Iliah Borg
Forum Pro
That's an odd thing to ask, as nothing substitutes first-hand experience.
-
Welcome to the new forums! Please read this first. For known issues we are working to resolve, click here.
RAW vs. JPEG
- Thread starter Brad Bohland
- Start date
Iliah Borg
Forum Pro
But midtones in such a JPEG are now in need of editing, 'cause they are "underexposed". Try and see.
Next, there are no pure ISO-less cameras, intermediate ISO settings are often digital, and sometimes they are clipped from the next "even" ISO setting. Most of the current Canon cameras are fully ISO-dependent.
And let's assume the sensor is not ISO-invariant which is the case with pretty much everybody's cameras. And I regularly increase exposure 2/3 to 1 1/2 stops above the exposure that's ideal for a JPEG to get a best raw file. And again the word is BEST.
Depends on what you want and what you're shooting. One more time I'm going to write this word and make it bold: BEST, any difference, even "not that great" will be less than best.
Let's see what happens when what you're shooting is a little demanding:
Here's the two files Nick doesn't have the expertise to create. He's welcome to them:
DSCF4643.RAF
Alley_jpeg.jpg
The difference between the exposures is 2/3 stop. The blue channel in the JPEG is still clipped but the red and green channels are not so if Nick is good he can recover the JPEG clipping. I could give you a JPEG that's 1 full stop less exposed than the raw without the blue channel clipped, but you really don't want to see that. Here's what this JPEG looks like SOOC:
And here's my processing of the raw file that received 2/3 stop more exposure:
Go for it, process the JPEG, let's see how the shadows look when you open them up. The photo I took here could not be done at all shooting JPEG -- simply impossible.
Iliah beat me to this one: "I should very much like to know the answer to this question, but unfortunately I lack the expertise to capture pairs of raw and JPEG images in which both captures are optimally exposed for their respective capture format." Says it all.
Attachments
Last edited:
Brad Bohland
Veteran Member
- Messages
- 6,356
- Solutions
- 1
- Reaction score
- 2,566
"I should very much like to know the answer to this question, but unfortunately I lack the expertise to capture pairs of raw and JPEG images in which both captures are optimally exposed for their respective capture format. Perhaps someone who has the necessary expertise will produce some such pairs to allow this issue to be explored in practical terms rather than theoretically as has so far been the case?"
http://www.dpreview.com/forums/post/57661805
Iliah Borg
Forum Pro
Bracketing and FastRawViewer is all that is needed.
Brad Bohland
Veteran Member
- Messages
- 6,356
- Solutions
- 1
- Reaction score
- 2,566
It says he's inquiring and doesn't know it all.
Brad Bohland
Veteran Member
- Messages
- 6,356
- Solutions
- 1
- Reaction score
- 2,566
That I don't know.
There's three threads on this series now and I haven't even begun to absorb it. And my specific question (best processed RAW vs. best processed JPEG) seems to be not front and center.
OK -- been there, done that and happy to help.
Brad Bohland
Veteran Member
- Messages
- 6,356
- Solutions
- 1
- Reaction score
- 2,566
Thanks.
I appreciate your help. And I think your post with the two alley shots could provide the answer I'm looking for, though I'm not able to process the JPEG.
Part of my problem is the software choice for processing both JPEGs and RAW. I have E9, which is evidently fairly worthless for this kind of thing.
Iliah Borg
Forum Pro
JPEG is already heavily processed in camera, re-processing is limited after that. How much is lost during the conversion to JPEG in a camera depends on the scene, in-camera JPEG processing engine, and JPEG settings.
A simple test would be to shoot low contrast text (similar to what DPR are using in their new studio scene, but adding snaps in different colours) and see how much underexposure you can recover from JPEG and from raw.
Brad Bohland
Veteran Member
- Messages
- 6,356
- Solutions
- 1
- Reaction score
- 2,566
Can that be done with E9?
Time for better raw processing software. It doesn't have to cost: RawTherapp, and DarkTable are free. What came with the camera?
Brad Bohland
Veteran Member
- Messages
- 6,356
- Solutions
- 1
- Reaction score
- 2,566
Steen Bay
Veteran Member
Guess you're right that raw has the advantage here. Lifting shadows is one of the cases where it's nice to have the raw file. Though, wouldn't say that it's impossible to do if shooting JPEG. Most cameras have features like ADL (Nikon) or DRO (Sony) that'll lift the shadows in-camera before saving the JPEG. Or even better, you could use a HDR mode that automatically will shoot, merge and align several images in-camera. Not impossible that such a mode could give you a result with JPEG that's better than what's possible with a single raw exposure.Depends on what you want and what you're shooting. One more time I'm going to write this word and make it bold: BEST, any difference, even "not that great" will be less than best.
Let's see what happens when what you're shooting is a little demanding:
Here's the two files Nick doesn't have the expertise to create. He's welcome to them:
DSCF4643.RAF
Alley_jpeg.jpg
The difference between the exposures is 2/3 stop. The blue channel in the JPEG is still clipped but the red and green channels are not so if Nick is good he can recover the JPEG clipping. I could give you a JPEG that's 1 full stop less exposed than the raw without the blue channel clipped, but you really don't want to see that. Here's what this JPEG looks like SOOC:
And here's my processing of the raw file that received 2/3 stop more exposure:
Go for it, process the JPEG, let's see how the shadows look when you open them up. The photo I took here could not be done at all shooting JPEG -- simply impossible.
Old Greenlander
Veteran Member
GuysGuess you're right that raw has the advantage here. Lifting shadows is one of the cases where it's nice to have the raw file. Though, wouldn't say that it's impossible to do if shooting JPEG. Most cameras have features like ADL (Nikon) or DRO (Sony) that'll lift the shadows in-camera before saving the JPEG. Or even better, you could use a HDR mode that automatically will shoot, merge and align several images in-camera. Not impossible that such a mode could give you a result with JPEG that's better than what's possible with a single raw exposure.Depends on what you want and what you're shooting. One more time I'm going to write this word and make it bold: BEST, any difference, even "not that great" will be less than best.
Let's see what happens when what you're shooting is a little demanding:
Here's the two files Nick doesn't have the expertise to create. He's welcome to them:
DSCF4643.RAF
Alley_jpeg.jpg
The difference between the exposures is 2/3 stop. The blue channel in the JPEG is still clipped but the red and green channels are not so if Nick is good he can recover the JPEG clipping. I could give you a JPEG that's 1 full stop less exposed than the raw without the blue channel clipped, but you really don't want to see that. Here's what this JPEG looks like SOOC:
And here's my processing of the raw file that received 2/3 stop more exposure:
Go for it, process the JPEG, let's see how the shadows look when you open them up. The photo I took here could not be done at all shooting JPEG -- simply impossible.
you show overexposed or underexposed RAW of JPEGs
I am not in this game. I try to expose correctly but I take 2000 JPEG pictures at the occasion therefore very little incentive for PP
I need to see the superior RAW exposed the best over the best exposed JPEG of the same frame.
Is it a crushing difference?
This is the question.
--
Old Greenlander
"I show the world the way I see it"
35 years of photography and still learning
gardenersassistant
Veteran Member
I am the author about whom some comments have been made in this thread in relation to some posts that I made in another forum and assumptions and conclusions therefrom about my beliefs and competencies.
You might or might not be interested to know the following.
I have been shooting raw exclusively for some time because I believe I can get better results with raw. This is not just a matter of being able better to recover highlights and/or shadows, but involves more subtle differences in the rendition of textures and colours. (This matters a lot to me as I photograph a lot of botanical subjects.) I discussed these differences in this post in the referenced set of posts, based on what I saw as between the processed raw images, the processed JPEGs and the out of the camera JPEGs in the large number of animated 3-way comparisons that I posted.
I do not believe in "one-click" post processing for my own images. I always process them individually. The quoted text in this post is not from me.
I shot raw and JPEG solely for the purpose of the exercise, not because I have any intention of deriving my processed images from JPEG originals in future. I mentioned that I would be continuing to shoot raw in the pre-penultimate paragraph of the above linked post.
My use of batch processing during the exercise was initially to let me do some quick initial comparisons between some processed raw, processed JPEG and out of the camera JPEG images for the 1100 or so raw+JPEG test captures from the first phase of the exercise. Because of the (to me surprisingly good) results from the batch processing I became interested in batch processing from two perspectives:
Would batch processing be useful to at least some people who did not want to get involved in the details of post processing and/or raw file processing, but who might be able to get better results from a batch processed image, raw or JPEG, than from an out of the camera JPEG?
From the reaction I got to the examples I posted, the answer to that seems to be in the affirmative.
Would it be useful to me to use one or more batch processes in the early stages of my image processing workflow?
I often have to 500 to 1000 and sometimes more images to deal with from a day's and/or night's shooting. I shoot mainly close-ups and macros - up to 10:1 or so in full frame terms - and, like other macro shooters, I have a high failure rate. I also have a high repetition and similarity rate for successful shots. I under-expose a great deal (as in often, and by relatively large amounts) in order to preserve highlights in shots of invertebrates, which often have highly reflective surfaces which can defeat the best flash diffusers, to guard against colour and detail bleaching in natural light botanical shots of subjects with delicately coloured and/or textured petals, and/or to conserve flash battery power and/or reduce flash recycle time.
I am conservative in my exposures (as in, not exposing to the right) because I often have to work quickly when opportunities first arise, and continuously when capturing behavioural sequences or tracking moving subjects, during the course of which tracking the scene's reflectance etc can vary considerably with me having little and sometimes no time to make adjustments. In addition I may only have one opportunity to capture a shot because the act of shooting may cause the subject to exit the scene, or the subject may move away at any time irrespective of my activity.
With botanical subjects under-exposure is often more considered, being done on a scene-specific basis, based on reviewing on the LCD captured images of subjects which are susceptible to bleaching even before any channel is clipped (eg delicate pink petals) or channel clipping (eg bright red petals), but sometimes there may be a random element of exposure fluctuation as when photographing flowers in dappled sunlight beneath moving foliage, in which case I can decrease the probability of over-exposure by "under-exposing the average".
Being successful with the approach of reviewing captured images requires an appreciation, based on experience, of the relationship between what is seen on the LCD and how the (raw) files will respond to post processing. I do not make any use of the camera histogram or exposure meter. The conservatism in my exposures stems from my preference to have underexposed shots that I can massage up in post processing rather than risk over-exposed shots which have one or more channels blown and irretrievably lost details or colour integrity. I "play the probabilities"; that is, I don't underexpose enough to guarantee that I never get any bleaching or channel clipping, because of the implications that would have for retrievable image quality. I try to balance the probability of shots lost from bleaching or clipping against the issues of image recovery and rendering.
Especially because of the frequent under-exposure, the shots from a session need some processing in order to get them into a consistent state where I can see enough to assess which of them are likely to be useful and select them for inclusion in the next round of assessment/rejection. The evidence from the exercise suggests that batch processing can have a useful role for me in this regard.These might be batch processed JPEG images because the processed JPEG images are close enough in appearance to the processed raw files to act as surrogates for the raw files at this stage in the process, and useful in that role because they (batch) process significantly faster than the raw files, which is a consideration when dealing with large numbers of images.
The exercise also suggested that batch processing the raw files that emerge from the selection process may be useful as the first stage in the processing pipeline, prior to image-specific adjustments.
By the content and tone of various comments in this thread it appears that some of you have already come to settled conclusions about my ignorance and incompetence which, being settled, need no further discussion, but in case there are others who still have some lingering doubts on that score it may be helpful if I give you some additional information which should put the matter to rest once and for all. You may like to note that:
You might or might not be interested to know the following.
I have been shooting raw exclusively for some time because I believe I can get better results with raw. This is not just a matter of being able better to recover highlights and/or shadows, but involves more subtle differences in the rendition of textures and colours. (This matters a lot to me as I photograph a lot of botanical subjects.) I discussed these differences in this post in the referenced set of posts, based on what I saw as between the processed raw images, the processed JPEGs and the out of the camera JPEGs in the large number of animated 3-way comparisons that I posted.
I do not believe in "one-click" post processing for my own images. I always process them individually. The quoted text in this post is not from me.
I shot raw and JPEG solely for the purpose of the exercise, not because I have any intention of deriving my processed images from JPEG originals in future. I mentioned that I would be continuing to shoot raw in the pre-penultimate paragraph of the above linked post.
My use of batch processing during the exercise was initially to let me do some quick initial comparisons between some processed raw, processed JPEG and out of the camera JPEG images for the 1100 or so raw+JPEG test captures from the first phase of the exercise. Because of the (to me surprisingly good) results from the batch processing I became interested in batch processing from two perspectives:
Would batch processing be useful to at least some people who did not want to get involved in the details of post processing and/or raw file processing, but who might be able to get better results from a batch processed image, raw or JPEG, than from an out of the camera JPEG?
From the reaction I got to the examples I posted, the answer to that seems to be in the affirmative.
Would it be useful to me to use one or more batch processes in the early stages of my image processing workflow?
I often have to 500 to 1000 and sometimes more images to deal with from a day's and/or night's shooting. I shoot mainly close-ups and macros - up to 10:1 or so in full frame terms - and, like other macro shooters, I have a high failure rate. I also have a high repetition and similarity rate for successful shots. I under-expose a great deal (as in often, and by relatively large amounts) in order to preserve highlights in shots of invertebrates, which often have highly reflective surfaces which can defeat the best flash diffusers, to guard against colour and detail bleaching in natural light botanical shots of subjects with delicately coloured and/or textured petals, and/or to conserve flash battery power and/or reduce flash recycle time.
I am conservative in my exposures (as in, not exposing to the right) because I often have to work quickly when opportunities first arise, and continuously when capturing behavioural sequences or tracking moving subjects, during the course of which tracking the scene's reflectance etc can vary considerably with me having little and sometimes no time to make adjustments. In addition I may only have one opportunity to capture a shot because the act of shooting may cause the subject to exit the scene, or the subject may move away at any time irrespective of my activity.
With botanical subjects under-exposure is often more considered, being done on a scene-specific basis, based on reviewing on the LCD captured images of subjects which are susceptible to bleaching even before any channel is clipped (eg delicate pink petals) or channel clipping (eg bright red petals), but sometimes there may be a random element of exposure fluctuation as when photographing flowers in dappled sunlight beneath moving foliage, in which case I can decrease the probability of over-exposure by "under-exposing the average".
Being successful with the approach of reviewing captured images requires an appreciation, based on experience, of the relationship between what is seen on the LCD and how the (raw) files will respond to post processing. I do not make any use of the camera histogram or exposure meter. The conservatism in my exposures stems from my preference to have underexposed shots that I can massage up in post processing rather than risk over-exposed shots which have one or more channels blown and irretrievably lost details or colour integrity. I "play the probabilities"; that is, I don't underexpose enough to guarantee that I never get any bleaching or channel clipping, because of the implications that would have for retrievable image quality. I try to balance the probability of shots lost from bleaching or clipping against the issues of image recovery and rendering.
Especially because of the frequent under-exposure, the shots from a session need some processing in order to get them into a consistent state where I can see enough to assess which of them are likely to be useful and select them for inclusion in the next round of assessment/rejection. The evidence from the exercise suggests that batch processing can have a useful role for me in this regard.These might be batch processed JPEG images because the processed JPEG images are close enough in appearance to the processed raw files to act as surrogates for the raw files at this stage in the process, and useful in that role because they (batch) process significantly faster than the raw files, which is a consideration when dealing with large numbers of images.
The exercise also suggested that batch processing the raw files that emerge from the selection process may be useful as the first stage in the processing pipeline, prior to image-specific adjustments.
By the content and tone of various comments in this thread it appears that some of you have already come to settled conclusions about my ignorance and incompetence which, being settled, need no further discussion, but in case there are others who still have some lingering doubts on that score it may be helpful if I give you some additional information which should put the matter to rest once and for all. You may like to note that:
- I use the smallest available aperture for most of my invertebrate photography, which as everyone (including me) knows means a considerable loss of detail from diffraction (and bear in mind that invertebrate photography is largely about detail);
- by choice after extensive (and doubtless methodologically flawed) comparisons I, by choice, use a bridge camera with a small (1/2.3") sensor for the majority of my invertebrate photography apart from large and extremely small invertebrate subjects, for which I use an APS-C and micro four thirds camera respectively;
- Having assessed the results I obtained from, and working practices I needed to use with, prime macro lenses including the Canon 100L Macro and the Canon MPE-65, I rejected prime macro lenses in favour of continuing to use close-up lenses on my interchangeable lens cameras. For the avoidance of doubt, with all my cameras, fixed and interchangeable lens, I am using relatively inexpensive, optically unsophisticated close-up lenses on the front of relatively inexpensive general purpose telephoto zoom lenses which are of course optically suboptimal;
- I use auto-focus down to about 5:1 in full frame terms, which as many people know does not work for macros even at 1:1 or less magnification;
- I use contrast detect focusing rather than (when available) phase detect focusing, even with moving subjects apart from birds in flight, for which I do use phase detect focusing; and
- I use live view and the LCD rather than the viewfinder, even with my dSLR (Canon 70D).
Brad Bohland
Veteran Member
- Messages
- 6,356
- Solutions
- 1
- Reaction score
- 2,566
Just found this thread. I don't understand this at all. What is it?
"While I prefer doing processing on my Mac, I used the camera RAW Processing and voila. I can see where this could be useful at times."
http://www.dpreview.com/forums/post/54271544
"The RAW to JPG option allows adjustments to Exposure, shadows, contrast, sharpness, various modes (vivid, B&W etc.), highlights, saturation and so on. In fact you'd have more dynamic range than what you'd get with JPGS."
http://www.dpreview.com/forums/post/54272809
I would and it is.Guess you're right that raw has the advantage here. Lifting shadows is one of the cases where it's nice to have the raw file. Though, wouldn't say that it's impossible to do if shooting JPEG.Depends on what you want and what you're shooting. One more time I'm going to write this word and make it bold: BEST, any difference, even "not that great" will be less than best.
Let's see what happens when what you're shooting is a little demanding:
Here's the two files Nick doesn't have the expertise to create. He's welcome to them:
DSCF4643.RAF
Alley_jpeg.jpg
The difference between the exposures is 2/3 stop. The blue channel in the JPEG is still clipped but the red and green channels are not so if Nick is good he can recover the JPEG clipping. I could give you a JPEG that's 1 full stop less exposed than the raw without the blue channel clipped, but you really don't want to see that. Here's what this JPEG looks like SOOC:
And here's my processing of the raw file that received 2/3 stop more exposure:
Go for it, process the JPEG, let's see how the shadows look when you open them up. The photo I took here could not be done at all shooting JPEG -- simply impossible.
Yep, got the same function in the Fuji I used to take the photo above. Those functions are self-defeating. They get the "expanded DR" JPEGs by forcing an underexposure of the raw file. For example the DR function on my Fuji is disabled unless you raise the ISO above base. At that point as the camera meters, for example an exposure at ISO 400, the camera electronics withhold the ISO gain and in fact expose the sensor at ISO 200 forcing an underexposure -- the last thing needed!
In the above photo and in the photo of the truck I posted earlier I used every last bit of the 10 stops of DR my camera sensor can deliver. Those functions would have forced an underexposure and likewise forced a DR reduction in the sensor's capacity. Those photos are impossible shooting JPEG.
Do I have to dig up my high contrast direct-sun photos of cyclists going 30 mph. for you? I shoot exclusively raw because it's not only the best way, it's also the easy way.
I've seen some of those in-camera HDR functions. They make some really ugly stuff. The final result has to be brought down to a printable or displayable output range and automated software isn't going to do that without getting ugly. It's still limited to a single tone curve applied to the entire final image. Then you're left trying to repair the JPEG and it's already compressed 8 bit and you're screwed.
With still subjects I can shoot HDR raw if I want to and again do the processing right -- rarely needed with today's cameras. One well-exposed raw file brings it home.
Why try and make it hard when it's not?
IanYorke
Veteran Member
Just interested what do you do with your 2000 pics? Wedding photographer?GuysGuess you're right that raw has the advantage here. Lifting shadows is one of the cases where it's nice to have the raw file. Though, wouldn't say that it's impossible to do if shooting JPEG. Most cameras have features like ADL (Nikon) or DRO (Sony) that'll lift the shadows in-camera before saving the JPEG. Or even better, you could use a HDR mode that automatically will shoot, merge and align several images in-camera. Not impossible that such a mode could give you a result with JPEG that's better than what's possible with a single raw exposure.Depends on what you want and what you're shooting. One more time I'm going to write this word and make it bold: BEST, any difference, even "not that great" will be less than best.
Let's see what happens when what you're shooting is a little demanding:
Here's the two files Nick doesn't have the expertise to create. He's welcome to them:
DSCF4643.RAF
Alley_jpeg.jpg
The difference between the exposures is 2/3 stop. The blue channel in the JPEG is still clipped but the red and green channels are not so if Nick is good he can recover the JPEG clipping. I could give you a JPEG that's 1 full stop less exposed than the raw without the blue channel clipped, but you really don't want to see that. Here's what this JPEG looks like SOOC:
And here's my processing of the raw file that received 2/3 stop more exposure:
Go for it, process the JPEG, let's see how the shadows look when you open them up. The photo I took here could not be done at all shooting JPEG -- simply impossible.
you show overexposed or underexposed RAW of JPEGs
I am not in this game. I try to expose correctly but I take 2000 JPEG pictures at the occasion therefore very little incentive for PP
I need to see the superior RAW exposed the best over the best exposed JPEG of the same frame.
Is it a crushing difference?
This is the question.
--
Old Greenlander
"I show the world the way I see it"
35 years of photography and still learning
Ian
Similar threads
Keyboard shortcuts
- f
- Forum