Go Back   Steve's Digicams Forums > Digital Cameras (Point and Shoot) > Fujifilm

Reply
 
Thread Tools Search this Thread
Old May 5, 2006, 10:05 AM   #21
Senior Member
 
Join Date: May 2005
Posts: 1,063
Default

Betty,

I must say that I find it extremely curious that you would prefer s7raw over Adobe Camera Raw, even with Photoshop CS. There is an excellent book, "Real World Camera Raw with Photoshop CS" by Bruce Fraser that is really informative and excellent reading. There is a separate version for CS2 as well. But the book explains the unique characteristics of the raw format, and then teaches you how and why you do certain things in Camera Raw and other things in Photoshop. There is a full chapter on automating your workflow that is really worth reading. I tell people that s7raw is a pretty good raw converter, but I'm also usually quick to add that I don't think it is in the same league with Adobe Camera Raw.

Jim
jphess is offline   Reply With Quote
Old May 5, 2006, 12:31 PM   #22
Junior Member
 
THE_LoneWolf's Avatar
 
Join Date: Apr 2006
Posts: 21
Default

Apparently you are either under the impression that you have psychic abilities and can thus divine my level of understanding and experience with the raw workflow process without knowing a damned thing about me or you have difficulty reading and comprehending written english.

Yes I am fully aware of the advantages and nuances of processing the raw image. What you are apparently unaware of and would be if you cared to fully read what I have previously posted in this thread is that in the conversion from CCD-RAW to Exif-TIFF there is no alteration of the data in the conversion process.

From that point it is simply a matter of loading the unprocessed ( for all intents and purposes RAW ) tif image into PSP, setting your workspace to 16 bit depth from 8, and proceeding with ALL the rest of your post processing. No need to "bother with a raw converter that provides editing capabilities" at all which is exactly my point. Thank you for pointing out the obvious.


jphess wrote:
Quote:
Apparently you haven't studied the advantages of using a raw converter to its fullest capabilities. The preferred workflow in Adobe Camera Raw is to do white balance adjustments, curve adjustments, levels, color temperature, and other adjustments in camera raw. I'm not going to waste your time trying to explain it to you because judging from your attitude you wouldn't believe me anyway. But, yes, some adjustments are done better in the raw converter. If you don't understand that, then you need to be doing some reading. Why do you suppose all those adjustments are available in s7raw if all you are going to do is convert the file? If that IS all you are going to do, then you might as well not even bother with s7raw and use the very simple converter that comes with the camera. That will produce a TIF image that you can load into Paintshop Pro. Why even bother with a raw converter that provides editing capabilities?
THE_LoneWolf is offline   Reply With Quote
Old May 5, 2006, 2:35 PM   #23
Senior Member
 
Join Date: May 2005
Posts: 1,063
Default

I went back and read your comments again, and I stand by what I said. I'm not going to argue the point. Since you were the first to start pointing fingers, I will step aside and not comment in this thread anymore. It isn't worth it.
jphess is offline   Reply With Quote
Old May 5, 2006, 3:29 PM   #24
Senior Member
 
bernabeu's Avatar
 
Join Date: Jan 2005
Posts: 990
Default

CHILLLLLDREN: Play nice!
bernabeu is offline   Reply With Quote
Old May 5, 2006, 5:58 PM   #25
Junior Member
 
Join Date: Apr 2006
Posts: 7
Default

i prefer s7raw because it works!!!

Adobe does not,be it my setup here or my version of photoshop,no matter how i use adobe software the end result is,well,rubbish!!

And my god this thread has just got out of hand so can i just add



:?:shock::sad::?:P
bettyswolloks is offline   Reply With Quote
Old May 5, 2006, 8:57 PM   #26
Senior Member
 
Join Date: May 2005
Posts: 1,063
Default

Betty,

I was only trying to find out what it is that you saw that is superior with s7raw because Adobe Camera Raw works so smoothly on my computers and is clearly a better choice. It was not my intention to cause such a problem with any of you. But, apparently it is not my right to ask a question. I honestly thought I was contributing something. It is now clear that I have only antagonized some of you. My apologies to all of you.

Jim
jphess is offline   Reply With Quote
Old May 9, 2006, 7:37 PM   #27
Member
 
honus402's Avatar
 
Join Date: Mar 2006
Posts: 38
Default

I prefer the S7RAW over the ACR. Not because it's any better, just because I like the layout and options.
honus402 is offline   Reply With Quote
Old May 9, 2006, 9:05 PM   #28
Administrator
 
Join Date: Jun 2003
Location: Savannah, GA (USA)
Posts: 22,378
Default

THE_LoneWolf wrote:
Quote:
Yes I am fully aware of the advantages and nuances of processing the raw image. What you are apparently unaware of and would be if you cared to fully read what I have previously posted in this thread is that in the conversion from CCD-RAW to Exif-TIFF there is no alteration of the data in the conversion process.
Actually, the conversion process does modify the data, and there can be a big difference between the demosaic algorithms used by different raw converters.

Remember, the invidual photosites in a sensor are only sensitive to one color each, and with most Bayer Pattern CCD designs, you have twice as many sensitive to green.

The raw conversion algorithms take the values from the red, green, and blue photosites and combine them via sophisticated interpolation techniques so that all 3 colors are stored for each pixel.

The raw file has not combined the photosites in any way.

That's what the raw conversion process does. There are a number of different algorithms used, and some are better than others. You can see some of the common ones discussed here (requires Adobe Acrobat Reader):

http://www.ece.gatech.edu/research/labs/MCCL/pubs/dwnlds/bahadir05.pdf

When you convert to TIFF using a raw converter, 16 bit or not, you've gone through this interpolation process combining the values from the photosites.

Most converters are doing some additional processing, too (sharpening, contrast, etc.).

In Fuji's case, they are using some relatively sophisticated noise reduction techniques. So, the TIFF files are rather heavily processed from newer models like the S9000 using the Fuji supplied Raw Converter LE.

Personally, I'd save the raw files, no matter your preference in raw converters, as technology is continuing to improve and you may want to reprocess cherished images later using better tools.

For example, b
eginning with dcraw.c version 7.60, David Coffin started using AHD (Adaptive Homogeniety Directed) for the interpolation algorithms.

This was in part due to Paul J. Lee, whocollaborated with Hirakawa Keigo (the original co-author of AHD algorithm as part of his Ph.D thesis).

Once Paul fully understood the algorithm,he modifieddcraw.c to use it and gave a prototypeto Dave. After some comparison studies, Dave was convinced that AHDwas superior to VNG and other demosaicing algorithms.

Dave converted and optimized the prototype created by Paul J. Lee, integrating it into dcraw.c versions 7.60 and later (he's continuing to refine it).


So, products using some or all of David's code to perform the raw conversion also benefit from these improvements, while adding their own enhancements and features.

I personally like the edge transitions better with AHD. So, I sometimes use dcraw.c (or products using David's code).

Other times, I may use ACR. It's image dependent which raw converter I choose.

I can remember reading some of Eric Hyman's comments a while back, discussing a raw conversion review that put Bibble in a bad light compared to some of the other converters. He took some time to explain why a particular image didn't work well with his algorithms, and why the test was slanted against him, explainng that the ones he used at the time were a compromise (primarly for better speed).

He's continued to refine them to reach a good compromise between speed and quality.

Anyway, a raw converted to TIFF does not contain the same data as the raw file. This data goes through a lot of changes during the conversion process, and all converters are not the same. Fuji's probably does a lot more processing than many, because of the noise reduction being applied.

JimC is offline   Reply With Quote
Old May 10, 2006, 1:50 AM   #29
Junior Member
 
THE_LoneWolf's Avatar
 
Join Date: Apr 2006
Posts: 21
Default

Insofar as Fuji's LE raw conversion goes I have discovered that, despite the huge 51MB file size, it is only an 8 bit conversion so I was wrong about there not being any data loss in the translation anyway. I have since then found other converters that do a 16 bit conversion within that same 51MB.

Arguably, whether or not the processing takes place in-camera or on the PC the image data that gets recorded after it's presented from the sensor is the result of some interpolation method or other and not the true representation of the light stream striking the sensor anyway. Two thirds of the recorded data is not 100% true at all being only approximations. The fact remains however that for every pixel location 3 values for color will be stored or 3 bits of data of the recorded 12 bits of information per pixel. The article you reference deals entirely with the varying methods of performing those estimations from which the raw image is recorded and which I would point out is not within the control of the end user, and therefore irrelevant.

In converting the entire 12 bit pixel map to any 16 bit format there can be no data loss. All the pixels are mapped to their correct respective physical locations on the data plane. That's a given otherwise you wouldn't get anything resembling the original subject of the photograph. The only difference is that there will be 4 bits of null data left over in the process.

So whether you postprocess the raw image within the raw converter software if it offers that capabilty and then save it out as some other format or immediately save it out as a 16 bit TIFF from the raw converter software and performing your postprocessing with Photoshop, or PaintShop Pro, or whatever you'll still be working with all the data that was available in the original raw file. How could you possibly not be?

THE_LoneWolf is offline   Reply With Quote
Old May 10, 2006, 7:18 AM   #30
Administrator
 
Join Date: Jun 2003
Location: Savannah, GA (USA)
Posts: 22,378
Default

THE_LoneWolf wrote:
Quote:
Insofar as Fuji's LE raw conversion goes I have discovered that, despite the huge 51MB file size, it is only an 8 bit conversion so I was wrong about there not being any data loss in the translation anyway. I have since then found other converters that do a 16 bit conversion within that same 51MB.
It's doing a lot more... If you compare it's output to some of the other converters, you'll see that a significant amount of noise reduction is being applied to the image. It's actually pretty darn good at it. Ditto for the in camera processing.

Fuji has apparently worked very hard to reduce the appearance of noise to allow the use of higher ISO speeds with a relatively small sensor, even though the sensor itself is producing significantly more noise that you get in the output from the camera or Fuji's raw converter.

Quote:
Arguably, whether or not the processing takes place in-camera or on the PC the image data that gets recorded after it's presented from the sensor is the result of some interpolation method or other and not the true representation of the light stream striking the sensor anyway.
That's why you have competition. ;-) It's a constant battle to try and produce an image that is representative of what the human eye sees without unwanted artifacts. Some converters are better than it compared to others.

Quote:
The article you reference deals entirely with the varying methods of performing those estimations from which the raw image is recorded and which I would point out is not within the control of the end user, and therefore irrelevant.
Of course it's within your control since you have a choice of raw converters. You choose the raw converter that works the best, for how you want to perceive the image.

Or, you develop one that works for you. I once spent two full days doing nothing but modifying David Coffin's dcraw.c code trying to match the output of a camera's in camera processing. I added a Konica KD-510z to it, before David added it (this camera has a hidden raw mode).

David and another developer got involved and did a better job than I did, and they also tried to match the camera's output, and gave me suggestions for "tweaking" some of the process myself for the desired results.

So, when I want something close to what the camera produces, I've got a customized version of dcraw.c that I can use with it. It's based on his older algorithms, before he redesigned the matrix to support Adobe DNG, and before he switched to the AHD algorithm. I haven't tried to port the customized changes to any of the newer versions, since I rarely shoot in raw with that camera (it's pretty slow writing raw files). His new versions still support this camera (just using different algorithms).

David supports more cameras than anyone (and if you see another raw converter with as many cameras supported, you can bet they're using some or all of David's code). I feel confident that David has more experience developing raw conversion code compared to anyone else. David Coffin released the first version of dcraw.c on February 23, 1997 (years before Adobe's first release of camera raw in 2002).

The pioneers at improving raw conversion are really David Coffin (the author of dcraw.c), mchaney (the author ofQImage Pro, ) and Eric Hyman (the author of Bibble

David's first efforts were primarily to improve images from his own Canon models. Mike and Eric were the first to work on raw conversion code to improve images from DSLR models (for example, the Nikon D1).

We owe a lot of thanks to all of them for leading the way, since the manufacturers' code left a lot to be desired for raw conversion from early models (and in many cases, the manufacturers' code (either in camera conversion or using their raw converter) still leaves a lot to be desired.

Technology has improved a lot over the years, and I believe that it will continue to improve (which is one reason I keep my raw files, so that I can take advantage of better raw converters later).

Quote:
So whether you postprocess the raw image within the raw converter software if it offers that capabilty and then save it out as some other format or immediately save it out as a 16 bit TIFF from the raw converter software and performing your postprocessing with Photoshop, or PaintShop Pro, or whatever you'll still be working with all the data that was available in the original raw file. How could you possibly not be?
The output of a raw converter is not even close to what the sensor produced. The converter is guessing (or if you prefer, estimating) the color (stored as RGB values) a pixel needs to be (since each photosite is only responsive to red, green or blue), based on the values of adjacent photosites, using a variety of different algorithms, depending on the raw converter used.

How that conversion process is performed can make a big difference in the way the image looks, especially at edge transitions between different colors (how well a converter handles it without unwanted artifacts or smoothing of detail).

We're continuing to see refinements in the algorithms used to better estimate each pixel's value. In the future, I'd expect far more sophisticated algorithms that do an even better job. As PCs get more powerful, it wouldn't surprise me to see very advanced scene recognition techniques being used to improve on this process, rather than trying to rely on what adjacent photosites recorded alone.

We're already seeing a lot of progress being made in the highlight retention area, with Eric Hyman and Thomas Knoll (Chief Engineer for Adobe Camera Raw adn the original author of Photoshop) leading the way.

Their newer algorithms attempt to reconstruct the highlight detail, even when the values are maxed out (blown) for some photosites, by looking at adjacent photosites and better predicting the values needed to get detail back that was otherwise lost during the exposure.

Noise reduction is another area we're seeing a lot of progress in of late. Eric Hyman has recently incorporated some of the technology from Noise Ninja into his products. Eric believes that by integrating this technology into the earliest part of the raw conversion pipeline, he can more effectively eliminate noise.

Once the demosaic process has taken place, you limit your options, since you have no way to know what values were recorded by the individual photosites in the sensor, so that you can modify how you're combining the data into the final stored values for each pixel.

Fuji's image processing is also an excellent example of reducing noise early in the image processing pipeline (both the in camera processing and their raw converters). Unfortunately, they don't publish their code (no doubt to help maintain a competitive advantage for low apparent noise from relatively small sensors). Their newer sensors are not as clean they appear (but, their image processing algorithms to reduce the appearance of noise are very good).


JimC is offline   Reply With Quote
 
Reply


Thread Tools Search this Thread
Search this Thread:

Advanced Search

Posting Rules
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts

BB code is On
Smilies are On
[IMG] code is On
HTML code is Off
Trackbacks are Off
Pingbacks are Off
Refbacks are Off



All times are GMT -5. The time now is 8:18 PM.