JPG Degradation

[CatherineA]

I wondered if anyone else was shooting in JPG and, if so, whether they are concerned with image degradation. Are there easy steps to take to minimize the loss of quality?

[DanK]

What you mostly lose is two things. First, it is harder to edit a JPEG and easier to create artifacts like posterization. Second, JPEGs are either sRGB or Adobe RGB, which are smaller color gamuts than most camera's raw files can produce. sRGB is the smaller of the two, but if you are just going to display online, this doesn't matter, as most computers display in sRGB. If you are going to print via a lab, some will accept Adobe RGB.

On the cameras I have, there is nothing you can do to alter this, other than choosing sRGB or Adobe RGB. The loss of data occurs when the camera converts the raw image to a JPEG.

[CatherineA]

Thank you for this Dan!

[SSSdave]

Well no. I've only been shooting digital since 2014 when I switched from 4x5 film to the Sony A6000 system. So not familiar with all the early digital years when arguments for working in RAW had stronger validity. Newer camera system jpg outputs are reportedly much improved. I'm no authority on that. On other forums there are certainly good numbers of pros shooting mostly in jpg with comments they don't see enough of a difference for most subjects that it makes the extra work worthwhile. Then again there are other advocates that always shoot RAW that disagree though neither side of that argument is convincing.

I'm not shooting in Adobe RGB either but rather just sRGB. The intent of my landscape work is to make large images though I also do many single frame smaller images depending on subject like wildlflower close-ups that are better printed small. Landscapes are often much larger than the largest DSLR sensors deliver by using focus stack blending multi column row stitch blending processes. For instance last month shot using the high resolution Sigma 60mm F2.8 DN lens mostly at its best aperture F5.6 an 8 column by 3 row 24 frame composite that processed into 23400 by 12000 pixels.

When I do final processing of serious images with CS6, I save at quality level 11 or 12, loss less. And what do the outputs look like. Well on my 24 inch 4k Dell Ultrasharp monitor, a transmissive medium, incredible versus anything one is going to end up with on even highest quality printers, on reflective media, that is less than the sRGB gamut. But then again there is valid logic for a workflow with more bits before the output stage especially for smooth gradient subjects like blue skies. My intended output is the coming large 8k displays and from what I've already seen on my two 4k displays a lot of jaws are going to drop...looking at those jpg's.

[dje]

Are there easy steps to take to minimize the loss of quality?

Catherine probably the main things are to set the camera to the finest image quality and best quality jpeg compression. The exact terminology for these settings varies from one camera model to another. Also take care to avoid blown highlights as there is no scope for recovery of these with jpegs..

Dave

[MrB]

Are there easy steps to take to minimize the loss of quality?

At the image capture stage:

Get your camera to work for you by ensuring that it is set to record the largest image size (i.e. Megapixels) and the highest quality JPEG files.

The aim should then be to get the captured image as close as possible to your desired product. Select the Image Style that appeals to you (I set Natural or Standard for most shots), and there will probably be other settings that can be adjusted within that choice, e.g. contrast, saturation, sharpness. It is probably best to avoid the more extreme options as they might require more severe adjustments in PP and, of course, if you select Mono there is no going back to colour later for that image.

Make use of the camera's rear screen - "chimping" - it displays the JPEG image recorded by the camera.
Depending on the camera, the screen (or viewfinder) might also display the histogram for the image, before and/or after shooting. Similarly it might also flash in areas of the image that are under or over-exposed. Make use of these features to adjust the exposure settings, if necessary. Details lost by over-exposure of important bright areas (e.g. the clouds in a landscape) cannot be retrieved from a JPEG image file. However, if reducing the exposure makes the shadow areas rather dark (by 1 or 2 stops), they can be brightened in PP or, in close-by scenes, by lighting - e.g. using a reflector or fill-in flash.

Before any editing:

Make a new folder on your computer and in that new folder make copies of the JPEG files for editing. (I re-name them by adding a1 to the original file name, a2 if I make another copy of the same image for editing differently, etc.). Never edit the out-of-camera files - keep them as the back-up originals.

During editing:

The image will have lost some of the original image data when it was compressed to form the camera JPEG file. When the JPEG is loaded into a photo editor, it will be uncompressed by the computer software to produce the full image that you see on the screen, but minus the data originally lost during compression in the camera.

While making adjustments, that image is held in the computer's memory. When the image is then saved as a JPEG it will be compressed again, so some more data will be lost. If this cycle - load, edit, save as JPEG - is repeated, the loss of data will degrade the image quality some more at each iteration. Therefore, to maintain high quality, it is best to do all the editing in one session, so that the data is compressed only once more into the saved final JPEG. If you want to continue editing in different sessions, then save the image in a non-compressed file format, e.g. TIFF then, when finished, you can save it as a JPEG. Also to maintain image quality, in the editing software set the JPEG compression to low (or the JPEG image quality to high, e.g. at least level 10 in Photoshop).

There should be few problems editing well-captured high quality JPEGs, which are already fairly close to what you intended. Any extra noise or artifacts introduced by moderate adjustments will probably not be discernible by those who view images from sensible viewing distances rather than viewing details at 100% or more magnification. However, care is needed with areas of solid but slightly graduated colour, e.g. a blue sky. Some adjustments might lead to obvious colour-banding in those areas, so perhaps mask those areas off in Layers, or select only the area where you need to apply that adjustment.

Cheers.
Philip

[xpatUSA]

I wondered if anyone else was shooting in JPG and, if so, whether they are concerned with image degradation. Are there easy steps to take to minimize the loss of quality?

If one shoots JPEG only, it is worth finding out exactly what your camera gives you when you select "FINE" or "BEST" quality or whatever your camera calls it.

In the world of JPEG, part of the quality loss is due to the amount of compression applied - and JPEG compression is "lossy" which means that anything lost during a save is not regained when viewing the result, either on-screen or in the print.

The other part of the quality loss is more technical. The colors can be saved at half-resolution (chroma sub-sampling). Sub-sampled color has to be interpolated - again when viewing the result either on-screen or in the print. But a JPEG can also be saved with no sub-sampling which is the best quality offer by most editors. There are codes for sub-sampling: 4:4:4 (1 1) for best. 4:4:2 (2 1) for next best and 4:2:0 (2 2) for worst. They can be found in EXIF usually written as YCbCr = <the code>.

See also https://en.wikipedia.org/wiki/Chroma_subsampling

My cameras use "next best 4:4:2" if I select FINE JPEGs. That means lost color data, albeit not really noticeable on my screen. Which is why I shoot raw and use my editor to output JPEGs, because I'm a pedant.

Off-topic but little known to many of us is the ability to save a JPEG as RGB (no messing with chroma) and no compression. See FastStone Viewer.

[DanK]

My apologies. I don't generally shoot JPEG, and I neglected to put in some important information that others have added. In particular, I just assumed that you were using the highest-quality JPEG setting your camera has. Storage is cheap, so there is no reason to use lower quality/ smaller files.

It might be worth separating the good advice in this thread into two bins. One bin is comments that are not specifically about JPEG shooting and will apply even if you start shooting raw, e.g.,

Make use of the camera's rear screen - "chimping" - it displays the JPEG image recorded by the camera.
Depending on the camera, the screen (or viewfinder) might also display the histogram for the image, before and/or after shooting. Similarly it might also flash in areas of the image that are under or over-exposed. Make use of these features to adjust the exposure settings, if necessary.

The other bin is things specific to JPEG shooting, of which there are not many.

[CatherineA]

thank you SSSdave!

[CatherineA]

Catherine probably the main things are to set the camera to the finest image quality and best quality jpeg compression. The exact terminology for these settings varies from one camera model to another. Also take care to avoid blown highlights as there is no scope for recovery of these with jpegs..

Thank you, Dave. I've gone ahead and taken those steps for my JPGs....My camera can't shoot RAW, so I know even less about those images. I didn't know that you could recover blown highlights with them.

[CatherineA]

Thanks Philip! I'm just starting learning more about pp and this is really helpful. I had a trial with Photoshop and two images disappeared to skinny lines. I couldn't understand it but now wonder whether I edited them to death. I think the images would be left untouched by Lightroom but if I were to use Photoshop again then I would make a TIFF (something that would be new to me). At least I would make TIFFs as I learn to edit to cover the possibility of multiple sessions.

Thaks again,
I'll keep this on hand as I learn pp!

[CatherineA]

Thanks Ted for this information. I don't understand all the science presented here but you have given me the links to pursue. And I will also look into FastStone Viewer and whether it is easy to save a JPG without compression. (No matter how easy, I expect I would struggle with it but its worth looking into)>

[shreds]

Pro's I know who shoot jpeg files generally do so due to time pressures and the need to very quickly process and output shots, where speed has a commercial element to it. They probably are already proficient in taking shots that need little post production manipulation anyway. As a very good wedding professional I know often says to me…

'I don't want to spend hours sitting in front of a computer. I would rather try my best to get as strong an image captured first time, them I can get on with the next wedding'.

Similarly most commercial sports photographers are on a very tight shoot schedule and will send the shots back to the editor direct from the venue as they are taken and not even do any manipulation of the images themselves. These will usually be jpeg files.

A jpeg is already a compressed and reduced size file by its nature, based on information built into the software of the camera. Saving it as a TIFF will not increase the 'colour space' or the quality of the shot. It is a format however that does not reduce in quality because it is not compressed, unlike each time a jpeg is subsequently saved. Lightroom (and other some other softwares) are 'non destructive' in their editing process unlike Photoshop.

I personally would not bother saving any already jpeg format files as a TIFF, I only do so from RAW format files that my camera also produces.

For someone starting out with all these issues, I would suggest you don't get bogged down with the technicalities at this stage; you will pick these up as you move on in your photographic journey. A good rule from film and transparency film days was to 'try and get as much right in camera first time' so that your shots need little, if any, post shooting work on them. That way you will not have a need to edit and ultimately destroy your work by over-manipulation of the jpegs.

[CatherineA]

Thanks Ian, that makes good sense. I suppose at this stage jpegs are in a sense freeing for me - they strip down technicalities. I'll try to remember to work with that and not against it.

[Manfred M]

Catherine - I sometimes feel that the "downsides" of JPEGs are overblown. I take a somewhat more practical view of the whole subject. Consider the following:

1. Much of the data loss of going from an image capture (raw) to a JPEG is taking the 14-bit data and converting it to 8-bit data. Most people end up editing and displaying their image on an 8-bit per channel screen. Nicely said with higher bit data, much of the information is thrown away by the screen driver, so no one will ever miss it.

If one is using a higher end 10-bit per channel screen, then there can be a noticeable quality difference, but the percentage of 10-bit displays being sold is quite low when compared to the 8-bit ones.

2. Much is made of the "lossy" data loss, but again, their is a bit of misunderstanding as to what happens here too. It is true that data is lost, but for the most part the reason that the 4:4:2 is often used is that the human visual system does not discern a lot of difference between it and the "best case" 4:4:4 sub-sampling. Most HD video we watch uses 4:4:2.

The other issue with the data loss is that a lot of people seem to think this is cumulative. My testing and confirms what I have read is that this is not the case. There is no additional noticeable quality or data loss after the initial compression takes place.

For my purposes, which are unique to the way that I work, there is an advantage in using high-bit value data. I am someone who prints a lot. The printer and papers that I use have a wider gamut than my computer screen, so I can and do make use of the extra data in my work. I use the ProPhoto RGB colour space, and because is can handle many more colours than either sRGB or Adobe RGB, 16-bit is the recommended bit-depth for that colour space.

There are two main advantages to me to shoot raw. The first one is that my camera does not produce ProPhoto images as JPEGs. Secondly, if I shoot JPEG, the colour temperature / white balance is "baked" into the image file. I often tweak my colour temperature in post to get the look that I want. JPEG files do not offer that flexibility.

If you do not need to use a wide-gamut colour space and you are comfortable with the colour balance out of your camera and you mostly post on the internet; I don't see a lot of advantage for you to shoot raw at this point. On the other hand, if you suspect that you will switch over to using raw at some point as your skill level increase, I would suggest that you consider shooting JPEG + raw. You will have to store and backup more data, but storage is fairly inexpensive.

[xpatUSA]

I will also look into FastStone Viewer and whether it is easy to save a JPG without compression. (No matter how easy, I expect I would struggle with it but its worth looking into)

It's easy because the screen is easy to follow and has before and after panes and tells before and after file sizes and lots more. Better yet, the app is free (small donation appreciated).

I'll be glad to help with any questions ...

I can not recommend it for serious work, unfortunately. Files saved after editing do not have embedded ICC color profiles - that means that any browser will treat them as sRGB with it's own profile.

[pnodrog]

Catherine it is difficult to add to the advice all ready given without knowing how you intend to use your favorite photographs. When I was doing photography for real estate I used Jpeg for the bulk of photographs as it was convenient and all that was needed for the size they were going to be reproduced at. If a large billboard was required I would take all the photographs as RAW.

For my personal use in the hope that one day I will capture the perfect image I always use RAW. There is nothing more frustrating than having a potentially great image to find that one of the adjustments you would like to make is limited by the recorded file type, size or quality.

[xpatUSA]

For all those advising Catherine to shoot raw or implying that she should, her camera can't do it:

JPG Degradation
.

[CatherineA]

Thank you for your reply Manfred, it helps me understand a bit more of the physics. I had been reading that the data loss was in fact cumulative though and that is what prompted me to post the question. Perhaps it isn't though. I did something weird in my trial of PhotoShop to somehow turn two images into skinny lines but perhaps it wsn't over-editing that did it. I won't worry about it, I will just make a virtual copy should I venture
into PS in the future.

[CatherineA]

Thanks again Ted! And, I appreciate the heads up about the colour profile. That's a consideration that I haven't thought about - haven't really been aware of - until reading the responses to this post. It's good to know for the future should I get to another level in photography.