Raw is better how? 2

[BrianHuffman]

UPDIG is the recommended resource for image gatherers and image
purchasers. It is an association of numerous creative groups;
photographers, artists, retouchers, pre-press, etc., etc.

Here is the 'quick guide' to UPDIG protocol for image workflow with
digital gear. In case you want a hint, it is basically shoot RAW and
deliver TIFF. But for those of you with thick skulls, here is the
guide - http://www.updig.org/guidelines/quick.php

Argue with them. Not me.

Very informative, I quote:

"For printing, uncompressed TIFF is often preferred, although high-quality JPEGs (Level 10-12] can be visually indistinguishable from TIFFs and some printers prefer their smaller file size."

I guess all the information didn't get through your skull either. Also jury is still out on DNG, but they recommend. Frankly there is better information in these forums than on that site.

Again, I reiterate, what is wrong with the concept of you do it your way, I'll do it my way. (incidently I shoot raw, I just don't agree with calling folks thick skulled and then not even quoting your source correctly.

--
BrianH

 

[nanook s]

And being a good photographer is just about clicking a button once in a while

You obbviously never had experience living with the multiple 750gig USB drive solution that you are proposing . . . and not imaginative enough to ponder what happens when one of them gets a knock or overheat. My guess is that, you never had a situation where a client from 5 years ago want a print from her wedding. Can you imagine digging up 60 gig USB1.0 hard drives (from 5 years ago) and try to pick out a print?

I solve this problem by running three RAID-5 systems, backing each other up in three different rooms. As the drive network geometry complexity increases in order to accommodate more and more data, the fault tolerance level decreases. That's why life is about trade-offs.

 

[jimcreative]

Pardon me for using the term 'thick skulled' but if you had read most of what has been posted on this topic I think you would be less inclined to add your name to the long list of people who are short on information and long on insults.

UPDIG most definitely recommends RAW capture and TIFF delivery. They mention that a 10-12 level JPEG may be used and even preferred where size is an issue. I have not problem with that, but JPEG is still a lossy format and will not hold up well to edits.

Keep in mind, as I have stated in many of my posts, I am not producing 2000 images per day shooting weddings. I am shooting 12 image per day and probably do not produce 2000 images in a YEAR. For the optimal output quality, RAW capture and TIFF finalization is the preferred workflow.

If others want to do something else then that is fine. But making up baseless claims that there is nothing lost in shooting JPEGS, editing JPEGS and delivering JPEGS is pure unadulterated horse hockey.

Anyone who makes claims to the contrary is thick skulled. There, I said it again.

If someone like nanook will just state that "I prefer JPEG shooting because it saves space. I realize that I am givng up a lot of flexibility and the ability to recover highlights and shadows, but I am willing to trade those luxuries for the space that I save and for the time I can save by not having to deal with RAW"... I could respect that. But instead he and others claim, falsely, that photographer who shoot RAW do so because they are not skilled enough to get it right, or that there is no appreciable difference between a well executed JPEG and a well executed RAW file.

They are making excuses for their lack of willingness to spend the time to get the absolutely best image.

Oh, I can live with being called 'thick skulled' when I know that my stubborness is backed up with facts.

--

'We spend all of our lives pushing the buttons and pulling the levers found on the front panel of reality. How can we be so certain that there is also not a rear panel... one that only God can reach, and when He does flip an unseen switch or turn a dial that is out of our reach we see it as a Miracle?' JR
http://www.jimroofcreative.com

 

[nanook s]

1. Relatively minor, but still bigger than the minor adjustment that can be gained from shooting RAW . . . when you learn to expose correctly.

2. So why aren't you using 512Meg cards instead of 4GIG cards? After all, a 512Meg card can accommodate more than 36 RAW frames too. So why are you using cards bigger than 512Meg? The answer is quite simple: the less switching makes for less interruption and less chance of loss/mistake. As for ineptness, how many weddings do you cover in a year, again? Do we really want to go over the difference between ponit-and-shoot vs. photography?

3. So you do discard, after all. So what if your assistant makes a mistake in the culling and RAW conversion work for you? and toss the "motion" picture that you love so much but she deems it dud due to motion blur? Oh, wait, I forgot, you spend so much time online bashing other members that you don't have a business big enough to hire full-time assistants. For what it's worth, I actually show clients 95-98% of what I shoot. The shots from assistants get tossed a lot, but out of the 1500-2000 pictures that I personally get from a wedding, 95-98% are keepers.

 

[nanook s]

Absolute best image is judged by what the print shows, and its importance to the client. A missed shot in the middle of a card change will never be an "absolute best image"; a lost image in the culling and never backed up will never make an "absolute best image"; being over-run by the data bandwidth in order to preserve all-raw may lose a lot of "absolute best images."

JPEG can capture far more than what an "absolute best image" can deliver in print.

Since when is any DSLR the right tool for making "absolute best image" of still subjects? Why aren't you using large format film? Do you at least remove the anti-aliasing filter and Bayer filter from your sensor? Each one of them remove far more high spatial frequency data than JPEG filters.

 

[nanook s]

Actually, the best highlight/shadow recovery software out there use fractal interpolation to generate/guess the "details" that was never really there in the RAW file itself. That's the reality of life behind the anti-aliasing filter and Bayer filter (sensors only "see" the world through the filters). That's why different RAW conversion software have very deifferent detail recovery results.

 

[MICHAEL_61]

• The built in program dos in fact use a raw file info to make the jpeg, it converts it in cam then let the raw file drop. And saves the jpeg image made from it, now cam makers got a pretty good program to do this.

• You actually trust this program to be smarter than YOU?

• leave out the junk bits that not really needed.

-That's right, let's cut one of your ears off - it's just a junk, isn't it?

• With a raw file you use what ever software you want, and it has set default in most cases.

• What do you mean "default"? I develop every raw file INDIVIDUALLY.

• But that does not mean raw files have so much more info, because that not the case at all, it has very little more info over a jpeg.

• Yeah, right, very little - just 2/3 of the file.

• I'm pretty sure I got this stuff right

• You seem to think that no other file formats even exist. For good quality we save a TIFF file, not a jpeg.

http://lordofthelens.smugmug.com/

 

[nanook s]

"You actually trust this program to be smarter than YOU? "

The choice is between trusting the proprietary hardware built into the camera vs. trusting the software running on the desktop. I presume, not even the extreme fool-hardy masochists are into interpreting pixel by pixel by hand, using a byte editor. For quite a few years now, the best software RAW converter for image quality are the ones from the manufacturers . . . and that software essentially simulates what's done in the camera hardware.

 

[nanook s]

The purpose is to weed out the wannabes when wading through a sea of submissions. A similar guide line used to say something like we only accept 2 1/2 and 4x5 . . . that doesn't mean all pros should dump their 35mm equipment.

 

[MICHAEL_61]

• For quite a few years now, the best software RAW converter for image quality are the ones from the manufacturers . . . and that software essentially simulates what's done in the camera hardware.

• Right, I'm using DPP for RAW development. But I decide how to adjust WB in it, for instance - how can the camera possibly know that better than me?

http://lordofthelens.smugmug.com/

 

[Joe Pa]

I have done that test numerous times. There isn't nearly enough
difference to warrant several times the data bandwidth, unless I know
the person can not get exposure right within reasonable margin. The
biggest killer in image quality is the printing process, that and the
observer's visual accuity. Both toss away far more data than the
JPEG compression process. It's like comparing cars with 2000hp vs.
2500hp when the transmission can only handle 500lb-ft, and the tires
are only W-rated. The difference in engine power at that extreme is
quite pointless . . . little of the difference can be put to the
ground.

Actually JPEG is more like taking a 2000hp engine and putting a restricter plate on it, so that it cannot output maximum hp. I've personally run the test, and JPEG files suffer visually in saturation in as few as 3 iterations of compression.

Since you are such a stickler for the theoretical maximum quality of
capture, why aren't you taking every shot using large format film?
Percentage-wise, the anti-aliasing filter and the Bayer filter each
toss away far more data than JPEG passes. So why don't you? The
answer is quite simple, there are far more important factors in play
for the pro than the theoretical maximum capture quality.

There are laws of diminishing returns, for my use, a 35mm camera and RAW are more than ample for my needs. For some a medium format or large format camera with a 36MP digitial back is required.

Have you ever wondered whether the JPEG from a 1DsIII preserves more
detail or the RAW from a lower Nyquist frequency camera like 1DIII?
(or 1DsII JPEG vs. 1DII RAW, 1Ds JPEG vs. 1D RAW). JPEG is a great
tool for maximizing the efficient use of a given bandwidth.

The reason for this is simple, JPEG compression is dependant upon the amount of uncompressed data fed into it. It's the same reason the same image compressed in JPEG file from a 10MP camera is larger than one from a 5MP camera if the quality settings are kept the same. That doesn't change the fact that data is discarded based on the cameras programming of what is and isn't visibile or detectable by a human eye. No matter how efficient you claim it to be, it still discards data, data which might be criticial to pulling out shadows or highlights in an image.

 

[Joe Pa]

You throw out alot of terms and phrases but don't seem to understand what they really mean.

Fractal interpolation "creates" artificial details or data based on the available data. Fractal interpolation can generate new and higher spatial frequencies, but the additional detail that is added was not available in the source image, though it looks plausible. As an FYI, Vector Quantization based interpolation has been demonstrated to be superior and faster in some cases to fractal interpolation.

Regardless, interpolation of any kind is quite different from adjusting curves or exposure on a RAW file where all the data is present and you are not adding artificial data to improve the image or add details. Fractal interpolation is typically used to add detail back to JPEGS or poor quality images when the original data is not present.

As for anti-aliasing filters and Bayer filters, these are built into most cameras to avoid aliasing (too much detail to the sensor which could result in moire or jaggies in an image). The filters are designed to prevent / reduce aliasing from occuring by keeping the sampling frequency to half of the sensors sensel pitch (Nyquist frequency). The fact that some RAW viewers may show a moire effect while others dont has no relation to lost data or available detail. True the anti-alias filter and Bayer filter might cause images to appear softer and less sharp, but the impact would still be greater to the JPEG than RAW file so I don't understand your point for bringing it up in this discussion.

Actually, the best highlight/shadow recovery software out there use
fractal interpolation to generate/guess the "details" that was never
really there in the RAW file itself. That's the reality of life
behind the anti-aliasing filter and Bayer filter (sensors only "see"
the world through the filters). That's why different RAW conversion
software have very deifferent detail recovery results.

 

[Joe Pa]

The manufacturers RAW converter simulates what's done in the camera if you utilize the automatic options for processing. If you manually adjust the file, the result may have no resemblence to what the camera produces, and there lies the power of RAW processing.

• For quite a few years now, the best software RAW converter for

image quality are the ones from the manufacturers . . . and that
software essentially simulates what's done in the camera hardware.

 

[Joe Pa]

I don't really care if someone shoots RAW or JPEG. The OP'ers question was which is better, and from a purely technical standpoint there is no question processing RAW files will produce superior results to JPEG if the person is willing to spend the time to properly learn how and process them.

If someone hates sitting in front of computers or considers themselves a super pro that doesn't need to shoot in RAW then let them shoot JPEG. Those reasons aren't the technical basis for an argument which justifies JPEG as being superior to RAW, they are merely personal biases, and should be represented as such.

 

[Joe Pa]

Given the cost of all your RAID storage, servers, electricity and time, why not offload old weddings to an optical jukebox? Optical storage gives you true archival safety, on-line or near on-line access to all your images, and eliminates your need to back up and sync your servers constantly.

 

[nanook s]

Have you ever tried an optic jukebox first hand? Clunky and brittle are the first adjectives coming to mind. To come up with 2.5T from an optical jukebox, one would have to have a 500-DVD rolling stock. If shooting the same amount in RAW, that would be 1000 to 2000 discs, with each wedding spread over several discs. A basic discipline of data security is that maitenance only touches devices for replacement . . . the data retrieval process itself should involve as little mechanical intervention as possible. Somehow I think the jukebox solution violates that rule even if managing several discs in a jukebox were feasible.

 

[nanook s]

Please re-read my post. The fabulous highlight and shadow recovery that some poster attributes to their RAW conversion software is actually the fractal interpolation built into their algorithms. The two of us obviously agree that fractal interpolation is making up data . . . for RAW and for JPG.

JPG compression is based on spatial frequency analysis . . . the loss is in the high frequency domain. Anti-alias filter does the same thing; and the Bayer filter does the same for each color. It's ironic that all RAW converters use "sharpening," which essentially artifically concoct high frequency components into the image that had been lost due to the anti-alias and Bayer filters . . . yet the advocates would do the pot-calling-the-kettle dance when JPG users use similar techniques to get visual accuity back into the image.

 

[nanook s]

1. Shoot a grey card, and manual color balancing; that's a function built into most cameras;

2. As one gain experience, the Kelvin setting works even better.

For what it's worth, I often find the color balancing choice made by some manually converted RAW captures rather disturbing because the person sitting in front of the computer may not know which part of the scene was supposed to be true white on the day.

 

[nanook s]

I think it belongs to the power of the Jukebox Nifty idea for someone who is imagining things based on the first-hand knowledge with a few dozen discs (and the commensurate number of images), but a nightmare for anyone who actually tries it on a much larger scale.

 

[nanook s]

Shooting RAW vs. shooting JPEG are workflow issues. As such, they can only be judged in the context of an integral workflow. As for single shots, the far more relevant questions are: which cameras, which lenses and what kind prints . . . all of which are far more important tool decisions than RAW vs. JPEG.