The great K-5 RAW vs JPG debate - (fun and games inside)

RioRico · 04-09-2012, 03:58 AM

#2 looks better, quickly noticeable as @TOUGEFC said. But...

Originally posted by timh

It's totally irrelevant. You're effectively comparing two prints from the same negative.

"JPEG vs. raw" is a lot like "'Take your film to the local lab who never give it back' vs. 'keep your film and develop it yourself'".

There are circumstances where the former is better, usually related in some way to time. Usually it's not.

There is no great debate. There is understanding raw, and not understanding raw.

This is quite true. RAW sources look better when they're carefully and individually developed. I have batch-developed sets of similarly-exposed RAW files and had them mostly look good enough, but the resultant images still usually require a bit more PP than do custom-developed RAWs. If I'm in a *real* hurry, I can make a batch of RAWs look slightly worse than OOC JPGs.

But even though the development is a bit of work, I still prefer shooting RAW than for OOC JPG.

Philoslothical · 04-09-2012, 04:17 AM

Originally posted by JohnBee

As can be seen, both files were shot in identical settings. That is to say that both were setup under the same lighting, with matching camera and lens settings(focus). However... to help make things as objective as possible, I've chosen a JPG sample yielding the most contrast and detail out of all of the camera profiles. However, if for any anyone should feel that the test would have been better conducted using another OOC candidate, then I would be more than happy to revise the sample.

#1 was shot at f./5.6 and 1/20th. #2 was shot at f./8 and 1/10th. I have to wonder if the DoF difference is causing a noticeable change in the results. It could affect sharpness slightly, too. Many lenses are at their sharpest at about f./8. The same settings should really have been used.

The other thing I wonder about is that the camera passes various jpeg defaults to the raw editor via EXIF, and depending on the raw editor used, some, all or none of these may be applied on opening the raw file. I haven't used RAW Therapee, so I don't know its behaviour on this front. It would fuzz the results quite a bit if some of the jpeg defaults were being applied.

**Schraubstock** · 04-09-2012, 04:46 AM

Originally posted by keithg

My 27 inch monitor is set by a spyder pro and no 2 walks all over no 1. There is a big difference in favour of no 2. Contrast and sharpness.

Which images are we looking at, the posted crops at the top or the ones down loaded from here:

CANDIDATE 1: http://www.bertin.ca/tmp/CANDIDATE_1.jpg
CANDIDATE 2: http://www.bertin.ca/tmp/CANDIDATE_2.jpg The Candidate2 from here is most definitely not better.

Greetings

Philoslothical · 04-09-2012, 04:58 AM

Totally uncalibrated monitor, a cheap one at that, but I agree. In the linked images, #2 looks like crap, as you said, like it's through dirty glass. In the crops in the first post, #2 looks a lot better because the colours are more controlled, giving better contrast. A lot more detail in the fabric is readily apparent.

There's a huge difference between the two #2 files, something doesn't seem right.

**Schraubstock** · 04-09-2012, 05:17 AM

Originally posted by Philoslothical

Totally uncalibrated monitor, a cheap one at that, but I agree. In the linked images, #2 looks like crap, as you said, like it's through dirty glass. In the crops in the first post, #2 looks a lot better because the colours are more controlled, giving better contrast. A lot more detail in the fabric is readily apparent.

There's a huge difference between the two #2 files, something doesn't seem right.

Thanks for that, for a moment I thought I had lost it.

You are right, there is something crook in Tallarook. (Australian rhyming idiom - Meanings and origins of Australian words and idioms - Australian National Dictionary Centre - ANU )
http://en.wikipedia.org/wiki/Tallarook,_Victoria

Greetings

**NaClH2O** · 04-09-2012, 05:26 AM

While I agree that the #2 crop is better looking than the #1 crop, I'm of a mind with RioRico. I usually shoot RAW+ mainly for two reasons. 1) it's easier to sort the 'keepers' out using jpegs, any oof, badly overexposed, badly framed etc shots can be more quickly eliminated simply using windows jpeg viewer. 2) Approximately 30% of the time the images I've decided to keep need no post processing outside of cropping, I then use the jpegs. The other 70% of the time I use the raw file. Since storage is now cheap and plentiful, I see no reason why I should limit myself to either one format or the other. By shooting RAW+ I get the best of both worlds.

NaCl(what is better raw or jpeg? the answer for myself is both)H2O

cbope · 04-09-2012, 05:40 AM

Originally posted by keithg

My 27 inch monitor is set by a spyder pro and no 2 walks all over no 1. There is a big difference in favour of no 2. Contrast and sharpness.

Same here, Dell IPS professional monitor, calibrated with Spyder4 Pro. Number 2 is no contest.

**Schraubstock** · 04-09-2012, 05:51 AM

Originally posted by cbope

Same here, Dell IPS professional monitor, calibrated with Spyder4 Pro. Number 2 is no contest.

Hi

You are looking at the posted crop

But look here, it is a different story
CANDIDATE 1: http://www.bertin.ca/tmp/CANDIDATE_1.jpg
CANDIDATE 2: http://www.bertin.ca/tmp/CANDIDATE_2.jpg

prsjnb · 04-09-2012, 05:59 AM

IMHO, the second image has better colour rendition/contrast and more fine detail.

Apropos the K5 jpeg vs raw debate, I have found the in-camera processing of images to be a great improvement over that seen in the K10D and K20D. In fact, at ISO settings between 100 and 800 ASA if the exposure/WB are optimal, then there is little to choose between the two formats when images are viewed on screen or printed on A4 or smaller (can't comment on larger sizes as I don't have an A3 printer). When required, however, I have been able to obtain small but definitely worthwhile improvements in tonal range, clarity and sharpness by using the raw file and post-processing in ACR or LR.

Jon

Nuff · 04-09-2012, 06:04 AM

I'm not sure how RT works, but the 2nd image if I was working in LR I would add a bit more contrast and blacks. Especially the blacks, they are lacking and look too grey. Needs more sharpening, but that's my personal taste. That's why the threads on right side look a bit soft. In the JPG the reds are blown.

I use a 27" imac, also colour calibrated.

Todd Adamson · 04-09-2012, 06:18 AM

I'm also not seeing the relevance of looking at it this way. An OOC jpg is one engineer's (or team of engineers) opinion on what a finished file should look like. A RAW file contains some fundamentals of that opinion, but is intended to contain more data, for pliability (whether that pliability is is needed for subtle fine-tuning or error correction). I may be missing the point, but I think if we want to ask whether there are important differences between RAW and jpg output, we'd compare heavily processed files, or, for example, files that were two stops underexposed.

**Rondec** · 04-09-2012, 07:45 AM

It is just a comparison of two different ways of developing the same file, right? I think the jpeg engine in the K5 is pretty good, but I do think that it struggles with reds quite a bit (reds are tough, even with RAW files). Anyway, I do think the second file it a little better, but both could use a little work to get them to there best.

I never found shooting RAW+ saved me much time. I review my files in lightroom and can delete (most) files that aren't worth keeping. Then I work on the remaining files. With lightroom there is no benefit (and some significant down side) to working with jpegs.

bikecoboss · 04-09-2012, 08:33 AM

I'm not sure I get it... the full size candidates say what they are right on them... #1 is jpg, #2 is a DNG. The unprocessed DNG is softer, as expected. The crops seem to be reversed, with #2 being cleaner, but still labelled a DNG.

As someone pointed out, the exposure settings are different. Why not just use RAW+JPG and use the exact same original exposure? I think you may have to revisit this one, John.

rfortson · 04-09-2012, 09:16 AM

Originally posted by timh

It's totally irrelevant. You're effectively comparing two prints from the same negative.

"JPEG vs. raw" is a lot like "'Take your film to the local lab who never give it back' vs. 'keep your film and develop it yourself'".

There are circumstances where the former is better, usually related in some way to time. Usually it's not.

There is no great debate. There is understanding raw, and not understanding raw.

Agreed. Raw is not like the negative. Raw is like the undeveloped film. I've gone back to files several years old and reprocessed them with newer versions of LR and each time I feel like I get more out of the file.

Of the two shots, the second one is exposed slightly better, giving better contrast. Since you didn't do anything to the raw file, I'll say the second one is the jpeg and is more "developed".

JohnBee · 04-09-2012, 09:30 AM

Originally posted by bikecoboss

I think you may have to revisit this one, John.

I think you're right.

I put this together late last night and in the rush of things, renamed some of the samples to mask the originals. Needless to say, I think I made a mess of things as some files are reversed whereas others are the wrong ones altogether

I'm correcting them as I write this, I just hope it isn't to late to save the experiment.

View Poll Results: Which image sample do you feel is better?
CANDIDATE 1	9	8.11%
CANDIDATE 2	101	90.99%
I can't tell the difference	1	0.90%
Voters: 111. You may not vote on this poll

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
ALEC the fun and games begin	jeffkrol	General Talk	19	04-22-2012 09:10 AM
want RAW+ along with JPG on Pentax K-7	penties rider	Troubleshooting and Beginner Help	68	10-10-2011 04:30 PM
How to change from JPG to RAW ?	isagold	Pentax K-5 & K-5 II	6	08-05-2011 10:36 PM
The great 2.8 zoom debate!!!!	Twin-Reverb	Pentax SLR Lens Discussion	41	06-01-2011 05:31 AM
The great Prime Debate	Gennatay	Pentax SLR Lens Discussion	62	05-29-2011 01:11 AM

04-09-2012, 03:58 AM	#16
RioRico Veteran Member Join Date: Jun 2008 Location: Limbo, California Posts: 11,263	#2 looks better, quickly noticeable as @TOUGEFC said. But... Originally posted by timh It's totally irrelevant. You're effectively comparing two prints from the same negative. "JPEG vs. raw" is a lot like "'Take your film to the local lab who never give it back' vs. 'keep your film and develop it yourself'". There are circumstances where the former is better, usually related in some way to time. Usually it's not. There is no great debate. There is understanding raw, and not understanding raw. This is quite true. RAW sources look better when they're carefully and individually developed. I have batch-developed sets of similarly-exposed RAW files and had them mostly look good enough, but the resultant images still usually require a bit more PP than do custom-developed RAWs. If I'm in a real hurry, I can make a batch of RAWs look slightly worse than OOC JPGs. But even though the development is a bit of work, I still prefer shooting RAW than for OOC JPG.

04-09-2012, 04:17 AM	#17
Philoslothical Veteran Member Join Date: May 2011 Posts: 1,889	Originally posted by JohnBee As can be seen, both files were shot in identical settings. That is to say that both were setup under the same lighting, with matching camera and lens settings(focus). However... to help make things as objective as possible, I've chosen a JPG sample yielding the most contrast and detail out of all of the camera profiles. However, if for any anyone should feel that the test would have been better conducted using another OOC candidate, then I would be more than happy to revise the sample. #1 was shot at f./5.6 and 1/20th. #2 was shot at f./8 and 1/10th. I have to wonder if the DoF difference is causing a noticeable change in the results. It could affect sharpness slightly, too. Many lenses are at their sharpest at about f./8. The same settings should really have been used. The other thing I wonder about is that the camera passes various jpeg defaults to the raw editor via EXIF, and depending on the raw editor used, some, all or none of these may be applied on opening the raw file. I haven't used RAW Therapee, so I don't know its behaviour on this front. It would fuzz the results quite a bit if some of the jpeg defaults were being applied.

04-09-2012, 04:46 AM	#18
Schraubstock Site Supporter Join Date: Oct 2010 Location: Melbourne Photos: Gallery Posts: 3,386	Originally posted by keithg My 27 inch monitor is set by a spyder pro and no 2 walks all over no 1. There is a big difference in favour of no 2. Contrast and sharpness. Which images are we looking at, the posted crops at the top or the ones down loaded from here: CANDIDATE 1: http://www.bertin.ca/tmp/CANDIDATE_1.jpg CANDIDATE 2: http://www.bertin.ca/tmp/CANDIDATE_2.jpg The Candidate2 from here is most definitely not better. Greetings

04-09-2012, 04:58 AM	#19
Philoslothical Veteran Member Join Date: May 2011 Posts: 1,889	Totally uncalibrated monitor, a cheap one at that, but I agree. In the linked images, #2 looks like crap, as you said, like it's through dirty glass. In the crops in the first post, #2 looks a lot better because the colours are more controlled, giving better contrast. A lot more detail in the fabric is readily apparent. There's a huge difference between the two #2 files, something doesn't seem right.

04-09-2012, 05:17 AM	#20
Schraubstock Site Supporter Join Date: Oct 2010 Location: Melbourne Photos: Gallery Posts: 3,386	Originally posted by Philoslothical Totally uncalibrated monitor, a cheap one at that, but I agree. In the linked images, #2 looks like crap, as you said, like it's through dirty glass. In the crops in the first post, #2 looks a lot better because the colours are more controlled, giving better contrast. A lot more detail in the fabric is readily apparent. There's a huge difference between the two #2 files, something doesn't seem right. Thanks for that, for a moment I thought I had lost it. You are right, there is something crook in Tallarook. (Australian rhyming idiom - Meanings and origins of Australian words and idioms - Australian National Dictionary Centre - ANU ) http://en.wikipedia.org/wiki/Tallarook,_Victoria Greetings

04-09-2012, 05:26 AM	#21
NaClH2O Site Supporter Join Date: Sep 2006 Location: Bronx NY Photos: Gallery \| Albums Posts: 5,631	While I agree that the #2 crop is better looking than the #1 crop, I'm of a mind with RioRico. I usually shoot RAW+ mainly for two reasons. 1) it's easier to sort the 'keepers' out using jpegs, any oof, badly overexposed, badly framed etc shots can be more quickly eliminated simply using windows jpeg viewer. 2) Approximately 30% of the time the images I've decided to keep need no post processing outside of cropping, I then use the jpegs. The other 70% of the time I use the raw file. Since storage is now cheap and plentiful, I see no reason why I should limit myself to either one format or the other. By shooting RAW+ I get the best of both worlds. NaCl(what is better raw or jpeg? the answer for myself is both)H2O

04-09-2012, 05:40 AM	#22
cbope Veteran Member Join Date: Dec 2011 Location: Helsinki Posts: 664	Originally posted by keithg My 27 inch monitor is set by a spyder pro and no 2 walks all over no 1. There is a big difference in favour of no 2. Contrast and sharpness. Same here, Dell IPS professional monitor, calibrated with Spyder4 Pro. Number 2 is no contest.

04-09-2012, 05:51 AM	#23
Schraubstock Site Supporter Join Date: Oct 2010 Location: Melbourne Photos: Gallery Posts: 3,386	Originally posted by cbope Same here, Dell IPS professional monitor, calibrated with Spyder4 Pro. Number 2 is no contest. Hi You are looking at the posted crop But look here, it is a different story CANDIDATE 1: http://www.bertin.ca/tmp/CANDIDATE_1.jpg CANDIDATE 2: http://www.bertin.ca/tmp/CANDIDATE_2.jpg

04-09-2012, 05:59 AM	#24
prsjnb Junior Member Join Date: Jun 2010 Location: Wiltshire Photos: Gallery \| Albums Posts: 32	IMHO, the second image has better colour rendition/contrast and more fine detail. Apropos the K5 jpeg vs raw debate, I have found the in-camera processing of images to be a great improvement over that seen in the K10D and K20D. In fact, at ISO settings between 100 and 800 ASA if the exposure/WB are optimal, then there is little to choose between the two formats when images are viewed on screen or printed on A4 or smaller (can't comment on larger sizes as I don't have an A3 printer). When required, however, I have been able to obtain small but definitely worthwhile improvements in tonal range, clarity and sharpness by using the raw file and post-processing in ACR or LR. Jon

04-09-2012, 06:04 AM	#25
Nuff Veteran Member Join Date: Mar 2011 Location: Sydney Posts: 671	I'm not sure how RT works, but the 2nd image if I was working in LR I would add a bit more contrast and blacks. Especially the blacks, they are lacking and look too grey. Needs more sharpening, but that's my personal taste. That's why the threads on right side look a bit soft. In the JPG the reds are blown. I use a 27" imac, also colour calibrated.

Thread Tools	Search this Thread
	Search this Thread: Advanced Search

04-09-2012, 06:18 AM	#26
Todd Adamson Veteran Member Join Date: Oct 2010 Location: Iowa Posts: 722	I'm also not seeing the relevance of looking at it this way. An OOC jpg is one engineer's (or team of engineers) opinion on what a finished file should look like. A RAW file contains some fundamentals of that opinion, but is intended to contain more data, for pliability (whether that pliability is is needed for subtle fine-tuning or error correction). I may be missing the point, but I think if we want to ask whether there are important differences between RAW and jpg output, we'd compare heavily processed files, or, for example, files that were two stops underexposed.

04-09-2012, 07:45 AM	#27
Rondec Loyal Site Supporter Join Date: Mar 2009 Location: Gladys, Virginia Photos: Gallery Posts: 27,652	It is just a comparison of two different ways of developing the same file, right? I think the jpeg engine in the K5 is pretty good, but I do think that it struggles with reds quite a bit (reds are tough, even with RAW files). Anyway, I do think the second file it a little better, but both could use a little work to get them to there best. I never found shooting RAW+ saved me much time. I review my files in lightroom and can delete (most) files that aren't worth keeping. Then I work on the remaining files. With lightroom there is no benefit (and some significant down side) to working with jpegs.

04-09-2012, 08:33 AM	#28
bikecoboss Senior Member Join Date: Jan 2010 Location: Pemberton BC Posts: 238	I'm not sure I get it... the full size candidates say what they are right on them... #1 is jpg, #2 is a DNG. The unprocessed DNG is softer, as expected. The crops seem to be reversed, with #2 being cleaner, but still labelled a DNG. As someone pointed out, the exposure settings are different. Why not just use RAW+JPG and use the exact same original exposure? I think you may have to revisit this one, John.

04-09-2012, 09:16 AM	#29
rfortson Veteran Member Join Date: Sep 2006 Location: Houston TX Photos: Gallery Posts: 2,129	Originally posted by timh It's totally irrelevant. You're effectively comparing two prints from the same negative. "JPEG vs. raw" is a lot like "'Take your film to the local lab who never give it back' vs. 'keep your film and develop it yourself'". There are circumstances where the former is better, usually related in some way to time. Usually it's not. There is no great debate. There is understanding raw, and not understanding raw. Agreed. Raw is not like the negative. Raw is like the undeveloped film. I've gone back to files several years old and reprocessed them with newer versions of LR and each time I feel like I get more out of the file. Of the two shots, the second one is exposed slightly better, giving better contrast. Since you didn't do anything to the raw file, I'll say the second one is the jpeg and is more "developed".

04-09-2012, 09:30 AM	#30
JohnBee Veteran Member Join Date: Sep 2008 Location: Newrfoundland Photos: Albums Posts: 4,667 Original Poster	Originally posted by bikecoboss I think you may have to revisit this one, John. I think you're right. I put this together late last night and in the rush of things, renamed some of the samples to mask the originals. Needless to say, I think I made a mess of things as some files are reversed whereas others are the wrong ones altogether I'm correcting them as I write this, I just hope it isn't to late to save the experiment.