RAW files, a gimmick trick...?

... are always inspiring.
They are :-)
For some reasons, I assumed, that it cannot be sooner than page 31 of
my Profile.
Yes, a reasonable assumption if you hadn't noticed that the "months ago" counter starts at "0" instead of "1".

I'm sure it must have seemed like a good idea at the time when Phil wrote the code :-(

--
John Bean [BST/GMT+1] ('British Stupid Time')

PAW 2007 Week 39:
http://waterfoot.smugmug.com/gallery/2321711/3/201701188/Large



Index page: http://waterfoot.smugmug.com
Latest walkabout (30 July 2007):
http://waterfoot.smugmug.com/gallery/3247039
 
to RIP ;-)
My notes were solely based on the few thoughts from Ben Willmore's "Photoshop CS2 Studio Techniques" , for long years (and previous editions) one of the most universal and comprehensive books around on the subject. I thought it will be pointed immediately after my post.
I am still surprised, how emotional the RAW issue is...
(--)
 
My notes were solely based on the few thoughts from Ben Willmore's
"Photoshop CS2 Studio Techniques" , for long years (and previous
editions) one of the most universal and comprehensive books around on
the subject. I thought it will be pointed immediately after my post.
And that's its problem - it's an old book based on an ancient version of Photoshop brought up to date in a less than perfect way, giving a false impression of modern versions. I've been using PS since version 5 and support for 16-bit editing has increased with every new version.

I currently use CS2 and my whole workflow is 16-bit; there are a few things I can't use in 16-bit mode but nothing I'd normally use anyway - most are just 8-bit filters and plugins that have carried forward from old versions for backward compatibility and in most cases have much better new replacements in CS2. All core tools and features work in any mode except 32-bit, but including 16-bit of course.
I am still surprised, how emotional the RAW issue is...
Yes, it becomes almost religious. I too find it odd that discussions get so heated over a simple user choice; choice is always A Good Thing in my book.

--
John Bean [BST/GMT+1] ('British Stupid Time')

PAW 2007 Week 39:
http://waterfoot.smugmug.com/gallery/2321711/3/201701188/Large



Index page: http://waterfoot.smugmug.com
Latest walkabout (30 July 2007):
http://waterfoot.smugmug.com/gallery/3247039
 
Once you get a serious camera and get serious, you'll know.

Raw is the digital negative of the photo you've taken . When you
have the negative, you can re-edit the photo. Raw files hold ALL the
data uncompressed. But its up to you to finish the photo using a
Raw program.

RAW is not for everyone.

==========================
Hi guys, before I get to the point let me enlist some characteristics
between raw files and jpeg....
 
1. I agree. CS2 seems to be pretty well prepared for 16-bit channels, however it cannot save 16-bit file as jpg and gif.
2. What kind of immediate advantage do you have of working with 16-bit files?

3. In general nothing to argue about the expanded possibilities of the RAW file as long as the feature is not addressed to people, who would never have reason and tools to use it . Slightly more than 99 percent of us.
(--)
 
1. I agree. CS2 seems to be pretty well prepared for 16-bit
channels, however it cannot save 16-bit file as jpg and gif.
Thats probably because 16 Bit jpeg are not official standard and 16 bit gifs are an abomination.
2. What kind of immediate advantage do you have of working with
16-bit files?
You have greater control over how the 12 - 14 Bits of data from the sensor are pressed into the 8 Bits of a JPEG. This includes better mapping / gamma curves, less data loss due to rounding (if you stretch the 12 Bit to 16 Bit), no data lost due to clipping during WB and a couple of other things
3. In general nothing to argue about the expanded possibilities of
the RAW file as long as the feature is not addressed to people, who
would never have reason and tools to use it . Slightly more than 99
percent of us.
I agree that everyone should use what he likes. If you find that at some point JPEG limits you, its nice to know that there also is RAW instead.

I also agree that probably more than 99% never use raw, but this includes millions of P&S and Cellphone photographers.

In this forum, which leans a bit more to the DSLR side and is frequented by more technical people i'm pretty sure that at least 10 - 20% use raw at least some of the time and probably more than 1% use it most of the time.

Bye,
Philip
 
1. I agree. CS2 seems to be pretty well prepared for 16-bit
channels, however it cannot save 16-bit file as jpg and gif.
Well, if GIF and JPEG were 16-bit formats I'm sure it could - but since they're 8-bit formats then of course it can't ;-)

However it's not exactly hard to change the image mode to 8-bit (Image Menu: Mode-> 8-bit) before saving to an 8-bit file format.
2. What kind of immediate advantage do you have of working with
16-bit files?
No gaps in the histogram (and no banding) when applying tone curves, especially when recovering shadow detail and in areas of smooth tone like skies.

To be honest I can't see the point in not using it since it does no harm and can avoid some numeric pitfalls of working in 8-bit mode. Some modern editors always use 16-bit (and sometimes 32-bit) internally even on 8-bit files and convert back to 8-bit on-the-fly for saving for this very reason.
3. In general nothing to argue about the expanded possibilities of
the RAW file as long as the feature is not addressed to people, who
would never have reason and tools to use it . Slightly more than 99
percent of us.
And as long as those non-users also avoid the use of misinformation as a means of "disproving" the advantages raw offers to those who do use it.

--
John Bean [BST/GMT+1] ('British Stupid Time')

PAW 2007 Week 39:
http://waterfoot.smugmug.com/gallery/2321711/3/201701188/Large



Index page: http://waterfoot.smugmug.com
Latest walkabout (30 July 2007):
http://waterfoot.smugmug.com/gallery/3247039
 
In the days of film many people shot their film and then turned it in to the local photo lab and accepted the results that came back.

Others spent hours in the darkroom developing and printing their images, with full control over the results.

Same difference between jpeg and RAW.
--
'I don't take snaps - I paint with light' - Tony Hancock
 
So, is raw feature merely a gimmick trick of camera makers ...
Or is it really a need...?
nor are cameras, or photography. these are instruments of our desires.

however if your question was not meant to be purely esoteric -- what prison of time have you been locked in? i can understand this kind of enquiry back in 2003, maybe 2004 -- but the ecosystem has massively shifted. we are on the cusp of 2008. search the archives. search google. perspective is relative. vision is deep. jpegs are shallow. the possibilities for life are infinite. :-)

the born 2B esoteric
design guy
 
Can't we use today's software on older jpeg images we shot six or
seven years ago?

I'm sure that Photoshop CS3 has better jpeg tools than Photoshop 1 .
. . they just aren't making software better for RAW files only!
True, BUT the jpg does NOT contain ALL the information in the raw file. JPG is a LOSSY format. Plus that jpg was produced years ago by a not as advanced algorythm as is available today, so who knows who much was really lost.

No, this is a case where having the raw will (and does) produce the better result.

As I have said before, use raw or jpg, whichever you PERSONALLY prefer. trying to argue that no one should use raw is just plain silly, not to mention pointless.
--
J. Daniels
Colorful Colorado
Panasonic FZ10, FZ50
Olympus E-510
Fuji S602Z, A825

Remember . . . always keep the box and everything that came in it!
--

The greatest of mankind's criminals are those who delude themselves into thinking they have done 'the right thing.'
  • Rayna Butler
 
As I have said before, use raw or jpg, whichever you PERSONALLY
prefer. trying to argue that no one should use raw is just plain
silly, not to mention pointless.
If you go back and reread every one of my posts you'll find that I never said that no one should use RAW!

I did say it really isn't as necessary as many would try to make us believe, and to say that using RAW all the time is best for everyone is just as silly and pointless!

And, regardless of all that extra information that RAW files have, it isn't always a necessity to have to use it!

Again, I'm not arguing . . . I'm just stating my point of view as to the way I like to take photos . . .

I'm not trying to push my photo idiology on to anyone . . . if anything, this whole topic thread seems to be the full-time RAW users trying to push RAW on those who prefer jpegs.

--
J. Daniels
Colorful Colorado
Panasonic FZ10, FZ50
Olympus E-510
Fuji S602Z, A825



Remember . . . always keep the box and everything that came in it!
 
Holy cow.
So, is raw feature merely a gimmick trick of camera makers ...
Or is it really a need...?
nor are cameras, or photography. these are instruments of our desires.

however if your question was not meant to be purely esoteric -- what
prison of time have you been locked in? i can understand this kind of
enquiry back in 2003, maybe 2004 -- but the ecosystem has massively
shifted. we are on the cusp of 2008. search the archives. search
google. perspective is relative. vision is deep. jpegs are shallow.
the possibilities for life are infinite. :-)

the born 2B esoteric
design guy
 
Talk about ACR, DCRAW, or else, put those SOFTWAREs in your computer. You carefully notice differences of each clipped channel datas between in-camera jpegs and raws. You see numerical deviations in each R, G, and B, and correct them into what's called "numerically correct", which is: "zero deviation" or "deviation: 0%".

Let me put again my point: "I want my in-camera jpeg to be as close as or be exactly like what I can do my jpeg by post processing my raw files. That's because I believe camera makers are not trying hard enough in making the most optimum quality as a standard. I want it badly because we have paid them so much, paid for what we think is providing the best".

Now with the term of "numerical measurement", all computers are made for doing computations, and also numerical measurements. If you visually perceive a "clipped highlight, clipped shadow" and explain it as a "deviation R=2.2% G=1.2% B=0.5%", then computers should do just fine in defining such visual quality and numerical values.
How would you compare your brain's agile speed vs that of DIGIC's?

A DIGIC can draw megapixel data from sensor, apply in-camera processings I've mentioned above, and save it in just under a blink of your eyes.

And what if the DIGIC fed with a new extra processing algorithm to simulate just like what you do in your SOFTWAREs and your RAW files? The DIGIC will slow down about... maybe a little bit more than half blink of an eye...

Whatever your term is, processing your raw files, exploiting the raw, using the raw, or another words you may say, they all come from one thing: "PROCESSING". By processing, you turn the sliders right or left. In other word, you are applying algorithms onto the file. Good or bad the processing results are depend on your PROCESSING METHOD.

That's why I think that camera makers make those in-camera jpegs decent but not that good since they can squeeze more of it with a better method.

And, again, please notice that THIS IS NOT ABOUT RAW vs JPEG. It may be hard to discern since I use a thread title that "attacks" raw files. But I never intended it to be.

Yes jpeg is an industry standard not made by any camera makers, but the way they choose processing method make the output jpeg different from each other.

For example, in the same scene Panasonic's jpeg would smear heavily on the color while Nikon's or Canon's don't.

Raw lovers should humbly admit that:

1. no matter what, jpeg is a standard in all cameras, it is the main output intended
2. no matter what, raw workflow is not instant
3. when jpegs are ok, there's no further need in editing the raw

Jpeg lovers should humbly admit that:

1. it's not as flexible as raw, in term of extensive editability and data bit depth retention
2. it's mostly not as good as post-processed raw's jpeg
3. it is not good, since it can be squeezed more with better methods

Shoot raw, shoot jpeg, it's not a religion, it's a preference based on particular conditions. When you consider it as a religion, then you can't shoot otherwise for the whole of your life. No reasons for things like this heated again and again (I guess being in both sides, I'd make enemies in each...)

It's just about how camera makers should be able to push further their in-camera jpeg processing quality, and not leaving most quality processing to the user. That's why I said "gimmick", that unsatisfied picky people of their in-camera jpegs can still dig more from the raw. But like I said in "raw admission #3", if jpegs are ok, no further need of digging more. And how to "ok"-ing the jpegs?? I'll get dizzy explaining it again and again and again.

Huuuge responses... really exciting:)
 
Anyway, GUI's raw-jpeg comparison of blownout bikini girl should add my suspicions to camera makers (no offense to you GUI), that the undeveloped raw data has a completely different histogram response from the in-camera jpeg (in the middle vs completely in the right of histogram).

What I would expect from such histogram is that they should look almost similar since the exposure value (whether numerical or perceptual) of the camera is set by Shutter Speed, and Aperture, NOT by internal camera processing (I don't know if any digital camera would do otherwise, tell me then...)

So at least the undeveloped raw and in-camera jpeg have almost default identical histogram response.

What I expect from the raw is, while it has almost identical clipped histogram response, that it can be dragged into the middle from the right histogram wall safely since it contains deeper bit of data.

Like I said before, if you can measure numbers in histogram data (in this case the RGB deviation percentage), a dedicated computer like DIGIC should be able to outpower you, or your brain (in this case in making final undeviated-jpeg output).

Human brain is the greatest that it can make an artificial brain works faster than the real one.

If Jpeg couldn't be better today, at least it will tomorrow.

Remember, I never hate raws, I only hate half-heartly-processed jpegs ;)
 
No, the histograms from raw and jpegs vary greatly. RAW files are linear, which are not pleasing to the eye. They have to go through a tone mapping to look correct. Half the data in the raw file goes to the brightest f-stop.
 
Don't forget how much data is being processed. The JPG is converted 8-bit from a nominal 12, or now 14, bit batch of raw data. The conversion that is done "in camera" is fixed, what you can do in your computer is infinite. Raw speed, as you note, is far from the only consideration when talking about conversion and the use of RAW data.

You note, correctly, that the CPU in your computer is "general purpose", but you also don't take into account what else is using cycles on your computer when you calculate speed. But again, speed isn't everything, not by a long shot.

Can you honestly say that a JPG out of your 20D cannot be improved upon from the RAW data when both are processed appropriately? Simply looking at the difference between 8 and 12 bit data in regards to color and tonal gradations shows that working from any higher bit count data will yield better results.

--
Bill Dewey
http://www.deweydrive.com
 
How powerful is the processor in your computer? Answer, much more
powerful than the processor in your camera. Does your computer have
more RAM than your camera? Answer, unless your computer is really
old and moldy :-). Does your camera have severe time constraints on
how long it can take to process that JPG? Answer, absolutely.
The processors in cameras are made for the express purpose of
converting RAW data into JPEG image files. The processor in your
computer isn't. This makes a big difference. Canon supplies a SDK
(Software Development Kit) that performs the exact same processing
that the in camera processor does. My 20D can process 5 images per
second. My 2.4 GHz Core 2 with 2 gigs RAM takes approx. 2 seconds to
process each image when running a RAW converter that uses the Canon
SDK (BreezeBrowser.)

Rough calculations of the above data show that the processor in my
20D is at least 10 times faster than my 2.4 GHz core 2 CPU. For
doing RAW conversions.

(Are there any RAW converters that can beat the 20D's 5
images/per/second rate? Running on a reasonable PC.)

Wayne
You mean you don't have the 20D hack that lets you run photoshop on your camera?!?!
--
http://www.pbase.com/ewhalen

 
Let me put again my point: "I want my in-camera jpeg to be as close
as or be exactly like what I can do my jpeg by post processing my raw
files.
cookie cutter algorithms can only get you the same cookie.
for everything else there is Photoshop.
That's because I believe camera makers are not trying hard
enough in making the most optimum quality as a standard.
not hard enough? or is that just another rhetorical deception?

"optimum quality" will differ per each person's taste and style. when photographers took responsibility for IQ, they'd pop in a roll of film with the characteristics they favored. sometimes those characteristics could be further massaged by their choice of developer and developing process. there were trade offs. "optimum quality" was relative. same principles apply today. this is why I suggested you research the topic a little more. you seem to be coming out of an "egg" - and a little late.
And what if the DIGIC fed with a new extra processing algorithm to
simulate just like what you do in your SOFTWAREs and your RAW files?
The DIGIC will slow down about... maybe a little bit more than half
blink of an eye...
you would still need to work within the scope of what you are given. that can vary from scene to scene. is it better to futz with this before the shot or after? why complicate the decision making? complication adds uncertainty -- not to mention expense. trying to come up with the ultimate metamorphic cookie cutter can ruin your jpeg output as much as it could enhance them. there are no second chances with jpeg. either you nailed it -- or not.
That's why I think that camera makers make those in-camera jpegs
decent but not that good since they can squeeze more of it with a
better method.
it's becoming clear that you don't understand raw processing. jpegs are much like OREO cookies. some cameras allow you to modify the filling a little. you seem to want to modify it a lot. consumers like OREOs. they know what they look like. most don't want to complicate their choices down to the type of wheat gluten used. if they did, they'd bake their cookies at home with Photoshop.
It's just about how camera makers should be able to push further
their in-camera jpeg processing quality, and not leaving most quality
processing to the user.
again it seems you don't clearly understand the difference between formats. why would a camera maker build a custom software/hardware interface for your detailed pre-processing choices and bill every camera owner for it? we can't anticipate every decision in advance. neither can the camera maker. RAW output gives you the option of a second chance, JPEG gives you the convenience of a cookie cutter. some cameras permit you to bake both at the same time.

what i really think you are saying, and correct me if i'm wrong, is that you don't feel comfortable post-processing, and you'd prefer to have the camera come with a built-in version of Photoshop, and a built in art director. it might save you from taking responsibility for what you don't know. then again, it might not. you're obviously not content with the current art director.

if that's the case -- best of luck! :-D

the born 2 crack eggs ... ;-)
design guy
 
Anyway, GUI's raw-jpeg comparison of blownout bikini girl should add
my suspicions to camera makers (no offense to you GUI), that the
undeveloped raw data has a completely different histogram response
from the in-camera jpeg (in the middle vs completely in the right of
histogram).
What I would expect from such histogram is that they should look
almost similar since the exposure value (whether numerical or
perceptual) of the camera is set by Shutter Speed, and Aperture, NOT
by internal camera processing (I don't know if any digital camera
would do otherwise, tell me then...)
You are making a conceptual mistake here ikari: the JPEG image was RIGHT, as it was that way according to the user's settings. It's simply that HE made a mistake when chosing the exposure value. The camera cannot think for you, it doesn't know whether you want an underexposed, neutral, or overexposed appearance for your image. So if you ask the camera for an overexposed image (and this is what the user did in this case) an overexposed JPEG you will get.

A different question is if the RAW data still contained enough information to obtain a better image through exposure correction down, i.e. setting an exposure value less than that asked by the user at shooting time. And indeed it did.

But I insist: the camera was not wrong, it was the user who made the mistake. That's why RAW is good, allows among other things, correct USER'S MISTAKES in exposure which could never be corrected in the JPEG.

You cannot compare the RAW histogram I showed, which is linear (i.e. non gamma corrected) and had no white balance applied, with the final image histogram. If you look at the image of a RAW prior to gamma and white balance corrections they will be tremendously dark and green pictures. Something like this:



for this image:



I calculated it just to find out what real information was the RAW file containing.

--
http://www.guillermoluijk.com
 
I'm just wondering why you have to use someone elses images instead of your own to try and prove your point.

--
J. Daniels
Colorful Colorado
Panasonic FZ10, FZ50
Olympus E-510
Fuji S602Z, A825



Remember . . . always keep the box and everything that came in it!
 

Keyboard shortcuts

Back
Top