitor.

So what are the big differences between "raw" and "jpeg"? Bitdepth and compression. The out off the converter created RGB-raster image has a bitdepth of 12 or 14 bits, depending on the camera and its settings. A jpeg is always a compressed 8-bits RGB-rasterimage saved in a special diskfile structure.

Converters do have a lot of PP-tools nowadays. All of them use the RGB-raster image. I'm only still not sure about the WB corrections.

The advance of a higher bitdepth is that in PP the changing of colors etc. is smoother.

Working on a jpeg means you start with a 8 bit image which has been compressed once when it was written to disk.

Since I started in 2008 with a DSLR I saved all my pictures in raw. I'm now able to correct those images with my knowledge I've now. And can re-correct them as many times I want. That's for me the advantage of raw.

George

[Rent]

Now that Adobe Photoshop CS5 is available from Adobe as a free download (no support of updates though) I may download it and play around with raw.

Pat How do you get Adobe cs5 for free ? Iv'e got a free cs2 but can't find a free cs5
Thanks Roy

[DanK]

I realized later that I missed something about his post. I wrote that his comparison of the detail in a raw and jpeg shot missed the point. I realized later that it is not even meaningful, and he gave the reason himself. If you look at this comparison of the "raw" and jpeg images, he correctly notes:

actually it's a jpeg from the RAW

So what he is really showing is that the level of detail in two jpegs--one converted from raw by processing in the camera, and the other converted by raw by his software--is essentially the same. Hardly a surprise. I don't think I have ever heard any one argue otherwise. The issue is how much information is lost before you get access to the file yourself.

[xpatUSA]

I shoot both and my only "concern" with jpeg is the loss of quality (author mentioned and suggested it is minimal) and the problem with WB adjustments, other than that there is no debate. I will say that I often try to minimize any loss by converting the edited jpeg file to TIFF format, don't know if that eliminates the quality loss but it makes me feel comfortable.

I would add that there are more options to JPEG files than many people realize.

For example, FastStone Viewer's (FSV) JPEG save options are:

Photometric: RGB, Grayscale, YCbCr, CMYK, YCbCrK

Sub-sampling: none, medium, high

Of course, up to 100% Quality

Not claiming that one's Viewer or Browser can cope with all combinations of the above but they are legal.

The issue is how much information is lost before you get access to the file yourself.

So, if I opened a raw file in FSV, selected JPEG options of RGB, no sub-sampling and 100% quality, very little "information" would be lost at all. Probably couldn't tell the difference from a non-compressed TIFF or PNG!

[george013]

I realized later that I missed something about his post. I wrote that his comparison of the detail in a raw and jpeg shot missed the point. I realized later that it is not even meaningful, and he gave the reason himself. If you look at this comparison of the "raw" and jpeg images, he correctly notes:



So what he is really showing is that the level of detail in two jpegs--one converted from raw by processing in the camera, and the other converted by raw by his software--is essentially the same. Hardly a surprise. I don't think I have ever heard any one argue otherwise. The issue is how much information is lost before you get access to the file yourself.

The jpeg is made from the RGB-raster image, either in-camera or out-camera. That's the essence of jpeg-compression.

George

[ajohnw]

So what are the big differences between "raw" and "jpeg"? Bitdepth and compression. The out off the converter created RGB-raster image has a bitdepth of 12 or 14 bits, depending on the camera and its settings. A jpeg is always a compressed 8-bits RGB-rasterimage saved in a special diskfile structure.

The advance of a higher bitdepth is that in PP the changing of colors etc. is smoother.

Working on a jpeg means you start with a 8 bit image which has been compressed once when it was written to disk.

George

The output you see when a raw converter is used to display an image is generally a 3 x 8 bit rasters one for each colour or 10 bit if some one has a complete 10 bit work flow right up to the monitor. The same is true of a jpg. It will normally be displayed on 3 x 8 bit rasters. All that happens with a raw file is that it's bit depth is compressed in some way to make it fit into the output colour space. How that appears on the screen is purely dependent on the camera profile that is used so in many cases it will be one of the ways Adobe feels is suitable or maybe Capture One or the same for any other raw converter even the ones built into the camera that produce the jpg's.

There is no need to even mention raster displays really. The important part is that colours are formed from 3 8bit channels so that limits the dynamic range that can be achieved and the number of possible colours. It also limits the steps in tones that can be reproduced on the screen. That aspect doesn't change when raw is used.

This page shows what light levels monitors can actually show - if looked at rather carefully. LCD / TFT panels are off to a bad start due to back lighting and the bright end has other problems. Many displays out there would be completely incapable of showing rather a lot of the split squares. If shots are for general consumption that aspect limits the tonal range that can be used. Some argue that gamma that makes the whole thing work also limits the achievable dynamic range. Calibration can too as maybe blue or some other colour is too powerful so always has to be mixed in at a reduced level.

https://www.drycreekphoto.com/Learn/...nsitivity.html

It's a pretty sobering image to look at. We all generally get round the problems by messing with colours and tonal range but the output resolution steps in both always have to be the same. Raw can offer more control of colour and tone after a fashion but only by shifting it into a colour or tone level that the display can actually show. Where it traditionally comes into force is in recovering shadows and highlights that might have been blown in the jpg but only from moving them from where they were to somewhere that can be displayed on a monitor - or in the case of say prophoto a colour that can be printed.

Moving the dark stuff up to visibilty - well D-Light does that and some camera curves compress the highlights and the dark end. As the camera does this on jpg's they can be manipulated further by editing,

John
-

[george013]

The output you see when a raw converter is used to display an image is generally a 3 x 8 bit rasters one for each colour or 10 bit if some one has a complete 10 bit work flow right up to the monitor. The same is true of a jpg. It will normally be displayed on 3 x 8 bit rasters. All that happens with a raw file is that it's bit depth is compressed in some way to make it fit into the output colour space. How that appears on the screen is purely dependent on the camera profile that is used so in many cases it will be one of the ways Adobe feels is suitable or maybe Capture One or the same for any other raw converter even the ones built into the camera that produce the jpg's.

The difference is not what you see but how you edit the image. Editing a raster image converted directly out off a raw file is 12 or 14 bits. That's where the converter is calculating with. What you see is what your output device is capable for. When no editing has been done, you don't see a difference between the "raw""and "jpeg".

There is no need to even mention raster displays really. The important part is that colours are formed from 3 8bit channels so that limits the dynamic range that can be achieved and the number of possible colours. It also limits the steps in tones that can be reproduced on the screen. That aspect doesn't change when raw is used.

There is no relation between the dynamic range and the bitdepth.
I think you are talking about monitors, but that isn't the subject of this thread.

This page shows what light levels monitors can actually show - if looked at rather carefully. LCD / TFT panels are off to a bad start due to back lighting and the bright end has other problems. Many displays out there would be completely incapable of showing rather a lot of the split squares. If shots are for general consumption that aspect limits the tonal range that can be used. Some argue that gamma that makes the whole thing work also limits the achievable dynamic range. Calibration can too as maybe blue or some other colour is too powerful so always has to be mixed in at a reduced level.

https://www.drycreekphoto.com/Learn/...nsitivity.html

It's a pretty sobering image to look at. We all generally get round the problems by messing with colours and tonal range but the output resolution steps in both always have to be the same. Raw can offer more control of colour and tone after a fashion but only by shifting it into a colour or tone level that the display can actually show. Where it traditionally comes into force is in recovering shadows and highlights that might have been blown in the jpg but only from moving them from where they were to somewhere that can be displayed on a monitor - or in the case of say prophoto a colour that can be printed.

Moving the dark stuff up to visibilty - well D-Light does that and some camera curves compress the highlights and the dark end. As the camera does this on jpg's they can be manipulated further by editing,

John
-

George

[leprechaun]

Hi Rent,
I haven't tried to download it yet. I was browsing the web looking for the best free photo editing software and came across a photo site that mentioned that now that adobe Photoshop was in the cloud and no longer distributed that Adobe was allowing free down loads.
Ops!! Sorry. just looked at my scribble note and it wasCS2. Cant read my own scribble, The old eyes are not as good as they once were.

Boy, I sure created a lot of discussion on this topic. I sure am learning a lot about RAW. Thanks all.
Pat

[xpatUSA]

The important part is that colours are formed from 3 8bit channels so that limits the dynamic range that can be achieved

A popular misconception, John, depending on your personal definition of "dynamic range" of course.

Some people, not necessarily your good self, think that the dynamic range of 3 8bit channels is only 255/1 = 8EV. On the other hand, if your monitor has a 1000:1 contrast for 0-255 then you will see a dynamic range (in cd/m2) of 10EV - albeit in the limited number of steps that you mentioned.