RAW benifits/costs?

[John Sheehy]

You guys crack me up... :-)  wrong approach:

Presuming you meant (0, 0, 0) to (255, 255, 128),

Yes, that's what I meant to type, but isn't what came out.

otherwise it'd be blue, try this:

- create the image with S-curve and all.
- go to curve dialog
- select blue channel
- move the white point horizontally toward the blackpoint side until the midtones are gray
- add a midpoint to fix the curve
- move the white point back (curve will automatically clip with a shoulder)
[a href=\"index.php?act=findpost&pid=85239\"][{POST_SNAPBACK}][/a]

It will never be exact, though, and any further processing will bring that out.  Real RAW conversions and JPEGs also have other issues, much more complex than a curve applied to an RGB image; hues are shifted and saturation is altered based on RAW hue and saturation, black is not black, etc.

In an already-converted image, you do the best you can to WB.  In a yet-to-be-converted one, you do it right.

My demo was not intended to be a hopeless case that could not be adjusted, at least coarsely; it was a visual demonstration of the effects of inconsistent gamma on color balance.

[Wayne Lorentz]

Do you think RAW makes folks lazier?
[a href=\"index.php?act=findpost&pid=84738\"][{POST_SNAPBACK}][/a]

A lazy photographer will be a lazy photographer no matter what equipment or technology he has.

That said, I made the switch from JPEG to RAW about two years ago and will never go back.  Here's why:

Not everyone shoots in a studio.  If you're like me, and shoot outside, you don't have the time that a studio photographer has to compose your subject.  You can't tell a deer to "gimme that look again" after it bolts off through the woods.  Capturing the moment in RAW means that I have the moment, even if I have to extract it from the data with a computer.  At least I have an image I can work with.  Yes, I might have that same image in JPEG.  Or I might not.  If the exposure or some other factor wasn't quite right, I could end up with a garbage JPEG instead of a salvagable RAW file.

[John Sheehy]

I'd be interested in someone describing a test that demonstrates the shortcomings of adjusting JPG WB in Lightroom, not PS.
[a href=\"index.php?act=findpost&pid=85251\"][{POST_SNAPBACK}][/a]

That was about mutliple gammas used in an image, not about PS per se.

[32BT]

It will never be exact, though, and any further processing will bring that out.  Real RAW conversions and JPEGs also have other issues, much more complex than a curve applied to an RGB image; hues are shifted and saturation is altered based on RAW hue and saturation, black is not black, etc.

In an already-converted image, you do the best you can to WB.  In a yet-to-be-converted one, you do it right.

Agreed. In that respect I'm still not sure what to make of the white balance implementation in Aperture (and some of the other converters). Is the WB corrected after the fact? (As the WB adjustment module is the last in the row?)

[32BT]

Knowing the law of gravitation, I don’t have to drop all the dishes to see it work.

Sure, but varying a single channel only, as per the original example, does not introduce the problematic hue shifts. The hue shifts introduced there are a result of applying the cure (gamma curve instead of a linear curve). In other words, the dishes didn't break because of gravity, but because of trying to nail the jelly to the dishes afterwards... ;-)

[PeterLange]

In other words, the dishes didn't break because of gravity, but because of trying to nail the jelly to the dishes afterwards... ;-)

Or this way round.

Tell you what - I simply join your agreement with John:

>>  In an already-converted image, you do the best you can to WB.  In a yet-to-be-converted one, you do it right.  <<

Peter

--

[Mark D Segal]

Or this way round.

Tell you what - I simply join your agreement with John:

>>  In an already-converted image, you do the best you can to WB.  In a yet-to-be-converted one, you do it right.  <<

Peter

--
[a href=\"index.php?act=findpost&pid=85384\"][{POST_SNAPBACK}][/a]

Peter it isn't clear to me what you are saying here. First  when is the image converted - by the time it is downloaded from the camera and opened in Camera Raw, or after youo finish in Camera Raw and open it in Photoshop. Usually when we speak of "conversion" it is the latter. Now if we're on the same page with that, I'm not sure what " doing it right" means in this context. We can white-balance either in the raw converter (say Camera Raw or Lightroom) or in Photoshop in largely the same way. I know the data differs between these two platforms, but setting theory aside I haven't seen what difference it makes to the quality of real-world (16-bit) images whether one makes this adjustment in the one program or the other.

[Tim Gray]

At at least one point in the thread the question was how well (or how poorly) the "one click" WB tool in LR could deal with WB error in an 8 bit file.  The concensus was "nailing jello to a tree" - but that was in the context of PS, not LR.

If the issue is that the WB adjustment in 8 bit space causes clipping (or can cause more clipping than adjusting in 12/14bit space) then that's fine, but I'm just wondering when saying that the benefits of RAW are better dynamic range/headroom AND the ability to adjust WB after the fact - if the reference to WB is redundant.

I'll play around this weekend with some shots of a Gretag card, solux light and see what happens. For example, RAW, WB based on the dropper, convert to 8 bit tif.  Then compare to the same RAW file that's been WB'd based on, say tungsten, convert to 8 bit tif, bring back into Lightroom and re-white balance with the dropper, export to 8 bit tif and compare to the first file.

[jani]

At at least one point in the thread the question was how well (or how poorly) the "one click" WB tool in LR could deal with WB error in an 8 bit file.  The concensus was "nailing jello to a tree" - but that was in the context of PS, not LR.

If the issue is that the WB adjustment in 8 bit space causes clipping (or can cause more clipping than adjusting in 12/14bit space) then that's fine, but I'm just wondering when saying that the benefits of RAW are better dynamic range/headroom AND the ability to adjust WB after the fact - if the reference to WB is redundant.

I'll play around this weekend with some shots of a Gretag card, solux light and see what happens. For example, RAW, WB based on the dropper, convert to 8 bit tif.  Then compare to the same RAW file that's been WB'd based on, say tungsten, convert to 8 bit tif, bring back into Lightroom and re-white balance with the dropper, export to 8 bit tif and compare to the first file.
[{POST_SNAPBACK}][/a]

I tried to test JPEG vs RAW in Lightroom a real world example that I've struggled with earlier, without being completely satisfied with either alternative.

The lighting conditions in the scene are pretty bad; tungsten and fluorescent, spotlights bounced out of store windows, and it's night time. The JPEG was shot with auto white balance.

[a href=\"http://folk.uio.no/jani/hobbies/photo/test/high+shadow/_MG_8285.JPG]Original JPEG[/url], 2544x1696, ca. 1.75 MB
DNG from raw file, 3504x2336, ca. 7 MB

Feel free to download the above unprocessed files and experiment a bit for yourself for the sake of this discussion; sorry about the different sizes, but I don't shoot full size JPEGs (not that the quality is such that full size would be defendable, anyway).

The wall of the building in the background is painted concrete, I don't recall which colour exactly. The street sign has an almost perfectly white plastic bag over it. The girls' tops were almost perfectly white, the jeans are of course jeans-blue. The concrete stairs were close to neutral grey, the tiles on the ground are perhaps a shade warmer than neutral.

The results by using the WB tool on the JPEG are abysmal; skin tones are either too pink, or other parts of the image are way too blue, or a combination. No matter how I tweak colour temperature or tint, I can't get close to what I get from the raw conversion.

(my interpretation)

Is that worth it? Not to me.

FWIW, Lightroom makes the raw adjustments significantly easier than ACR ever did, I feel like I have better control of the light curve. But I'd still like to have even more control over the light curve, similar to that of the curves tool in Photoshop, or perhaps something even better, whatever that would be.

[Mark D Segal]

At at least one point in the thread the question was how well (or how poorly) the "one click" WB tool in LR could deal with WB error in an 8 bit file.  The concensus was "nailing jello to a tree" - but that was in the context of PS, not LR.

If the issue is that the WB adjustment in 8 bit space causes clipping (or can cause more clipping than adjusting in 12/14bit space) then that's fine, but I'm just wondering when saying that the benefits of RAW are better dynamic range/headroom AND the ability to adjust WB after the fact - if the reference to WB is redundant.

I'll play around this weekend with some shots of a Gretag card, solux light and see what happens. For example, RAW, WB based on the dropper, convert to 8 bit tif.  Then compare to the same RAW file that's been WB'd based on, say tungsten, convert to 8 bit tif, bring back into Lightroom and re-white balance with the dropper, export to 8 bit tif and compare to the first file.
[a href=\"index.php?act=findpost&pid=85399\"][{POST_SNAPBACK}][/a]

Tim, unless I don't understand properly - always possible - it seems to me that too many variables are being comingled here. If you are going to test anything, best to sort out the questions and the procedures to isolate one variable at a time.

I doubt that 8 or 16 bit image mode has anything to do with WB or the ability to achieve an acceptable WB by making adjustments in any of these programs. But that is one question to test all by itself.

As far as I understand, the WB dropper in Lightroom is about the same as the WB dropper in Camera Raw, is about the same thing as the middle grey dropper in Curves in PSCS 2. You point all of them to something you think or are sure is neutral and it will make that hue neutral and adjust everything else "pro rata". So the second question is whether for a given bit depth any one of these programs do it better IN REALITY than the others.

I believe that in the process of applying a WB adjustment, the larger that adjustment needs to be the more assurance you will have from both a RAW file and higher bit depth that the risk of image break-up is minimized, but testing the exact value of that "insurance" (RAW versus JPG and/or 8 bit versus 16 bit) is yet another question.

[dlashier]

If the issue is that the WB adjustment in 8 bit space causes clipping (or can cause more clipping than adjusting in 12/14bit space) then that's fine, but I'm just wondering when saying that the benefits of RAW are better dynamic range/headroom AND the ability to adjust WB after the fact - if the reference to WB is redundant.
[a href=\"index.php?act=findpost&pid=85399\"][{POST_SNAPBACK}][/a]

The clipping issue I'm referring to has nothing to do with 8/16 bit space but rather the fact that with an image like the example I provided earlier, WB choice at raw conversion time affects clipping at both ends of the spectrum, and once this data is gone there is no way to recover it for a different WB without guessing. This is aside from the fact applying gamma and tone curves have been applied that further stretch or compress parts of the data. I would think that a very smart algorithm that was aware of preceding tonal/gamma adjustments could accurately do after the fact WB adjustment (preferably on a 16 bit image), but I don't see how it could account for clipped data.

Again, non-believers can try the example I posted. The original impetus for that example was because Bernd ran into exactly this sort of problem when trying to adjust the WB in Photoshop.

- DL

[PeterLange]

[{POST_SNAPBACK}][/a]

[a href=\"index.php?act=findpost&pid=85447\"][{POST_SNAPBACK}][/a]

Mark,

I’m not sure what I could further contribute. Your last post seems to answer your previous one perfectly.

The second aspect of WB Raw vs JPG (beside channel clipping) is Uniformity.  Without indenting to advocate for neutral white, for example, when you adjust one single bright gray patch to R=G=B, does it lead to neutrals all along the grayscale, or would a brighter gray still show e.g. a green color cast while a darker gray patch might reveal a complementary magenta hue.  For illustration, above suggested tests (as they read now) should have their validity.  Relevance will certainly depend on the image content, light conditions, personal tolerance and the WB approach in detail, etc.  Let’s hope that the WB dropper in LR at least switches from linear scaling via single R/G/B ‘highlights’ (perfect with Raw data) to something more complex & curved for already-processed JPGs; as it was good [a href=\"http://www.creativepro.com/printerfriendly/story/13811.html]tradition[/url] in Photoshop.

Peter

--

[howiesmith]

I may have missed it, but I can't find any info here on the cost of RAW.

IF you could get two identical cameras except for RAW (one with, one without), how much difference in price should there be?  How much more would you actually pay (not might be interested at this price)?

[Tim Gray]

I don't think the difference in camera cost is the issue.  All cameras actually capture RAW in the beginning regardless of what happens later - maybe a bigger buffer etc.  The incremental cost is in needing more field storage, higher horsepower processing and more archival storage, as well as potentially conversion time (although that's easier to manage/mitigate).

[howiesmith]

I don't think the difference in camera cost is the issue.  All cameras actually capture RAW in the beginning regardless of what happens later - maybe a bigger buffer etc.  The incremental cost is in needing more field storage, higher horsepower processing and more archival storage, as well as potentially conversion time (although that's easier to manage/mitigate).
[a href=\"index.php?act=findpost&pid=85624\"][{POST_SNAPBACK}][/a]

Understood.  But all that comes with a $$$ cost.  What is it?  And how much are you willing to spend to get it - even if it is there and just discarded.

Horsepower is your example.  I can buy a car that is essentially equivalent to another but for horsepower (and the stuff that comes with more or less horsepower, like mpg.)  How much more will I pay for a car with just more power?

Let's say I have a kit that you can install on a G7 to get RAW output.  How much would you pay for it?

[Tim Gray]

Understood.  But all that comes with a $$$ cost.  What is it?  And how much are you willing to spend to get it - even if it is there and just discarded.

Horsepower is your example.  I can buy a car that is essentially equivalent to another but for horsepower (and the stuff that comes with more or less horsepower, like mpg.)  How much more will I pay for a car with just more power?

Let's say I have a kit that you can install on a G7 to get RAW output.  How much would you pay for it?
[a href=\"index.php?act=findpost&pid=85632\"][{POST_SNAPBACK}][/a]

From strictly the camera perspective I would be irritated to have to pay anything.  As I said, the camera already creates a RAW file.  But to specifically answer the question, I'd pay - say $100 to get a 3 or 4 shot buffer with, say 1 second between shots.  I'd pay $200 to get a 5 shot buffer with no noticeable lag between shots.  I'd pay $300 to get a 15 shot buffer etc. etc. etc.

"Getting it" is more than just the cost of the camera.  If I was shooting JPG and was considering RAW (I shoot roughly 20,000 frames per year) I'd need to add another 4 gig in CF cards (assuming my 1d2) and 8 gig if it was a 1ds2.  I'd need a bigger drive in my Hyperdrive - say $50 incremental difference between a 60 gig and 120 gig.  I'd want 2 gig ram in my pc over 1 - say another $100. ($ CDN)  Archiving depends.  I only save the raw files that I work up and I only save the raw and final jpg, not multilayer 16 bit tif files, so your mileage will vary.  If I upgrade to the mythical 1d3 I'd probably get an entire new PC system.

[howiesmith]

From strictly the camera perspective I would be irritated to have to pay anything.  As I said, the camera already creates a RAW file.  But to specifically answer the question, I'd pay - say $100 to get a 3 or 4 shot buffer with, say 1 second between shots.  I'd pay $200 to get a 5 shot buffer with no noticeable lag between shots.  I'd pay $300 to get a 15 shot buffer etc. etc. etc.


[a href=\"index.php?act=findpost&pid=85662\"][{POST_SNAPBACK}][/a]

I think you know you are paying to get RAW now.  There is no free lunch.

But, at about $450 for a RAWless G7, you would pay 20 to 60% more to get  a G7 with RAW?  That is a lot of money to some folks who just want a "nice camera."

[Tim Gray]

Well, I wouldn't buy a G7 at any price since my Sony R1 (with RAW) is perfect for my needs as a second camera.  But, if a "P&S" had everything I wanted except RAW, I'd pay $100 to get basic raw capability, IF I wanted a bigger buffer and faster read times, I'd pay for that as well.  Regardless, a P&S with everything I want except RAW is going to cost significantly more than $500.

The lions share of the cost attributable to RAW in my 1d2 isn't having raw per se, rather its the cost of the buffer and speed with which 8 mpx files can be captured.  As I said before, RAW is captured by every digital camera made.  What you're paying for in a camera with no raw option, is the programming that processes  raw, ie the in camera jpg processing and associated options - sharpening, saturation, file size, degree of compression etc.  There's no free jpg lunch either.  The creation of the raw file is always there, it's not an option.

[thomashoven]

I am a lazy photographer - I shoot all my stuff in RAW even if I only use a digicam. Then I don't need to think about what format when I shoot. From my Digicam, $100 worth of memorycards can hold 100s of pics - no need to run out of memory these days.

At home, I batch-convert everything into JPG, and keep both the RAW and JPG files (DVD discs are also cheap these days).

With family and friends, I fool around with my JPG files. If I find a smashing shot where the JPG look terrible, I go back to the RAW of that image for a customized conversion/editing.

Rgds,
Thomas

[rainer_v]

mcgyver,- may i ask what this tread has  to do in any way with mf backs?
i am not often unpolied, but if you were shooting jpegs till now, you might find really better forums for such fundamental steps.