LF jpgs vs LSF jpgs

I'm not denying that at the pixel level there will be differences, after all one file is 3 times the size of another, what has struck me is that for all intents and purposes I can't see any difference at any sensible level of magnification or print size and viewing distance.

PP will introduce more differences as pris has shown, but this thread was originally about Andy Westlakes assertation that he didn't think there was any tangible benefit in using LSF over LF if you're not going to PP. I'd go further and say that as far as I can see there is no effective difference between LN and LSF if you're not going to PP

I'll reiterate that I'm not trying to push people into shooting one over another, or saying one way is better than another, I'm just posting the results that I've seen with my testing - people are free to make up their own minds one way or another. It's been interesting to me to see how little difference I've seen between the files and hopefully it's been of interest to others as well.

This might also be camera specific, there seems to be evidence that the differences are more visible with E-5 files, in which case what I've posted isn't applicable to other Oly models - can't test for this as I only have the E-3 and E-1. Perhaps others might be tempted to share examples of their cameras, but I'm not holding my breath on this one

Thanks for your thoughts,

Nick

You're welcome to download them and play with them.

I'm at work at the moment so can't do any comparisons myself.

If you get the chance to do side by side comparisons I'd be interested to see what differences you've seen. I'd not looked everywhere when I did my comparisons so there's every chance that you've seen something I've missed,

Nick

nick_webster said:

PP will introduce more differences as pris has shown, but this thread was originally about Andy Westlakes assertation that he didn't think there was any tangible benefit in using LSF over LF if you're not going to PP.

Click to expand...

Don't forget that while initial claim went along those lines, one of the following statements as GBs quote has shown was "LF is enough for 99% of photographers in 99% of cases." That's a bit of a stretch, IMO. It's not like 1% PPs 1% of their photos.

Although you showed there was a difference in LSF and LF after PP you haven't necessarily shown that there is any tangible difference for 99% of users

I don't think it is possible to prove or disprove the 99% number as it will depend on extent of PP, what size the resulting image is viewed at, personal levels of what is acceptable etc. Certainly it's beyond what I'm prepared to do to say if it's true or not

Did you see any visible differences in the 2 jpgs you PP'd ? I know the difference between them increased but was that visible in final prints/ on your monitor ? No agenda here I'm just curious,

Nick

nick_webster said:

Although you showed there was a difference in LSF and LF after PP you haven't necessarily shown that there is any tangible difference for 99% of users

Click to expand...

No... It's not like anyone knows exact % of photographers that posprocess their images, right? And it's not like you took DPR to the task on their claim about 99% in 99% of cases? My thesis was: the degree of compression is likely to influence the IQ as PP is applied; and progressively worsen it as PP becomes more extensive. It's a common sense, it's not a new concept, and a simple test has shown just what was expected.

nick_webster said:

I don't think it is possible to prove or disprove the 99% number as it will depend on extent of PP, what size the resulting image is viewed at, personal levels of what is acceptable etc. Certainly it's beyond what I'm prepared to do to say if it's true or not

Click to expand...

So can we agree that 99% applying no PP in 99% of their images is intuitively a serious stretch? Let's not turn this into circus, it was nice, informative and surprisingly civil

nick_webster said:

Did you see any visible differences in the 2 jpgs you PP'd ? I know the difference between them increased but was that visible in final prints/ on your monitor ? No agenda here I'm just curious,

Click to expand...

I haven't even looked. My purpose was to get to overlay which I did. To seek visible difference and find out when it appears, the test must be set in a very different manner and on different kind of images. It also will become much more time consuming.

krh_57 said:

Do you have any thing to add?
http://forums.dpreview.com/forums/read.asp?forum=1022&message=38311209

Click to expand...

To be honest, I don't know what you were trying to say in the post above. Your initial comments on this matter, was that:

http://forums.dpreview.com/forums/read.asp?forum=1022&message=38291391

But here is a great example of DPR and their web pundits finding any excuse they could possibly wring out of a situation, where DPR was clearly wrong.

Lets face it, they couldn't give a cr@p about Olympus, and tried everything they could to disprove Olympus's worth.

I am constantly amazed how DPR and others think can better engineer, even a company like Olympus... You guys think your can, but your only kidding yourselves!

And some like DPR, have to lie and find cover to get their points across. What a freaking joke.

Yet, in the link which you ask me if I have anything to add, you say:

http://forums.dpreview.com/forums/read.asp?forum=1022&message=38311209

No I do not, and I didn't expect to see any difference between between LSF and LF.

I don't know how to reconcile these two posts. Did, or did not, DPR try "everything they could to disprove Olympus's worth"? Does, or does not, DPR "have to lie and find cover to get their points across"?

Or was it instead simply a matter of:

http://forums.dpreview.com/forums/read.asp?forum=1022&message=28419296

I think 'hard' is an understatement; in a proper double-blind test I don't think anyone could tell them apart.

Shortly after the whole E3 HQ/SHQ 'issue' blew up, I reconverted a number of image files in-camera from RAW to both HQ and SHQ (including some detailed landscapes and macros which I shot for the E3 gallery) and compared them in Photoshop, using the layer mode set to 'difference'. And there truly is a difference, but at most it's 5 bits at any one pixel, and limited to a small proportion of the total pixels. I also did a direct side-by-side comparison of the two JPEGs to see if I could identify any visible difference between them, and saw nothing even though I knew exactly where to look. So if you're intending to use JPEGs directly as final output, SHQ really is just a waste of space (perhaps there's a reason why Olympus don't offer it as the default option, after all?).

pris said:

pris said:

Hey pris -- I'm here to bait you on that whole RAW vs jpg thing.

You said in the other thread that you can just as easily change the WB and tone curve of a jpg as with a RAW with pretty much the same results. I'm curious about that. You also said, "There are of course non-arguable advantages to raw, but these are not of much significance." What are the "non-arguable advantages to RAW", then, if WB and tone curve mapping are "not of much significance"?

Click to expand...

Ugh... I am biting just a little bit, ready to let the bait go the moment I see it turning into classic Raw-JPEG debate

Click to expand...

At last! I'm getting a nibble.

pris said:

Ability to apply an art filter, decide on a gradation, global noise filtering, contrast etc after the fact is non-arguable advantage in a situation where photographer has to work fast and may not have time to set optimal settings. Mixed difficult lighting where you just don't know what to set for WB, better shoot Raw and think later. Sharply divided WB (example: a room with artificial lighting and a large opening into the outdoor with natural light) - much easier to shoot Raw, produce two JPEGs with different WBs and combine them into one. Scene with very wide DR which must be bracketed - better do it in Raw. Shooting a whole series in similar conditions that are likely to require identical treatment - easier to shoot in Raw, fine-tune the settings and then run whole series through the batch treatment (although this one can be dione in JPEG almost as easy, but still).

Click to expand...

Gotcha.

pris said:

I am sure you understand that saying "WB and tone curve are not of much significance" I meant that they are fairly easy to change in JPEG; not that they do not matter

Click to expand...

Yes -- I understood that's what you meant. What I didn't understand was why changing the tone curve from a jpg would not result in a "significant" loss of quality as opposed to changing the tone curve in the RAW conversion. Of course, "significant" is in quotes because what constitutes "significant" depends on many factors.

pris said:

pris said:

In terms of the tone curve, well, a jpg is 8 bits, so it can muster at most 8 levels within the DR. However, most cameras have more than 8 stops of DR. So a jpg, by "virtue" of it's 8 bit files, must either cut down to a maximum of 8 stops of DR, or represent multiple levels within the DR indistinguishably.

Click to expand...

Sorry, this above just doesn't sound right to me. Look at any DR measurement in any of the reviews, you will see much wider DR in JPEGs. I am not sure where this is coming from, nor am I proficient in the academic side of things enough for the discussion in these terms.

Click to expand...

DR -- dynamic range -- is the number of stops from the read noise to the saturation point of a pixel. If the bit depth is less than the DR, then it simply gives the maximum number of levels of the DR that can be distinctly recorded. For example, an 8 bit jpg can render 14 stops of DR, but it can only show 8 distinct levels of that DR.

If the bit-depth is greater than the DR, then it's simply an inefficient use of memory. For example, if a sensor has 11 stops of DR, then a 14 bit ADC has no advantage over a 12 bit ADC, and simply makes for larger files with no increase in IQ.

The E5 has 10.5 stops of DR, a jpg can render 8 levels of those 10.5 stops. That is, it can capture 8 stops of DR, each level being distinguishable, or it can capture all 10.5 stops, but some levels won't be distinguishable.

Since most display media (prints, monitors) only have 8-9 stops of DR, a jpg will work well for both prints and monitor display, so long as the jpg had the desired tone curve applied at the time of capture.

The advantage of RAW, in this respect, is that all 10.5 stops of the DR are recorded, and we can decide after the fact what range and what levels are to be represented in the jpg. If the initial capture is a jpg, then we are "stuck" with the range of the initial capture, which can be rendered with at most 8 levels.

I hope that makes some kind of sense.

Gidday Pris

pris said:

nick_webster said:

For people who want to do extensive PP then surely RAW is the most sensible answer, which is what I shoot

Click to expand...

"Surely" eh... It's a somewhat too narrow view of PP, no offense. There is more than one PP workflow, PP purpose and the way to reach it.

The kind of PP I do requires LAB colorspace which takes one right into JPEG. Don't think for a second that this some kind of marginal exotic treatment - although not as mainstream as RGB or CMYK, this colorspace has quite a following.

Click to expand...

One can use LAB colour space on PSD-16 files. No need to use it only on JPEG files. ??

--
Regards, john from Melbourne, Australia.
(see profile for current gear)
Please do not embed images from my web site without prior permission
I consider this to be a breach of my copyright.
-- -- --

The Camera doth not make the Man (or Woman) ...
Perhaps being kind to cats, dogs & children does ...

Gallery: http://canopuscomputing.com.au/gallery2/main.php



Bird Control Officers on active service.

Member of UK (and abroad) Photo Safari Group

Great Bustard said:

pris said:

Great Bustard said:

Hey pris -- I'm here to bait you on that whole RAW vs jpg thing.

You said in the other thread that you can just as easily change the WB and tone curve of a jpg as with a RAW with pretty much the same results. I'm curious about that. You also said, "There are of course non-arguable advantages to raw, but these are not of much significance." What are the "non-arguable advantages to RAW", then, if WB and tone curve mapping are "not of much significance"?

Click to expand...

Ugh... I am biting just a little bit, ready to let the bait go the moment I see it turning into classic Raw-JPEG debate

Click to expand...

At last! I'm getting a nibble.

Click to expand...

Very reluctant and conditional

Great Bustard said:

Great Bustard said:

Ability to apply an art filter, decide on a gradation, global noise filtering, contrast etc after the fact is non-arguable advantage in a situation where photographer has to work fast and may not have time to set optimal settings. Mixed difficult lighting where you just don't know what to set for WB, better shoot Raw and think later. Sharply divided WB (example: a room with artificial lighting and a large opening into the outdoor with natural light) - much easier to shoot Raw, produce two JPEGs with different WBs and combine them into one. Scene with very wide DR which must be bracketed - better do it in Raw. Shooting a whole series in similar conditions that are likely to require identical treatment - easier to shoot in Raw, fine-tune the settings and then run whole series through the batch treatment (although this one can be dione in JPEG almost as easy, but still).

Click to expand...

Gotcha.

Great Bustard said:

I am sure you understand that saying "WB and tone curve are not of much significance" I meant that they are fairly easy to change in JPEG; not that they do not matter

Click to expand...

Yes -- I understood that's what you meant. What I didn't understand was why changing the tone curve from a jpg would not result in a "significant" loss of quality as opposed to changing the tone curve in the RAW conversion. Of course, "significant" is in quotes because what constitutes "significant" depends on many factors.

Click to expand...

Never observed it to a significance that would reveal itself in any kind of inspection. Like I said before, I had posted images processed both ways with 100% crops and asked viewers to tell them apart. No luck. More than that - unless I make a note, I can't tell them apart myself. Now, considering that I can do in JPEG editing that I simply can't do in Raw, why would I trade real advantage for a theoretical one...

Great Bustard said:

Great Bustard said:

Great Bustard said:

In terms of the tone curve, well, a jpg is 8 bits, so it can muster at most 8 levels within the DR. However, most cameras have more than 8 stops of DR. So a jpg, by "virtue" of it's 8 bit files, must either cut down to a maximum of 8 stops of DR, or represent multiple levels within the DR indistinguishably.

Click to expand...

Sorry, this above just doesn't sound right to me. Look at any DR measurement in any of the reviews, you will see much wider DR in JPEGs. I am not sure where this is coming from, nor am I proficient in the academic side of things enough for the discussion in these terms.

Click to expand...

DR -- dynamic range -- is the number of stops from the read noise to the saturation point of a pixel. If the bit depth is less than the DR, then it simply gives the maximum number of levels of the DR that can be distinctly recorded. For example, an 8 bit jpg can render 14 stops of DR, but it can only show 8 distinct levels of that DR.

If the bit-depth is greater than the DR, then it's simply an inefficient use of memory. For example, if a sensor has 11 stops of DR, then a 14 bit ADC has no advantage over a 12 bit ADC, and simply makes for larger files with no increase in IQ.

The E5 has 10.5 stops of DR, a jpg can render 8 levels of those 10.5 stops. That is, it can capture 8 stops of DR, each level being distinguishable, or it can capture all 10.5 stops, but some levels won't be distinguishable.

Since most display media (prints, monitors) only have 8-9 stops of DR, a jpg will work well for both prints and monitor display, so long as the jpg had the desired tone curve applied at the time of capture.

The advantage of RAW, in this respect, is that all 10.5 stops of the DR are recorded, and we can decide after the fact what range and what levels are to be represented in the jpg. If the initial capture is a jpg, then we are "stuck" with the range of the initial capture, which can be rendered with at most 8 levels.

I hope that makes some kind of sense.

Click to expand...

It does. Got it now, thank you for explanation. In my case, one of three things (or combination of them) takes care of that:

1. Envisioning intent prior to shooting and making sure I got it right. I realize the desire of some to have this decision postponed for later, but it's a personal matter - some want to do it in front of their computer and some - during shooting, getting it in camera close to intended result. I am of latter kind.

2. Preserving highlights and correcting shadows in PP - works fine if scene DR is just a tad wider than what I need from the image.

3. Bracketing - for the cases when scene DR is significantly wider. Although I refer to switch to Raw for that but it's not mandatory.

John King said:

pris said:

nick_webster said:

For people who want to do extensive PP then surely RAW is the most sensible answer, which is what I shoot

Click to expand...

"Surely" eh... It's a somewhat too narrow view of PP, no offense. There is more than one PP workflow, PP purpose and the way to reach it.

The kind of PP I do requires LAB colorspace which takes one right into JPEG. Don't think for a second that this some kind of marginal exotic treatment - although not as mainstream as RGB or CMYK, this colorspace has quite a following.

Click to expand...

One can use LAB colour space on PSD-16 files. No need to use it only on JPEG files. ??

Click to expand...

No idea, never tried to convert JPEG to PSD before starting PP to see if there is any advantage to it. All materials I use as guidance for LAB work use JPEG. PSD as saving format - sure, that's pretty much standard workflow. When reopening and continuing working on it, it's PSD of course, up until to eventual saving. Eventual format for displaying and printing will be JPEG anyway.

G'day again Pris

pris said:

John King said:

pris said:

nick_webster said:

For people who want to do extensive PP then surely RAW is the most sensible answer, which is what I shoot

Click to expand...

"Surely" eh... It's a somewhat too narrow view of PP, no offense. There is more than one PP workflow, PP purpose and the way to reach it.

The kind of PP I do requires LAB colorspace which takes one right into JPEG. Don't think for a second that this some kind of marginal exotic treatment - although not as mainstream as RGB or CMYK, this colorspace has quite a following.

Click to expand...

One can use LAB colour space on PSD-16 files. No need to use it only on JPEG files. ??

Click to expand...

No idea, never tried to convert JPEG to PSD before starting PP to see if there is any advantage to it. All materials I use as guidance for LAB work use JPEG. PSD as saving format - sure, that's pretty much standard workflow. When reopening and continuing working on it, it's PSD of course, up until to eventual saving. Eventual format for displaying and printing will be JPEG anyway.

Click to expand...

I always shoot RAW + SHQ/LF JPEG. I never use the JPEGs for any manual editing or printing. As someone else mentioned, I do consider them a further line of defence against possible loss of the RAW - better a decent JPEG than no image at all ...

I also never convert to JPEG. I print straight from the PSD file (ProPhotoRGB, 16 bit), and allow PS to manage the translation into the printer's CMYK and 8 bit colour space. Turn off the printer driver's colour management, of course ...

Seems to work well with all the printers I have worked with, whether dye based or pigment. All give (relatively) comparable results. Needless to say, there were some reasons as to why I bought an Epson R3880 ... .

--
Regards, john from Melbourne, Australia.
(see profile for current gear)
Please do not embed images from my web site without prior permission
I consider this to be a breach of my copyright.
-- -- --

The Camera doth not make the Man (or Woman) ...
Perhaps being kind to cats, dogs & children does ...

Gallery: http://canopuscomputing.com.au/gallery2/main.php



Bird Control Officers on active service.

Member of UK (and abroad) Photo Safari Group

Here's a 100% comparison and a 400% comparison - I don't see any difference - perhaps you could let me know what the plain differences are ?









I can possibly see a bit less smoothness in the background in the 400% crop ( not sure as I'm really tired and about to go to bed ), but I can't see this impacting on any real life situation. I see no more detail in the hairs.

As I mentioned in my previous reply you are welcome to download the full size jpgs from my gallery to compare yourself and post the differences you see,

Nick

Nick, in order to make any comparison its hard to find somewhere sharp on those images in the firstplace. Not criticizing your images, just saying they arent the best guide for what you are trying to discover..

--
Riley

any similarity to persons living or dead is coincidental and unintended

Great Bustard said:

krh_57 said:

Do you have any thing to add?
http://forums.dpreview.com/forums/read.asp?forum=1022&message=38311209

Click to expand...

To be honest, I don't know what you were trying to say in the post above.

Click to expand...

Let me help you then.

Snippet: " But anyone doing a camera or lens review, using those same LF settings for both O/N cameras. You would see a considerable difference, an unfair one at that, between the two in both the resolution chart patterns and LPH figures."

So to use an Oly 1/4 compression test file, to compare against a 1/2.7 compression test file for C/N by a review site, would be perfectly fine by you?

Great Bustard said:

Your initial comments on this matter, was that:

Click to expand...

Great Bustard said:

http://forums.dpreview.com/forums/read.asp?forum=1022&message=38291391

Click to expand...

And?

--
Ken

krh_57 said:

Great Bustard said:

krh_57 said:

Do you have any thing to add?
http://forums.dpreview.com/forums/read.asp?forum=1022&message=38311209

Click to expand...

To be honest, I don't know what you were trying to say in the post above.

Click to expand...

Let me help you then.

Snippet: " But anyone doing a camera or lens review, using those same LF settings for both O/N cameras. You would see a considerable difference, an unfair one at that, between the two in both the resolution chart patterns and LPH figures."

So to use an Oly 1/4 compression test file, to compare against a 1/2.7 compression test file for C/N by a review site, would be perfectly fine by you?

Click to expand...

If it didn't affect the outcome of the test, sure. And all the evidence in this thread is that it would make no difference in the outcome of their tests.

That said, how could they have known it would make no difference if they didn't shoot both compressions and compare? But why go to the trouble to do that rather than just shoot in the highest quality setting from the start?

My guess is that the review staff initially did not know about the LSF setting, as it was buried in the menus (and why would it be buried, rather than right there with the other compression settings?). After they found the LSF setting, they took a few shots in LSF and compared to the LF setting and found it made no difference, so they just kept the LF pics and results for the review.

After the uproar, DPR was sure to use LSF, but still maintained that it made no difference:

http://forums.dpreview.com/forums/read.asp?forum=1022&message=38302069

That's just a guess, though.

krh_57 said:

krh_57 said:

Your initial comments on this matter, was that:

Click to expand...

krh_57 said:

http://forums.dpreview.com/forums/read.asp?forum=1022&message=38291391

Click to expand...

And?

Click to expand...

And it seems a much more reasonable guess than what was written in the link above:

But here is a great example of DPR and their web pundits finding any excuse they could possibly wring out of a situation, where DPR was clearly wrong.

Lets face it, they couldn't give a cr@p about Olympus, and tried everything they could to disprove Olympus's worth.

I am constantly amazed how DPR and others think can better engineer, even a company like Olympus... You guys think your can, but your only kidding yourselves!

And some like DPR, have to lie and find cover to get their points across. What a freaking joke.

Or do you still feel that my "big head" is still just "coming up with excuses"?

he could plainly see a difference and asked for crops even though I've put the images fullsize in my gallery for anybody to download and play with. So I've posted some crops for him and I can't see any significant difference, perhaps he'll get back to me and show me what I'm missing - that's why I posted them.

Some of the other shots and crops I've posted have been sharper - look at the 3rd sample I posted : http://forums.dpreview.com/forums/read.asp?forum=1022&message=38308460 where you can see individual pollen grains or the 800% crops : http://forums.dpreview.com/forums/read.asp?forum=1022&message=38311232 and even at that pixel level I'm not seeing anything that would show up in any real life situation I can imagine.

I've got no agenda here, just driven by curiousity and surprise at my findings. Initially I thought that it might be subject dependant, but I've taken numerous shots that I've not posted that show exactly the same thing - I can't see any tangible difference between LN and LF or LSF jpgs. Perhaps there are some situations where the differences are tangible, costas ear's examples from an E-5 showed some difference but his methodology was flawed and he didn't bother to repeat the test converting the same one RAW image in camera. It might also be an E-3 specific finding, but again nobody has posted examples to show this one way or another so it remains a theory only. You're more than welcome to join in with examples

Really have to go to bed now, I'll look in again when I'm up,

Nick

pris said:

pris said:

Yes -- I understood that's what you meant. What I didn't understand was why changing the tone curve from a jpg would not result in a "significant" loss of quality as opposed to changing the tone curve in the RAW conversion. Of course, "significant" is in quotes because what constitutes "significant" depends on many factors.

Click to expand...

Never observed it to a significance that would reveal itself in any kind of inspection. Like I said before, I had posted images processed both ways with 100% crops and asked viewers to tell them apart. No luck. More than that - unless I make a note, I can't tell them apart myself. Now, considering that I can do in JPEG editing that I simply can't do in Raw, why would I trade real advantage for a theoretical one...

Click to expand...

I make some adjustments in the RAW conversion, but convert the RAW to a tiff, and do further editing there, which is lossless. We discussed the difference between tiff and jpg a bit back:

http://forums.dpreview.com/forums/read.asp?forum=1022&message=37714880

Sometimes, the jpg compression really messes up a pic compared to the original tiff, such as with this shot here...

So, I only use jpg for web display, and print the tiff.

pris said:

pris said:

DR -- dynamic range -- is the number of stops from the read noise to the saturation point of a pixel. If the bit depth is less than the DR, then it simply gives the maximum number of levels of the DR that can be distinctly recorded. For example, an 8 bit jpg can render 14 stops of DR, but it can only show 8 distinct levels of that DR.

If the bit-depth is greater than the DR, then it's simply an inefficient use of memory. For example, if a sensor has 11 stops of DR, then a 14 bit ADC has no advantage over a 12 bit ADC, and simply makes for larger files with no increase in IQ.

The E5 has 10.5 stops of DR, a jpg can render 8 levels of those 10.5 stops. That is, it can capture 8 stops of DR, each level being distinguishable, or it can capture all 10.5 stops, but some levels won't be distinguishable.

Since most display media (prints, monitors) only have 8-9 stops of DR, a jpg will work well for both prints and monitor display, so long as the jpg had the desired tone curve applied at the time of capture.

The advantage of RAW, in this respect, is that all 10.5 stops of the DR are recorded, and we can decide after the fact what range and what levels are to be represented in the jpg. If the initial capture is a jpg, then we are "stuck" with the range of the initial capture, which can be rendered with at most 8 levels.

I hope that makes some kind of sense.

Click to expand...

It does. Got it now, thank you for explanation. In my case, one of three things (or combination of them) takes care of that:

1. Envisioning intent prior to shooting and making sure I got it right. I realize the desire of some to have this decision postponed for later, but it's a personal matter - some want to do it in front of their computer and some - during shooting, getting it in camera close to intended result. I am of latter kind.

2. Preserving highlights and correcting shadows in PP - works fine if scene DR is just a tad wider than what I need from the image.

3. Bracketing - for the cases when scene DR is significantly wider. Although I refer to switch to Raw for that but it's not mandatory.

Click to expand...

If the camera has the tone curve you want for the shot, and you select everything right in-camera, then the in-camera jpg should be pretty much on the money, since the display medium (print or monitor) pretty much has the same DR as a jpg.

However, the lattitude that RAW allows over a jpg in terms of after-the-fact decisions about the tone curve, WB, etc., simply makes RAW well worth the extra memory and processing time, in my opinion.

In other words, why shoot jpg for convenience, only to find that 199 pics were just what you wanted, but the most important one really would have benefited from RAW, and you'd gladly trade the other 199 for the one?

To each their own, of course. I'm not saying anyone's an idiot for not using RAW. But I am saying that I, personally, shot jpg for some time, and now really regret not having used RAW from day one. It's not like it mattered for most pics (most of my pics suck, anyway), but for those few memorable shots when it did matter, well, it's too late now.

Great Bustard said:

krh_57 said:

Great Bustard said:

krh_57 said:

Do you have any thing to add?
http://forums.dpreview.com/forums/read.asp?forum=1022&message=38311209

Click to expand...

To be honest, I don't know what you were trying to say in the post above.

Click to expand...

Let me help you then.

Snippet: " But anyone doing a camera or lens review, using those same LF settings for both O/N cameras. You would see a considerable difference, an unfair one at that, between the two in both the resolution chart patterns and LPH figures."

So to use an Oly 1/4 compression test file, to compare against a 1/2.7 compression test file for C/N by a review site, would be perfectly fine by you?

Click to expand...

If it didn't affect the outcome of the test, sure. And all the evidence in this thread is that it would make no difference in the outcome of their tests.

That said, how could they have known it would make no difference if they didn't shoot both compressions and compare? But why go to the trouble to do that rather than just shoot in the highest quality setting from the start?

My guess is that the review staff initially did not know about the LSF setting, as it was buried in the menus (and why would it be buried, rather than right there with the other compression settings?). After they found the LSF setting, they took a few shots in LSF and compared to the LF setting and found it made no difference, so they just kept the LF pics and results for the review.

After the uproar, DPR was sure to use LSF, but still maintained that it made no difference:

http://forums.dpreview.com/forums/read.asp?forum=1022&message=38302069

That's just a guess, though.

Great Bustard said:

Great Bustard said:

Your initial comments on this matter, was that:

Click to expand...

Great Bustard said:

http://forums.dpreview.com/forums/read.asp?forum=1022&message=38291391

Click to expand...

And?

Click to expand...

And it seems a much more reasonable guess than what was written in the link above:

But here is a great example of DPR and their web pundits finding any excuse they could possibly wring out of a situation, where DPR was clearly wrong.

Lets face it, they couldn't give a cr@p about Olympus, and tried everything they could to disprove Olympus's worth.

I am constantly amazed how DPR and others think can better engineer, even a company like Olympus... You guys think your can, but your only kidding yourselves!

And some like DPR, have to lie and find cover to get their points across. What a freaking joke.

Or do you still feel that my "big head" is still just "coming up with excuses"?

Click to expand...

--
Ken

GB
My above post was screwed up by a glitch at DPR, I won't have time to rewrite.
--
Ken

I also use the LAB colour space extensively myself as well as the full range of Photoshop tools but this has nothing to do with jpegs at all (you should know that Photoshop uses LAB internally for everything). The point of starting with RAW rather than jpegs is twofold - firstly one can retain 12 bits/channel during processing thus reducing the risk of posterization and secondly avoid the data loss which comes with the generation of jpegs in camera followed by saving once again as a jpeg after PP.

For myself I very rarely produce jpegs - just for on-line use such as challenges etc.

If you prefer for whatever reason to use jpegs as a starting point for PP, that's your choice. However nick_webster is correct in saying that RAW is the most sensible choice for extensive PP.

Frank

pris said:

nick_webster said:

For people who want to do extensive PP then surely RAW is the most sensible answer, which is what I shoot

Click to expand...

"Surely" eh... It's a somewhat too narrow view of PP, no offense. There is more than one PP workflow, PP purpose and the way to reach it.

The kind of PP I do requires LAB colorspace which takes one right into JPEG. Don't think for a second that this some kind of marginal exotic treatment - although not as mainstream as RGB or CMYK, this colorspace has quite a following.

I also make an extensive use of layers, masks and selections. Although Raw made good strides in this direction with local adjustments introduced in latest versions, it's still way less precise and convenient that Photoshop's implementation. I am not even speaking of the gamut of possibilities presented by layer blending modes.

For my work, I would have to generate JPEG from Raw as a first step of my PP. Thus to me, Raw will simply introduce unneeded step and slow down the process.

Going back to the original question of LSF vs LF: Multiple quotes from original statements by DPR cited over the course of last days include this (bold emphasis is mine):

if you're intending to use JPEGs directly as final output

This sounds to me like the impact of compression on PP process was admitted from the get-go as possible and likely factor, right? If so, then for a photographer intending such PP using lower compression makes full sense. Whether he should use Raw or not is whole other matter - for the purposes of LSF vs LF debate it's nothing but a red herring.

Click to expand...