Does "Gamma Correction" harm JPEGS quality?

[zereshk]

Hi. I am trying to understand "gamma correction" . I tried to simulate gamma correction process with photoshop but the result was terrible. first I opened a raw file and applied 2.2 gamma and saved it then opened it agian and applied 0.45 but result as I said was not good. Now could someone pleas tell me if gamma correction harms image quality or not?

[Shadowman]

Question is why are you applying gamma correction? The Photoshop utility uses your eyes to make the adjustments, why are you applying it twice to the same image? It doesn't harm the image per se but the adjustments you make affect what you see. The 2.2 setting is often used with a specific white balance (e.g. 6500 K) and a specific luminance. I think you'd get better results if you calibrate your monitor using a program such as ColorMunki, after applying the changes the software will adjust the monitor as the time of day changes.

[dje]

Now could someone pleas tell me if gamma correction harms image quality or not?

Hi zereshk

I think the simple answer to this question is "not really".

When you edit a raw file in Adobe Camera Raw (ACR), which is the raw development plug-in for Photoshop, all editing is done in a linear data space (preferably in 16 bit rather than 8 bit).

It's only when the file is sent to Photoshop proper and converted to a jpeg that gamma encoding is "burnt in" to the image. There is actually some benefit when gamma encoding is applied to a 16 bit file before conversion to 8 bit (for a jpeg) as this causes an effective re-distribution of sample points to the lower end of the tonal range - where more tones are useful in minimising visual banding.

Don't get too hung up on gamma, it's a complicated subject but it's all managed seamlessly in Photoshop and the colour managemnt software associated with your display.

Dave

[Rent]

zereshk I suspect that this a typical case of "I've got this I must use it" , don't, What is wrong with your RAW file or your JPEG for that matter that you think it needs altering ? You will gather that I am anti post processing as far as possible. The camera should be able to give a reasonable result. Try to keep everything simple. ( Please don't take any offence at my remarks, they are made with the best intentions )
Roy

[xpatUSA]

<> You will gather that I an anti post processing as far as possible. The camera should be able to give a reasonable result.

Been thinking that way myself lately. At least to the extent that it would be nice to not re-sample (re-size) the captured image. Easy for me because I don't print and my Sigma cameras are not greatly larger than my monitor (pixel-wise) e.g. 2268x1512px vs. 1920x1080px. Therefore, a quick crop and sharpen of the perfectly composed and exposed capture - job done!

In theory that is . .

[Manfred M]

You will gather that I am anti post processing as far as possible.

I'm not, because the only time I am truly happy with my shoots straight out of camera is when I shoot in a studio setting and I can control the light. I can get shots that are "good enough" if I'm not too picky, and most of my camera raw shots get the same basic processing that the camera does to create jpegs.

To each his or her own.

[Manfred M]

Hi. I am trying to understand "gamma correction" . I tried to simulate gamma correction process with photoshop but the result was terrible. first I opened a raw file and applied 2.2 gamma and saved it then opened it agian and applied 0.45 but result as I said was not good. Now could someone pleas tell me if gamma correction harms image quality or not?

Gamma correction is primarily to take the linear output from a digital camera and then to map it to the non-linear human vision.

I guess I'm trying to figure out why you are changing the gamma off the standard 2.2?

[Black Pearl]

You will gather that I am anti post processing as far as possible. The camera should be able to give a reasonable result.
Roy

I look at it this way - all digital photographs are processed to one degree or another so its not a case of being anti processing its a case of what degree of processing do you feel happy doing yourself to ensure the image is your own and what degree are you happy to leave to the person who programmed your camera.

I suppose the camera should be able to give you a reasonable result but that leads to a few caveats; what is reasonable to one photographer won't be to another, what camera are you using as not all sensors have the same capabilities, what do you shoot as not all subjects are able to be captured in an uncontrolled way, was do you do with the final images as there is a whole world of difference between shots for a bit of fun and commercial work for a paying client.

[xpatUSA]

Gamma correction is primarily to take the linear output from a digital camera and then to map it to the non-linear human vision.

I guess I'm trying to figure out why you are changing the gamma off the standard 2.2?

My guess is an attempt at a "round trip", not clearly expressed, since 1/2.2 = 0.45. I'm not guessing as to why that would be done, though.

A snipe at Adobe - if their standard is 2.2 - that certainly ain't 2.4 + a straight bit (see sRGB, BTU.709).

[DanK]

I look at it this way - all digital photographs are processed to one degree or another so its not a case of being anti processing its a case of what degree of processing do you feel happy doing yourself to ensure the image is your own and what degree are you happy to leave to the person who programmed your camera.

We've wandered off the OP's topic, but I agree.

For those who think that it's somehow bad to deviate substantially from the processing the camera or the defaults of the editing program do to render the image, I recommend watching the video at this link. It's a discussion of how Ansel Adams created "Moonrise, Hernandez, New Mexico," which many people consider his most iconic image. At one point in the video, his son shows a print made from the negative with no further editing at all. It's dramatically different from the final print, and I doubt it would have been received anywhere nearly as well.

[Black Pearl]

We've wandered off the OP's topic, but I agree.

For those who think that it's somehow bad to deviate substantially from the processing the camera or the defaults of the editing program do to render the image, I recommend watching the video at this link. It's a discussion of how Ansel Adams created "Moonrise, Hernandez, New Mexico," which many people consider his most iconic image. At one point in the video, his son shows a print made from the negative with no further editing at all. It's dramatically different from the final print, and I doubt it would have been received anywhere nearly as well.

Thats a great link Dan

[Manfred M]

zereshk I suspect that this a typical case of "I've got this I must use it" , don't, What is wrong with your RAW file or your JPEG for that matter that you think it needs altering ? You will gather that I am anti post processing as far as possible. The camera should be able to give a reasonable result. Try to keep everything simple. ( Please don't take any offence at my remarks, they are made with the best intentions )
Roy

Roy - you thoroughly misunderstand cameras and processing. The raw data that your camera collects has to be turned into image data and that happens either in camera, using settings that the camera manufacturer has set for you (in the default settings) or in whatever post processing tool you happen to choose. When people refer to that "Canon look" or "Nikon look", that is exactly what they are writing about - the post-processing choices some camera design team has made for you and output as jpeg. Your camera lets you override these default settings if you want (read your camera manual, that's what some of those built in presets and some of the custom functions do for you). The easiest to understand example is the camera's B&W mode - you get B&W images based on a default conversion algorithm. If you shoot jpeg + raw, you can go back and do what you want with the raw data after the fact as the raw data has been preserved.

In the B&W wet darkroom days, we used to do something similar and that was through the selection of film. Agfa Agfapan 400 versus Ilford HP4 versus Kodak Tri-X would all give you a different look. I could process them in Kodak D-76, Kodak Microdol X, Agfa Rodinal, etc. and get different looking output on the same film stock. Different printing papers would give me different looks, just as using different paper developers would influence the final output. Negatives could be bleached and enhanced through post-processing, just as Ansel Adams did.

There simply never has been a "standard" approach to produce what our cameras show, so suggesting that people straight-out-of-camera is somehow superior is simply shows that you don't understand how digital imagery works. Suggesting that the PP group is somehow "cheating" shows a remarkable misunderstanding of what is really going on.

There is no right or wrong approach here - it's purely a personal choice. Remember, choosing the camera defaults is a choice too!

[Rent]

GrumpyDiver

There is no right or wrong approach here - it's purely a personal choice. Remember, choosing the camera defaults is a choice too!

I rest my case
Roy

[Shanghai Steve]

Roy - you thoroughly misunderstand cameras and processing. The raw data that your camera collects has to be turned into image data and that happens either in camera, using settings that the camera manufacturer has set for you (in the default settings) or in whatever post processing tool you happen to choose. When people refer to that "Canon look" or "Nikon look", that is exactly what they are writing about - the post-processing choices some camera design team has made for you and output as jpeg. Your camera lets you override these default settings if you want (read your camera manual, that's what some of those built in presets and some of the custom functions do for you). The easiest to understand example is the camera's B&W mode - you get B&W images based on a default conversion algorithm. If you shoot jpeg + raw, you can go back and do what you want with the raw data after the fact as the raw data has been preserved.

In the B&W wet darkroom days, we used to do something similar and that was through the selection of film. Agfa Agfapan 400 versus Ilford HP4 versus Kodak Tri-X would all give you a different look. I could process them in Kodak D-76, Kodak Microdol X, Agfa Rodinal, etc. and get different looking output on the same film stock. Different printing papers would give me different looks, just as using different paper developers would influence the final output. Negatives could be bleached and enhanced through post-processing, just as Ansel Adams did.

There simply never has been a "standard" approach to produce what our cameras show, so suggesting that people straight-out-of-camera is somehow superior is simply shows that you don't understand how digital imagery works. Suggesting that the PP group is somehow "cheating" shows a remarkable misunderstanding of what is really going on.

We have all heard those arguments before and some of us find them disingenuous and self-justifying, not to mention patronising.

Some PP is certainly cheating. Some of it has been shown to be incredibly harmful, e.g. fashion and cosmetics images causing anorexia and psychological problems. Altering images can lead to miscarriages of justice, riots or war.

[Manfred M]

We have all heard those arguments before and some of us find them disingenuous and self-justifying, not to mention patronising.

Steve - suggesting that my comments are disingenuous and self-justifying without offering anything to back up your comments suggests to me that I might have hit a nerve here.

If you have some information to suggest anything I have written is incorrect, please let me know specifically what is not correct. I suspect you will have a hard time doing so because it represents my personal knowledge and experience as a photographer who has spent a lot of time behind the camera and a long time working in both the wet and dry darkroom.


Let's look at three or four of my favourite photographers:

1. Ansel Adams - who would spend hours and days reworking his negative, dodging and burning his prints. Each of his images was a one-off custom print, even when done from the same negative. Most landscape photographers I know look to him as an example, in spite of manipulating his images both chemically and during the printing process.

2. Yousef Karsh - employed a negative retoucher and printer to produce those brilliant portraits. Examine his images closely and you will see hints that the final print had a lot of expertise in post processing applied to the final product. I don't hear him being called anything other than a great photographer, even though he had help in creating those masterful images.

3. Henri Cartier-Bresson - he just turned his rolls of film over to his favourite lab and let them produce the final image. He did admit to cropping (once) but there has always been a suspicion that some of his shots were staged. Much like Steve McCurry who has recently admitted that his images are not straight out of the camera...

While we are at it, perhaps we can discuss careful framing, using specific focal lengths, depth of field or other methods we can use to influence the viewer using in-camera techniques.

Just as an aside. I do post a lot of SOOC images. I have no issues at all with shooting jpegs, just like I have no issues working raw data.

[Manfred M]

Steve - having re-read your comment, I think we need to clarify our semantics.

When I write about post-processing, I include anything done to an image, even if all I do is resize it to ensure faster load times. The moment I use any editing tool to change the original image, whether it is a jpeg or I work a raw file, these are all PP to me. If you convert your images from a colour capture to B&W, that too is post-processed.

If you are looking at more significant edits, removing something in an image, or drastically changing the physical appearance, that is PP work as well.

If you are railing against misrepresenting the truth, especially if it leads to harm, I can sympathize with that position, but on the other hand, is this an absolute position or is it a relative one. Am I allowed to remove sensor dust, or is that cheating and I should be cleaning my sensor? What about removing a temporary facial blemish, like acne or perhaps a stray hair. I know people that feel removing acne using a post-processing tool is wrong, but have no issues of hiding it with makeup.

To me the issue is the ethics of misleading someone, and frankly, you don't need Photoshop to do that. Take a look at real estate advertising that make tiny apartments look huge, courtesy of using a wide angle lens, or pictures of a resort, where the camera is placed to ensure that we don't see the noisy, major highway running past the property or the open sewer that is just out of frame.

For now, I will leave the ethical arguments to the philosophers, and when they come to an agreement, I'll be prepared to listen, but until that happens, I will make my own judgments on these matters.

[Shanghai Steve]

Hi Manfred.
I have been thinking about this today and yes, there is a problem with such a broad term. Processing of course, technically means any change at all. But neither I nor Roy nor almost anyone else who is keen to keep photography as a record of what was in front of the camera at the time of capture, would object to, say, cropping or straightening.

However, manipulating images to misrepresent reality does cause psychological harm to millions of girls and women exposed to manipulative, un-natural images of already gorgeous humans who have been photoshopped beyond all possible human perfection, simply to increase profit for multi-billion dollar companies. This has been proven to cause human suffering and it is wrong. Unequivocally harmful and wrong.

For me and I believe for Roy and of course many other people all over the world, it is this misrepresentation which is the problem. The last time I replied to this thread, I was timed-out after spending a long time expanding on my argument, so I'll keep it at this.

[Rent]

Up to now nobody has answered the original post

[Shanghai Steve]

I just did a very quick experiment on Irfanview with a colour JPeg. I took the gamma down to ridiculously low levels and it looked very dark but OK, then I undid it and the file was fine. I hadn't saved it in the interim but I don't think it would have been a problem. On the other hand, I increased the gamma significantly and the file looked messy, with inconsistent, blotchy noise in most parts of the frame. So I wouldn't increase it unles I had to.

[george013]

I don't understand the original question. Jpg is a disk file.

And again this flow.



All editing is done on the RGB raster file in memory. This file comes from either the converter or from a diskfile. If there's a difference between loading a jpg-file again, it's due to the compression.
You can repeat this with TIFF, uncompressed.

George